Methods and materials for making and using transgenic dicamba-degrading organisms

FIELD OF THE INVENTION

The present invention relates to transgenic organisms that are able to degrade the herbicide dicamba, including transgenic plants that have been made tolerant to dicamba. The invention also relates to dicamba-degrading enzymes and to DNA molecules and DNA constructs encoding dicamba-degrading enzymes. The invention further relates to a method of controlling weeds in fields of dicamba-tolerant transgenic plants and to a method of removing dicamba from materials contaminated with it (bioremediation). Finally, the invention relates to methods of selecting transformants based on dicamba tolerance or on detecting the fluorescence of 3,6-dichlorosalicylic acid which is generated as a result of dicamba degradation.

BACKGROUND

Herbicides are used routinely in agricultural production. Their effectiveness is often determined by their ability to kill weed growth in crop fields and the tolerance of the cash crop to the herbicide. If the cash crop is not tolerant to the herbicide, the herbicide will either diminish the productivity of the cash crop or kill it altogether. Conversely, if the herbicide is not strong enough, it may allow too much weed growth in the crop field which will, in turn, lessen the productivity of the cash crop. Therefore, it is desirable to produce economically important plants which are tolerant to herbicides. To protect the water and environmental quality of agricultural lands, it is also desirable to develop technologies to degrade herbicides in cases of accidental spills of the herbicide or in cases of unacceptably high levels of soil or water contamination.

Genes encoding enzymes which inactivate herbicides and other xenophobic compounds have previously been isolated from a variety of procaryotic and eucaryotic organisms. In some cases, these genes have been genetically engineered for successful expression in plants. Through this approach, plants have been developed which are tolerant to the herbicides 2,4-dichlorophenoxyacetic acid (Streber and Willmitzer (1989)

Bio/Technology 7:811-816; 2,4-D), bromoxynil (Stalker etal. (1988) Science 242:419-423; tradename Buctril), glyphosate (Comai et al. (1985) Nature 317:741-744; tradename RoundUp) and phosphinothricin (De Block et al. (1987) EMBO J. 6:2513-2518; tradename Basta).

WO 02/068607

PCT/US02/06310

Dicamba (tradename Banvel) is used as a pre-emergent and post-emergent herbicide for the control of annual and perennial broadleaf weeds and several grassy weeds in com, sorghum, small grains, pasture, hay, rangeland, sugarcane, asparagus, turf and grass seed crops. See Crop Protection Reference, pages 1803-1821 (Chemical & Pharmaceutical Press, Inc., New York, NY, 11th ed., 1995). Unfortunately, dicamba can injure many commercial crops (including beans, soybeans, cotton, peas, potatoes, sunflowers, tomatoes, tobacco, and fruit trees), ornamental plants and trees, and other broadleaf plants when it comes into contact with them. Id. Dicamba is chemically stable and can sometimes be persistent in the environment.

Dicamba is in the class of benzoic acid herbicides. It has been suggested that plants tolerant to benzoic acid herbicides, including dicamba, can be produced by incoiporating a 1-aminocyclopropane-l-carboxylic acid (ACC) synthase antisense gene, an ACC oxidase antisense gene, an ACC deaminase gene, or combinations thereof, into the plants. See Canadian Patent Application 2,165,036 (published June 16, 1996). However, no experimental data are presented in this application which demonstrate such tolerance.

Bacteria that metabolize dicamba are known. See U.S. PatentNo. 5,445,962; Krueger et al., J. Agric. Food Chem., 37,534-538 (1989); Cork and Krueger, Adv. Appl. Microbiol., 38,1-66 (1991); Cork and Khalil, Adv. Appl. Microbiol., 40,289-320 (1995). It has been suggested that the specific genes responsible for dicamba metabolism by these bacteria could be isolated and used to produce dicamba-resistant plants and other organisms. See id. and Yang et al., Anal. Biochem., 219:37-42 (1994). However, prior to the present invention, no such genes had been identified or isolated.

SUMMARY OF THE INVENTION

The invention provides an isolated and at least partially purified dicamba-degrading O-demethylase, an isolated and at least partially purified dicamba-degrading oxygenase, an isolated and at least partially purified dicamba-degrading ferredoxin, and an isolated and at least partially purified dicamba-degrading reductase, all as defined and described below.

The invention also provides an isolated DNA molecule comprising a DNA sequence coding for a dicamba-degrading oxygenase, an isolated DNA molecule comprising a DNA sequence coding for a dicamba-degrading ferredoxin, and an isolated DNA molecule

WO 02/068607

PCT/US02/06310 comprising a DNA sequence coding for a dicamba-degrading reductase. The invention further provides a DNA construct comprising a DNA sequence coding for a dicambadegrading oxygenase, a DNA sequence coding for a dicamba-degrading ferredoxin, or a DNA sequence coding for a dicamba-degrading reductase, each DNA coding sequence being operatively linked to expression control sequences. In addition, tbe invention provides a DNA construct comprising a DNA sequence coding for a dicamba-degrading oxygenase, a DNA sequence coding for a dicamba-degrading ferredoxin, and a DNA sequence coding for a dicamba-degrading reductase, each DNA coding sequence being operatively linked to expression control sequences.

The invention further provides any of the above-identified DNA constructs which also comprise a DNA sequence encoding a transitpeptide that targets tiie dicamba-degrading enzyme(s) to organelles of a plant cell or microorganism (chloroplast and/or mitochondria).

The invention further provides a transgenic host cell comprising DNA coding for a dicamba-degrading oxygenase, DNA coding for a dicamba-degrading ferredoxin, or DNA coding for a dicamba-degrading reductase, each DNA being operatively linked to expression control sequences. In addition, the invention provides a transgenic host cell comprising DNA coding for a dicamba-degrading oxygenase and DNA coding for a dicamba-degrading ferredoxin, DNA coding for a dicamba-degrading reductase, or DNA coding for a dicambadegrading ferredoxin and DNA coding for a dicamba-degrading reductase, each DNA being operatively linked to expression control sequences. Tbe DNA can, in some embodiments, also comprise a DNA sequence encoding a transit peptide that targets the dicamba-degrading enzyme(s) to organelles of a plant cell or microorganism (chloroplast and/or mitochondria). The transgenic host cell may be a plant cell or a prokaiyotic or eukaryotic microorganism.

The invention also provides a transgenic plant or plant part comprising one or more cells comprising DNA coding for a dicamba-degrading oxygenase, DNA coding for a dicamba-degrading ferredoxin, or DNA coding for a dicamba-degrading reductase, each DNA being operatively linked to expression control sequences. Tbe invention further provides a transgenic plant or plant part comprising one or more cells comprising DNA coding for a dicamba-degrading oxygenase and DNA coding for a dicamba-degrading ferredoxin, DNA coding for a dicamba-degrading reductase, or DNA coding for a dicambadegrading ferredoxin and DNA coding for a dicamba-degrading reductase, each DNA being

WO 02/068607

PCT/US02/06310 operatively linked to expression control sequences. In some embodiments, the DNA further comprises a DNA sequence encoding a transit peptide that targets the dicamba-degrading enzyme(s) to organelles of a plant cell. The transgenic plant or plant part is preferably tolerant to dicamba or has had its tolerance to dicamba increased as a result of the expression of the dicamba-degrading enzyme(s).

The invention also provides a method of controlling weeds in a field containing transgenic dicamba-tolerant plants. The method comprises applying an amount of dicamba to the field which is effective to control the weeds.

The invention further provides methods of decontaminating a material containing dicamba. In one embodiment, the method comprises applying an effective amount of a transgenic dicamba-degrading microorganism to the material. In another embodiment, the method comprises applying an effective amount of a dicamba-degrading O-demethylase or of a combination of a dicamba-degrading oxygenase, a dicamba-degrading ferredoxin and a dicamba-degrading reductase to the material.

The invention also provides a method of selecting transformed plant cells and transformed plants using dicamba tolerance as the selection marker. In one embodiment, the method comprises transforming at least some of the plant cells in a population of plant cells so that they are tolerant to dicamba and growing the resulting population of plant cells in a culture medium containing dicamba at a concentration selected so that transformed plant cells will grow and untransfoxmed plant cells will not grow. In another embodiment, the method comprising applying dicamba to a population of plants suspected of comprising plants that have been transformed so that they are tolerant to dicamba, tbe dicamba being applied in an amount selected so that transformed plants will grow, and the growth of untransformed plants will be inhibited.

Finally, the invention provides a method of selecting, or screening for, transformed host cells, intact organisms, and parts of organisms. The method comprises providing a population of host cells, intact organisms, or parts of organisms suspected of comprising host cells, intact organisms, or parts of organisms that have been transformed so that they are able to degrade dicamba, contacting the population of host cells, intact organisms, or parts of organisms with dicamba, and ascertaining the presence or level of fluorescence due to 3,6dichlorosalicylic acid. The 3,6-dichlorosalicyclic acid is generated in transformed host cells, intact organisms, or parts of organisms as a result of the degradation of dicamba, but will not be generated in untransformed host cells, intact organisms, or parts of organisms.

Unless the context clearly requires otherwise, throughout the description and the claims, the words comprise “comprising” and the like, are to be construed in an inclusive sense as opposed to an exclusive or exhaustive sense, that is to say, in the sense of “including, but not limited to”.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1. A diagram of the proposed electron transport scheme for dicamba 0demethylase. Electrons from NADH are transferred sequentially from reductase_DIC to ferredoxin_DIC and then to oxygenase_DIC. The reaction of oxygen with the substrate dicamba to form 3,6-dichlorosalicylic acid is catalyzed by oxygenase_DIC. ox, oxidized; red, reduced.

Figure 2. Comparison of the derived amino acid sequence of the ferredoxin component of dicamba O-demethylase to the amino acid sequences of members of the adrenodoxin family of ferredoxins. In Figure 2, ferr di6 = ferredoxin component of dicamba O-demethylase from. Pseudomonas maltophilia DI-6 [SEQ ID NO:5]; fer6 rhoca=ferredoxin from Rhodobacter capsulatus [SEQ ID NO: 10]; fer caucr = ferredoxin from Caulobacter crescentus [SEQ IDNOril]; thcc rhocr = ferredoxin fromRhodococcus erythropolis [SEQ ID NO: 12]; putx psepu = ferredoxin from Pseudomonasputida [SEQ ID NO: 13]; terp psesp = ferredoxin from Pseudomonas sp.[SEQ ID NO:14]. Also in Figure 2, the numbers 1-3 designate the three conserved motifs of the adrenodoxin family of bacterial ferredoxins.

Figure 3. Comparison of the derived amino acid sequence of the two reductase components of dicamba O-demethylase to the amino acid sequences of members of the family of FAD-dependent pyridine nucleotide reductases. In Figure 3, redl di6 = reductase component of dicamba O-demethylase from?, maltophilia DI-6 [SEQ ID NO:7]; red2 di6 =reductase component of dicamba O-demethylase from? maltophilia DI-6 [SEQ ID NO:9]; AJ002606=reductase from Sphingomonas sp. [SEQ ID NO: 15]; thcd rhoer=reductase from R. erythropolis [SEQ IDNO:16]; cama psepu = reductase fromP. putida [SEQIDNO:17]; tera psesp = reductase from Pseudomonas sp.[SEQ ID NO: 18]. Also in Figure 3, the numbers 1-5 designate the five conserved motifs of FAD-dependent pyridine nucleotide reductases.

<img file="NZ528010A_D0001.tif" />

WO 02/068607

PCT/US02/06310

DETAILED DESCRIPTION OF THE PRESENTLY

PREFERRED EMBODIMENTS OF THE INVENTION

Prior studies (Cork and Kreuger, Advan. Appl. Microbiol. 36:1-56 and Yang et al. (1994) Anal. Biochem. 219:37-42) have shown that the soil bacterium, Pseudomonas maltophilia, strain DI-6, is capable of destroying the herbicidal activity of dicamba through a single step reaction in which dicamba (3,6-dichloro-2-methoxybenzoic acid) is converted to 3,6-dichlorosalicylic acid (3,6-DCSA). 3,6-DCSA has no herbicidal activity and has not been shown to have any detrimental effects on plants. In addition, 3,6-DCSA is readily degraded by the normal bacterial flora present in soil.

The experiments described herein confirm the hypothesis of Yang et al. (see id.) that an O-demethylase is responsible for the conversion of dicamba to 3,6-DCSA by P. maltophilia strain DI-6 and establish that the O-demethylase is a three-component enzyme system consisting of a reductase, a ferredoxin, and an oxygenase. See Examples 1 and 3 which describe in detail the isolation, purification and characterization of the P. maltophilia O-demethylase and its three components. The reaction scheme for the reaction catalyzed by the three components of dicamba O-demethylase is presented in Figure 1. As illustrated in Figure 1, electrons from NADH are shuttled through a short electron chain consisting of the reductase and ferredoxin to the terminal oxygenase which catalyzes the oxidation of dicamba to produce 3,6-DCSA.

In a first embodiment, the invention provides isolated and at least partially purified dicamba-degrading enzymes. Isolated is used herein to mean that the enzymes have at least been removed from the cells in which they are produced (i.e., they are contained in a cell lysate). At least partially purified is used herein to mean that they have been separated at least partially from the other components of the cell lysate. Preferably, the enzymes have been purified sufficiently so that the enzyme preparations are at least about 70% homogenous.

In particular, the invention provides an isolated and at least partially purified dicamba-degrading O-demethylase. Dicamba-degrading O-demethylase is defined herein to mean a combination of a dicamba-degrading oxygenase, a dicamba-degrading ferredoxin and a dicamba-degrading reductase, all as defined below.

WO 02/068607

PCT/US02/06310

The invention also provides an isolated and at least partially purified dicambadegrading oxygenase. Dicamba-degrading oxygenase is defined herein to mean the oxygenase purified from P. maltophilia strain DI-6 and oxygenases which have an amino acid sequence which is at least about 65% homologous (and preferably identical), and more preferably at least about 75% homologous (and preferably identical), and more preferably at least about 85% homologous (and preferably identical), and even more preferably at least about 95% homologous (and preferably identical), to that of the P. maltophilia oxygenase and which can participate in the degradation of dicamba. Dicamba-degrading oxygenases include mutant oxygenases having the amino acid sequence of the P. maltophilia oxygenase wherein one or more amino acids have been added to, deleted from, or substituted for, the amino acids of the P. maltophilia oxygenase sequence. Such oxygenases include enzymatically active fragments of the oxygenase, as well as oxygenases having a given percent homology or identity to the P. maltophilia oxygenase sequence. Activity of dicambadegrading oxygenases can be determined as described in Examples 1 and 3.

The invention further provides an isolated and at least partially purified dicambadegrading ferredoxin. Dicamba-degrading ferredoxin is defined herein to mean the ferredoxin purified from P. maltophilia strain DI-6 and feiredoxins which have an amino acid sequence which is at least about 65% homologous (and preferably identical), and more preferably at least about 75% homologous (and preferably identical), and more preferably at least about 85% homologous (and preferably identical), and even more preferably at least about 95% homologous (and preferably identical), to that of the P. maltophilia ferredoxin and which can participate in the degradation of dicamba. Dicamba-degrading ferredoxins include mutant ferredoxins having the amino acid sequence of the P. maltophilia ferredoxin wherein one or more amino acids have been added to, deleted from, or substituted for, the amino acids of the P. maltophilia ferredoxin sequence. Such ferredoxins include enzymatically active fragments of the ferredoxin, as well as ferredoxins having a given percent homology or identity to the P. maltophilia ferredoxin sequence. Activity of dicambadegrading ferredoxins can be determined as described in Examples 1 and 3.

Finally, the invention provides an isolated and at least partially purified dicambadegrading reductase. Dicamba-degrading reductase is defined herein to mean the reductases purified from P. maltophilia strain DI-6 and reductases which have an amino acid

WO 02/068607

PCT/US02/06310 sequence which is at least about 65% homologous (and preferably identical), and more preferably at least about 75% homologous (andpreferably identical), and more preferably at least about 85% homologous (and preferably identical), and even more preferably at least about 95% homologous (and preferably identical), to that of one of the P. maltophilia reductases and which can participate in the degradation of dicamba. Dicamba-degrading reductases include mutant reductases having tbe amino acid sequence of one of the P. maltophilia reductases wherein one or more amino acids have been added to, deleted from, or substituted for, the amino acids of the P. maltophilia reductase sequence. Such reductases include enzymatically active fragments of the reductase, as well as reductases having a given percent homology or identity to the P. maltophilia reductase sequence. Activity of dicambadegrading reductases can be determined as described in Examples 1 and 3.

Methods of determining the degree of homology of amino acid sequences are well known in the art. For instance, the FASTA program of the Genetics Computing Group (GCG) software package (University of Wisconsin, Madison,Wl) can be used to compare sequences in various protein databases such as the Swiss Protein Database.

The dicamba-degrading enzymes of the invention can be isolated and purified as described in Examples 1 and 3 from?, maltophilia or other organisms that degrade dicamba. Suitable other organisms include bacteria other than P. maltophilia strain DI-6 that degrade dicamba. Several strains of such bacteria are known and include many soil bacteria as well as fresh water bacteria. See, e.g., U.S. Patent No. 5,445,962; Krueger et al., J. Agric. Food Chem., 37,534-538 (1989); Cork and Krueger, Adv. Appl. Microbiol., 38,1-66 (1991); Cork and Khalil, Adv. Appl. Microbiol., 40,289-320 (1995). Dicamba-degrading bacteria (i.e., bacteria that use dicamba as a sole carbon source) are found in many genera of bacteria including, but not limited to, Pseudomonas, Sphingomonas, Aeromonas, Xanthomonas, Alcaligenes, and Moraxella. Other dicamba-degrading bacterial strains can be isolated as were these strains by methods well known in the art.

Preferably, however, the dicamba-degrading enzymes of the invention are prepared using recombinant DNA techniques (see below). In particular, mutant enzymes having the amino acid sequence of the P. maltophilia enzyme wherein one or more amino acids have been added to, deleted from, or substituted for, the amino acids of the P. maltophilia sequence are prepared in this manner using, for example, oligonucleotide-directed

WO 02/068607

PCT/US02/06310 mutagenesis, linker-scanning mutagenesis, mutagenesis using the polymerase chain reaction, and the like. See Ausubel et al. (eds.), Current Protocols In Molecular Biology (Wiley Interscience 1990) and McPherson (ed.), Directed Mutagenesis: A Practical Approach (IRL Press 1991).

In a second embodiment, the invention provides isolated DNA molecules coding for dicamba-degrading enzymes of the invention. Isolated is used herein to mean that the DNA molecule has been removed from its natural environment or is not a naturally-occurring DNA molecule. Methods of preparing these DNA molecules are well known in the art See, e.g., Maniatis et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY (1982), Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY (1989).

For instance, the DNA molecules of the invention maybe isolated cDNA or genomic clones, and may also include RNA molecules. The identification and isolation of clones coding for the dicamba-degrading enzymes of P. maltophilia strain DI-6 are described in Examples 2 and 4-5. Additional clones coding for dicamba-degrading enzymes can be obtained in a similar manner. The isolated clones, or portions of them, can be used as probes to identify and isolate additional clones from organisms other than the ones from which the clones were originally isolated. Suitable organisms include bacteria that degrade dicamba. As noted above, in addition to P. maltophilia strain DI-6, several strains of bacteria are known that degrade dicamba. See U.S. Patent No. 5,445,962; Krueger et al., J. Agric. Food Chem., 37,534-538 (1989); Cork and Krueger, Adv. Appl. Microbiol., 38,1-66 (1991); Cork and Khalil, Adv. Appl. Microbiol., 40,289-320 (1995).

The DNA molecules of the invention can also be chemically synthesized using the sequences of isolated clones. Such techniques are well known in the art. For instance, DNA sequences may be synthesized by phosphoamidite chemistry in an automated DNA synthesizer. Chemical synthesis has a number of advantages. In particular, chemical synthesis is desirable because codons preferred by the host in which the DNA sequence will be expressed may be used to optimize expression. Not all of the codons need to be altered to obtain improved expression, but preferably at least the codons rarely used in the host are changed to host-preferred codons. High levels of expression can be obtained by changing greater than about 50%, most preferably at least about 80%, of the codons to host-preferred

WO 02/068607

PCT/US02/06310 codons. The codon preferences of many host cells are known. See, e.g., Maximizing Gene Expression, pages 225-85 (Reznikoff &amp; Gold, eds., 1986), PCT WO 97/31115, PCT WO 97/11086, EP 646643, EP 553494, and U.S. Patents Nos. 5,689,052,5,567,862,5,567,600, 5,552,299 and 5,017,692. The codon preferences of other host cells can be deduced by methods known in the art. Also, using chemical synthesis, the sequence of the DNA molecule or its encoded protein can be readily changed to, e.g., optimize expression (e.g., eliminate mRNA secondary structures that interfere with transcription or translation), add unique restriction sites at convenient points, delete protease cleavage sites, etc.

In a third embodiment, the present invention provides DNA constructs comprising DNA coding for a dicamba-degrading enzyme operatively linked to expression control sequences or a plurality of DNA coding sequences, each coding for a dicamba-degrading enzyme and each being operatively linked to expression control sequences. DNA constructs are defined herein to he constructed (non-naturally occurring) DNA molecules useful for introducing DNA into host cells, and the term includes chimeric genes, expression cassettes, and vectors.

As used herein operatively linked refers to the linking of DNA sequences (including the order of the sequences, the orientation of the sequences, and the relative spacing of the various sequences) in such a manner that die encoded proteins are expressed. Methods of operatively linking expression control sequences to coding sequences are well known in the art. See, e.g., Maniatis et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY (1982), Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY (1989).

Expression control sequences are DNA sequences involved in any way in the control of transcription or translation in prokaryotes and eukaryotes. Suitable expression control sequences and methods of making and using them are well known in the art.

The expression control sequences must include a promoter. The promoter may be any DNA sequence which shows transcriptional activity in the chosen host cell or organism. The promoter may be inducible or constitutive. It may be naturally-occurring, may be composed of portions of various naturally-occuring promoters, or may be partially or totally synthetic. Guidance for the design of promoters is provided by studies of promoter structure, such as that of Harley and Reynolds, Nucleic Acids Res., 15,2343-61 (1987). Also, the location of

WO 02/068607

PCT/US02/06310 the promoter relative to the transcription start may be optimized. See, e.g., Roberts, et al., Proc. Natl Acad. Sci. USA, 76,760-4(1979). Many suitable promoters for use in prokaryotes and eukaryotes are well known in the art.

For instance, suitable constitutive promoters for use in plants include: the promoters from plant viruses, such as the 35S promoter from cauliflower mosaic virus (Odell et al., Nature 313:810-812 (1985), the full length transcript promoter with duplicated enhancer domains from peanut chlorotic streak caulimovirus (Maiti and Shepherd, BBRC 244:440-444 (1998)), promoters of Chlorella virus methyltransferase genes (U.S. Patent No. 5,563,328), and the foil-length transcript promoter from figwort mosaic virus (U.S. Patent No. 5,378,619); the promoters from such genes as rice actin (McElroy et al., Plant Cell 2:163-171 (1990)), ubiquitin (Christensen et al., Plant Mol. Biol. 12:619-632 (1989) and Christensen et al., Plant Mol. Biol. 18:675-689 (1992)), pEMU (Last et al., Theor. Appl. Genet. 81:581588 (1991)), MAS (Velten et al., EMBO J. 3:2723-2730 (1984)), maize H3 histone (Lepetit et al., Mol. Gen. Genet. 231:276-285 (1992) and Atanassova et al., Plant Journal 2(3):291300 (1992)), Brassica napus ALS3 (PCT application WO 97/41228); and promoters of various Agrobacterium genes (see U.S. Patents Nos. 4,771,002, 5,102,796, 5,182,200, 5,428,147).

Suitable inducible promoters for use in plants include: the promoter from the ACE1 system which responds to copper (Mett et al. PNAS 90:4567-4571 (1993)); the promoter of foe maize In2 gene which responds to benzenesulfonamide herbicide safeners (Hershey et al., Mol. Gen. Genetics 227:229-237 (1991) and Gatz et al., Mol. Gen. Genetics 243:32-38 (1994)), and foe promoter of foe Tet repressor from TnlO (Gatz et al., Mol. Gen. Genet. 227:229-237 (1991). A particularly preferred inducible promoter for use in plants is one that responds to an inducing agent to which plants do not normally respond. An exemplary inducible promoter of this type is the inducible promoter from a steroid hormone gene, foe transcriptional activity of which is induced by a ghicocorticosteroid hormone. Schena et al., Proc. Natl. Acad. Sci. USA 88:10421 (1991). Other inducible promoters for use in plants are described in EP 332104, PCT WO 93/21334 and PCT WO 97/06269.

Suitable promoters for use in bacteria include foe promoter of foe Bacillus stearothermophilus maltogenic amylase gene, foe Bacillus licheniformis alpha-amylase gene, foe Bacillus amyloliquefaciens BAN amylase gene, the Bacillus subtilis alkaline protease

WO 02/068607

PCT/US02/06310 gene, the Bacillus pumilus xylosidase gene, the phage lambda P_R and P_L promoters, and the Escherichia coli lac, trp and tac promoters. See PCT WO 96/23898 and PCT WO 97/42320.

Suitable promoters for use in yeast host cells include promoters from yeast glycolytic genes, promoters from alcohol dehydrogenase genes, the TPI1 promoter, and the ADH2-4c promoter. See PCT WO 96/23898.

Suitable promoters for use in filamentous fungi include the ADH3 promoter, the tpiA promoter, the promoters of the genes encoding Aspergillus oryzae TAKA amylase, Rhizomucor miehei aspartic proteinase, Aspergillus niger neutral alpha-amylase, A. niger acid stable alpha-amylase, A. niger ot Aspergillus awamori glucoamylase, R. miehei lipase, A. oryzae alkaline protease, A. oryzae triose phosphate isomerase, and Aspergillus nidulans acetamidase. See PCT WO 96/23898.

Suitable promoters for use in mammalian cells are the SV40 promoter, metallothionein gene promoter, murine mammary tumor virus promoter, Rous sarcoma virus promoter, and adenovirus 2 major late promoter. See PCT WO 96/23898 and PCT WO 97/42320.

Suitable promoters for use in insect cells include the polyhedrin promoter, P10 promoter, the Autographa califomica polyhedrosis virus basic protein promoter, the baculovirus immediate early gene 1 promoter and the baculovirus 39K delayed-early gene promoter. See PCT WO 96/23898.

Finally, promoters composed of portions of other promoters and partially or totally synthetic promoters can be used. See, e.g., Ni et al., Plant J., 7:661-676 (1995)andPCT WO 95/14098 describing such promoters for use in plants.

The promoter may include, or be modified to include, one or more enhancer elements. Preferably, the promoter will include a plurality of enhancer elements. Promoters containing enhancer elements provide for higher levels of transcription as compared to promoters which do not include them. Suitable enhancer elements for use in plants include the 35S enhancer element from cauliflower mosaic virus (U.S. Patents Nos. 5,106,739 and 5,164,316) andthe enhancer element from figwort mosaic virus (Maiti et al., Transgenic Res., 6, 143-156 (1997)). Other suitable enhancers for use in other cells are known. See PCT WO 96/23898 and Enhancers And Eukaryotic Expression (Cold Spring Harbor Press, Cold Spring Harbor, NY, 1983).

WO 02/068607

PCT/US02/06310

For efficient expression, the coding sequences are preferably also operatively linked to a 3' untranslated sequence. The 3' untranslated sequence contains transcription and/or translation termination sequences. The 3' untranslated regions can be obtained from the flanking regions of genes from bacterial, plant or other eukaryotic cells. For use in prokaryotes, the 3' untranslated region will include a transcription termination sequence. For use in plants and other eukaryotes, the 3' untranslated region will include a transcription termination sequence and a poiyadenylation sequence. Suitable 3' untranslated sequences for use in plants include those of the cauliflower mosaic virus 35S gene, the phaseolin seed storage protein gene, the pea ribulose biphosphate carboxylase small subunit E9 gene, the soybean 7S storage protein genes, the octopine synthase gene, and the nopaline synthase gene.

In plants and other eukaryotes, a 5' untranslated sequence is also employed. The 5' untranslated sequence is the portion of an mRNA which extends from the 5' CAP site to the translation initiation codon. This region of the mRNA is necessary for translation initiation in eukaryotes and plays a role in the regulation of gene expression. Suitable 5' untranslated regions for use in plants include those of alfalfa mosaic virus, cucumber mosaic virus coat protein gene, and tobacco mosaic virus.

As noted above, the DNA construct may be a vector. The vector may contain one or more replication systems which allow it to replicate in host cells. Self-replicating vectors include plasmids, cosmids and viral vectors. Alternatively, the vector may be an integrating vector which allows the integration into the host cell's chromosome of the sequences coding for foe dicamba-degrading enzymes of foe invention. The vector desirably also has unique restriction sites for foe insertion of DNA sequences. If a vector does not have unique restriction sites, it may be modified to introduce or eliminate restriction sites to make it more suitable for further manipulations.

The DNA constructs of foe invention can be used to transform a variety of host cells (see below). A genetic marker must be used for selecting transformed host cells (a selection marker). Selection markers typically allow transformed cells to be recovered by negative selection (/. e., inhibiting growth of cells that do not contain foe selection marker) or by screening for a product encoded by foe selection marker.

WO 02/068607

PCT/US02/06310

The most commonly used selectable marker gene for plant transformation is the neomycin phosphotransferase Π (nptH) gene, isolated from Tn5, which, when placed under the control of plant regulatory signals, confers resistance to kanamycin. Fraley et al., Proc. Natl. Acad. Sci. USA, 80:4803 (1983). Another commonly used selectable marker gene is the hygromycin phosphotransferase gene which confers resistance to the antibiotic hygromycin. Vanden Elzen et al., Plant Mol. Biol., 5:299 (1985).

Additional selectable marker genes of bacterial origin that confer resistance to antibiotics include gentamycin acetyl transferase, streptomycin phosphotransferase, aminoglycoside-3'-adenyl transferase, and the bleomycin resistance determinant. Hayford et al., Plant Physiol. 86:1216 (1988), Jones et al., Mol. Gen. Genet. 210:86 (1987), Svab et al., Plant Mol. Biol. 14:197 (1990), Hille et al., Plant Mol. Biol. 7:171 (1986). Other selectable marker genes confer resistance to herbicides such as glyphosate, glufosinate or bromoxynil. Comai et al., Nature 317:741-744 (1985), Gordon-Kamm et al., Plant Cell 2:603-618 (1990) and Stalker et al., Science 242:419-423 (1988).

Other selectable marker genes for plant transformation are not of bacterial origin. These genes include, for example, mouse dihydrofolate reductase, plant 5eno/pyruvylshikimate-3-phosphate synthase and plant acetolactate synthase. Eichholtz et al., Somatic Cell Mol. Genet. 13:67 (1987), Shah et al., Science 233:478 (1986), Charest et al., Plant Cell Rep. 8:643 (1990).

Commonly used genes for screening presumptively transformed cells include βglucuronidase (GUS), β-galactosidase, luciferase, and chloramphenicol acetyltransferase. Jefferson, R.A., Plant Mol. Biol. Rep. 5:387 (1987)., Teeri et al., EMBOJ. 8:343 (1989), Koncz et al., Proc. Natl. Acad. Sci. USA 84:131 (1987), De Block et al., EMBOJ. 3:1681 (1984), green fluorescent protein (GFP) (Chalfie et al., Science 263:802 (1994), Haseloff et al., TIG 11:328-329 (1995) and PCT application WO 97/41228). Another approach to the identification of relatively rare transformation events has been use of a gene that encodes a dominant constitutive regulator of the Zea mays anthocyanin pigmentation pathway. Ludwig et al., Science 247:449 (1990).

Suitable selection markers for use in prokaryotes and eukaryotes other than plants are also well known. See, e.g., PCT WO 96/23898 and PCT WO 97/42320. For instance,

WO 02/068607

PCT/US02/06310 resistance to antibiotics (ampicillin, kanamycin, tetracyiine, chloramphenicol, neomycin or hygromycin) may be used as the selection marker.

According to another aspect of the present invention, tolerance to dicamba can be used as a selection marker for plants and plant cells. Tolerance means that transformed plant cells are able to grow (survive and regenerate into plants) when placed in culture medium containing a level of dicamba that prevents untransformed cells from doing so. Tolerance also means that transformed plants are able to grow after application of an amount of dicamba that inhibits the growth of untransformed plants.

Methods of selecting transformed plant cells are well known in the art. Briefly, at least some of the plant cells in a population of plant cells (e.g., an explant or an embryonic suspension culture) are transformed with a DNA construct or combination of DNA constructs providing for dicamba degradation. The resulting population of plant cells is placed in culture medium containing dicamba at a concentration selected so that transformed plant cells will grow, whereas untransformed plant cells will not. Suitable concentrations of dicamba can be determined empirically as is known in the art.

Methods of selecting transformed plants are also known in the art. Briefly, dicamba is applied to a population of plants suspected of comprising a DNA construct or a combination of DNA constructs providing for dicamba degradation. The amount of dicamba is selected so that transformed plants will grow, and the growth of untransformed plants will be inhibited. The level of inhibition must be sufficient so that transformed and untransformed plants can be readily distinguished (z.e., inhibition must be statistically significant). Such amounts can be determined empirically as is known in the art.

Further, the generation of 3,6-DCSA as a result of the degradation of dicamba can be used for selection and screening. The generation of 3,6-DCSA can be readily ascertained by observing the fluorescence of this compound, allowing selection and screening of transformed host cells, intact organisms, and parts of organisms (e.g., microorganisms, plants, plant parts, and plant cells). In this regard, the invention allows for selection and screening of transformed host cells, intact organisms, and parts of organisms in the same manner as for green fluorescent protein (GFP). See U.S. Patents Nos. 5,162,227 and 5,491,084 and PCT applications WO 96/27675, WO 97/11094, WO 97/41228 and WO 97/42320, all of which are incorporated herein by reference. In particular, 3,6-DCSA can be

WO 02/068607

PCT/US02/06310 detected in transformed host cells, intact organisms, and parts of organisms using conventional spectrophotometric methods. For instance, microscopes can be fitted with appropriate fitter combinations for fluorescence excitation and detection. A hand-held lamp may be used for bencbtop work or field work (see Example 1). Fluorescence-activated cell sorting can also be employed. 3,6-DCSA is excited at a wavelength of 312-313 nm, with a maximum emission wavelength of 424 nm.

Parts of organisms include organs, tissues, or any other part. Plant parts include seeds, pollen, embiyos, flowers, fruits, shoots, leaves, roots, stems, explants, etc.

Selection based on dicamba tolerance or dicamba degradation can be used in die production of dicamba-tolerant plants or dicamba-degrading microorganisms, in which case the use of another selection marker may not be necessary. Selection based on dicamba tolerance or dicamba degradation can also be used in tbe production of transgenic cells or organisms that express other genes of interest Many such genes are known and include genes coding for proteins of commercial value and genes that confer improved agronomic traits on plants (see, e.g., PCT WO 97/41228, the complete disclosure of which is incorporated herein by reference).

The DNA constructs of the invention can be used to transform a variety of host cells, including prokaryotes and eukaryotes. The DNA sequences coding for tbe dicambadegrading enzyme(s) and the selection marker, if a separate selection marker is used, maybe on the same or different DNA constructs. Preferably, they are arranged on a single DNA construct as a transcription unit so that all of the coding sequences are expressed together. Also, the gene(s) of interest and the DNA sequence(s) coding for the dicamba-degrading enzyme(s), when dicamba-tolerance or dicamba degradation is being used as a selection marker, may be on the same or different DNA constructs. Such constructs are prepared in tbe same manner as described above.

Suitable host cells include prokaryotic and eukaryotic microorganisms (e.g., bacteria (including Agrobacterium tumefaciens and Escherichia coli), yeast (including Saccharomyces cerevisiae) and other fungi (including Aspergillus sp.), plant cells, insect cells, and mammalian cells. Preferably, the host cell is one that does not normally degrade dicamba. However, the present invention can also be used to increase the level of dicamba degradation in host cells that normally degrade dicamba.

WO 02/068607

PCT/US02/06310

Thus, the transgenic cells and organisms of the invention include cells and organisms that do not normally degrade dicamba, but which have been transformed according to the invention so that they are able to degrade this herbicide. The transgenic cells and organisms of the invention also include cells and organisms that normally degrade dicamba, but which have been transformed according to the invention so that they are able to degrade more of this herbicide or to degrade the herbicide more efficiently.

Methods of transforming prokaryotic and eukaryotic host cells are well known in the art. See, e.g., Maniatis et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY (1982), Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor, NY (1989); PCT WO 96/23898 and PCT WO 97/42320.

For instance, numerous methods for plant transformation have been developed, including biological and physical transformation protocols. See, for example, Miki et al., Procedures for Introducing Foreign DNA into Plants in Methods in Plant Molecular Biology and Biotechnology, Glick, B.R. and Thompson, J.E. Eds. (CRC Press, Inc., Boca Raton, 1993) pp. 67-88. In addition, vectors and in vitro culture methods for plant cell or tissue transformation and regeneration of plants are available. See, for example, Gruber et al., Vectors for Plant Transformation in Methods in Plant Molecular Biology and Biotechnology, Glick, B.R. and Thompson, J.E. Eds. (CRC Press, Inc., Boca Raton, 1993) pp. 89-119. Example 6 provides a specific example ofthe transformation of plants according to the present invention.

The most widely utilized method for introducing an expression vector into plants is based on the natural transformation system of Agrobacterium. See, for example, Horsch et al., Science 227:1229 (1985). A. tumefaciens and A rhizogenes are plant pathogenic soil bacteria which genetically transform plant cells. The Ti and Ri plasmids of A. tumefaciens and A. rhizogenes, respectively, cany genes responsible for genetic transformation of the plant. See, for example, Kado, C.I., Crit. Rev. Plant. Sci. 10:1 (1991). Descriptions of Agrobacterium vector systems and methods for Agrobacterium-mediated gene transfer are provided by numerous references, including Gruber et al., supra, Miki et al., supra, Moloney et al., Plant Cell Reports 8:238 (1989), and U.S. Patents Nos. 4,940,838 and 5,464,763.

A generally applicable method of plant transformation is microprojectile-mediated transformation wherein DNA is carried on the surface of microprojectiles. The expression

WO 02/068607

PCT/US02/06310 vector is introduced into plant tissues with a biolistic device that accelerates the microprojectiles to speeds sufficient to penetrate plant cell walls and membranes. Sanford etal., Part. Sci. Technol. 5:27 (1987), Sanford, J.C., Trends Biotech. 6:299 (1988), Sanford,

J.C., Physiol. Plant 79:206 (1990), Klein et al., Biotechnology 10:268 (1992).

Another method for physical delivery of DNA to plants is sonication of target cells.

Zhang et al., Bio/Technology 9:996 (1991). Alternatively, liposome or spheroplast fusion have been used to introduce expression vectors into plants. Deshayes et al., EMBO J., 4:2131 (1985), Christou et al., Proc Natl. Acad. Sci. USA 84:3962 (1987). Direct uptake of DNA into protoplasts using CaCl₂ precipitation, polyvinyl alcohol or poly-L-omithine have also been reported. Hain et al., Mol. Gen. Genet. 199:161 (1985) and Draper et al., Plant Cell Physiol. 23:451 (1982). Electroporation of protoplasts and whole cells and tissues have also been described. Donn et al., In Abstracts of Vllth International Congress on Plant Cell and Tissue Culture IAPTC, A2-38, p. 53 (1990); D’Halluin et al., Plant Cell 4:1495-1505 (1992) and Spencer et al., Plant Mol. Biol. 24:51-61 (1994).

Transgenic dicamba-tolerant plants of any type may be produced according to the invention. In particular, broadleaf plants (including beans, soybeans, cotton, peas, potatoes, sunflowers, tomatoes, tobacco, fruit trees, and ornamental plants and trees) that are currently known to be injured by dicamba can be transformed so that they become tolerant to the herbicide. Other plants (such as com, sorghum, small grains, sugarcane, asparagus, and grass) which are currently considered tolerant to dicamba can be transformed to increase their tolerance to the herbicide. Specifically, any dicotyledonous or monocotyledonous plant can be transformed with the DNA constructs of the present invention. Tolerant means that foe transformed plants can grow in foe presence of an amount of dicamba which inhibits foe growth of untransformed plants

It is anticipated that foe dicamba-degrading oxygenases of foe invention can function with endogenous reductases and ferredoxins found in transgenic host cells and organisms to degrade dicamba. Plant chloroplasts are particularly rich in reductases and ferredoxins. Accordingly, a preferred embodiment for foe production of transgenic dicamba-tolerant plants is the use of a sequence coding for a peptide (e.g., a transit peptide) that will direct foe dicamba-degrading oxygenase into chloroplasts (a chloroplast targeting sequence). DNA coding for the chloroplast targeting sequence is preferably placed upstream (5') of the

WO 02/068607

PCT/US02/06310 sequence coding for the dicamba-degrading oxygenase, but may also be placed downstream (3') of the coding sequence, or both upstream and downstream of the coding sequence. Any suitable chloroplast targeting sequence can be used. Exemplary chloroplast targeting sequences include the maize cab-m7 signal sequence (see Becker et al., Plant Mol. Biol. 20:49 (1992) and PCT WO 97/41228), the pea glutathione reductase signal sequence (Creissen et al., Plant J. 2:129 (1991) and PCT WO 97/41228) and the pea ribulose 1,5bisphosphate carboxylase small subunit signal sequence (nucleic acid sequence represented by SEQ ID NO:19) (Fluhr etal., EMBO J. 5,2063-2071,(1986)). An alternative preferred embodiment is the direct transformation of chloroplasts using a construct comprising a promoter functional in chloroplasts to obtain expression of the oxygenase in chloroplasts. See, e.g., PCT application WO 95/24492 and U.S. Patent No. 5,545,818. Of course, if a selected transgenic host cell or organism does not produce sufficient endogenous reductase, ferredoxin, or both, the host cell or organism can be transformed so that it produces one or both of these enzymes as well as the oxygenase.

In yet another embodiment, the invention provides a method of controlling weeds in a field where transgenic dicamba-tolerant plants are growing. The method comprises applying an effective amount of dicamba to the field to control the weeds. Methods of applying dicamba and amounts of dicamba effective to control various types of weeds are known. See Crop Protection Reference, pages 1803-1821 (Chemical &amp; Pharmaceutical Press, Inc., New York, NY, 11th ed., 1995).

In another embodiment, the invention provides a method of degrading dicamba present in a material, such as soil, water, or waste products of a dicamba manufacturing facility. Such degradation can be accomplished using the dicamba-degrading enzymes of the invention. The enzymes can be purified from microorganisms naturally expressing them (see Examples 1 and 3) or can be purified from transgenic host cells producing them. If the enzymes are used in such methods, then appropriate cofactors must also be provided (see Example 1). Effective amounts can be determined empirically as is known in the art (see Example 1). Alternatively, transgenic prokaryotic and eukaryotic microorganisms can he used to degrade dicamba in such materials. Transgenic prokaryotic and eukaryotic microorganisms can be produced as described above, and effective amounts can be determined empirically as is known in the art.

WO 02/068607

PCT/US02/06310

<img file="NZ528010A_D0002.tif" />

Dicamba is introduced into the environment in large quantities on a continuing basis. The elimination of dicamba is dependent in large part on the action of enzyme systems which are found in microorganisms inhabiting the soil and water of the planet. An understanding of these enzyme systems, including dicamba-degrading O-demethylases and their three components, is important in efforts to exploit natural and genetically modified microbes for bioremediation and the restoration of contaminated soil, water and other materials. Thus, the dicamba-degrading enzymes, DNA molecules, DNA constructs, etc., ofthe invention can be used as research tools for the study of dicamba degradation and bioremediation.

Finally, the dicamba-degrading enzymes of the invention can be used in an assay for dicamba. A sample suspected of containing dicamba is mixed with a dicamba-degrading 0demethylase or a combination of a dicamba-degrading oxygenase, a dicamba-degrading ferredoxin and a dicamba-degrading reductase. Suitable assays are described in Examples 1 and 3. In particular, detecting or quantitating the fluorescence due to the generation of 3,6DCSA makes for a convenient assay.

WO 02/068607

PCT/US02/06310

EXAMPLES

EXAMPLE 1: Purification And Characterization Of The Components

Of Dicamba O-Demethvlase Of Pseudomonas maltophilia DI-6

METHODS AND MATERIALS:

Bacterium and growth conditions. Pseudomonas maltophilia, strain DI-6 (Kreuger, et al., (1989) J. Agric. Food Chem., 37:534-538) was isolated from a soil site persistently contaminated with dicamba. The bacterium was provided by Dr. Douglas Cork ofthe Illinois Institute of Technology (Chicago, EL), and was maintained on reduced chloride medium (Kreuger, J.P., (1989) Ph.D. thesis, Illinois Institute of Technology, Chicago, IL), supplemented with either dicamba (2 mg/ml) or a mixture of glucose (2 mg/ml) and Casamino Acids (2 mg/ml). The carbon sources were filter-sterilized and added to the medium after it was autoclaved. Solid media were prepared by the addition of 1% (wt/vol) Gelrite (Scott Laboratories, West Warwick, R.I.).

Chemicals and reagents. Dicamba, 3,6-DCSA and [¹⁴C]dicamba (U-phenyl-¹⁴C,

42.4 mCi/mmol, radiochemical purity greater than 98%) were supplied by Sandoz Agro, Inc. (Des Plaines, EL). To increase solubility, the dicamba and 3,6-DCSA stock solutions were prepared by titration with KOH to pH 7.0. All chemicals were purchased from Sigma Chemical Co. (St. Louis, MO), unless otherwise stated. Superose 12, Mono Q, Q-Sepharose (Fast Flow) andPhenyl-Sepharose (CL-4B) column packings forthe FPLC (fastperformance liquid chromatography) apparatus were obtained from Pharmacia (Milwaukee, WI). Ampholyte pH 4-6 and ampholyte pH 4-9 were purchased from Serva (Heidelberg, FRG). Acrylamide, β-mercaptoethanol, N,N,N',N'-tetramethylethylenediamine (TBMED) and ammonium persulfate (APS) were from Bio-Rad Laboratories (Hercules, CA). Thin layer chromatography (TLC) plates were silica gel (250 pm thickness) with UV 254 indicator, and were purchased from J.T. Baker Chemical Co. (Phillipsburg, NJ).

Enzyme assays. Dicamba O-demethylase activity was assayed by measuring the formation of [¹⁴C]3,6-DCSA from [^l4C]dicamba. Briefly, the activity in mixtures of enzyme components was measured at 30°C in a standard 300 μΐ reaction mixture containing 25 mM potassium phosphate buffer (pH 7.0), 10 mM MgCl₂,0.5 mM NADH (beta-nicotinamide adenine dinucleotide, reduced form), 0.5 mM ferrous sulfate, 50 μΜ cold dicamba, 2.5 μΜ

WO 02/068607

PCT/US02/06310 [¹⁴CJ dicamba (the final specific activity of the radioactive dicamba was 1.9 mCi/mmol), and different amounts of cell lysate or partially purified enzyme. All enzyme assays during the final purification steps were conducted in phosphate buffer because the pH optimum for dicamba O-demethylase activity was found to be in the mid range of phosphate buffers, and higher enzyme activity was observed with phosphate buffer compared to Tris-HCl [tris(hydroxymethyl)aminomeihane hydrochloride] buffer at pH 7.0. Reactions were initiated by the addition of the substrate, dicamba. At specific times, the reactions were stopped by adding 50 μί of 5% (vol/vol) H₂SO₄. Dicamba and dicamba metabolites were then extracted twice with one volume of ether, and the extracts were evaporated to dryness. The efficiencies of recovery (means + standard deviations) for the extraction procedure were 87% + 2% for dicamba and 85% ± 3% for 3,6-DCSA (Yang et al., Anal. Biochem. 219:37-42 (1994)).

[¹⁴C]dicamba and ¹⁴C-labeled metabolites were separated by thin layer chromatography (TLC). The ether-extracted dicamba and its metabolites were redissolved in 50 μί of ether prior to being spotted onto a TLC plate. The solvent system for running the

TLC was chloroform-ethanol-acetic acid (85:10:5 [vol/vol/vol]). The resolved reaction products were visualized and quantified by exposing the TLC plate to a phosphor screen for 24 hours and then scanning the screen in a Phosphorimager SF (Molecular Dynamics, Sunnyvale, CA). Estimates of the amount of radioactivity in a particular spot on the TLC plate were determined by comparing the total pixel count in that spot relative to a spot on the same plate containing a known amount of [¹⁴C] dicamba. One unit of activity was defined as the amount of enzyme that catalyzes die formation of 1 nmol of 3,6-DCSA from dicamba per minute at 30°C. Specific activities were based on the total protein concentration of the assay mixture.

The activity of the reductase component of dicamba demethylase was assayed by measuring reduction of 2,6-dichlorophenolindophenol (DCIP) with a Hitachi U-2000 spectrophotometer. The reaction contained 0.5 mM NADH, 0.2 mM FAD (flavin adenine dinucleotide), 50 μΜ DCIP, 20 mM Tris buffer (pH 8.0), and 10-100 μί of enzyme sample in a total volume of 1 ml. The change in absorbance at 600 nm with time was measured at room temperature. Specific activity was calculated using an extinction coefficient at 600 nm of 21.0 xnM’^lcm·¹ for reduced DCIP. Specific activity was expressed as nmol DCIP reduced min·* mg¹ of protein.

WO 02/068607

PCT/US02/06310 hi addition, an in situ DCIP assay was used to detect and locate the reductase activity in protein preparations separated on isoelectric focusing (IEF) gels. After electrophoresis of the proteins on an IEF gel, lanes sliced from the gel were washed with 20 ml of cold 20 mM Tris-HCl buffer (pH 8.0). Low melting agarose was dissolved by heating in 10 ml of 20 mM Tris-HCl buffer (pH 8.0) at a final concentration of 1.5% (w/v). When the agarose cooled to near room temperature, it was supplemented with 0.2 mM FAD, 50 μΜ DCIP, and 0.5 mMNADH. The assay mixture was poured onto a glass plate and allowed to solidify. The gel piece was placed on top of the solidified reaction mixture and allowed to set at room temperature for 15 minutes. If the gel slice contained a protein with reductase activity, a colorless band of reduced DCIP was generated in the blue background of DCIP.

Cell lysates. Cells were grown to an optical density at 550 nm of 1.3 to 1.5 in liquid reduced chloride medium containing a mixture of glucose and Casamino Acids on a rotary shaker (250 rpm at 30 °C). The cells were harvested by centrifugation, washed twice with cold 100 mM MgCly and centrifuged again. Cell pastes were either used immediately or quickly frozen in liquid nitrogen and stored at -80°C. At the time of enzyme purification, 25 g of frozen cells were thawed and resuspended in 50 ml of isolation buffer containing 25 mM Tris buffer (pH 7.0), 10 mM MgCl₂, and 0.5 mM EDTA. Phenylmethylsulfonyl fluoride and dithiothreitol were added to final concentrations of 0.5 mM and 1 mM, respectively. After addition of 10 mg of lysozyme and 1 mg of DNase, cells were stirred for 10 min on ice and broken by sonication (model XL2020 sonicator; Heat Systems) on ice at a medium setting (setting 5) with twelve 20-second bursts and 40-second resting intervals. The resulting cell lysates were diluted to 90 ml with isolation buffer and centrifuged at 76,000 x g for 1 h at 4 °C. The supernatant was used as the source of cleared cell lysate.

Enzyme purification. All procedures were performed at 4°C, unless otherwise stated. Solid ammonium sulfate was slowly added to a 90-ml volume of cleared cell lysate to 40% (wt/vol) saturation, with constant stirring. After 15 minutes of stirring, the mixtures were centrifuged at 15,400 x g for 15 minutes, and the precipitate was discarded. Additional solid ammonium sulfate was added to 70% (wt/vol) saturation, with constant stirring of the supernatant. After 15 min of stirring, the mixtures were centrifuged under the conditions described above. The supernatant was discarded, and the precipitate was resuspended in a

WO 02/068607

PCT/US02/06310 minimal volume of buffer A (20 mM Tris [pH 8.0], 2.5 mM MgCl₂, 0.5 mM EDTA, 5% (vol/vol) glycerol, and 1 mM dithiothreitol).

The 40%-70% ammonium sulfate cut was then loaded onto a Phenyl-Sepharose column (2.5 by 10 cm) connected to a FPLC apparatus (Pharmacia) and eluted with a decreasing linear gradient of (NH₄)₂SO₄ from 10% (w/v) to 0% (w/v). The column was preequilibrated with buffer A containing 10% (wt/vol) ammonium sulfate. The flow rate was 1 ml/min. Protein concentrations were continuously monitored at A₂₈₀ during column elution. The column was washed with 120 ml of buffer A containing 10% (wt/vol) ammonium sulfate until baseline A₂₈₀ readings were obtained. Bound proteins were eluted with a decreasing gradient of (NH₄)₂SO₄ in buffer A [ 10 to 0% (wt/vol) (NH₄)₂SO₄ in a total volume of 210 ml]. Fractions of 2 ml were collected. Aliquots of 10 μϊ were taken from each fraction and added to the standard dicamba O-demethylase assaymixture (see above), except thatnonradioactive dicamba was used as the substrate. Dicamba O-demethylase activity was detected by monitoring the conversion of dicamba to the highly fluorescent reaction product 3,6-DCSA with a hand-held UV lamp (312 nm, Fotodyne) in a darkened room.

This procedure allowed resolution of dicamba O-demethylase into three pools containing the separated components (designated components I, H and ΠΓ). Each component was essentia] for dicamba O-demethylase activity (see below). When a single component was assayed, the other two components were provided in excess. Fractions containing a single type of activity were pooled (component I, fractions 128-145; component Π, unbound fractions 12-33; component HI, fractions 62-92).

(i) Purification of component I. Fractions containing component I activity (eluting from a Phenyl-Sepharose column at 0 M (NH₄)₂SO₄, fractions 128-145) were pooled to provide a total volume of 34 ml. Ihe pooled samples were concentrated to 10 ml by centrifugation in a Centriprep-10 device (Amicon) and then applied to a Q-Sepharose (Fast Flow) FPLC column (Pharmacia) (2.5 by 6 cm) equilibrated with buffer A and washed with 80 ml of buffer A. Proteins bound to the column were eluted with a 100 ml linear gradient of 0 to 0.6 M KCI in buffer A at a flow rate of 1 ml/min. Fractions were collected at 1.5 minute intervals. Those fractions exhibiting component I activity (fractions 29-37) were pooled, dialyzed against buffer A overnight at 4°C and applied to a Mono Q HR 5/5 FPLC anion-exchange column in buffer A. Proteins were eluted at 1 ml/min by using a 50 ml

WO 02/068607

PCT/US02/06310 gradient of increasing KCI concentration (0 to 0.5 M). Fractions showing component I activity (fractions 19 to 25) were pooled and concentrated to 0.4 ml by centrifugation in a Centricon-10 device. The concentrated sample was then subjected to chromatography on a Superose 12 FPLC column (1.6 by 50 cm) at a flow rate of 0.2 ml/min with buffer A containing 100 mM KCI. Fractions 7-10 showing component I activity were pooled and concentrated by centrifugation in a Centricon-10 device.

The partially purified component I was diluted with cold 1% (w/v) glycine and concentrated by centrifugation in a Centricon-10 device three more times to desalt it in preparation for IEF electrophoresis. The desalted and concentrated sample was then applied to a 6% (w/v) IEF (pH 4-6) gel and subjected to electrophoresis for 1.5 hours at 4°C (see below). After electrophoresis, the gel was washed with 25 mM cold phosphate buffer (pH 7.0) for 5 minutes and then each slice of the gel lane was diced into small (6 mm x 4 mm) pieces. Protein was eluted from the diced gel fragments by grinding them with a pipette tip in the presence of 10 μί of 25 mM phosphate buffer (pH 7.0). Protein from each segment was mixed with an excess of components H and ΠΙ and assayed for dicamba O-demethylase activity. The gel segment which showed component I activity (which was also reddish brown in color) was loaded onto a 12.5% (w/v) sodium dodecyl sulfate polyacrylamide gel (SDSPAGE) to check sample purity.

(ii) Purification of component II. Component Π obtained by Phenyl-Sepharose column chromatography was dialyzed against buffer A overnight at 4 °C and applied to a FPLC Q-Sepharose column (2.5 by 6 cm). Sample elution conditions were identical to those described above for component I except that the elution gradient was 0 to 1M KCI in buffer A. FractionsexhibitingcomponentIIactivity(ffactions30-37)werepooled,dialyzedagainst buffer A, concentrated to 0.4 ml and applied to a FPLC Superose 12 column (1.6 by 50 cm).

The procedures for sample application and elution were identical to those described above for component I. Fractions exhibiting component Π activity (fractions 3-6) were pooled, diluted with an equal volume of buffer A, and applied to a FPLC Mono Q column. Proteins were eluted from the column using the same KCI gradient as for component I. Fractions 2025 showed component]! activity. Partiallypurified component Π was furtherpurifiedbylEF (pH 4-6) electrophoresis using the same conditions as described for component I. The gel

WO 02/068607

PCT/US02/06310 segment which showed component Π activity was loaded onto a 12.5% (w/v) SDS-PAGE for further analysis.

(iii) Purification of component III. Component ΙΠ obtained by Phenyl-Sepharose column chromatography was dialyzed against buffer A overnight at 4°C and applied to a FPLC Q-Sepharose column (2.5 by 6 cm). Conditions were identical to those described above for component I. Fractions exhibiting component HI activity (fractions 26-38) were dialyzed against buffer B [10 mM Tris-HCl (pH 7.5), 2.5 mM MgCl₂,5% (v/v) glycerol, 1 mM dithiothreitol] and concentrated to 5 ml. Blue dye affinity matrix [Cibacron Blue 3 GA type 3000 (Sigma)] was packed into a FPLC column (1x5 cm) and pre-equilibrated with 20 ml of buffer B. Concentrated component ΙΠ was loaded onto the blue dye column and washed with 20 ml of buffer B at a flow rate of 0.2 ml/min until the A._28O of the column effluent reached baseline levels. Bound protein was then eluted with 5 mM NADH in buffer

B. Fractions containing reductase activity were detected by assaying for dicamba Odemethylase activity in the presence of an excess of components I and Π and also by the ability of each fraction to reduce DCIP in the presence of NADH. Fractions having strong reductase activity in both assays were pooled, dialyzed against buffer A containing 100 mM KCI, concentrated to 0.2 ml, and applied to a FPLC Superose 12 column. The same conditions were used for running the Superose column as described for component I. Fractions containing proteins which catalyzed DCIP reduction were pooled, dialyzed against buffer A and applied to a FPLC Mono Q column. Proteins were eluted using the same conditions as for component I. Partially purified component HI was further purified by IEF (pH 4-6) gel electrophoresis. The reductase activity of proteins within the IEF gel was detected by assaying for DCIP reduction in an agarose gel overlay as described above. The gel segment which showed componentH activity was loaded onto a 12.5% (w/v) SDS-PAGE for further analysis.

Determination ofNH₂-terinitial amino acid sequences. Protein bands were excised from IEF gels and placed in the wells of a 12.5% (w/v) SDS polyacrylamide gel. After electrophoresis, the gel slices containing the purified proteins were transblotted onto a PVDF (polyvinylidene difluoride) membrane (Millipore) in a Trans-Blot cell (Bio-Rad, Richmond, CA) at 25 volts for 16 hours. The blotting buffer was a solution of 20% (v/v) methanol with 10 mM CAPS [3-(cyclohexylamino)-l-propanesulfonic acid], pH 10.0. Sequencing was

WO 02/068607

PCT/US02/06310 performed using an Applied Biosystems Inc. 420 H machine by Edman degradation (Edman and Henschen (1975) pages 232-279 in S.B. Needleman (ed.), Protein sequence determination, 2nd ed., Springer-Verlage, New York).

Determination of protein concentration. Protein concentrations were determined by the method of Bradford (1976) Anal. Biochem. 72:248-254, with bovine serum albumin as the standard.

SDS-PAGE. Sodium dodecyl sulfate-polyacrylamide gel eletrophoresis (SDS-PAGE) was performed according to modified methods of Laemmli (Laemmli (1970) Nature, 227:680-685). 12.5% (w/v) SDS gels of 85 x 65 x 0.75 mm were made as follows: running gel: 2.5 ml 40% (w/v) acrylamide/bis solution (37:5:1), 1 ml running buffer solution [3M Tris-HCl (pH 8.8), 0.8% (w/v) SDS], 4.5 ml H₂O,5 μΐ TEMED, and 40 μΐ 10% (w/v) APS;

. stacking gel: 0.5 ml 40% (w/v) acrylamide/bis, 0.5 ml stacking buffer solution [1M Tris-HCl (pH 6.8), 0.8% (w/v) SDS], 3 ml H₂O, 5 μΐ TEMED, and 12.5 μΐ 10% (w/v) APS. The composition of the running buffer was 25 mM Tris-HCl (pH 8.3), 0.2 M glycine, and 0.1% (w/v) SDS. The sample buffer contained 0.25 ml stacking buffer, 0.6 ml 20% (w/v) SDS,

0.2 ml β-mercaptoethanol, and 0.95 ml 0.1% bromphenol blue (w/v) in 50% (v/v) glycerol. Electrophoresis was performed at 80 volts in a Bio-Rad Mini Gel apparatus until the tracking dye was 0.5 cm from the anode end of the gel. Proteins were stained with 0.1% (w/v) Coomassie Brilliant Blue R-250 in a mixture of isopropanol, water, and acetic acid at a ratio of 3:6:1 (v/v/v). Destaining was performed in a mixture of methanol, water, and acetic acid at aratio of 7:83:10 (v/v/v). Standard proteins (Gibco BRL) included: myosin (214.2 kDa), phosphorylase B (111.4 kDa), bovine serum albumin (74.25 kDa), ovalbumin (45.5 kDa), carbonic anhydrase (29.5 kDa), Prlactoglobulin (18.3 kDa), and lysozyme (15.4 kDa).

Determination of molecular weight. The molecular weight (Λξ) of peptides under denaturing conditions was estimated using SDS-PAGE analysis. The molecular weights of the native components I, Π and ΠΙ were estimated by gel filtration through a Superose 12 HR 10/3 0 FPLC column (Pharmacia) at a flow rate of 0.2 ml/min in buffer A containing 100 mM KCI. Calibration proteins were gel filtration standards from Bio-Rad. They were: bovine thyroglobulin (670 kDa), bovine gamma globulin (158 kDa), chicken ovalbumin (44 kDa), horse myoglobin (17 kDa) and vitamin B-12 (1.35 kDa). The void volume of the Superose column was calculated using Blue Dextran (M_t 2,000,000, Sigma).

WO 02/068607

PCT/US02/06310

IEF. Isoelectric focusing (IEF) gel electrpphoresis was performed in a vertical minigel apparatus (Model #MGV-100) from C.B.S. Scientific Co. (Del Mar, CA). IEF gels with 6% (w/v) polyacrylamide (70 x 90 χ 1 mm) were made by mixing the following: 1.6 ml 30% (w/v) acrylamide/bis (37:5:1), 0.8 g glycerol, 0.32 ml ampholyte pH 4-6 (Serva), 0.08 ml ampholyte pH 4-9 (Serva), 5.2 ml H₂O, 10 μϊ TEMED, and 80 μϊ 10% (w/v) APS. The cathode buffer was 100 mM β-alanine and the anode buffer was 100 mM acetic acid. Protein samples in approximately 1 to 10 μϊ of 1% (w/v) glycine were mixed with an equal volume of sample buffer [50% (v/v) glycerol, 1.6% (v/v) ampholyte pH 4-9,2.4% (v/v) ampholyte pH 4-6]. Samples were loaded at the cathode end of the gel and allowed to migrate at 200 volts for 1.5 hours and 400 volts for another 1.5 hours. Proteins were stained with Coomassie Brilliant Blue R-250 using the procedure described above for SDS polyacrylamide gels. IEF markers (Sigma) were: amyloglucosodase, pi 3.6; glucose oxidase, pi 4.2; trypsin inhibitor, pi 4.6; β-lactoglobulin A, pi 5.1; carbonic anhydrase Π, pi 5.4; carbonic anhydrase Π, pi 5.9 and carbonic anhydrase I, pi 6.6

Kinetic analysis. The kinetics of the demethylation reaction catalyzed by dicamba

O-demethylase were studied by analyzing the initial rates of the reaction in the presence of a constant concentration of the enzyme and increasing concentrations of the substrate, dicamba. Reaction mixtures contained 25 mM potassium phosphate buffer (pH 7.0), lOmM MgCl₂,0.5 mMNADH, 0.5 mM FeSO₄,25 pg of partially purified O-demethylase enzyme [the 40%-70% (w/v) (NH₄)₂SO₄ fraction from a cleared cell lysate], various concentrations (0.5 to 50 μΜ) of dicamba and various concentrations (0.025 to 2.5 μΜ) of [¹⁴C]dicamba (Uphenyl-¹⁴C, 42.4 mCi/mmol) in a total volume of 300 μϊ. For assays with dicamba concentrations of 0.5 μΜ and 1 μΜ, the reaction volume was increased to 900 μϊ to ensure that sufficient amounts of radioactive dicamba and its metabolites were present to allow detection. In these reactions, the amounts of all other components in the reaction were tripled. The conversion of [¹⁴C]dicamba to [¹⁴C]3,6-DCSA was determined for different time points at each concentration of dicamba using a Phosphorlmager SF to scan radioactivity on phosphor screens which had been exposed to TLC plates for 24 hours. One unit of activity was defined as the amount of enzyme that forms 1 nmol of 3,6-DCSA per minute at 30°C. The initial rates of each reaction were determined by plotting the reaction rate versus time at each substrate concentration. Data were modeled to Michaelis-Menten

WO 02/068607

PCT/US02/06310 kinetics and values of and V_max were determined by fitting to Lineweaver-Burk plots using SigmaPlot® (Jandel Scientific, Corte Madera, CA).

Oxygen requirement. Preliminary experiments using a Garic oxygen electrode indicated oxygen consumption during a standard dicamba O-demethylase assay with dicamba as a substrate. To verify a requirement for oxygen in the O demethylation of dicamba by dicamba O-demethylase, reactions were conducted in an anaerobic chamber which contained less than 1 ppm of oxygen. The procedures for displacement of oxygen from the reaction mixture were performed at 4⁰ C. Reaction mixtures lacking enzyme were placed in a vial and sealed with a rubber stopper. For displacement of oxygen, the vial was evacuated twice by vacuum and flushed each time with nitrogen. After a third evacuation, the vial was flushed with 90% nitrogen plus 10% hydrogen. The enzyme solution was likewise purged of oxygen (with care taken not to bubble the enzyme solution). Both the reaction mixtures and enzyme solutions were transferred into an anaerobic chamber (95% N₂-5% H₂ atmosphere). Two hundred forty microliters of cleared cell lysate was injected through the rubber stopper with a microsyringe and gently mixed with 960 μί of oxygen-free reaction mixture. Reactions were carried out at 30⁰C.

An examination of the reaction products on TLC plates showed that the rate of [¹⁴C]3,6-DCSA production from [¹⁴C]dicamba under anaerobic conditions was significantly lower than the rate of reactions with the same amount of enzyme under aerobic conditions. Under anaerobic conditions, there was virtually no conversion of dicamba to 3,6-DCSA within 1 hour. However, when a parallel reaction mixture was taken from the anaerobic chamber after 30 min and incubated with air, a significant quantity of one of the components of the dicamba O-demethylase enzyme complex was an oxygenase.

It may be noted that the in vitro conversion of [^l4C]dicamba to [¹⁴C]3,6-DCSA mimics the in vivo conversion pathway documented earlier (Cork and Kreuger, Adv. Appl. Microbiol. 36:1-66 (1991); Yang et si., Anal. Biochem. 219:37-42 (1994)). In these studies, DCSA was identified as a reaction product by both TLC and capillary electrophoresis. Stringent identification of the first major product of dicamba degradation as DCSA both in vivo and in vitro has been obtained by gas chromatography-mass spectrometry analyses.

Component and cofactor requirements. After the initial separation of the three components of dicamba O-demethylase by phenyl-Sepharose column chromatography, the

WO 02/068607

PCT/US02/06310 partially purified preparations were taken individually through one additional purification on a Q-Sepharose column (2.5 by 6 cm). Samples were applied to a Q-Sepharose (Fast Flow) fast protein liquid chromatography column (Pharmacia) in buffer A and eluted with a 100-ml linear gradient of 0 to 0.6 M KCI (for the oxygenase component) or 0 to 1.0 M KCI (for the ferredoxin and reductase components) in 1.5-ml fractions. Appropriate pooled fractions from separate columns for oxygenase purification (fractions 29 to 37), for ferredoxin purification (fractions 30 to 37), or for reductase purification (fractions 26 to 38) were used for the determination of component and cofactor requirements.

The three components were assayed for O-demethylase activity in various combinations to determine component requirements.

To determine cofactor requirements, O-demethylase activity was assayed using a mixture of the three components with [¹⁴C]dicamba for 30 minutes at 30°C. The amounts of partially purified protein (provided in an optimized ratio) in the reaction mixtures were 85 μg of oxygenase, 55 μg of ferredoxin and 50 μg of reductase. The concentration of cofactors used in the reaction mixtures were 0.5 mM NADH, 0.2 mM FAD, 0.5 mM FeSO₄,10 mM MgCL, 0.5 mM NADPH, and 0.2 mM FMN.

RESULTS

Component I. The component of dicamba O-demethylase which bound most tightly to the Phenyl-Sepharose column (designated initially as component I and subsequently identified as an oxygenase) was distinctly reddish brown in color. This indicated the potential presence of a protein(s) containing an iron-sulfur clusters) or a heme group(s). The fractions with component I activity from the Phenyl-Sepharose column were subjected to further purification by Q-Sepharose (Fast Flow) and Mono Q chromatography and then to separation on a Superose 12 size exclusion column. The component I protein was then further purified on an IEF gel.

Protein from the major band on the IEF gel (with an apparentpl of approximately 4.6) was excised and separated from any remaining minor contaminants by SDS-PAGE. The major component I protein obtained after purification by IEF was greater than 90% pure as judged by densitometric analysis of this SDS-polyacrylamide gel stained with Coomassie

Blue.

WO 02/068607

PCT/US02/06310

The N-terminal amino acid sequence of the dominant protein with an apparent molecular mass of approximately 40,000 Daltons was determined. Results of amino acid sequencing indicated that the first 29 amino acids of the N-terminal region were present in the following sequence (residues in parentheses are best guesses):

Thr Phe Val Arg Asn Ala Trp Tyr Val Ala Ala Leu Pro Glu Glu Leu Ser Glu Lys Pro Leu Gly Arg Thr De Leu Asp (Asp or Thr) (Pro) [SEQ ID NO: 1].

Comparison with amino acid sequences in various databases indicated little or no homology withNH₂-terminal sequences reported for other monoxygenases or dioxygenases.

Component II. The protein fraction which did not bind to a Phenyl-Sepharose column was designated as component Π. Because this yellowish colored fraction could be replaced by ferredoxin from Clostridiumpasteurianum (but with slower reaction rates) when assays were performed in combination with components I and HI, it was tentatively designated as a ferredoxin-containing fraction. The Clostridium ferredoxin clearly functioned in place of component Π, but given the highly impure nature of the component Π used in these experiments, the efficiency of the Clostridium enzyme was significantly lower than that ofthe putative ferredoxin from strain DI-6. In particular, 55 pg of partially purified component Π mixed with excess amounts of components I and HI catalyzed the conversion of dicamba to 3,6-DCSA at a rate of approximately 5 nmol min¹ mg‘^l, while 100 pg of highly purified ferredoxin from Clostridium resulted in an activity of only 0.6 nmol min’¹ mg'¹.

Purification steps involving Q-Sepharose (Fast Flow) chromatography, Superose 12 gel filtration and Mono Q chromatography yielded approximately one milligram of purified protein from an initial 25 grams of cell paste. This fraction was purified in a similar manner to the oxygenase component by electrophoresis on an IEF gel and subsequent electrophoresis of the active IEF fraction on an SDS-polyacrylamide gel.

Analysis of component Π activity in proteins eluted from segments of the IEF gel indicated that a fraction with a pi of approximately 3.0 contained the active protein in component Π. Protein from this gel slice was eluted and subjected to SDS-PAGE. Staining

WO 02/068607

PCT/US02/06310 of the gel with Coomassie Blue revealed one prominent band of protein (molecular weight of about 28,000 Daltons) along with a smear of lower molecular weight proteins.

ComponentlU. Component HI of dicamba O-demethylase was retained on a PhenylSepharose column in high concentrations of (NH₄)₂SO₄ and eluted at approximately 4% (w/v) (NH₄)₂SO₄. This light yellow fraction was tentatively identified as a reductase-containing fraction based on its ability to reduce oxidized cytochrome c and DdP in the presence of NADH and because it could be replaced by cytochrome c reductase from porcine heart (Type 1, Sigma) in assays with components I and Π. In this set of reactions, the use of 50 gg of partially purified component ΙΠ produced a reaction rate of approximately 5 nmol min'¹ mg'¹ when mixed with an excess of components I and Π. The highly purified cytochrome c reductase showed a specific activity of approximately 2.5 nmol min'¹ mg'¹ in the reaction, an activity well below that provided by component ΙΠ when one considers the impurity of the crude component ΠΙ used in these assays. In addition, component ΙΠ exhibited reductase activity when incubated with cytochrome c or 2,6-dichIorophenol-indophenol (DCPIP) in the presence of NADH. Neither component I nor component Π showed activity in either of these two reductase assays.

Additional purification of this fraction by chromatography on columns containing.QSepharose (Fast Flow), blue dye affinity matrix, Superose 12, andMono Q packings resulted in low amounts of protein in the fractions with reductase activity. The component IH protein was about 70% pure as judged by densitometric analysis of the active protein fraction after separation by SDS-PAGE and staining with Coomassie Blue.

To further exacerbate purification of component ΙΠ, it was found that two different protein fractions from the Mono Q column step contained reductase activity when assayed with the ferredoxin and oxygenase components. Further purification of these two fractions by eletrophoresis on an IEF gel revealed that the reductase activities of the two fractions had distinctly different isoelectric points. This was demonstrated by excising lanes containing each of the two reductase fractions from the IEF gel and laying the slices on top of a pad of low melt agarose containing a DCIP reaction mixture. Reductase activity in both gel slices was identified by ±e NADH-dependent reduction of DCIP to its colorless, reduced form.

The reductase in fraction 35 had an apparent pi of approximately 5.6 while the reductase in fraction 27 possessed an apparent pi of approximately 4.8.

WO 02/068607

PCT/US02/06310

Both reductase activities isolated from the IEF gel slices were unstable and present in low amounts. Indeed, only the reductase from fraction 35 from the Mono Q column fractionation retained sufficient protein concentration and activity to allow further purification and characterization. A slice from an IEF gel containing this reductase activity was eluted and separated from contaminating proteins by SDS-PAGE. The predominant protein in this gel was one with a mass of approximately 45,000 Daltons. Size exclusion chromatography had indicated an approximate molecular mass of 50,000 Daltons for component HI in its native state.

Biochemical characteristics of dicamba O-demethylase. Dicamba O-demethylase activity was measured during incubations invitro at temperatures ranging from 20°Cto 50°C and at pH values from approximately 6 to 9. Activity peaked sharply at 30°C and broadly at pH values between 6.5 and 7.5. Enzymatic activity was dependent on the type of pH buffer employed. At pH 7, for example, activity was approximately 40% lower in Triscontaining buffers than in phosphate-containing buffers.

Values for K,,, and V_max for dicamba O-demethylase were estimated using SigmaPlot® to generate best fit curves from Michaelis-Menten and Lineweaver-Burk plots of data from duplicate experiments. The Κ„ for dicamba was estimated to be approximately 9.9 ±3.9 μΜ and the V_max for the reaction was estimated to be approximately 108 ± 12 nmol/min/mg.

The three components were assayed for dicamba O-demethylase activity in various combinations. None of the components showed enzyme activity when assayed alone. Indeed, a significant amount of O-demethylase activity could be detected only when all three components were combined. A mixture of components I and Π exhibited small amounts of enzyme activity, probably due to traces of component HI contaminating the component I fractions.

Both NADH and NADPH supported enzyme activity, with NADH being markedly more effective than NADPH. Mg²⁺was necessary for enzyme activity. Fe²*, flavin adenine dinucleotide (FAD), and flavin mononucleotide (FMN) produced little or no stimulation of enzymatic activity with the partially purified protein preparations in these experiments. The highest activity was obtained using a combination of NADH, Fe²*, Mg²⁺, and FAD.

WO 02/068607

PCT/US02/06310

DISCUSSION

Dicamba O-demethylase from Pseudomonas maltophilia, strain DI-6, is a three component oxygenase (Wang, X-Z (1996) Ph.D. thesis, University of Nebraska-Lincoln, Lincoln, NE) responsible for the conversion of the herbicide, dicamba (2-methoxy-3,6dichlorobenzoic acid), to 3,6-dichlorosalicylic acid (3,6-DCSA; 2-hydroxy-3,6dichlorobenzoic acid). Purification schemes have been devised which have allowed the isolation of each of the three components to a homogeneous or near-homogeneous state.

initial separation of the three components was achieved by chromatography on a Phenyl-Sepharose column. Enzymatic activities and other characteristics of the partially purified components allowed a tentative identification of the components as a reductase, a ferredoxin and an oxygenase - a composition similar to that found in a number of other previously studied heme-containing and nonheme-containing, multicomponent oxygenases (Batie, et al. (1992) pages 543-565, In F. Muller (ed.), Chemistry and biochemistry of flavoenzymes, vol. HI, CRC Press, Boca Raton; Harayama, et al. (1992) Annu. Rev. Microbiol. 46:565-601; Mason and Cammack (1992) Annu. Rev. Microbiol. 46:277-305; Rosche et al. (1995) Biochem. Biophys. Acta 1252:177-179). Component HI isolated from the Phenyl-Sepharose column catalyzed theNADH-dependent reduction of both cytochrome c and the dye, DCIP. In addition, its ability to support conversion of dicamba to 3,6-DCSA when combined with components I and Π could be replaced in part by cytochrome c reductase. Component Π could be replaced by the addition of ferredoxin from Clostridium pasteurianum to reactions containing components I and HL Tbe absolute need for molecular oxygen to support the O-demethylation reaction indicated that the remaining component was an oxygenase.

Uxy^e/zfls&lt;’_DIC. Component I of dicamba O-demethylase (designated as oxygenase_DIC) has been purified to homogeneity and subjected to N-terminal amino acid sequencing. The resulting sequence of twenty nine amino acid residues showed no significant homology to other protein sequences in the various data banks. However, the information obtained from this amino acid sequence permitted the design of degenerate oligonucleotide probes which have been successfully used to detect and clone the component I gene (see Example 2). Furthermore, a comparison of the amino acid sequence derived from the nucleotide sequence

WO 02/068607

PCT/US02/06310 of this clone with that of the protein sequences in the data base showed strong homology to other oxygenases (see Example 2).

The apparent molecular mass of oxygenase^, estimated from its migration in SDSpolyacrylamide gels, is approximately 40,000 Daltons. Purified preparations of the oxygenase exhibited only one major band on SDS-polyacrylamide gels stained with Coomassie Blue and Edman degradation of the protein contained in that band indicated the presence of only one N-tenninal species. Estimates derived from the behavior of native oxygenase_DIC on size exclusion columns suggests a molecular size of approximately 90,000 Daltons. All of these results suggest that the native oxygenase exists as a homodimer.

The oxygenase/hydroxylase component of a number of multicomponent systems is composed of an (tfo)_n-type subunit arrangement in which the larger a subunit is approximately 50,000 Daltons in size and the smaller β subunit is approximately 20,000 Daltons in molecular mass (Harayama, et al. (1992) Annu. Rev. Microbiol. 46:565-601). In contrast, the oxygenase component of dicamba O-demethylase appears to possess a single subunit of approximately 40 kDa in molecular mass which may exist as a dimer in its native state. This (a)_n-type subunit arrangement is similar to that found in other well characterized oxygenases such as 4-chlorophenylacetate 3,4-dioxygenase from Pseudomonas sp. strain CBS (Markus, et al. (1986) J. Biol. Chem. 261:12883-12888), phthalate dioxygenase from Pseudomonas cepacia (Batie, et al. (1987) J. Biol. Chem. 262:1510-1518), 4-sulphobenzoate

3,4-dioxygenase from Comamonas testosteroni (Locher, et al. (1991) Biochem. J., 274:833842), 2-oxo-1,2-dihydroquinoline 8-monooxygenase fromPseudomonas putida 86 (Rosche et al. (1995) Biochem. Biophys. Acta 1252:177-179), 4-carboxydiphenyl ether dioxygenase from Pseudomonaspseudoalcaligenes (Dehmel, et al. (1995) Arch. Microbiol. 163:35-41), and 3-chlorobenzoate 3,4-dioxygenase from Pseudomonas putida, (Nakatsu, et al. (1995)

Microbiology (Reading) 141:485-495).

Ferredoxin_DIC. Component Π (ferredoxin_rac) of dicamba O-demethylase waspurified by several steps of column chromatography and IEF. Final purification by SDS-PAGE produced one major band of protein (M_r~28,000) and a smear of slightly smaller proteins.

The N-terminal amino acid sequence ofthe protein with an apparent molecular weight of approximately 28,000 Daltons was determined. This amino acid sequence permitted the preparation of degenerate oligonucleotide probes, and these probes were used to isolate a

WO 02/068607

PCT/US02/06310 genomic clone coding for this protein. Although the protein produced by this clone was a ferredoxin (ferrodoxin,^.), it was subsequently determined not to be active in the degradation of dicamba when combined with the other two components of dicamba 0demethylase (data not shown).

Other evidence supports the conclusion that ferrodoxin_2gkDa is not the ferredoxin component of dicamba O-demethylase. First, the molecular mass of this protein(28 kDa) protein is significantly higher than that of the other ferredoxins found in multicomponent oxygenases from bacteria (i.e., 8-13 kDa) (Batie, et al. (1992) pages 543-565, In F. Muller (ed.), Chemistry and biochemistry of flavoenzymes, vol. ΙΠ, CRC Press, Boca Raton;

Harayama, et al. (1992) Annu. Rev. Microbiol. 46:565-601).

Second, a comparison of the N-terminal sequence of 20 amino acid residues obtained from ferredoxin_2gkDa to other amino acid sequences in die various protein data banks using Genetics Computing Group (GCG) software package (University of Wisconsin, Madison, WI) revealed strong homology (80-85% identity compared to the most likely N-terminal sequence of ferredoxin_28kDa) to a number of dicluster bacterial ferredoxins (those from Pseudomonas stutzeri, Pseudomonas putida, Rhodobacter capsulatus and Azotobacter vinelandii). The four dicluster ferredoxins which showed strong homology to ferredoxin_2gkDahave a [3Fe-4S] cluster followed by a [4Fe-4S] cluster at the N-terminus of the protein. This suggests that ferredoxin_2gkDa is distinctly different from the ferredoxin components with [2Fe-2S] clusters which are usually associated with non-heme multicomponent oxygenases (Harayama, et al. (1992) Annu. Rev. Microbiol. 46:565-601; Mason and Cammack (1992) Annu.Rev. Microbiol. 46:277-305; Rosche, etal. (1995) Biochem. Biophys. Acta 1252:177179).

Reductase_Dlc. Component IH of dicamba O-demethylase (designated as reductaseD_IC) has been the most recalcitrant of the three components to purify. This is due in part to its apparent instability and low abundance in lysates of strain DI-6. Nonetheless, sufficient protein has been purified to assign a tentative molecular mass of 45,000 Daltons. This is similar to the molecular mass of approximately50,000 Daltons obtained from size exclusion chromatography and suggests that reductaseDIC exists in its native form as a monomer. The purification of the reductase component has been further complicated by the fact that chromatography on a Mono Q column and IEF resolves purified reductase preparations into

WO 02/068607

PCT/US02/06310 two activities with apparently distinct pi values. Both fractions from the Mono Q column functioned in combination with purified ferredoxin_DIC and oxygenase_rac to produce dicamba O-demethylase activity. The presence in Sphingomonas sp. strain RW1 of two similar flavoproteins which function equally well as reductase components in the three component dibenzofuran 4,4a-dioxygenase has recently been reported by Βΰηζ and Cook (Βΰηζ and Cook (1993) J. Bacteriol. 175:6467-6475). Interestingly, both reductases were 44,000 Daltons in molecular mass, quite similar to that of the 45,000 Dalton reductase^. Multiple components of leghemoglobin reductase have also been observed in lupin root nodules using isoelectric focusing techniques (Topunov, et al. (1982) Biokhimiya (English edition) 162:378-379). In Ibis case, IEF revealed four separate components with NADH-dependent reductase activity. The resolution of the question of whether there is only one reductase^ which exists in two forms or two distinct reductases in strain DI-6 will rely on the development of improved procedures for isolating larger quantities of the proteins and/or on the cloning and sequencing of the gene(s) involved (see Examples 3 and 5).

Dicamba O-demethylase characteristics. In addition to the physical and biochemical properties of the individual components noted above, analyses of enzymatic activity have shown that the O-demethylase system has a strong affinity (Βζ,=~ 10μΜ) for its substrate and a V_max of approximately 100-110 nmol/min/mg. As expected for a soil bacterium collected in a temperate climatic zone, the maximal enzymatic activity was observed at temperatures near 30 °C. While the pH optima for the enzyme system was in the range from pH 6.5 to pH 7.5, the activity measured with a given preparation of enzyme was strongly affected by the type of pH buffering system employed. Activity in the presence of Tris buffers was at least 40% lower than with phosphate buffers at the same pH.

The reaction scheme for the reaction catalyzed by the three components of dicamba O-demethylase is presented in Figure 1. Electrons from NADH are shuttled through a short electron chain consisting of the reductase and ferredoxin to the terminal oxygenase which catalyzes the oxidation of dicamba. The similarities between dicamba O-demethylase and several multicomponent dioxygenases suggest that dicamba O-demethylase may potentially possess cryptic dioxygenase activity. It is clear, however, that this enzyme is not in the class of dioxygenases which split O₂ and incorporate one atom of oxygen into the major substrate and the other into a small organic substrate such as α-ketoglutarate (Fukumori and Hansinger

WO 02/068607

PCT/US02/06310 (1993) J. Biol. Chem. 268:24311-24317). Indeed, combinations of highly purified reductase_DIC, ferredoxin_DIC, and oxygenase_rac require only O₂, NADH, Mg²*, Fe²⁺, and dicamba for activity.

EXAMPLE 2: Identification And Sequencing Of A Clone Coding For The Oxygenase Of Dicamba O-Demethylase Of Pseudomonas maltophilia DI-6

As noted in Example 1, the first 29 amino acids of the N-terminal amino acid sequence of oxygenase_DIC had been determined to be (residues in parentheses are best guesses):

Thr Phe Val Arg Asn Ala Trp Tyr Val Ala Ala Leu Pro Glu Glu Leu Ser Glu Lys Pro Leu Gly Arg Thr De Leu Asp (Asp or Thr) (Pro) [SEQIDNO:1].

This sequence permitted the design of degenerate oligonucleotide probes which were synthesized by Operon, Alameda, CA. In particular, a mixture of 32 probes, each of which was 17 nucleotides in length, and contained all of the possible nucleotide sequences which could encode the amino acid sequence highlighted in bold above, was used. The oligonucleotide probes were 3'-end-laheled with digoxigenin (DIG) according to instructions provided by Boehringer Mannheim, Indianapolis, IN.

The DIG-labeled probes were first hybridized to P. maltophilia DI-6 genomic DNA which had been digested with various combinations of restriction enzymes, resolved on a 1 % agarose gel, and blotted to a nylon filter. Based on these results, a size-fractionated genomic library was constructed in the pBluescript Π KS+ vector and transformed into Escherichia coli DH5a competent cells. The genomic library contained 1 -2 kb Xho I/Hindin fragments. The DIG-labeled oligonucleotide probes were hybridized to an array of bacterial colonies streaked on nylon filters. Plasmid DNA was isolated from positive colonies and suhcloned. Both strands of each subclone were sequenced by the DNA Sequencing Facility at the

University of Nebraska-Lincoln. Hybridization and detection of DIG-labeled probes were performed according to protocols provided by Boehringer Mannheim.

WO 02/068607

PCT/US02/06310

A genomic DNA clone coding for the oxygenase_DIC was identified. The nucleotide sequence and the deduced amino acid sequence of the entire oxygenase^ are given in the

Sequence Listing below as SEQ ID NO:2 and SEQ ID NO:3, respectively.

A comparison of the amino acid sequence derived from the nucleotide sequence of this clone with that of the protein sequences in the Swiss Protein Database showed homology to other oxygenases. Homology was determined using the FASTA program of the GCG software package. The strongest homology was with the oxygenase component of vanillate demethylase (from Pseudomonas sp., ATCC strain 19151) which showed 33.8% identity.

EXAMPLE 3: Purification And Characterization Of Other Components Of

Dicamba O-Demethylase Of Pseudomonas maltophilia DI-6

Bacterial cultures and preparation of cleared cell lysates. Pseudomonas maltophilia, strain DI-6, was inoculated into six two-liter Erlenmeyer flasks containing one liter of reduced chloride medium (Kreuger, J.P. 1989. Ph.D. thesis. Illinois Institute of Technology, Chicago) supplemented with glucose (2.0 mg/ml) and Casamino acids (2.0 mg/ml) as the carbon sources. Cultures were incubated on an orbital shaker (225 rpm at 30°C). Cells were harvested at an A₆₀₀ ranging from 1.5 to 2.0 using a JLA-10.500 rotor in a Beckman Avanti J-25I Centrifuge at 4,000 x g for 20 minutes. Pelleted cells were stored at -80°C. The frozen cells were resuspended in 40 ml of 100 mM MgCl₂, pelleted again using the same conditions as above, and then resuspended in breaking buffer (100 mM 3-[Nmorpholino]propanesulfonic acid (MOPS) (pH 7.2), 1 mM dithiothreitol, 5% glycerol) in a ratio of 2 ml breaking buffer per gram of cells wet weight. Lysozyme was added in a ratio of 80 μϊ per gram of cells along with a Protease Inhibitor Cocktail for bacterial extracts (Sigma, P 8465) in a ratio of 5 ml per 20 grams of cells wet weight. Finally, phenylmethylsulfonyl fluoride (0.1 M stock solution in 100% ethanol) was added in a ratio οί'250μ1 per 50 ml of breaking buffer. The cells were disrupted with a sonicator(Sonics and Materials Inc., Model VCX 600) in 9.0 second bursts, with 3.0 second resting periods, for 30 minutes at an amplitude of 50%. Lysed cells were centrifuged for 75 minutes at 56,000 x g in a JA-25.50 rotor of a Beckman Avanti J-25I Centrifuge at 4°C. The supernatant (cleared cell lysate) was decanted and glycerol was added to a final concentration of 15% prior to storage at -80 °C.

WO 02/068607

PCT/US02/06310

Initial purification of dicamba O-demethylase components. An aliquot of the cleared cell lysate containing approximately 2.7 grams of protein was applied to a Pharmacia XK 26/60 column containing 25 ml of DEAE-Sepharose Fast Flow equilibrated with 50 mM MOPS (pH7.2), 1 mM dithiothreitol, and 15% (v/v) glycerol (buffer A). The column was connected to a Bio-CAD Perfusion Chromatography Workstation (USDA NRICGP Grant # 9504266) and run at a flow rate of 5.0 ml/min. After the column was loaded, it was washed with buffer A until the absorbance reading at 280 nm decreased to below 0.1. All three components of dicamba O-demethylase were bound to the DEAE column under these conditions. The column was developed with a linear gradient of 0 to 500 mM NaCl in buffer A. This resulted in the elution of the ferredoxin at 400 mM NaCl and the co-elution of the reductase and oxygenase components at 250 mM NaCl.

Purification of the ferredoxin_BIC. Fractions containing the ferredoxin_DIC eluted from the DEAE-Sepharose column were pooled and buffer exchanged into 50 mM MOPS (pH 7.2), 5% glycerol (v/v) and 200 mM NaCl (buffer B). They were then concentrated to approximately 2 ml using a Amicon Cell Concentrator with a YM10 membrane and a Centricon 10 concentrator. This sample was then applied to a pre-packed Pharmacia HiPrep 26/60 Sephacryl S-100 column equilibrated with buffer B and run at a flow rate of 0.5 ml/min on a Pharmacia FPLC apparatus. Fractions that showed activity were pooled and buffer exchanged into 50 mM MOPS (pH 7.2), 1 mM dithiothreitol and 5% glycerol (v/v) (buffer C). The fractions were concentrated to approximately 2ml and then loaded onto a Pharmacia Mono Q HR 5/5 column equilibrated in buffer C. The column was developed with a linear gradient of 0 - 2.0 M NaCl in buffer C. Fractions that contained ferredoxin activity were assayed for protein content and stored at -80°C.

Purification of the reductase_DIC Fractions from the initial DEAE-Sepharose column containing the oxygenase/reductase components were pooled, and ammonium sulfate was added to a final concentration of 1.5 M. After incubating at 4°C for 1.5 hours, the samples were centrifuged for 75 minutes at 56,000 x g at 4°C in a JA-25.50 rotor of a Beckman Avanti J-25I centrifuge. The supernatant was retained and loaded at a flow rate of 5.0 ml/min onto a Pharmacia XK 26/20 column containing 25 ml of Phenyl-Sepharose 6 Fast Flow (high sub) that was equilibrated in buffer A containing 1.5 M (NH₄)₂SO₄. Fractions that contained the reductase component were pooled, buffer exchanged into buffer B, and

WO 02/068607

PCT/US02/06310 concentrated to approximately 2 ml using the Amicon concentrator with an YM30 membrane and a CentriconlO concentrator. The 2 ml sample was applied to a pre-packed Pharmacia HiPrep 26/60 Sephacryl S-100 column equilibrated in buffer B and run at 0.5 ml/min. Fractions containing reductase activity were pooled, buffer exchanged into buffer C, and concentrated down to approximately 2 ml. The 2 ml sample was loaded onto a pre-packed Pharmacia Mono Q HR 5/5 column that was equilibrated in buffer C. The column was developed with a linear gradient of 0 - 2.0 M NaCI in buffer C at a flow rate of 0.5 ml/min. Fractions that showed reductase activity were assayed for protein content and stored at 80°C.

Rapid enzyme assays. Activity for each of the three individual components was monitored in reactions using ¹⁴C-labeled dicamba in a reaction containing an excess of the remaining two components. For each reaction, a buffer solution composed of 24 mM potassium phosphate (KPj) buffer (pH 7.0), 0.48 mM NADH, 0.48 mM FeSO4 and 9.6 mM MgCl2 was added to 200 μΐ of protein sample along with 30 μΐ of master mix (562.4 μΐ sterile water, 12.0 μΐ 50 mM dicamba and 25.6 μΐ ¹⁴C-dicamba stock solution [1.28pCi]) for a total volume of 311 μΐ and a ¹⁴C-dicamba specific activity of 4.12 pC/ml. After 60 min, each reaction was stopped by the addition of 50 μΐ 5% H2SO₄ (v/v) and 500 μΐ ether. Reaction tubes were vortexed and centrifuged in a microfuge (Eppendorf, Model 5415C) at 14,000 x g for 2 minutes. For a quick visual appraisal of enzymatic activity, each reaction tube was placed under a hand-held UV light (Fotodyne Inc., Model 3-6000) to detect fluorescence of the reaction product, DCSA. For more accurate, semi-quantitative measurements of enzymatic activities, reaction products were separated using thin layer chromatography as described in Example 1.

Protein concentration determinations. The fractions obtained at different stages of the purification protocol were assayed for protein concentration using the Bradford assay (standard protocol; Bio-Rad; see Example 1).

EXAMPLE 4: Identification and Sequencing of a Clone Coding for Ferredoxinr,T_C

The N-terminal sequence obtained from the purified ferredoxin protein (purification described in Example 3) was 29 amino acids in length (sequencing of proteins was performed by the Protein Core Facility at the University of Nebraska-Lincoln using a standard Edman

WO 02/068607

PCT/US02/06310 degradation procedure; see Example 1). A comparison of this sequence to the Genbank database showed that it was 35% identical in a 26 amino acid overlap to a terpredoxin from Pseudomonas sp., a bacterial [2Fe-2S] ferredoxin in the adrenodoxin family. The sequence information was used to design three degenerate oligonucleotide primers (two forward and one reverse). The sequence of the two 17mer forward primers was based on the N-terminal amino acid sequence obtained from the purified ferredoxin protein. The sequence of the 17mer reverse primer was based on a conserved sequence of six amino acids near the C-terminal end of six previously sequenced bacterial adrenodoxin-type ferredoxins. The primers were used in a nested PCR reaction to amplify a 191 bp product from total P. maltophilia DNA. The product was cloned into the pGEM-T Easy vector (Promega, Madison, WI) and sequenced. DNA sequencing was performed by the DNA Sequencing Core Facility at the University of Nebraska-Lincoln using a standard dideoxy-mediated chain termination procedure. An analysis of the predicted amino acid sequence of this clone confirmed that it matched the N-terminal and internal amino acid sequence previously obtained from the purified ferredoxin protein. Furthermore, the derived amino acid sequence had 48% identity over its entire length with a [2Fe-2S] ferredoxin from Rhodobacter capsulatus. The cloned fragment was labeled with digoxigenin (DIG) (Roche Diagnostics) using a standard PCR protocol (DIG/Genius System User’s Guide) and hybridized by a Southern blot to total P. maltophilia DNA that had been digested with a number of restriction enzymes. A map of the restriction sites surrounding the gene was constructed based on the sizes of the restriction fragments that hybridized to the probe. This initial experiment showed that the gene was contained on a Xho 11 Pst 1 fragment that was approximately 1 kb in length. Subsequently, total P. maltophilia DNA was digested with Xho I and Pst I, and the restriction fragments were resolved on a gel. Fragments between 0.5 and 1.5 kb in length were excised from the gel, ligated into the vector pBluescript Π KS+ (Stratagene, Inc.) and transformed into DH5a cells (Gibco BRL, Inc.). Bacterial colonies containing the size-fractionated library were screened with the DIG-labeled probe and a 900 bp Xho 1/Pst I fragment was identified. Sequence analysis showed that this clone contained a fiill-length 318 bp ferredoxin gene [SEQ ID NO :4] that encoded an 11.4 kDa protein composed of 105 amino acid residues [SEQ ID NO:5]. The amino acid sequence predicted by the cloned gene matched the N-terminal and internal amino acid sequence previously obtained from the

WO 02/068607

PCT/U S02/06310 purified ferredoxin protein. Furtheimore, the predicted amino acid sequence was homologous over its entire length to five other members of the adrenodoxin family of [2Fe2S] bacterial ferredoxins, ranging from 42% identity with a ferredoxin from Rhodobacter capsulatus to 35% identity with a ferredoxin from Pseudomonas (see Figure 2). The other three ferredoxins were from Caulobacter crescentus, Rhodococcus erythropolis, and Pseudomonas putida. Proteins in this iamily bind a single 2Fe-2S iron-sulfor cluster and have three conserved motifs. Motif 1 includes three conserved cysteines which are 2Fe-2S ligands. Motif 2 contains a cluster of negatively charged residues. Motif 3 includes the fourth conserved cysteine of the 2Fe-2S cluster.

EXAMPLE 5: Identification and Sequencing Of Clones Coding Reductase™?

Two reductase genes were cloned by the same approach that was used in Example 4 to clone the ferredoxin gene. The N-terminal sequence obtained from the purified reductase protein (purification described in Example 3) was 25 amino acids in length. A comparison of this sequence to the Genbank database showed that it was 90% identical in a 20 amino acid overlap to a cytochrome P450-type reductase component of dioxin dioxygenase, a three component enzyme previously isolated from Sphingomonas sp. RW 1. An internal sequence of 10 amino acid residues was also obtained from tryptic digests of the purified protein. The internal sequence had 87.5% identity with residues 62 through 69 of the same cytochrome P450-type reductase from Sphingomonas sp. RW1. This sequence information was used to design three degenerate oligonucleotide primers (two forward and one reverse). The sequence of the two 17mer forward primers was based on the N-terminal amino acid sequence and the sequence of the 17mer reverse primer was based on the internal amino acid sequence. The primers were used in a nested PCR reaction to amplify a 180 bp product from total P. maltophilia DNA. The product was cloned into the pGEM-T Easy vector and sequenced. An analysis of the predicted amino acid sequence of Ibis clone confirmed that it encoded a protein that matched the N-terminal and internal amino acid sequence obtained from the purified reductase protein. Furthermore, the predicted sequence had 80% identity over its entire length with the cytochrome P-450 type reductase component of the dioxin dioxygenase from Sphingomonas sp. RW1.

WO 02/068607

PCT/US02/06310

The cloned fragment was labeled with DIG and hybridized by a Southern blot to total P. maltophilia DNA that had been digested with a number of restriction enzymes. The Southern blot showed that the DIG-labeled probe recognized two distinct loci in the various restriction digests ofP. maltophilia total DNA. This observation suggested that there are two reductase genes located at different positions in the genome of P. maltophilia. It was possible that the two genes are identical duplications that encode identical reductase proteins. Alternatively, one of the genes could encode a truncated protein with no activity or a fulllength protein with low activity in our dicamba O-demethylase assay. Because it was necessary to clone the gene that encodes a protein with optimal activity, the DIG-labeled probe was used to retrieve both reductase genes.

When total P. maltophilia DNA was digested with Apra I andFcoR I, the DIG-labeled probe hybridized to one restriction fragment that was approximately 4.0 kb in length and to another larger fragment with a size of approximately 20 kb. A map of a number of restriction sites surrounding the gene located on the 4.0 kb Kpn VEcoR I fragment was constructed based on the sizes of different restriction fragments that hybridized to the probe. The restriction map indicated that the entire gene should be located on this 4.0 kb fragment. Subsequently, total P. maltophilia DNA was digested with Kpn I and EcoR I, and the restriction fragments were resolved on a gel. Fragments between 3.0 and 5.0 kb in length were excised from the gel, ligated into the vector pBluescript H KS+, and transformed into DH5a cells. Bacterial colonies containing the size-fractionated library were screened with the DIG-labeled probe and a 4.3 kb Kpn REcoR I fragment was identified. Sequence analysis showed that this clone contained a full-length 1224 bp reductase gene [SEQ ID NO:6] and encoded a 43.7 kDa protein consisting of408 amino acids [SEQ ID NO:7]. The amino acid sequence predicted by the cloned gene matched the N-terminal and internal amino acid sequence previously obtained from the purified reductase protein. Furthermore, the predicted amino acid sequence was homologous over its entire length to at least four other FAD-dependent pyridine nucleotide reductases, ranging from 69% identity with a cytochrome P450-type reductase component of dioxin dioxygenase from Sphingomonas sp. RW1 to 36% identity with the teipredoxin reductase iromsPseudomonas species (see Figure

3). The two other FAD-dependent pyridine nucleotide reductases were from R. erythropolis andP. putida. Proteins in this family ofFAD-dependentpyridinenucleotidereductases have

WO 02/068607

PCT/US02/06310 five conserved motifs. Motifs 1 and 3 contain three conserved glycine residues and correspond to the ADP binding site for FAD and NAD(P) respectively. Motif 5 corresponds to the site binding the FAD flavin moiety.

To clone the second gene, total P. maltophilia DNA was digested with Kpn I and EcoR. I and the resulting restriction fragments were resolved on an agarose gel. Fragments with a size of approximately 20 kb were excised from the gel, digested with a number of restriction enzymes, and then hybridized by Southern blot to the DIG-labeled probe. A map of the restriction sites surrounding the second gene was constructed based on the sizes of the restriction fragments that hybridized to the probe. These experiments showed that the fulllength second reductase gene was contained on an Apa I fragment that was approximately 3.0 kb in length. Subsequently, total P. maltophilia DNA was digested with Apa I and the restriction fragments were resolved on a gel. Fragments between 2.0 and 4.0 kb in length were excised from the gel, ligated into the vector pBluescript Π KS+, and transformed into DH5a cells. Bacterial colonies containing the size-fractionated library were screened with the DIG-labeled probe and a 3.0 kb Apa I fragment was identified. Sequence analysis showed that the 3.0 kb clone contained an open reading frame of 1227 bp [SEQ ID NO:8] and encoded a 43.9 kDa protein consisting of409 amino acids [SEQ ID NO:9]. The amino acid sequence encoded by the second reductase gene is almost identical (98.8% identity) to the sequence of the first gene.

EXAMPLE 6: Transformation Of Plants

In order to place each of the three genes that encode the components of dicamba Odemethylase individually into cassettes suitable for expression in plants, the following steps were taken. Oligonucleotide primers were designed to generate a Neo I site at the 5' end and an Xba I site at the 3' end of each of the three genes by PCR amplification. The authenticity of the resulting PCR products was confirmed by sequencing, and each gene was then cloned individually into the pRTL2 vector (provided by Dr. Tom Clemente of the Plant Transformation Core Research Facility, University of Nebraska, Lincoln, Nebraska). This vector contains a 144 bp translation enhancer sequence from tobacco etch virus (TEV) (Carrington and Freed, J. Virology, 64:1590-1597 (1990)) at the 5' end ofthe polylinker. The oxygenase, reductase, and ferredoxin genes, each with a 5' translation enhancer, were then

WO 02/068607

PCT/US02/06310 cloned individually as Xho VXba I fragments into the plant expression vector pKLP36 (obtained from Indu Maiti, University of Kentucky, Lexington, Kentucky) (Maiti and Shepard, Biochem. Biophys. Res. Commun., 244:440-444 (1998)). This binary vector contains the peanut chlorotic virus full-length promoter (PC1SV FLt36) with a duplicated enhancer domain for constitutive expression in plants and the pea rbcS 3' sequence for efficient transcript termination (Maiti and Shepherd, Biochem. Biophys. Res. Commun., 244:440-444 (1998)). Constructs with all three genes on one binary vector can be produced using combinations of the three genes in a number of different orders and orientations.

The following methods were employed to move the oxygenase, ferredoxin, and reductase constructs individually into Agrobacterium tumefaciens and to transform each construct into Arabidopsis and tobacco. The three constructs were moved into the A. tumefaciens strain C58C1 by a modified triparental mating procedure routinely used by the Plant Transformation Core Research Facility. This involved incubating Escherichia coli cells carrying each construct with a mixture of A. tumefaciens cells and£. coli cells carrying the helper plasmid pRK2013 (for plasmid mobilization). A. tumefaciens cells containing each of the constructs were then used to transform tobacco and Arabidopsis with the assistance of the Plant Transformation Core Research Facility. For tobacco, leaf explants were incubated with a suspension of A. tumefaciens cells containing each of the constructs, and shoots were regenerated on solid medium containing kanamycin (Horsch et al., Science, 227:1229-1231 (1985)). The details of the tobacco transformation protocol are as follows. Preparation of Tobacco explants:

3-4 inch leaves were selected from 1 month old plants (top 2 leaves). This is sufficient for 30-40 explants. The leaves were placed into a large beaker and covered with dH₂0 for20-30 minutes. The water was drained and the leaves covered with 10% Chlorox™ plus Tween 20™ (1 drop per 50 ml). Leaves were allowed to soak for 15 minutes, and rinsed 3X-5X with sterile dH₂O. The outside margin of the leaf along with the tip and petiole were trimmed, the midrib was dissected out, and the remaining tissue was sliced into 0.5 x 0.5 cm squares. The explants were precultured, 30 per plate, ad axial side up, for 1 day prior to inoculation.

WO 02/068607

PCT/US02/06310

Agrobacterium inoculum preparation:

A 2 ml culture of Agrobacterium was initiated in YEP medium (10 g/l peptone, 10 g/l NaCl and 5 g/l yeast extract, pH 7.0) amended with appropriate antibiotics. The culture was allowed to become saturated. The 2 ml culture was subcultured into 50 mis of YEP medium and allowed to grow for 6-8 hrs (28 °C with constant shaking). The cells were harvested by centrifugation (3,000-4,000 rpm). The bacterial pellets were resuspended to a final OD ₆₆₀ - 0.3-1.0, in co-cultivation medium.

The prepared Agrobacterium inoculum was transferred to petri plates. The precultured explants were inoculated for 30 minutes. The explants were blotted on sterile filter paper and placed on to co-cultivation medium (10 explants per plate) solidified with 0.8% agar, which was overlaid with a sterile filter paper (Co-culture explants ad axial side up. The explants were allowed to co-cultivate for 3 days. Following the co-cultivation period, the explants were briefly washed in regeneration medium.

The washed explants were blotted on sterile filter paper, and the tissue (10 explants per plate) was transferred to regeneration medium solidified with 0.8% agar, plus the appropriate selective agent (Kanamycin 150 mg/1). The explants were cultured ad axial side up. The tissue was transferred every 2 weeks to fresh regeneration medium, and shoots were excised (&gt;3 cm) and sub,cultured to rooting medium. The rooted shoots were then acclimated to soil.

Reagents:

Preculture medium (PM):

MS salts with B5 vitamins, 3% sucrose amended with 1 mg/1 BAP, 0.1 mg/1

NAA and 8-pg/l pCPA (p-Chlorophenoxyacetic acid), pH 5.7 (Note: filter sterilize all growth regulators).

Co-cultivation medium (CM):

1/10 MS/B5 medium, amended with 3% sucrose, 1.0 mg/1 BAP, 0.1 mg/1

NAA and 8 gg/l pCPA.

Regeneration medium (RM):

MS/B5 medium amended with 1.0 mg/1 BAP, 0.1 mg/1 NAA and 3% sucrose, pH 5.7. Utilize the antibiotics carbenicillin (500 mg/1) and cefotaxime (100 mg/1).

WO 02/068607

PCT/US02/06310

Rooting medium:

MS salts with full strength B5 vitamins, amended with 0.1 mg/1 NAA, and 1 % sucrose. Solidify the medium with 0.8% agar. Maintain the antibiotics (if using npt Π 75 mg/1) carbenicillin (500 mg/1) and cefotaxime (100 mg/1).

Ten shoots were selected from each of the three transformation experiments, placed on rooting medium for a few weeks, and then moved to pots in the greenhouse. For Arabidopsis, a pot of plants with flowers was incubated with a suspension of A. tumefaciens cells containing each of the constructs, and the plants were then allowed to set seed (Bechtold et al., C.R. Acad. Sci. Paris, Sciences de la vie/Life sciences, 316:1194-1199 (1993); Clough and Bent, Plant J., 16(6):735-743 (1998)). The seed was collected and germinated on medium with kanamycin. After the seedlings had developed an adequate root system, several plants were selected from each transformation experiment and moved to pots in a growth chamber.

For evaluation of the expression of each gene in the transformed plants, Western blots of leaf lysates from several transformed plants were prepared and probed with polyclonal antibodies that detect the three components of dicamba O-demethylase. To determine enzyme activities for each enzyme in plant extracts, the approach was to combine leaf lysates from transformed plants expressing foe oxygenase, ferredoxin or reductase proteins with excess amounts of foe other O-demefoylase components purified from P. maltophilia strain DI-6. These mixtures were tested for dicamba O-demefoylase activity with bofo foe standard ^l4C-labeled dicamba assay (see Examples 1 and 3) and an HPLC assay employing nonradioactive dicamba as substrate (dicamba and 3,6-DCSA elute at different points from foe HPLC and can be quantitated).

To test expression of gene products in the chloroplast compartment, constructs were made with a transit peptide sequence at foe 5' end of foe oxygenase, ferredoxin, and reductase genes. Such an approach was successful for introducing tolerance to foe sulfonylurea herbicides into tobacco plants (O’Keefe et al., Plant Physiol., 105:473-478 (1994)). More specifically, foe pea ribulose 1,5-bisphosphate carboxylase small subunit transit peptide (signal) sequence (nucleic acid sequence represented by SEQ ID NO: 19) was genetically engineered into an intermediate pRTL2 vector containing foe oxygenase, ferredoxin or foe reductase genes, each with its own expression control sequence (see above). This was

WO 02/068607

PCT/US02/06310 accomplished by creating Ncol sites at both ends of the transit peptide using oligonucleotide primers and the PCR protocol. The Ncol sites facilitate the introduction of the transit peptide between the TEV leader sequence and the reporter gene (oxygenase, reductase or ferredoxin) of the pRTL2 vector. The entire cassette (TEV leader, the transit peptide and the reporter gene) was then cut out of the pRTL2 vector as an Xhol/Xbal fragment and introduced into the binaiy vector pKLP3 6 containing the PC1S V FLt36 promoter and the rbcS 3 ’ terminator.

To test the possibility that the codon usage of the bacterial ferredoxin gene was not fully optimal for efficient translation in a plant cell, a synthetic ferredoxin gene encoding the same amino acid sequence as the P. maltophilia strain DI-6 ferredoxin, but with optimized codon bias for dicot plants was synthesized by Entelechon, GmbH (Regensberg, Germany). It has been well documented that changes in codon usage are essential for optimal expression of the bacterial B.t. toxin genes in plant cells (see, e.g., Diehn et al., Genetic Engineering, 18:83-99 (1996)).

EXAMPLE 7: Spraying of Plants

To test the resistance of the transgenic tobacco plants to dicamba, seeds from the T„ generation described in Example 6 above were germinated and grown to the 10-12-leaf stage. With the assistance of the Department of Agronomy, these plants were subjected to ‘constant rate spraying’ using a custom made sprayer (Burnside, Weed Science Vol.17, No.l, 102104,1969), manufactured by ISCO. The sprayer has a TEEJET nozzle and screen, uses a delivery speed of 1.87 mph and a pressure of 42 PSI. The nozzle height is set at 8” above canopy. Dicamba (commercial product-Clarity), at various concentrations, was sprayed on the plants mixed with a non-ionic surfactant at a rate of 20 gallons per acre under compressed air.

To establish an initial kill curve, wild type tobacco plants were treated with increasing concentrations of dicamba ranging from 0 to 0.5 lbs/acre. Having established 100% sensitivity of the wild type plants to 0.5 lbs/acre of dicamba, transgenic plants containing the oxygenase gene alone with or without the transit peptide, were treated with increasing concentrations of dicamba ranging from 0.5 lbs/acre to 5 lbs/acre. Transgenic plants bearing oxygenase gene constructs but lacking the transit peptide coding sequence displayed resistance (i.e., grew and were healthy) to dicamba treatment of at least 2.5. lbs/acre. The

WO 02/068607

PCT/US02/06310

<img file="NZ528010A_D0003.tif" />

transgenic tobacco plants bearing the oxygenase gene constructs with the transitpeptide grew and were healthy when sprayed with a concentration of dicamba of up to 5 lbs/acre. That is, transgenic plants containing the oxygenase gene with or without a transit peptide coding sequence appeared similar to the untreated control plants up to at least the indicated amounts of dicamba, whereas the non-transgenic plants treated with the same amounts of dicamba showed reduced growth, withering and severe necrosis.

In a similar experiment, transgenic plants containing the three gene construct (oxygenase, ferredoxin and reductase genes), were sprayed with 0.125 lb/acre dicamba. Transgenic plants containing the three gene construct grew and were healthy when sprayed with this concentration of dicamba, whereas the non-transgenic plants treated with the same amount of dicamba showed reduced growth, withering and severe necrosis. Since transgenic plants with the oxygenase construct alone displayed dicamba tolerance up to 2.5 lb/acre dicamba, it is predicted that the transgenic plants containing the three gene construct will display similar resistance up to 2.5 lb/acre or even greater.

These results clearly demonstrate that the transgenic plants transformed with the dicamba-degrading oxygenase or the three gene construct of the present invention have acquired resistance to dicamba.

WO 02/068607 A3

PCT/US02/06310