Structure-based designed herbicide resistant products

Abstract

Disclosed herein are structure-based modelling methods for the preparation of acetohydroxy acid synthase (AHAS) variants, including those that exhibit selectively increased resistance to herbicides such as imidazoline herbicides and AHAS inhibiting herbicides. The invention encompasses isolated DNAs encoding such variants, vectors that include the DNAs, and methods for producing the variant polypeptides and herbicide resistant plants containing specific AHAS gene mutations. Methods for weed control in crops are also provided.

Description

FIELD OF THE INVENTION

This invention pertains to structure-based modelling and design of variants of acetohydroxy acid synthase (AHAS) that are resistant to imidazolinones and other herbicides, the AHAS inhibiting herbicides, AHAS variants themselves, DNA encoding these variants, plants expressing these variants, and methods of weed management.

BACKGROUND OF THE INVENTION

Acetohydroxy acid synthase (AHAS) is an enzyme that catalyzes the initial step in the biosynthesis of isoleucine, leucine, and valine in bacteria, yeast, and plants. For example, the mature AHAS from Zea Mays is approximately a 599-amino acid protein that is localized in the chloroplast (see FIG. 1). The enzyme utilizes thiamine pyrophosphate (TPP) and flavin adenine dinucleotide (FAD) as cofactors and pyruvate as a substrate to form acetolactate. The enzyme also catalyzes the condensation of pyruvate and 2-ketobutyrate to form acetohydroxybutyrate. AHAS is also known as acetolactate synthase or acetolactate pyruvate lyase (carboxylating), and is designated EC 4.1.3.18. The active enzyme is probably at least a homodimer. Ibdah et al. (Protein Science, 3:479-S, 1994), in an abstract, disclose one model for the active site of AHAS.

A variety of herbicides including imidazolinone compounds such as imazethapyr (PURSUIT®—American Cyanamid Company—Wayne, N.J.), sulfonylurea-based compounds such as sulfometuron methyl (OUST®—E.I. du Pont de Nemours and Company-Wilmington, Del.), triazolopyrimidine sulfonamides (Broadstrike™—Dow Elanco; see Gerwick, et al., Pestic. Sci. 29:357-364, 1990), sulfamoylureas (Rodaway et al., Mechanisms of Selectively of Ac 322,140 in Paddy Rice, Wheat and Barley, Proceedings of the Brighton Crop Protection Conference—Weeds, 1993), pyrimidyl-oxy-benzoic acids (STABLE®—Kumiai Chemical Industry Company, E.I. du Pont de Nemours and Company; see, The Pesticide Manual 10th Ed. pp. 688-889, Clive Tomlin, Ed., British Crop Protection Council, 49 Downing Street, Farmham, Surrey G49 7PH, UNITED KINGDOM), and sulfonylcarboxamides (Alvarado et al., U.S. Pat. No. 4,883,914) act by inhibiting AHAS enzymatic activity. (See, Chaleff et al., Science 224:1443, 1984; LaRossa et al., J. Biol. Chem. 259:8753, 1984; Ray, Plant Physiol. 7:827, 11984; Shaner et al., Plant Physiol. 11:545, 1984). These herbicides are highly effective and environmentally benign. Their use in agriculture, however, is limited by their lack of selectivity, since crops as well as undesirable weeds are sensitive to the phytotoxic effects of these herbicides.

Bedbrook et al., U.S. Pat. Nos. 5,013,659, 5,141,870, and 5,378,824, disclose several sulfonylurea resistant AHAS variants. However, these variants were either obtained by mutagenizing plants, seeds, or cells and selecting for herbicide-resistant mutants, or were derived from such mutants. This approach is unpredictable in that it relies (at least initially) on the random chance introduction of a relevant mutation, rather than a rational design approach based on a structural model of the target protein.

Thus, there is still a need in the art for methods and compositions that provide selective wide spectrum and/or specific herbicide resistance in cultivated crops. The present inventors have discovered that selective herbicide resistant variant forms of AHAS and plants containing the same can be prepared by structure-based modelling of AHAS against pyruvate oxidase (POX), identifying an herbicide binding pocket or pockets on the AHAS model, and designing specific mutations that alter the affinity of the herbicide for the binding pocket. These variants and plants are not inhibited or killed by one or more classes of herbicides and retain sufficient AHAS enzymatic activity to support crop growth.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is an illustration of a 600 amino acid sequence corresponding to the approximately 599 amino acid sequence of acetohydroxy acid synthase (AHAS) from Zea Mays which is given as an example of a plant AHAS enzyme. The sequence does not include a transit sequence, and the extra glycine is vestigial from a thrombin cleavage site. Residues Met53, Arg128, and Phe135 are shown in bold.

FIG. 2 is an illustration of the alignment of the sequence of maize AHAS and pyruvate oxidase (POX) from Lactobacillus planarum.

FIG. 3 is a schematic representation of the secondary structure of an AHAS subunit. Regular secondary structure elements, α-helices and β-sheets, are depicted as circles and ellipses, respectively, and are numbered separately for each of the three domains within a subunit. Loops and coiled regions are represented by black lines, with numbers representing the approximate beginnings and ends of the elements. The locations of cofactor binding sites and known mutation sites are indicated by octahedrons and stars, respectively.

FIG. 4 is an illustration of a computer-generated model of the active site of maize AHAS with imazethapyr (PURSUIT® herbicide) modeled into the binding pocket.

FIG. 5 is an illustration of the homology among AHAS amino acid sequences derived from different plant species. pAC 751 is maize als 2 AHAS isozyme as expressed from the pAC 751 E. Coli expression vector as in FIG. 1; Maize als 2 is the maize als 2 AHAS isozyme; Maize als 1 is the maize als 1 AHAS isozyme; Tobac 1 is the tobacco AHAS SuRA isozyme; Tobac 2 is the tobacco AHAS SuRB isozyme; Athcsr 12 is the Arabidopsis thaliana Csr 1.2 AHAS gene; Bnaal 3 is the Brassica napus AHAS III isozyme; and Bnaal 2 is the Brassica napus AHAS II isozyme.

pAC 751 and Maize als 2 are identical genes except that Maize als 2 starts at the beginning of the transit sequence and pAC 751 starts at the putative mature N-terminal site with an additional glycine at the N-terminal due to the thrombin recognition sequence in the pGEX-2T expression vector. The N-terminal glycine is not a natural amino acid at that position.

Amino acid sequence alignments of the AHAS proteins were generated by PILEUP (GCG Package—Genetics Computer Group, Inc.,—University Research Park—Madison, Wis.). The consensus sequence was generated by PRETTY GCG Package.

FIG. 6 is a photographic illustration of an SDS-polyacrylamide gel stained for protein showing purification of maize AHAS. The lanes contain (from left to right): A, Molecular weight markers; B, Crude E. coli cell extract; C, Glutathione-agarose affinity purified preparation; D, Thrombin digest of the affinity purified preparation; E, Second pass through glutathione-agarose column and Sephacryl S-100 gel filtration.

FIG. 7 is a graphic illustration of the results of in vitro assays of the enzymatic activity of wild-type and mutant AHAS proteins in the absence and in the presence of increasing concentrations of imazethapyr (PURSUIT^∘ herbicide) The Y axis represents the % of activity of the mutant enzyme, wherein the 100% value is measured in the absence of inhibitor.

FIG. 8 is a graphic illustration of the results of in vitro assays of the enzymatic activity of wild-type and mutant AHAS proteins in the absence and presence of increasing concentrations of sulfometuron methyl (OUST® herbicide). The Y axis represents the % of activity of the mutant enzyme, wherein the 100% value is measured in the absence of inhibitor.

FIG. 9 is a graphic illustration of in vitro assays of the enzymatic activity of wild-type Arabidopsis AHAS protein and the Met124Ile mutant Arabidopsis AHAS protein in the absence and presence of increasing concentrations of imazethapyr (PURSUIT® herbicide) and sulfometuron methyl (OUST® herbicide). The Y axis represents the % activity of the mutant enzyme, wherein the 100% value is measured in the absence of inhibitor.

FIG. 10 is a graphic illustration of in vitro assays of the enzymatic activity of wild-type Arabidopsis AHAS protein and the Met124His mutant Arabidopsis AHAS protein in the absence and presence of increasing concentrations of imazethapyr (PURSUIT® herbicide) and sulfometuron methyl (OUST® herbicide). The Y axis represents the X activity of the mutant enzyme, wherein the 100% value is measured in the absence of inhibitor.

FIG. 11 is a graphic illustration of in vitro assays of the enzymatic activity of wild-type Arabidopois AHAS protein and Arg199Glu mutant Arabidopsis AHAS protein in the absence and presence of increasing concentrations of imazethapyr (PURSUIT® herbicide) and sulfometuron methyl (OUST® herbicide). The Y axis represents the % activity of the mutant enzyme, wherein the 100% value is measured in the absence of inhibitor.

FIG. 12 is a schematic illustration of a DNA vector used for plant transformation, which contains the nptII gene (encoding kanamycin resistance) under the control of the 35S promoter and an AHAS gene (wild type or variant) under the control of the Arabidopsis AHAS promoter.

FIG. 13 is a photograph showing the root development of tobacco plants transformed with the Arabidopsis AHAS gene containing either the Met124Ile or Arg199Glu mutation and a non-transformed control. Plants were grown for 18 days after transfer into medium containing 0.25 μM imazethapyr.

FIG. 14 is a photograph showing tobacco plants transformed with the Arabidopsis AHAS gene containing either the Met124Ile, Met 124His, or Arg199Glu mutation and a non-transformed control, which had been sprayed with twice the field rate (100 g/ha) of imazethapyr.

FIG. 15 is a photograph showing the results of germination tests performed in the presence of the herbicide CL 299,263 (imazamox), which were performed on seeds harvested from primary tobacco plant transformants that had been transformed with the Arabidopsis AHAS gene containing either the Met124Ile, Met 124His, or Arg199Glu mutation.

The present invention provides a structure-based modelling method for the production of herbicide resistant AHAS variant protein. The method includes:

• • (a) aligning a target AHAS protein on pyruvate oxidase template or an AHAS modelling equivalent thereof to derive the three-dimensional structure of the target AHAS protein;
  • (b) modelling one or more herbicides into the three-dimensional structure to localize an herbicide binding pocket in the target AHAS protein;
  • (c) selecting as a target for a mutation, at least one amino acid position in the target AHAS protein, wherein the mutation alters the affinity of at least one herbicide for the binding pocket;
  • (d) mutating DNA encoding the target AHAS protein to produce a mutated DNA encoding a variant AHAS containing the mutation, such as, for example, at least one different amino acid, at the position; and
  • (e) expressing the mutated DNA in a first cell, under conditions in which the variant AHAS containing the mutation, such as, for example, the different amino acid(s), at the position is produced.

The method further may include:

• • (f) expressing DNA encoding wild-type AHAS protein parallel in a second cell;
  • (g) purifying the wild-type and the variant AHAS proteins from the cells;
  • (h) assaying the wild-type and the variant AHAS proteins for catalytic activity in conversion of pyruvate to acetolactate or in the condensation of pyruvate and 2-ketobutyrate to form acetohydroxybutyrate, in the absence and in the presence of the herbicide; and
  • (i) repeating steps (c)-(h), wherein the DNA encoding the AHAS variant of step (e) is used as the AHAS-encoding DNA in step (c) until a first herbicide resistant AHAS variant protein is identified having:
    • (i) in the absence of the at least one herbicide,
      • (a) catalytic activity alone sufficient to maintain the viability of a cell in which it is expressed; or
      • (b) catalytic activity in combination with any herbicide resistant AHAS variant protein also expressed in the cell, which may be the same as or different than the first AHAS variant protein, sufficient to maintain the viability of a cell in which it is expressed;
      • wherein the cell requires AHAS activity for viability; and
    • (ii) catalytic activity that is more resistant to the at least one herbicide than is wild-type AHAS.

An alternate structure-based modelling method for the production of herbicide resistant AHAS variant protein is also provided. This method includes:

• • (a) aligning a target AHAS protein on a first AHAS template derived from a polypeptide having the sequence of FIG. 1 or a functional equivalent thereof to derive the three-dimensional structure of the target AHAS protein;
  • (b) modelling one or more herbicides into the three-dimensional structure to localize an herbicide binding pocket in the target AHAS protein;
  • (c) selecting as a target for a mutation, at least one amino acid position in the target AHAS protein, wherein the mutation alters the affinity of at least one herbicide for the binding pocket;
  • (d) mutating DNA encoding the target AHAS protein to produce a mutated DNA encoding a variant AHAS containing the mutation at the position; and
  • (e) expressing the mutated DNA in a first cell, under conditions in which the variant AHAS containing the mutation at the position is produced.

This method can further include:

• • (f) expressing DNA encoding wild-type AHAS protein in parallel in a second cell;
  • (g) purifying the wild-type and the variant AHAS protein from the cells;
  • (h) assaying the wild-type and the variant AHAS protein for catalytic activity in conversion of pyruvate to acetolactate or in the condensation of pyruvate and 2-ketobutyrate to form acetohydroxybutyrate, in the absence and in the presence of the herbicide; and
  • (i) repeating steps (c)-(h), wherein the DNA encoding the AHAS variant of step (e) is used as the AHAS-encoding DNA in step (c) until a first herbicide resistant AHAS variant protein is identified having:
    • (i) in the absence of the at least one herbicide,
      • (a) catalytic activity alone sufficient to maintain the viability of a cell in which it is expressed; or
      • (b) catalytic activity in combination with any herbicide resistant AHAS variant protein also expressed in the cell, which may be the same as or different than the first AHAS variant protein, sufficient to maintain the viability of a cell in which it is expressed;
      • wherein the cell requires AHAS activity for viability; and
    • (ii) catalytic activity that is more resistant to the at least one herbicide than is wild-type AHAS.

In another alternate embodiment, the method includes:

• • (a) aligning a target AHAS protein on a first AHAS template having an identified herbicide binding pocket and having the sequence of FIG. 1 or a functional equivalent thereof to derive the three-dimensional structure of the target AHAS protein;
  • (b) selecting as a target for a mutation, at least one amino acid position in the target AHAS protein, wherein the mutation alters the affinity of at least one herbicide for the binding pocket;
  • (c) mutating DNA encoding the target AHAS protein to produce a mutated DNA encoding a variant AHAS containing the mutation at the position; and
  • (d) expressing the mutated DNA in a first cell, under conditions in which the variant AHAS containing the mutation at the position is produced.

This method can further include:

• • (e) expressing DNA encoding wild-type target AHAS protein in parallel in a second cell;
  • (f) purifying the wild-type and the variant AHAS protein from the cells;
  • (g) assaying the wild-type and the variant AHAS protein for catalytic activity in conversion of pyruvate to acetolactate or in the condensation of pyruvate and 2-ketobutyrate to form acetohydroxybutyrate, in the absence and in the presence of the herbicide; and
  • (h) repeating steps (b)-(g), wherein the DNA encoding the AHAS variant of step (d) is used as the AHAS-encoding DNA in step (b) until a first herbicide resistant AHAS variant protein is identified having:
    • (i) in the absence of the at least one herbicide,
      • (a) catalytic activity alone sufficient to maintain the viability of a cell in which it is expressed; or
      • (b) catalytic activity in combination with any herbicide resistant AHAS variant protein also expressed in the cell, which may be the same as or different than the first AHAS variant protein, sufficient to maintain the viability of a cell in which it is expressed;
      • wherein the cell requires AWLS activity for viability; and
    • (ii) catalytic activity that is more resistant to the at least one herbicide than is wild-type AWLS.

In preferred embodiments of the above methods, the catalytic activity in the absence of the herbicide is at least about 5% and most preferably is more than about 20% of the catalytic activity of the wild-type AHAS. Where the herbicide is an imidazolinone herbicide, the herbicide resistant AHAS variant protein preferably has:

• • (i) catalytic activity in the absence of the herbicide of more than about 20% of the catalytic activity of the wild-type AHAS;
  • (ii) catalytic activity that is relatively more resistant to the presence of imidazolinone herbicides compared to wild-type AHAS; and
  • (iii) catalytic activity that is relatively more sensitive to the presence of sulfonylurea herbicides compared to imidazolinone herbicides.

The present invention further provides isolated DNA encoding acetohydroxy acid synthase (AHAS) variant proteins, the variant proteins comprising an AHAS protein modified by:

(i) substitution of at least one different amino acid residue at an amino acid residue of the sequence of FIG. 1 selected from the group consisting of P48, G49, S52, M53, E54, A84, A95, T96, S97, G98, P99, G100, A101, V125, R127, R128, M129, I130, G131, T132, D133, F135, Q136, D186, I187, T259, T260, L261, M262, G263, R276, M277, L278, G279, H281, G282, T283, V284, G300, V301, R302, F303, D304, R306, V307, T308, G309, K310, I311, E312, A313, F314, A315, S316, R317, A318, K319, I 320, E329, I330, K332, N333, K334, Q335, T404, G413, V414, G415, Q416, H417, Q418, M419, W420, A421, A422, L434, S435, S436, A437, G438, L439, G440, A441, M442, G443, D467, G468, S469, L471, N473, L477, M479, Q495, H496, L497, G498, M499, V501, Q502, Q504, D505, R506, Y508, K509, A510, N511, R512, A513, H514, T515, S524, H572, Q573, E574, H575, V576, L577, P578, M579, I580, P581, G583, G584, functional equivalents of any of the foregoing, and any combination of any of the foregoing;

(ii) deletion of up to 5 amino acid residues preceding, or up to 5 amino acid residues following at least one amino acid residue of the sequence of FIG. 1 selected from the group consisting of P48, G49, S52, M53, E54, A84, A95, T96, S97, G98, P99, G100, A101, V125, R127, R128, M129, I130, G131, T132, D133, F135, Q136, D186, I187, T259, T260, L261, M262, G263, R276, M277, L278, G279, H281, G282, T283, V284, G300, V301, R302, F303, D304, R306, V307, T308, G309, K310, I311, E312, A313, F314, A315, S316, R317, A318, K319, I320, E329, I330, K332, N333, K334, Q335, T404, G413, V414, G415, Q416, H417, Q418, M419, W420, A421, A422, L434, S435, S436, A437, G438, L439, G440, A441, M442, G443, D467, G468, S469, L471, N473, L477, M479, Q495, H496, L497, G498, M499, V501, Q502, Q504, D505, R506, Y508, K509, A510, N511, R512, A513, H514, T515, S524, H572, Q573, E574, H575, V576, L577, P578, M579, I580, P581, G583, G584, functional equivalents of any of the foregoing, and any combination of any of the foregoing;

(iii) deletion of at least one amino acid residue or a functional equivalent thereof between Q124 and H150 of the sequence of FIG. 1;

(iv) addition of at least one amino acid residue or a functional equivalent thereof between Q124 and H150 of the sequence of FIG. 1;

(v) deletion of at least one amino acid residue or a functional equivalent thereof between G300 and D324 of the sequence of FIG. 1;

(vi) addition of at least one amino acid residue or a functional equivalent thereof between G300 and D324 of the sequence of FIG. 1; or

(vii) any combination of any of the foregoing.

In this numbering system, residue #2 corresponds to the putative amino terminus of the mature protein, i.e., after removal of a chloroplast targeting peptide.

The above modifications are directed to altering the ability of an herbicide, and preferably an imidazolinone-based herbicide, to inhibit the enzymatic activity of the protein. In a preferred embodiment, the isolated DNA encodes an herbicide-resistant variant of AHAS. Also provided are DNA vectors comprising DNA encoding these AHAS variants, variant AHAS proteins themselves, and cells, grown either in vivo or in cell culture, that express the AHAS variants or comprise these vectors.

In another aspect, the present invention provides a method for conferring herbicide resistance on a cell or cells and particularly a plant cell or cells such as, for example, a seed. An AHAS gene, preferably the Arabidopsis thaliana AHAS gene, is mutated to alter the ability of an herbicide to inhibit the enzymatic activity of the AHAS. The mutant gene is cloned into a compatible expression vector, and the gene is transformed into an herbicide-sensitive cell under conditions in which it is expressed at sufficient levels to confer herbicide resistance on the cell.

Also contemplated are methods for weed control, wherein a crop containing an herbicide resistant AHAS gene according to the present invention is cultivated and treated with a weed-controlling effective amount of the herbicide.

Also disclosed is a structure-based modelling method for the preparation of a first herbicide which inhibits AHAS activity. The method comprises:

• • (a) aligning a target AHAS protein on pyruvate oxidase template or an AHAS modelling functional equivalent thereof to derive the three-dimensional structure of the target AHAS protein;
  • (b) modelling a second herbicide having AHAS inhibiting activity into the three-dimensional structure to derive the location, structure, or a combination thereof of an herbicide binding pocket in the target AHAS protein; and
  • (c) designing a non-peptidic first herbicide which will interact with, and preferably will bind to, an AHAS activity inhibiting effective portion of the binding pocket, wherein the first herbicide inhibits the AHAS activity sufficiently to destroy the viability of a cell which requires AHAS activity for viability.

An alternative structure-based modelling method for the production of a first herbicide which inhibits AHAS activity, is also enclosed. The method comprises:

• • (a) aligning a target AHAS protein on a first AHAS template derived from a polypeptide having the sequence of FIG. 1 or a functional equivalent thereof, to derive the three-dimensional structure of the target AHAS protein;
  • (b) modelling a second herbicide having AHAS inhibiting activity into the three-dimensional structure to derive the location, structure, or a combination thereof of an herbicide binding pocket in the target AHAS protein; and
  • (c) designing a non-peptidic first herbicide which will interact with, and preferably will bind to, an AHAS activity inhibiting effective portion of the binding pocket, wherein the first herbicide inhibits the AHAS activity sufficiently to destroy the viability of a cell which requires AHAS activity for viability.

Preferably in each method, the first herbicide contains at least one functional group that interacts with a functional group of the binding pocket.

DETAILED DESCRIPTION OF THE INVENTION

The present invention encompasses the rational design or structure-based molecular modelling of modified versions of the enzyme AHAS and AHAS inhibiting herbicides. These modified enzymes (AHAS variant proteins) are resistant to the action of herbicides. The present invention also encompasses DNAs that encode these variants, vectors that include these DNAs, the AHAS variant proteins, and cells that express these variants. Additionally provided are methods for producing herbicide resistance in plants by expressing these variants and methods of weed control. The DNA and the AHAS variants of the present invention were discovered in studies that were based on molecular modelling of the structure of AHAS.

Rational Structure-Based Design of AHAS Variants and AHAS Inhibiting Herbicides

Herbicide-resistant variants of AHAS according to the present invention are useful in conferring herbicide resistance in plants and can be designed with the POX model, AHAS model, or functional equivalents thereof, such as, for example, transketolases, carboligases, pyruvate decarboxylase, proteins that bind FAD and/or TPP as a cofactor, or any proteins which have structural features similar to POX and/or AHAS; with an AHAS model such as a model having the sequence of FIG. 1; or with a functional equivalent of the sequence of FIG. 1 including a variant modeled from a previous model. Proteins that can be used include any proteins having less than a root mean square deviation of less than 3.5 angstroms in their Cα carbons relative to any of the above-listed molecules. AHAS directed herbicides can be similarly modelled from these templates. A functional equivalent of an AHAS amino acid sequence is a sequence having substantial, i.e., 60-70%, homology, particularly in conserved regions such as, for example, a putative binding pocket. The degree of homology can be determined by simple alignment based on programs known in the art, such as, for example, GAP and PILEUP by GCG. Homology means identical amino acids or conservative substitutions. A functional equivalent of a particular amino acid residue in the AHAS protein of FIG. 1 is an amino acid residue of another AHAS protein which when aligned with the sequence of FIG. 1 by programs known in the art, such as, for example, GAP and PILEUP by GCG, is in the same position as the amino acid residue of FIG. 1.

Rational design steps typically include: (1) alignment of a target AHAS protein with a POX backbone or structure or an AHAS backbone or structure; (2) optionally, and if the AHAS backbone has an identified herbicide binding pocket, modelling one or more herbicides into the three-dimensional structure to localize an herbicide binding pocket in the target protein; (3) selection of a mutation based upon the model; (4) site-directed mutagenesis; and (5) expression and purification of the variants. Additional steps can include (6) assaying of enzymatic properties and (7) evaluation of suitable variants by comparison to the properties of the wild-type AHAS. Each step is discussed separately below.

1. Molecular Modelling

Molecular modelling (and particularly protein homology modelling) techniques can provide an understanding of the structure and activity of a given protein. The structural model of a protein can be determined directly from experimental data such as x-ray crystallography, indirectly by homology modelling or the like, or combinations thereof (See White, et al., Annu. Rev. Biophys. Biomol. Struct., 23:349, 1994). Elucidation of the three-dimensional structure of AHAS provides a basis for the development of a rational scheme for mutation of particular amino acid residues within AHAS that confer herbicide resistance on the polypeptide.

Molecular modelling of the structure of Zea mays AHAS, using as a template the known X-ray crystal structure of related pyruvate oxidase (POX) from Lactobacillus plantarum, provides a three-dimensional model of AHAS structure that is useful for the design of herbicide-resistant AHAS variants or AHAS inhibiting herbicides. This modelling procedure takes advantage of the fact that AHAS and POX share a number of biochemical characteristics and may be derived from a common ancestral gene (Chang et al., J. Bacteriol. 170:3937, 1988).

Because of the high degree of cross-species homology in AHAS the modelled AHAS described herein or functional equivalents thereof can also be used as templates for AHAS variant protein design.

Derivation of one model using interactive molecular graphics and alignments is described in detail below. The three-dimensional AHAS structure that results from this procedure predicts the approximate spatial organization of the active site of the enzyme and of the binding site or pocket of inhibitors such as herbicides including, but not limited to, imidazolinone herbicides. The model is then refined and re-interpreted based on biochemical studies which are also described below.

Protein homology modelling requires the alignment of the primary sequence of the protein understudy with a second protein whose crystal structure is known. Pyruvate oxidase (POX) was chosen for AHAS homology modelling because POX and AHAS share a number of biochemical characteristics. For example, both AHAS and POX share aspects of enzymatic reaction mechanisms, as well as cofactor and metal requirements. In both enzymes thiamine pyrophosphate (TPP), flavin adenine dinucleotide (FAD), and a divalent cation are required for enzymatic activity. FAD mediates a redox reaction during catalysis in POX but presumably has only a structural function in AHAS, which is possibly a vestigial remnant from the evolution of AHAS from POX. Both enzymes utilize pyruvate as a substrate and form hydroxyethyl thiamine pyrophosphate as a stable reaction intermediate (Schloss, J. V. et al. In Biosynthesis of branched chain amino acids, Barak, Z. J. M., Chipman, D. M., Schloss, J. V. (eds) VCH Publishers, Weinheim, Germany, 1990).

Additionally, AHAS activity is present in chimeric POX-AHAS proteins consisting of the N-terminal half of POX and the C-terminal half of AHAS, and there is a small degree of AHAS activity exhibited by POX itself. AHAS and POX also exhibit similar properties in solution (Risse, B. et al, Protein Sci. 1: 1699 and 1710, 1992; Singh, B. K., & Schmitt, G. K. (1989), FEBS Letters, 258: 113; Singh, B. K. et al. (1989) In: Prospects for Amino Acid Biosynthesis Inhibitors in Crop Protection and Pharmaceutical Chemistry, (Lopping, L. G., et al., eds., BCPC Monograph p. 87). With increasing protein concentration, both POX and AHAS undergo stepwise transitions from monomers to dimers and tetramers. Increases in FAD concentration also induce higher orders of subunit assembly. The tetrameric form of both proteins is most stable to heat and chemical denaturation.

Furthermore, the crystal structure of POX from Lactobacillus planarum had been solved by Muller et al., Science 259:965, 1993. The present inventors found that based in part upon the degree of physical, biochemical, and genetic homology between AHAS and POX, the X-ray crystal structure of POX could be used as a structural starting point for homology modelling of the AHAS structure.

AHAS and L. plantarum POX sequences were not similar enough for a completely computerized alignment, however. Overall, only about 20% of the amino acids are identical, while about 50% of the residues are of similar class (i.e. acidic, basic, aromatic, and the like). However, if the sequences are compared with respect to hydrophilic and hydrophobic residue classifications, over 500 of the 600 amino acids match. Secondary structure predictions for AHAS (Holley et al., Proc. Natl. Acad. Sci. USA 86:152, 1989) revealed a strong similarity to the actual secondary structure of POX. For nearly 70% of the residues, the predicted AHAS secondary structure matches that of POX.

POX monomers consist of three domains, all having a central, parallel β-sheet with crossovers consisting of α-helices and long loops. (Muller et al., Science 259:965, 1993). The topology of the sheets differs between the domains, i.e. in the first and third domains, the strands are assembled to the β-sheet in the sequence 2-1-3-4-6-5, while in the β-sheet of the second domain, the sequence reads 3-2-1-4-5-6.

Computer generated alignments were based on secondary structure prediction and sequence homology. The conventional pair-wise sequence alignment method described by Needleman and Wunch, J. Mol. Biol, 48: 443, 1970, was used. Two sequences were aligned to maximize the alignment score. The alignment score (homology score) is the sum of the scores for all pairs of aligned residues, plus an optional penalty for the introduction of gaps into the alignment. The score for the alignment of a pair of residues is a tabulated integer value. The homology scoring system is based on observing the frequency of divergence between a given pair of residues. (M O Dayhoff, R H Schwartz & B C Orcutt “Atlas of Protein Sequence and Structure” vol. 5 suppl. 3 pp. 345-362, 1978).

The alignments were further refined by repositioning gaps so as to conserve continuous regular secondary structures. Amino acid substitutions generated by evaluation of likely alignment schemes were compared by means of interactive molecular graphics. Alignments with the most conservative substitutions with respect to the particular functionality of the amino acids within a given site were chosen. The final alignment of both POX and AHAS is displayed in FIG. 2. Conserved clusters of residues were identified, in particular for the TPP binding site and for parts of the FAD binding site. The alignment revealed a high similarity between AHAS and POX for the first domain, for most parts of the second domain, and for about half of the third domain. Most of the regions that aligned poorly and may fold differently in POX and in AHAS were expected to be at the surface of the protein and were not involved in cofactor or inhibitor binding. The prediction of mutation sites is not substantially affected by small shifts in the alignment.

Most TPP binding residues are highly conserved between POX and AHAS (e.g. P48-G49-G50). In some cases, residues that were close to TPP differ between POX and AHAS but remain within a region that is highly conserved (for example, residues 90-110). On the other hand, the FAD binding site appeared to be less conserved. Although some FAD binding residues were strongly conserved (for example, D325-I326-D327-P328), others clearly differed between AWLS and POX (for example, residues in the loop from positions 278 to 285 are not homologous. A detailed analysis revealed that, at least for some of the less-conserved contact sites, the interactions were mediated by the polypeptide backbone rather than by the side chains. Hence, conservation was only required for the polypeptide fold and was not required for the amino acid sequence (for example, the backbone of residues 258-263 binds the ribitol chain of FAD). One half of the adenine and the isoalloxazine binding sites clearly differ.

After aligning the primary structure, a homology model was built by transposition of AHAS amino acid sequences to the POX template structure. Missing coordinates were built stepwise using templates of amino acid residues to complete undefined side chains. Data bank searches and energy minimization of small parts of the molecule were used to complete the conformations of undefined loop regions. The cofactors TPP and FAD were modeled into their binding pockets. This model was then subjected to a complete, 5000 cycle energy minimization. All computer modelling was performed in an IRIS Indigo Elan R4000 Workstation from Silicon Graphics Co. Interactive molecular modelling and energy-minimization were performed using Quanta/CHARMm 4.0 from Molecular Simulations Inc. During this step, the conformation was stable, indicating that no strongly disfavored interactions, such as, for example, close van der Waals contacts, had occurred. The results are shown schematically in FIG. 3.

Characteristics of Predicted AHAS Structure

Inspection of the modelled AHAS structure described above revealed that most of the protein folds with a backbone that is energetically reasonable, with most hydrophilic side chains accessible to the solvent. The surface of the β-sheets are smooth and accommodate the cross-over regions that are attached to them.

A model for dimeric AHAS was generated by duplicating the coordinates of the energy minimized monomeric AHAS and superimposing the two copies on two POX subunits using pairs of Cα coordinates as defined in the alignment scheme. The polypeptide chain of AHAS folds into three similarly folded domains composed of a six-stranded parallel β-sheet core surrounded by long “loops” and α-helices. Two subunits are assembled such that the first domain of one subunit is in close proximity to the cofactor-binding domains 2 and 3 of the other subunit. A solvent-filled space remains between the subunits at this site. This pocket, which is defined by the confluence of the three domains, is the proposed entry site for the substrate. It is also proposed to be the binding site for herbicides.

The inner surface of the binding pocket is outlined by the cofactors. The thiazol of TPP is positioned at the bottom of the pocket. Domain 3 contributes to the inner surface of the pocket with a short α-helix that points its axis towards the pyrophosphate of TPP, compensating the phosphate charges with its dipolar moment. This critical helix, which starts with G498, a “turn” residue in close contact with TPP, and which ends at F507, contains three known mutation sites for sulfonylurea resistance: V500, W503, and F507 (See, U.S. Pat. Nos. 5,013,659; 5,141,870; and 5,378,824). In domain 1, the loop defined as P48-S52 (between β-strand 2 and α-helix 2) faces W503, a mutation in which confers resistance to imidazolinones. Residues Y47 to G50 are also in contact with TPP. This loop is adjacent to P184-Q189, another turn, which connects the last strand of the β-sheet of domain 1 with a β-strand that connects with domain 2. Within the pocket, near its entrance, is a long region of domain 1 that interacts with a complementary stretch of domain 2. Residues 125-129 and 133-137 of domain 1 and residues 304-313 of domain 2 are at the surface of the pocket. A turn consisting of T96-G100 is between loop 125-129 and TPP. A further stretch of domain 3 and two regions of domain 2 that line the binding pocket are at the opposite corner of the pocket. Residues 572, 575, 582, and 583 of domain 3 define the pocket surface on one side. The remaining part of the interior of the pocket's surface is defined by FAD and by a loop, L278-G282, that contacts the isoalloxazine ring of FAD.

The structural models of the AHAS protein can also be used for the rational design of herbicides or AHAS inhibitors.

2. Modelling of Herbicides into Binding Sites

Imazethapyr, the active imidazolinone in PURSUIT®, was positioned into its proposed binding site using interactive molecular graphics (FIG. 4) and the software described above (FIG. 4). K185 was chosen as an “anchor” to interact with the charge of the carboxyl group. The imidazolinone's NH—CO unit was placed to form hydrogen bonds to G50 and A50. This positioned the methyl substitute of imazethapyr close to V500 on the backbone of the small α-helix. The isopropyl group is possibly bound by hydrophobic residues of the amino acids in the region of residues 125-135 that contribute to the inner surface of the pocket. The pyridine ring is most probably “sandwiched” between A134 or F135, F507 and W503. W503 also interacts with the imidazolinone ring system.

In a similar fashion, the sulfonylurea herbicides were modelled into a site that partially overlapped the described imidazolinone binding site. Overlap of sulfonylurea and imidazolinone binding sites was consistent with competition binding experiments and with established mutant data, which show that the same mutation in maize, W503L, can confer resistance to both herbicides. In these models, most of the known mutation sites that confer sulfonylurea herbicide resistance, i.e. G50, A51, K185, V500, W503, F507, are in close contact to the bound herbicides. P126 and A51 are required for keeping the K185 side chain in place by generating a hydrophobic pore. S582, a site for specific imidazolinone resistance, is distant from the binding region and is located in the region where the homology is so poor that a change in the fold is expected. The FAD binding site apparently has low homology between AHAS and POX in this region; S582 is a residue that confers resistance in maize, and that S582 and its adjacent residues are in close contact to the active site pocket. It is proposed that FAD and the loop region encompassing residues 278 to 285 move slightly away from the third domain, (downward in FIG. 4) and that a loop that contains S582 folds into the space between the helix at positions 499 to 507 and the loop at positions 278 to 285. D305, another known resistance site, is close to FAD and modulates the interaction between domains 1 and 2. M280 may either be involved in positioning of the helix at positions 498 to 507 or directly in inhibitor binding. M280 and D305 could also be directly involved in inhibitor binding if domains 1 and 2 move slightly closer to each other.

3. Selection of Mutations

Specific amino acid residues are pinpointed as sites for the introduction of mutations into the primary sequence of AHAS. These amino acids are selected based upon their position in that if that amino acid residue position is modified, there will be a resultant alteration (i.e. decline) in the affinity of an herbicide for the binding pocket. It is not necessary that the mutation position reside in the binding pocket as amino acid residues outside the pocket itself can alter the pocket charge or configuration. The selection of target sites for mutation is achieved using molecular models as described above. For example according to the model above, arginine at position 128 (designated R128 in FIG. 1 using the single-letter code for amino acids) is located near the entrance to the substrate- and herbicide-binding pocket and has a large degree of conformational freedom that may allow it to participate in transport of charged herbicides into the binding pocket. Therefore, this residue is substituted by alanine to remove both its charge and its long hydrophobic side chain. (The resulting mutation is designated R128A).

The mutations may comprise simple substitutions, which replace the wild-type sequence with any other amino acid. Alternatively, the mutations may comprise deletions or additions of one or more amino acids, preferably up to 5, at a given site. The added sequence may comprise an amino acid sequence known to exist in another protein, or may comprise a completely synthetic sequence. Furthermore, more than one mutation and/or more than one type of mutation may be introduced into a single polypeptide.

4. Site-Directed Mutagenesis

The DNA encoding AHAS can be manipulated so as to introduce the desired mutations. Mutagenesis is carried out using methods that are standard in the art, as described in, for example, Higuchi, R., Recombinant PCR, In M. A. Innis, et al., eds, PCR Protocols: A Guide to Methods and Applications, Academic Press, pp. 177-183, 1990.

5. Expression and Purification of Variants

The mutated or variant AHAS sequence is cloned into a DNA expression vector (Bee, e.g., Example 3) and is expressed in a suitable cell such as, for example, E. coli. Preferably, the DNA encoding AHAS is linked to a transcription regulatory element, and the variant AHAS is expressed as part of a fusion protein, for example, glutathione-S-transferase, to facilitate purification (see Example 3 below). The variant AHAS is then purified using affinity chromatography or any other suitable method known in the art. “Purification” of an AHAS polypeptide refers to the isolation of the AHAS polypeptide in a form that allows its enzymatic activity to be measured without interference by other components of the cell in which the polypeptide is expressed.

6. Assaying of Enzymatic Properties

The purified variant AHAS may be assayed for one or more of the following three properties:

• • (a) specific or catalytic activity for conversion of pyruvate to acetolactate (expressed as units/mg pure AHAS, wherein a unit of activity is defined as 1 μmole acetolactate produced/hour), or for condensation of pyruvate and 2-ketobutyrate to form acetohydroxybutyrate (expressed as units/mg pure AHAS, wherein a unit of activity is defined as 1 μmole acetohydroxybutyrate produced/hr.;
  • (b) level of inhibition by herbicide, such as, for example, imidazolinone (expressed as IC₅₀, the concentration at which 50% of the activity of the enzyme is inhibited); and
  • (c) selectivity of resistance to the selected herbicide vs. other herbicides. The selectivity index is defined as the fold resistance of the mutant to imidazolinones relative to the wild-type enzyme, divided by the fold resistance of the same mutant to other herbicides also relative to the wild-type). Fold resistance to an herbicide relative to the wild-type enzyme is expressed as the IC₅₀ of variant, divided by the IC₅₀ of the wild type. The selectivity index (S.I.) is thus represented by the following equation: $S.I.=\frac{{\mathrm{IC}}_{50}\mathrm{of}\mathrm{variant}\mathrm{for}\mathrm{herb}.A/{\mathrm{IC}}_{50}\mathrm{of}\mathrm{wild}\mathrm{type}\mathrm{for}\mathrm{herb}.A}{{\mathrm{IC}}_{50}\mathrm{of}\mathrm{variant}\mathrm{for}\mathrm{herb}.B/{\mathrm{IC}}_{50}\mathrm{of}\mathrm{wild}\mathrm{type}\mathrm{for}\mathrm{herb}.B}.$

S.I. IC₅₀ of variant for herb.A/IC₅₀ of wild type for herb.A herb.B. IC₅₀ of variant for herb.B/IC₅₀ of wild type for

Suitable assay systems for making these determinations include, but are not limited to, those described in detail in Example 4 below.

7.a. Evaluation of Suitable Variants

The enzymatic properties of variant AHAS polypeptides are compared to the wild-type AHAS. Preferably, a given mutation results in an AHAS variant polypeptide that retains in vitro enzymatic activity towards pyruvate or pyruvate and 2-ketobutyrate, i.e., the conversion of pyruvate to acetolactate or in the condensation of pyruvate and 2-ketobutyrate to form acetohydroxybutyrate (and thus is expected to be biologically active in vivo), while exhibiting catalytic activity that is relatively more resistant to the selected herbicide(s) than is wild-type AHAS. Preferably, the variant AHAS exhibits:

• • (i) in the absence of the at least one herbicide,
    • (a) catalytic activity alone sufficient to maintain the viability of a cell in which it is expressed; or
    • (b) catalytic activity in combination with any herbicide resistant AHAS variant protein also expressed in the cell, which may be the same as or different than the first AHAS variant protein, sufficient to maintain the viability of a cell in which it is expressed;
    • wherein the cell requires AHAS activity for viability; and
  • (ii) catalytic activity that is more resistant to the at least one herbicide than is wild type AHAS;

and that is relatively more resistant to the herbicide(s) than is wild-type AHAS.

Therefore, any one specific AHAS variant protein need not have the total catalytic activity necessary to maintain the viability of the cell, but must have some catalytic activity in an amount, alone or in combination with the catalytic activity of additional copies of the same AHAS variant and/or the catalytic activity of other AHAS variant protein(s), sufficient to maintain the viability of a cell that requires AHAS activity for viability. For example, catalytic activity may be increased to minimum acceptable levels by introducing multiple copies of a variant encoding gene into the cell or by introducing the gene which further includes a relatively strong promoter to enhance the production of the variant.

More resistant means that the catalytic activity of the variant is diminished by the herbicide(s), if at all, to a lesser degree than wild-type AHAS catalytic activity is diminished by the herbicide(s). Preferred more resistant variant AHAS retains sufficient catalytic to maintain the viability of a cell, plant, or organism wherein at the same concentration of the same herbicide(s), wild-type AHAS would not retain sufficient catalytic activity to maintain the viability of the cell, plant, or organism.

Preferably the catalytic activity in the absence of herbicide(s) is at least about 5% and, most preferably, is more than about 20% of the catalytic activity of the wild-type AHAS in the absence of herbicide(s). Most preferred AHAS variants are more resistant to imidazolinone herbicides than to other herbicides such as sulfonylurea-based herbicides, though in some applications selectivity is neither needed nor preferred.

In the case of imidazolinone-resistant variant AHAS, it is preferred that the AHAS variant protein has

(i) catalytic activity in the absence of said herbicide of more than about 20% of the catalytic activity of said wild-type AHAS;

(ii) catalytic activity that is relatively more resistant to presence of imidazolinone herbicides compared to wild type AHAS; and

(iii) catalytic activity that is relatively more sensitive to the presence of sulfonylurea herbicides compared to imidazolinone herbicides. Most preferred herbicide-resistant AHAS variants exhibit a minimum specific activity of about 20 units/mg, minimal or no inhibition by imidazolinone, and a selectivity index ranging from about 1.3 to about 3000 relative to other herbicides.

Without wishing to be bound by theory, it is believed that systematic and iterative application of this method to wild type or other target AHAS protein will result in the production of AHAS variants having the desired properties of high enzymatic activity as explained above and resistance to one or more classes of herbicides. For example, mutation of a wild-type AHAS sequence at a particular position to a given amino acid may result in a mutant that exhibits a high degree of herbicide resistance but a significant loss of enzymatic activity towards pyruvate or pyruvate and 2-ketobutyrate. In a second application of the above method, the starting or target AHAS polypeptide would then be this variant (in place of the wild-type AHAS). Rational design then involves substituting other amino acids at the originally mutated position and/or adding or deleting amino acids at selected points or ranges in the expectation of retaining herbicide resistance but also maintaining a higher level of enzymatic activity.

The structure-based rational design of herbicide resistant AHAS proteins offers many advantages over conventional approaches that rely on random mutagenesis and selection. For example, when substitution of a particular amino acid with another requires substitution of more than one nucleotide within the codon, the likelihood of this occurring randomly is so low as to be impractical. By contrast, even double or triple changes in nucleotide sequence within a codon can be easily implemented when suggested by a rational design approach. For example, one rationally designed mutation to confer selective imidazolinone resistance requires a change from arginine to glutamate. Arginine is encoded by CGT, CGC, CGA, CGG, AGA, AGG, while glutamate is encoded by GAA and GAG. Since none of the arginine codons begins with GA, this mutation would require a double substitution of adjacent nucleotides which would occur so rarely using random mutagenesis as to be unpredictable and unrepeatable with any certainty of success. Although mutation frequency can be increased during random mutagenesis, alterations in nucleotide sequence would have an equal probability of occurring throughout the AHAS gene, in the absence of prior site-direction of the mutations. This increases the chance of obtaining an irrelevant mutation that interferes with enzymatic activity. Similarly, it would be rare, using random mutagenesis, to find a multiple amino acid substitution, deletion, or substitution/deletion mutation that confers herbicide resistance while maintaining catalytic activity. Deletion mutations that confer herbicide resistance would also be unlikely using a random mutagenesis approach. Deletions would need to be limited to small regions and would have to occur in triplets so as to retain the AHAS reading frame in order to retain enzymatic activity.

However, with a rational structure-based approach, double amino acid substitution and/or deletion mutations are relatively easily achieved and precisely targeted. Furthermore, different mutagens used in random mutagenesis create specific types of mutations. For example, sodium azide creates point substitution mutations in plants, while radiation tends to create deletions. Accordingly, two mutagenesis protocols would have to be employed to obtain a multiple combination substitution/deletion.

Finally, the present structure-based method for rational design of herbicide-resistant AHAS variants allows for iterative improvement of herbicide resistance mutations, a step that is not facilitated by random mutagenesis. Identification of a mutation site for herbicide resistance by random mutagenesis may offer little, if any, predictive value for guiding further improvements in the characteristics of the mutant. The present structure-based approach, on the other hand, allows improvements to be implemented based on the position, environment, and function of the amino acid position in the structural model.

The iterative improvement method also allows the independent manipulation of three important properties of AHAS: level of resistance, selectivity of resistance, and catalytic efficiency. For example, compensatory mutations can be designed in a predictive manner. If a particular mutation has a deleterious effect on the activity of an enzyme, a second compensatory mutation may be used to restore activity. For example, a change in the net charge within a domain when a charged residue is introduced or lost due to a mutation can be compensated by introducing a second mutation. Prediction of the position and type of residue(s) to introduce, delete, or substitute at the second site in order to restore enzymatic activity requires a knowledge of structure-function relationships derived from a model such as that described herein.

7.b. Design of Non-Peptide Herbicides or AHAS Inhibitors

A chemical entity that alters and may fit into the active site of the target protein or bind in any position where it could inhibit activity may be designed by methods known in the art, such as, for example, computer design programs that assist in the design of compounds that specifically interact with a receptor site.

An example of such a program is LUDI (Biosym Technologies—San Diego, Calif.) (see also, Lam, et al., Science 263:380, 1994; Thompson, et al., J. Med. Chem., 37:3100, 1994).

The binding pocket and particularly the amino acid residues that have been identified as being involved as inhibitor binding can be used as anchor points for inhibitor design.

The design of site-specific herbicides is advantageous in the control of weed species that may spontaneously develop herbicide resistance in the field, particularly due to mutations in the AHAS gene.

Herbicide-Resistant AHAS Variants: DNA, Vectors, and Polypeptides

The present invention also encompasses isolated DNA molecules encoding variant herbicide-resistant AHAS polypeptides. Genes encoding AHAS polypeptides according to the present invention may be derived from any species and preferably a plant species, and mutations conferring herbicide resistance may be introduced at equivalent positions within any of these AHAS genes. The equivalence of a given codon position in different AHAS genes is a function of both the conservation of primary amino acid sequence and its protein and the retention of similar three-dimensional structure. For example, FIG. 5 illustrates the high degree of sequence homology between AHAS polypeptides derived from different plant species. These AHAS polypeptides exhibit at least about 60 to about 70% overall homology. Without wishing to be bound by theory, it is believed that in regions of the polypeptide having a highly conserved sequence, the polypeptide chain conformation will also be preserved. Thus, it is possible to use an AHAS-encoding sequence from one species for molecular modelling, to introduce mutations predictively into an AHAS gene from a second species for initial testing and iterative improvement, and finally, to introduce the optimized mutations into AHAS derived from yet a third plant species for expression in a transgenic plant.

In one series of embodiment, these AHAS DNAs encode variants of an AHAS polypeptide and preferably of the maize AHAS polypeptide of FIG. 1 in which the polypeptide is modified by substitution at or deletion preceding or following one or more of FIG. 1 amino acid residues P48, G49, S52, M53, E54, A84, A95, T96, S97, G98, P99, G100, A101, V125, R127, R128, M129, I130, G131, T132, D133, F135, Q136, D186, I187, T259, T260, L261, M262, G263, R276, M277, L278, G279, H281, G282, T283, V284, G300, V301, R302, F303, D304, R306, V307, T308, G309, K310, I311, E312, A313, F314, A315, S316, R317, A318, K319, I320, E329, 1330, K332, N333, K334, Q335, T404, G413, V414, G415, Q416, H417, Q418, M419, W420, A421, A422, L434, S435, S436, A437, G438, L439, G440, A441, M442, G443, D467, G468, S469, L471, N473, L477, M479, Q495, H496, L497, G498, M499, V501, Q502, Q504, D505S, R506, Y508, K509, A510, N511, R512, A513, H514, T515, S524, H572, Q573, E574, H575, V576, L577, P578, M579, I580, P581, G583, G584, functional equivalents of any of the foregoing; insertions or deletions between FIG. 1 Q124 and H150 or functional equivalents thereof; insertions or deletions between FIG. 1 G300 and D324 or functional equivalents thereof; and any combination of any of the foregoing thereof.

The mutations, whether introduced into the polypeptide of FIG. 1 or at equivalent positions in another plant AHAS gene, may comprise alterations in DNA sequence that result in a simple substitution of any one or more other amino acids or deletions of up to 5 amino acid residues proceeding or up to 5 amino acids residues following any of the residence listed above. Suitable amino acid substituents include, but are not limited to, naturally occurring amino acids.

Alternatively, the mutations may comprise alterations in DNA sequence such that one or more amino acids are added or deleted in frame at the above positions. Preferably, additions comprise about 3 to about 30 nucleotides, and deletions comprise about 3 to about 30 nucleotides. Furthermore, a single mutant polypeptide may contain more than one similar or different mutation.

The present invention encompasses DNA and corresponding RNA sequences, as well as sense and antisense sequences. Nucleic acid sequences encoding AHAS polypeptides may be flanked by natural AHAS regulatory sequences, or may be associated with heterologous sequences, including promoters, enhancers, response elements, signal sequences, polyadenylation sequences, introns, 5′- and 3′-noncoding regions, and the like. Furthermore, the nucleic acids can be modified to alter stability, solubility, binding affinity and specificity. For example, variant AHAS-encoding sequences can be selectively methylated. The nucleic acid sequences of the present invention may also be modified with a label capable of providing a detectable signal, either directly or indirectly. Exemplary labels include radioisotopes, fluorescent molecules, biotin, and the like.

The invention also provides vectors comprising nucleic acids encoding AHAS variants. A large number of vectors, including plasmid and fungal vectors, have been described for expression in a variety of eukaryotic and prokaryotic hosts. Advantageously, vectors may also include a promotor operably linked to the AHAS encoding portion. The encoded AHAS may be expressed by using any suitable vectors and host cells, using methods disclosed or cited herein or otherwise known to those skilled in the relevant art. Examples of suitable vectors include without limitation pBIN-based vectors, pBluescript vectors, and pGEM vectors.

The present invention also encompasses both variant herbicide-resistant AHAS polypeptides or peptide fragments thereof. As explained above, the variant AHAS polypeptides may be derived from the maize polypeptide shown in FIG. 1 or from any plant or microbial AHAS polypeptide, preferably plant AHAS polypeptide. The polypeptides may be further modified by, for example, phosphorylation, sulfation, acylation, glycosylation, or other protein modifications. The polypeptides may be isolated from plants, or from heterologous organisms or cells (including, but not limited to, bacteria, yeast, insect, plant, and mammalian cells) into which the gene encoding a variant AHAS polypeptide has been introduced and expressed. Furthermore, AHAS polypeptides may be modified with a label capable of providing a detectable signal, either directly or indirectly, including radioisotopes, fluorescent compounds, and the like.

Chemical-Resistant Plants and Plants Containing Variant AHAS Genes

The present invention encompasses transgenic cells, including, but not limited to seeds, organisms, and plants into which genes encoding herbicide-resistant AHAS variants have been introduced. Non-limiting examples of suitable recipient plants are listed in Table 1 below:

TABLE 1
RECIPIENT PLANTS
COMMON NAME	FAMILY	LATIN NAME
Maize	Gramineae	Zea mays
Maize, Dent	Gramineae	Zea mays dentiformis
Maize, Flint	Gramineae	Zea mays vulgaris
Maize, Pop	Gramineae	Zea mays microsperma
Maize, Soft	Gramineae	Zea mays amylacea
Maize, Sweet	Gramineae	Zea mays amyleasaccharata
Maize, Sweet	Gramineae	Zea mays saccharate
Maize, Waxy	Gramineae	Zea mays ceratina
Wheat, Dinkel	Pooideae	Triticum spelta
Wheat, Durum	Pooideae	Triticum durum
Wheat, English	Pooideae	Triticum turgidum
Wheat, Large Spelt	Pooideae	Triticum spelta
Wheat, Polish	Pooideae	Triticum polonium
Wheat, Poulard	Pooideae	Triticum turgidum
Wheat, Singlegrained	Pooideae	Triticum monococcum
Wheat, Small Spelt	Pooideae	Triticum monococcum
Wheat, Soft	Pooideae	Triticum aestivum
Rice	Gramineae	Oryza sativa
Rice, American Wild	Gramineae	Zizania aguatica
Rice, Australian	Gramineae	Oryza australiensis
Rice, Indian	Gramineae	Zizania aguatica
Rice, Red	Gramineae	Oryza glaberrima
Rice, Tuscarora	Gramineae	Zizania aguatica
Rice, West African	Gramineae	Oryza glaberrima
Barley	Pooideae	Hordeum vulgare
Barley, Abyssinian	Pooideae	Hordeum irregulare
Intermediate,
also Irregular
Barley, Ancestral	Pooideae	Hordeum spontaneum
Tworow
Barley. Beardless	Pooideae	Hordeum trifurcatum
Barley, Egyptian	Pooideae	Hordeum trifurcatum
Barley, fourrowed	Pooideae	Hordeum vulgare
		polystichon
Barley, sixrowed	Pooideae	Hordeum vulgare
		hexastichon
Barley, Tworowed	Pooideae	Hordeum distichon
Cotton, Abroma	Dicotyledoneae	Abroma augusta
Cotton, American	Malvaceae	Gossypium hirsutum
Upland
Cotton, Asiatic	Malvaceae	Gossypium arboreum
Tree, also Indian
Tree
Cotton, Brazilian,	Malvaceae	Gossypium barbadense
also, Kidney, and,		brasiliense
Pernambuco
Cotton, Levant	Malvaceae	Gossypium herbaceum
Cotton, Long Silk,	Malvaceae	Gossypium barbadense
also Long Staple,
Sea Island
Cotton, Mexican,	Malvaceae	Gossypium hirsutum
also Short Staple
Soybean, Soya	Leguminosae	Glycine max
Sugar beet	Chenopodiaceae	Beta vulgaris altissima
Sugar cane	Woody-plant	Arenga pinnata
Tomato	Solanaceae	Lycopersicon esculentum
Tomato, Cherry	Solanaceae	Lycopersicon esculentum
		cerasiforme
Tomato, Common	Solanaceae	Lycopersicon esculentum
		commune
Tomato, Currant	Solanaceae	Lycopersicon
		pimpinellifolium
Tomato, Husk	Solanaceae	Physalis ixocarpa
Tomato, Hyenas	Solanaceae	Solanum incanum
Tomato, Pear	Solanaceae	Lycopersicon esculentum
		pyriforme
Tomato, Tree	Solanaceae	Cyphomandra betacea
Potato	Solanaceae	Solanum tuberosum
Potato, Spanish,	Convolvulaceae	Ipomoea batatas
Sweet potato
Rye, Common	Pooideae	Secale cereale
Rye, Mountain	Pooideae	Secale montanum
Pepper, Bell	Solanaceae	Capsicum annuum grossum
Pepper, Bird, also	Solanaceae	Capsicum annuum minimum
Cayenne, Guinea
Pepper, Bonnet	Solanaceae	Capsicum sinense
Pepper, Bullnose,	Solanaceae	Capsicum annuum grossum
also Sweet
Pepper, Cherry	Solanaceae	Capsicum annuum
		cerasiforme
Pepper, Cluster,	Solanaceae	Capsicum annuum
also Red Cluster		fasciculatum
Pepper, Cone	Solanaceae	Capsicum annuum conoides
Pepper, Goat, also	Solanaceae	Capsicum frutescens
Spur
Pepper, Long	Solanaceae	Capsicum frutescens
		longum
Pepper, Oranamental	Solanaceae	Capsicum annuum
Red, also Wrinkled		abbreviatum
Pepper, Tabasco Red	Solanaceae	Capsicum annuum conoides
Lettuce, Garden	Compositae	Lactuca sativa
Lettuce, Asparagus,	Compositae	Lactuca sativa
also Celery		asparagina
Lettuce, Blue	Compositae	Lactuca perennis
Lettuce, Blue, also	Compositae	Lactuca pulchella
Chicory
Lettuce, Cabbage,	Compositae	Lactuca sativa capitata
also Head
Lettuce, Cos, also	Compositae	Lactuca sativa longifolia
Longleaf, Romaine
Lettuce, Crinkle,	Compositae	Lactuca sativa crispa
also Curled,
Cutting, Leaf
Celery	Umbelliferae	Apium graveolens dulce
Celery, Blanching,	Umbelliferae	Apium graveolens dulce
also Garden
Celery, Root, also	Umbelliferae	Apium graveolens
Turniprooted		rapaceum
Eggplant, Garden	Solanaceae	Solanum melongena
Sorghum	Sorghum	All crop species
Alfalfa	Leguminosae	Medicago sativum
Carrot	Umbelliferae	Daucus carota sativa
Bean, Climbing	Leguminosae	Phaseolus vulgaris
		vulgaris
Bean, Sprouts	Leguminosae	Phaseolus aureus
Bean, Brazilian	Leguminosae	Canavalia ensiformis
Broad
Bean, Broad	Leguminosae	Vicia faba
Bean, Common, also	Leguminosae	Phaseolus vulgaris
French, White,
Kidney
Bean, Egyptian	Leguminosae	Dolichos lablab
Bean, Long, also	Leguminosae	Vigna sesquipedalis
Yardlong
Bean, Winged	Leguminosae	Psophocarpus
		tetragonolobus
Oat, also Common,	Avena	Sativa
Side, Tree
Oat, Black, also	Avena	Strigosa
Bristle, Lopsided
Oat, Bristle	Avena
Pea, also Garden,	Leguminosae	Pisum, sativum sativum
Green, Shelling
Pea, Blackeyed	Leguminosae	Vigna sinensis
Pea, Edible Podded	Leguminosae	Pisum sativum axiphium
Pea, Grey	Leguminosae	Pisum sativum speciosum
Pea, Winged	Leguminosae	Tetragonolobus purpureus
Pea, Wrinkled	Leguminosae	Pisum sativum medullare
Sunflower	Compositae	Helianthus annuus
Squash, Autumn,	Dicotyledoneae	Cucurbita maxima
Winter
Squash, Bush, also	Dicotyledoneae	Cucurbita pepo melopepo
Summer
Squash, Turban	Dicotyledoneae	Cucurbita maxima
		turbaniformis
Cucumber	Dicotyledoneae	Cucumis sativus
Cucumber, African,		Momordica charantia
also Bitter
Cucumber, Squirting,		Ecballium elaterium
also Wild
Cucumber, Wild		Cucumis anguria
Poplar, California	Woody-Plant	Populus trichocarpa
Poplar, European		Populus nigra
Black
Poplar, Gray		Populus canescens
Poplar, Lombardy		Populus italica
Poplar, Silverleaf,		Populus alba
also White
Poplar, Western		Populus trichocarpa
Balsam
Tobacco	Solanaceae	Nicotiana
Arabidopsis	Cruciferae	Arabidopsis thaliana
Thaliana
Turfgrass	Lolium
Turfgrass	Agrostis
	Other families
	of turfgrass
Clover	Leguminosae

Expression of the variant AHAS polypeptides in transgenic plants confers a high level of resistance to herbicides including, but not limited to, imidazolinone herbicides such as, for example, imazethapyr (PURSUIT®), allowing the use of these herbicides during cultivation of the transgenic plants.

Methods for the introduction of foreign genes into plants are known in the art. Non-limiting examples of such methods include Agrobacterium infection, particle bombardment, polyethylene glycol (PEG) treatment of protoplasts, electroporation of protoplasts, microinjection, macroinjection, tiller injection, pollen tube pathway, dry seed imbibition, laser perforation, and electrophoresis. These methods are described in, for example, B. Jenes et al., and S. W. Ritchie et al. In Transgenic Plants, Vol. 1, Engineering and Utilization, ed. S.-D. Kung, R. Wu, Academic Press, Inc., Harcourt Brace Jovanovich 1993; and L. Mannonen et al., Critical Reviews in Biotechnology, 14:287-310, 1994.

In a preferred embodiment, the DNA encoding a variant AHAS is cloned into a DNA vector containing an antibiotic resistance marker gene, and the recombinant AHAS DNA-containing plasmid is introduced into Agrobacterium tumefaciens containing a Ti plasmid. This “binary vector system” is described in, for example, U.S. Pat. No. 4,490,838, and in An et al., Plant Mol. Biol. Manual A3:1-19 (1988). The transformed Agrobacterium is then co-cultivated with leaf disks from the recipient plant to allow infection and transformation of plant cells. Transformed plant cells are then cultivated in regeneration medium, which promotes the formation of shoots, first in the presence of the appropriate antibiotic to select for transformed cells, then in the presence of herbicide. In plant cells successfully transformed with DNA encoding herbicide-resistant AHAS, shoot formation occurs even in the presence of levels of herbicide that inhibit shoot formation from non-transformed cells. After confirming the presence of variant AHAS DNA using, for example, polymerase chain reaction (PCR) analysis, transformed plants are tested for their ability to withstand herbicide spraying and for their capabilities for seed germination and root initiation and proliferation in the presence of herbicide.

Other Applications

The methods and compositions of the present invention can be used in the structure-based rational design of herbicide-resistant ALAS variants, which can be incorporated into plants to confer selective herbicide resistance on the plants. Intermediate variants of ALAS (for example, variants that exhibit sub-optimal specific activity but high resistance and selectivity, or the converse) are useful as templates for the design of second-generation ALAS variants that retain adequate specific activity and high resistance and selectivity.

Herbicide resistant AHAS genes can be transformed into crop species in single or multiple copies to confer herbicide resistance. Genetic engineering of crop species with reduced sensitivity to herbicides can:

(1) Increase the spectrum and flexibility of application of specific effective and environmentally benign herbicides such as imidazolinone herbicides;

(2) Enhance the commercial value of these herbicides;

(3) Reduce weed pressure in crop fields by effective use of herbicides on herbicide resistant crop species and a corresponding increase in harvest yields;

(4) Increase sales of seed for herbicide-resistant plants;

(5) Increase resistance to crop damage from carry-over of herbicides applied in a previous planting;

(6) Decrease susceptibility to changes in herbicide characteristics due to adverse climate conditions; and

(7) Increase tolerance to unevenly or mis-applied herbicides.

For example, transgenic AHAS variant protein containing plants can be cultivated. The crop can be treated with a weed controlling effective amount of the herbicide to which the AHAS variant transgenic plant is resistant, resulting in weed control in the crop without detrimentally affecting the cultivated crop.

The DNA vectors described above that encode herbicide-resistant AHAS variants can be further utilized so that expression of the AHAS variant provides a selectable marker for transformation of cells by the vector. The intended recipient cells may be in culture or in situ, and the AHAS variant genes may be used alone or in combination with other selectable markers. The only requirement is that the recipient cell is sensitive to the cytotoxic effects of the cognate herbicide. This embodiment takes advantage of the relative low cost and lack of toxicity of, for example, imidazolinone-based herbicides, and may be applied in any system that requires DNA-mediated transformation.

EXEMPLIFICATION WITH RESPECT TO PREFERRED EMBODIMENTS

The following examples are intended to illustrate the present invention without limitation.

Example 1

Design of Herbicide-Resistant AHAS Variants

Residues located close to the proposed herbicide binding site of the model described in detail above were selected for mutagenesis in order to design an active AHAS polypeptide with decreased herbicide binding capacity. Each site at the surface of the pocket was considered in terms of potential interactions with other residues in the pocket, as well as with cofactors and herbicides. For example, addition of positively charged residue(s) is expected to interfere with the charge distribution within the binding site, resulting in a loss in affinity of binding of a negatively-charged herbicide.

Three residues were identified as most useful targets for mutagenesis:

(1) F135 was believed to interact with both the isoalloxazine ring of FAD and with the aromatic group of the herbicides. In accordance with the strategy of introducing more charged residues into the binding pocket, this residue was changed to arginine.

(2) M53 contacts helix 498-507. This helix contains known herbicide resistance mutation sites and is also implicated in TPP binding. Furthermore, substitution of glutamic acid at position 53 was believed to favor an interaction with K185, reducing the affinity of K185 for the carboxylate group of imazethapyr.

(3) R128 is located near the entrance to the pocket, where it was believed to be involved in the initial transport of charged herbicides into the binding pocket. This residue was changed to alanine to remove both its charge and its long hydrophobic side chain.

Example 2

Site-Directed Mutagenesis of AHAS to Produce Herbicide-Resistant Variants

The Arabidopsis AHAS gene was inserted in-frame to the 3′ end of the coding region of the glutathione S-transferase gene in the pGEX-2T vector (Pharmacia). Construction of the vector in this manner maintained the six amino acid thrombin recognition sequence at the junction of the expressed glutathione-S-transferase (GST)/AHAS fusion protein. Thrombin digestion of the expressed fusion protein results in an AHAS protein with an N-terminal starting position at the end of the transit peptide at a putative transit peptide processing site, with a residual N-terminal glycine derived from the thrombin recognition site. The final amino terminus of the cleaved AHAS protein consists of Gly-Ser-Ser-Ile-Ser. Site-directed mutations were introduced into the AHAS gene in this vector.

Site-directed mutations were constructed according to the PCR method of Higuchi (Recombinant PCR. In M A Innis, et al. PCR Protocols: A Guide to Methods and Applications, Academic Press, San Diego, pp. 177-183, 1990). Two PCR products, each of which overlap the mutation site, were amplified. The primers in the overlap region contained the mutation. The overlapping PCR amplified fragments were combined, denatured, and allowed to re-anneal together, producing two possible heteroduplex products with recessed 3′-ends. The recessed 3′-ends were extended by Taq DNA polymerase to produce a fragment that was the sum of the two overlapping PCR products containing the desired mutation. A subsequent re-amplification of this fragment with only the two “outside” primers resulted in the enrichment of the full-length product. The product containing the mutation was then re-introduced into the Arabidopsis AHAS gene in the pGEX-2T vector.

Example 3

Expression and Purification of AHAS Variants

A. Methods

E. coli (DH5α) cells transformed with the pGEX-2T vector containing either the maize wild type AHAS gene (vector designation pAC751), the Arabidopsis Ser653Asn mutant, or the Arabidopsis Ile401Phe mutant were grown overnight in LB broth containing 50 μg/mL ampicillin. The overnight culture of E. coli was diluted 1:10 in 1 L LB, 50 μg/mL ampicillin, and 0.1% v/v antifoam A. The culture was incubated at 37° C. with shaking until the OD₆₀₀ reached approximately 0.8. Isopropylthiogalactose (IPTG) was added to a final concentration of 1 mM and the culture was incubated for 3 more hours.

Cells were harvested by centrifugation at 8,670×g for 10 minutes in a JA-10 rotor and resuspended in 1/100th of the original culture volume in MTPBS (16 mM Na₂HPO₄, 4 mM NaPO₄, 150 mM NaCl, pH 7.3). Triton X-100 and lysozyme were added to a final concentration of 1% v/v and 100 μg/mL, respectively. Cells were incubated at 30° C. for 15 minutes cooled to 4° C. on ice, and were lysed by sonication for 10 seconds at level 7 with a Branson Sonifier Cell Disrupter equipped with a microtip probe. The cell free extract was centrifuged at 35,000×g for 10 min. at 4° C. The supernatant was decanted and the centrifugation step was repeated.

Purification of expressed fusion proteins was performed as modified from Smith and Johnson (Gene 67:31-40, 1988). The supernatant was warmed to room temperature and was passed through a 2 mL column of glutathione-agarose beads (sulfur linkage, Sigma) equilibrated in MTPBS. The column was subsequently washed with MTPBS at room temperature until the A₂₀₀ of eluant matched that of MTPBS. The fusion protein was then eluted using a solution containing 5 mM reduced glutathione in 50 mM Tris HCL, pH 8.0. The eluted fusion protein was treated with approximately 30 NIH units of thrombin and dialyzed against 50 mM citrate pH 6.5 and 150 mM NaCl.

The fusion protein was digested overnight at room temperature. Digested samples were dialyzed against MTPBS and passed twice through a glutathione-agarose column equilibrated in MTPBS to remove the released glutathione transferase protein. The protein fraction that did not bind to the column was collected and was concentrated by ultrafiltration on a YM10 filter (Amicon). The concentrated sample was loaded onto a 1.5×95 cm Sephacryl S-100 gel filtration column equilibrated in gel filtration buffer (50 mM HEPES, 150 mM NaCl, pH 7.0). Two mL fractions were collected at a flow rate of 0.14 mL/min. Enzyme stability was tested by storage of the enzyme at 4° C. in gel filtration buffer with the addition of 0.02% sodium azide and in the presence or absence of 2 mM thiamine pyrophosphate and 100 μM flavin adenine dinucleotide (FAD).

B. Results

E. coli transformed with the pAC751 plasmid containing the wide-type AHAS gene fused downstream and in-frame with the GST gene expressed a 91 kD protein when induced with IPTG. The 91 kD protein exhibited the predicted molecular mass of a GST/AHAS fusion protein (the sum of 26 kD and 65 kD, respectively). When the cell free extract of DH5α/pAC751 was passed through a glutathione-agarose affinity gel, washed, and eluted with free glutathione it yielded a preparation enriched in the 91 kD protein (FIG. 6, lane C). The six amino acid thrombin recognition site engineered in the junction of GST and AHAS was successfully cleaved by thrombin (FIG. 6, lane D). The cleaved fusion protein preparation consisted of the expected 26 kD GST protein and the 65 kD maize AWLS protein. Maize ALAS was purified to homogeneity by a second pass through the glutathione-agarose column to affinity subtract GST and subjected to a final Sephacryl S-100 gel filtration step to eliminated thrombin (FIG. 6, lane E). The 65 kD protein is recognized on western blots by a monoclonal antibody raised against a maize AHAS peptide.

Purified wild type maize AHAS was analyzed by electrospray mass spectrometry and was determined to have a molecular mass of 64,996 daltons (data not shown). The predicted mass, as calculated from the deduced amino acid sequence of the gene inserted into the pGEX-2T vector, is 65,058. The 0.096% discrepancy between the empirically determined and predicted mass was within tuning variability of the mass spectrometer. The close proximity of the two mass determinations suggests that there were no misincorporated nucleotides during construction of the expression vector, nor any post-translational modifications to the protein that would cause gross changes in molecular mass. Moreover, the lack of spurious peaks in the preparation of purified enzyme indicated that the sample was free of contamination.

Example 4

Enzymatic Properties of AHAS Variants

The enzymatic properties of wild-type and variant AHAS produced in E. coli were measured by a modification of the method of Singh et al. (Anal. Biochem 171:173-179, 1988) as follows:

A reaction mixture containing 1×AHAS assay buffer (50 mM HEPES pH 7.0, 100 mM pyruvate, 10 mm MgCl₂, 1 mM thiamine pyrophosphate (TPP), and 50 μM flavin adenine dinucleotide (FAD)) was obtained either by dilution of enzyme in 2× assay buffer or by addition of concentrated enzyme to 1×AHAS assay buffer. All assays containing imazethapyr and associated controls contained a final concentration of 5% DMSO due to addition of imazethapyr to assay mixtures as a 50% DMSO solution. Assays were performed in a final volume of 250 μL at 37° C. in microtiter plates. After allowing the reaction to proceed for 60 minutes, acetolactate accumulation was measured colorimetrically as described by Singh et al., Anal. Biochem 171:173-179, 1988.

Maize AHAS expressed and purified from pAC751 as described in Example 3 above is active in the conversion of pyruvate to acetolactate. Full AHAS activity is dependent on the presence of the cofactors FAD and TPP in the assay medium. No activity was detected when only FAD was added to the assay medium. The activity of the purified enzyme with TPP only, or with no cofactors, was less than 1% of the activity detected in the presence of both TPP and FAD. Normally, AHAS present in crude plant extracts is very labile, particularly in the absence of substrate and cofactors. In contrast, the purified AHAS from the bacterial expression system showed no loss in catalytic activity when stored for one month at 4° C. in 50 mM HEPES pH 7.0, 150 mM NaCl, 0.02% NaN₃ in the presence or absence of FAD and TPP. Furthermore, no degradation products were visible from these stored preparations when resolved in SDS-PAGE gels.

The specific activities of wild-type AHAS and the M124E, R199A, and F206R variants are shown in Table 2 below. As determined from the alignment in FIG. 5, the M124E mutation in Arabidopsis AHAS is the equivalent of the maize M53E mutation, the R199A mutation in Arabidopsis is the equivalent of the maize R128A mutation, and the F206R mutation in Arabidopsis is the equivalent of the maize F135R mutation. The mutations designed in the maize AHAS structural model were used to identify the equivalent amino acid in the dicot Arabidopsis AHAS gene and were incorporated and tested in the Arabidopsis AHAS gene. This translation and incorporation of rationally designed herbicide mutations into the dicot Arabidopsis AHAS gene can facilitate evaluation of herbicide resistance in plants of a dicot species.

TABLE 2
SPECIFIC ACTIVITY
		% Catalytic Activity as
	Specific Activity	Compared to Wild Type
	Wild-Type	99.8	100
	Met124Glu	9.15	9.16
	Arg199Ala	86.3	86.5
	Phe206Arg	5.07	5.1

The R199A mutation maintains a high level of catalytic activity (Table 2) while exhibiting a significant level of resistance to imazethapyr (FIG. 7). Notably, this variant retains complete sensitivity to sulfonylureas (FIG. 8). Thus, this variant fulfills the criteria of high specific activity and selective herbicide resistance. By contrast, the M124E substitution resulted in almost complete resistance to imazethapyr (FIG. 7) but also exhibited severely reduced catalytic activity (Table 2). Relative to imidazolinone resistance, this variant exhibits greater sensitivity to sulfonylurea (FIG. 8), suggesting that this residue is a good candidate for creating a mutation that confers selective resistance. Substitution of an amino acid other than glutamic acid may help to maintain catalytic activity. The F206R substitution yielded similar results to those observed with M124E variant, but lacked selectivity in resistance.

Example 5

Iterative Improvement of AHAS Herbicide-Resistant Variant Using a Rational Design Approach

Changing residue 124 in AHAS from Met to Glu as described in Example 4 above conferred imidazolinone resistance but also reduced enzymatic activity to 9.2% of the wild type value. The model of the maize AHAS structure described above suggested that Met53 (equivalent to the Arabidopsis Met124 residue) interacts with a series of hydrophobic residues on the face of an α-helix that is derived from a separate subunit but are in close proximity to Met53. Thus, the hydrophobic interaction between Met53 and the residues on the helix may stabilize both subunit/subunit association and the conformation of the active site. It was believed that the substitution of the hydrophobic Met residue with a charged glutamate residue most probably destabilizes the inter-subunit hydrophobic interaction and results in a loss of catalytic activity.

Based on this structure/function analysis, the activity of the original Arabidopsis Met124Glu (equivalent to maize Met53Glu) mutant enzyme was then iteratively improved by substituting a more hydrophobic amino acid (Ile) at this position. The hydrophobic nature of the Ile side chain resulted in restoration of activity to wild type levels (specific activity of 102, equivalent to 102% of the wild-type activity), but the greater bulk of the Ile side chain was still able to maintain a significant level of imidazolinone resistance (FIG. 9).

By comparison, substitution of a histidine residue at this position resulting in an AHAS variant exhibiting a specific activity of 42.5, equivalent to 42.6% of the wild-type activity. This mutant, nonetheless, exhibited a high degree of resistance to PURSUIT® (FIG. 10).

Example 6

Iterative Improvement of AHAS Herbicide-Resistant Variant Using a Rational Design Approach

Another example of iterative refinement using the methods of the present invention involves the Arg128Ala variant. The structural model of maize AHAS suggested that the Arg128 residue, which resides at the lip of the herbicide binding pocket, contributes to channeling charged substrates and herbicides into the herbicide binding pocket and into the active site. The Arg 128 residue is distant from the TPP moiety, which binds the initial pyruvate molecule in the reaction mechanism of AHAS, explaining why the substitution of Arabidopsis AHAS Arg199 (the equivalent to maize Arg128) to alanine had little effect on the catalytic activity of the enzyme. The structural model further indicated that a more radical change could be made at this position to raise the level of resistance while maintaining high levels of catalytic activity. On this basis, an iterative improvement of the mutation was made to substitute the positively charge arginine residue with a negatively charged glutamate residue. The enzyme thus mutated had improved levels of resistance to PURSUIT® while maintaining high levels of activity (specific activity of 114, equivalent to 114% of the wild-type activity) (FIG. 11).

Example 7

Interchangeability of AHAS Derived From Different Species in Structure-Based Rational Design of Herbicide-Resistant Variants

A structural model of the three-dimensional structure of AHAS is built with a monocot AHAS sequence such as that derived from maize, as described above. To introduce mutations into AHAS derived from a dicot species such as Arabidopsis, the sequences of AHAS derived from the monocot and dicot species are aligned using the GAP and PILEUP programs (Genetics Computer Group, 575 Sequence Drive, Madison, Wis. 53711). Equivalent positions are determined from the computer-generated alignment. The mutations are then introduced into the dicot AHAS gene as described above. Following expression of the mutant AHAS protein in E. coli and assessment of its biochemical properties (i.e., specific activity and resistance to herbicides), the mutant gene is introduced into a dicot plant by plant transformation methods as described above.

Example 8

Production of Herbicide-Resistant Plants by Transformation with Rationally Designed AHAS Gene

DNA Constructs:

Rationally designed AHAS variant genes contained within E. coli expression vectors were used as a source of DNA restriction fragments to replace the equivalent restriction fragment in a Arabidopsis AHAS gene. This gene is present in a 5.5 kb genomic DNA fragment which also contains the Arabidopsis AHAS promoter, the Arabidopsis AHAS termination sequence and 5′- and 3′-flanking DNA. After DNA sequencing through the mutation sites was performed to confirm the presence of the proper mutation, the entire 5.5 kb fragment from each plasmid was inserted into a pBIN based plant transformation vector (Mogen, Leiden, Netherlands). The plant transformation vector also contains the neomycin phosphotransferase II (nptII) kanamycin resistance gene driven by the 35S cauliflower mosaic virus promoter. The final vector construct is displayed in FIG. 12. Vectors containing Arabidopsis AHAS genes with Met124Ile, Met124His, and Arg199Glu mutations (corresponding to Met53Ile, Met53His, and Arg128Glu mutations in the maize AHAS sequence as shown in FIG. 1) were labeled pJK002, pJK003, and pJT004, respectively.

Each of these vectors was transformed into Agrobacterium tumefaciens strain LBA4404 (R&D Life Technologies, Gaithersburg, Md.) using the transformation method described in An et al., Plant Mol. Biol. Manual A3:1-19 (1988).

Plant Transformation:

Leaf disc transformation of Nicotiana tabacum cv. Wisconsin 38 was performed as described by Horsch et al. (Science, 227: 1229-1231, 1985) with slight modifications. Leaf discs were cut from plants grown under sterile conditions and co-cultivated upsidedown in Murashige Skoog media (Sigma Chemical Co., St. Louis, Mo.) for 2-3 days at 25° C. in darkness with Agrobacterium tumefaciens strains containing plasmids pJK002, pJK003, or pJK004. The discs were blotted dry and transferred to regeneration Murashige Skoog medium with B5 vitamins containing 1 mg/L benzyladenine and 0.1 mg/l 1-Napthyl Acetic Acid, 100 mg/L kanamycin, and 500 mg/L cefotaxime (all obtained from Sigma).

Initially, transformants were selected by kanamycin resistance conferred by the nptII gene present in the transformation vector. Shoots derived from the leaf discs were excised and placed on fresh Murashige Skoog hormone free media containing cefotaxime and kanamycin.

In Vivo Herbicide Resistance

Kanamycin-resistant tobacco shoots were transferred to medium containing a 0.25 μM imazethapyr. At this concentration of the imidazolinone herbicide, non-transformed tobacco shoots (containing endogenous wild-type AHAS) were not able to initiate root formation. By contrast, root initiation and growth were observed from tobacco shoots transformed with each of the mutant AHAS genes. Roots developed from shoots transformed with the Met124Ile and Arg199Glu mutant genes along with wild type are shown in FIG. 1. Furthermore, plants transformed with the Met124Ile or Arg199Glu mutant genes were resistant to spraying with twice the field rate (100 g/ha) of imazethapyr (FIG. 13). The patterns of root growth in transformed vs. non-transformed plants in the presence of herbicide, as well as the behavior after herbicide spraying suggest that expression of the rationally designed herbicide resistance genes confers herbicide resistance in vivo.

Detection of the Rationally Designed Genes in Herbicide Resistant Tobacco

Genomic DNA was isolated from the AHAS-transformed tobacco plants, and the presence of the Arabidopsis AHAS variant genes was verified by PCR analysis. Differences between the nucleotide sequences of the Arabidopsis AHAS gene and the two tobacco AHAS genes were exploited to design PCR primers that amplify only the Arabidopsis gene in a tobacco genomic DNA background. The rationally designed herbicide resistance genes were detected, as shown by amplification of a DNA fragment of the proper size, in a majority of the herbicide resistant plants. No PCR signal was seen from non-transformed tobacco plants.

Segregation of Transformed AHAS Genes:

To monitor segregation of rationally designed AHAS genes in transformed plants, germination tests were performed. Seeds were placed in hormone-free Murashige-Skoog medium containing up to 2.5 μM PURSUIT® and 100 μM kanamycin. The seedlings that resulted were visually scored for resistance or susceptibility to the herbicide.

Since tobacco plants are diploid, it would be expected that the progeny of self-pollinated plants should segregate 3:1 resistant:susceptible, reflecting the existence of 1 seedling homozygous for the resistant AHAS gene, 2 seedlings heterozygous for the resistant AHAS gene, and 1 seedling lacking a resistant AHAS gene.

The results indicate that resistant AHAS genes are segregating in the expected 3:1 ratio, supporting the conclusion that herbicide resistance is conferred by a single, dominant copy of a rationally designed AHAS gene.

These results indicate that rational design of herbicide-resistant AHAS genes can be used to produce plants that exhibit herbicide resistant growth in vivo.

Example 9

Production of Plants Cross-Resistant to Different Herbicides by Transformation with Rationally Designed AHAS Genes

Tobacco plants transformed with rationally designed AHAS genes as described in Example 8 above were also tested for cross-resistance to another herbicide, CL 299,263 (also known as imazamox). Germination tests were performed on seeds harvested from the primary transformants containing the Met124Ile, Met124His, and Arg199Glu Arabidopsis AHAS variant genes, in the absence or presence of 2.5 μM CL 299,263 (FIG. 15). This concentration of the herbicide causes severe stunting and bleaching of wild-type tobacco plants. Tobacco plants transformed with the Met124His AHAS gene showed the greatest level of resistance (FIG. 15). Arg199Glu transformants showed an intermediate level of resistance, while Met124Ile showed little resistance (FIG. 15).

All patents, applications, articles, publications, and test methods mentioned above are hereby incorporated by reference.

Many variations of the present invention will suggest themselves to those skilled in the art in light of the above detailed description. Such obvious variations are within the full intended scope of the appended claims.

Claims

1-84. (canceled)

85. A variant plant acetohydroxy acid synthase (AHAS) protein comprising at least one mutation at an amino acid residue corresponding to amino acid residue M53 of SEQ ID NO:1, wherein the mutation is Met53Thr, wherein said variant plant AHAS protein is more resistant to an herbicide than the wild-type AHAS protein.

86. A variant AHAS protein as defined in claim 85, wherein said herbicide is selected from the group consisting of an imidazolinones, sulfonylureas, triazolopyrimidine, sulfomamides, pyrimidyl-oxy-benzoic acids, sulfamoylureas, sulfonylcarboxamides, and combinations thereof.

87. A variant AHAS protein as defined in claim 85, wherein said AHAS protein is derived from Arabidopsis thaliana.

88. A variant AHAS protein as defined in claim 85, wherein said variant AHAS protein has catalytic activity that is more resistant to at least one herbicide than is wild type AHAS.

89. A variant AHAS protein as defined in claim 85, wherein said variant AHAS has more than about 20% of the catalytic activity of wild-type AHAS.

90. A variant AHAS protein as defined in claim 85, wherein said variant AHAS is at least 2-fold more resistant to imidazolinone-based herbicides than to sulfonylurea-based herbicides.