EP0795020A1 - Plant stearoyl-acp thioesterase sequences and methods to increase stearate content in plant seed oils - Google Patents

Abstract

This invention relates to plant thioesterases, specifically plant acyl-ACP thioesterases having significant activity on stearoyl-ACP. DNA constructs useful for the expression of a plant stearoyl-ACP thioesterase in a plant seed cell are described. Such constructs will contain a DNA sequence encoding the plant stearoyl-ACP thioesterase of interest under the control of regulatory elements capable of preferentially directing the expression of the plant stearoyl-ACP thioesterase in seed tissue, as compared with other plant tissues, when such a construct is expressed in a transgenic plant. This invention also relates to methods of using a DNA sequence encoding a plant stearoyl-ACP thioesterase for the modification of the stearate produced in a plant seed cell. A plant thioesterase having significant activity on stearoyl-ACP exemplified herein is from mangosteen. Modified stearate concentration may be obtained by expression of mangosteen thioesterase alone or in combination with antisense constructs for reduction of native stearoyl-ACP desaturase in plant seeds.

Description

PLANT STEAROYL-ACP THIOESTERASE SEQUENCES AND METHODS TO INCREASE STEARATE CONTENT IN PLANT SEED OILS

INTRODUCTION

Field of the Invention

This invention relates to the application of genetic engineering techniques to plants. More specifically, the invention relates to plant acyl-ACP thioesterase sequences having substantial activity on C18:0-ACP and methods for the use of such sequences to increase 18:0 levels in plant seed oils.

BACKGROUND

Fatty acids are organic acids having a hydrocarbon chain of from about 4 to 24 carbons. Many different kinds of fatty acids are known which differ from each other in chain length, and in the presence, number and position of double bonds. In cells, fatty acids typically exist in covalently bound forms, the carboxyl portion being referred to as a fatty acyl group. The chain length and degree of saturation of these molecules is often depicted by the formula CX:Y, where "X" indicates number of carbons and "Y" indicates number of double bands. Fatty acyl groups are major components of many lipids, and their long, non-polar hydrocarbon chain is responsible for the water-insoluble nature of these lipid molecules. The type of covalent linkage of the fatty acyl group to other factors can vary. For example, in biosynthetic reactions they may be covalently bound via a thioester linkage to an acyl carrier protein (ACP) or to CoenzymeA (CoA) , depending on the particular enzymatic reaction. In waxes, fatty acyl groups are linked to fatty alcohols via an ester linkage, and triacylglycerols have three fatty acyl groups linked to a glycerol molecule via an ester linkage.

The production of fatty acids in plants begins in the plastid with the reaction between acetyl-CoA and malonyl-ACP to produce butyryl-ACP catalyzed by the enzyme, β-ketoacyl-ACP synthase III. Elongation of acetyl-ACP to 16- and 18- carbon fatty acids involves the cyclical action of the following sequence of reactions: condensation with a two-carbon unit from malonyl-ACP to form a β-ketoacyl-ACP (β-ketoacyl-ACP synthase) , reduction of the keto-function to an alcohol (β-ketoacyl-ACP reductase) , dehydration to form an enoyl-ACP (β-hydroxyacyl-ACP dehydrase) , and finally reduction of the enoyl-ACP to form the elongated saturated acyl-ACP (enoyl-ACP reductase) . β-ketoacyl- ACP synthase I, catalyzes elongation up to palmitoyl-ACP (C16:0), whereas β-ketoacyl-ACP synthase II catalyzes the final elongation to stearoyl-ACP (C18.0). The longest chain fatty acids produced by the FAS are typically 18 carbons long. A further fatty acid biochemical step occurring in the plastid is the desaturation of stearoyl-ACP (C18.0) to form oleoyl-ACP (C18:l) in a reaction catalyzed by a Δ-9 desaturase, also often referred to as a "stearoyl-ACP desaturase" because of its high activity toward stearate the 18 carbon acyl-ACP.

Carbon-chain elongation in the plastids can be terminated by transfer of the acyl group to glycerol 3-phosphate, with the resulting glycerolipid retained in the plastidial,

"prokaryotic", lipid biosynthesis pathway. Alternatively, specific thioesterases can intercept the prokaryotic pathway by hydrolyzing the newly produced acyl-ACPs into free fatty acids and ACP. Subsequently, the free fatty acids are converted to fatty acyl-CoA's in the plastid envelope and exported to the cytoplasm. There, they are incorporated into the "eukaryotic" lipid biosynthesis pathway in the endoplasmic reticulum which is responsible for the formation of phospholipids, triglycerides and other neutral lipids. Following transport of fatty acyl CoA's to the endoplasmic reticulum, subsequent sequential steps for triglyceride production can occur. For example, polyunsaturated fatty acyl groups such as linoleoyl and <*-linolenoyl, are produced as the result of sequential desaturation of oleoyl acyl groups by the action of membrane- bound enzymes. Triglycerides are formed by action of the 1-, 2-, and 3- acyl-ACP transferase enzymes glycerol-3-phosphate acyltransferase, lysophosphatidic acid acyltransferase and diacylglycerol acyltransferase. The fatty acid composition of a plant cell is a reflection of the free fatty acid pool and the fatty acids (fatty acyl groups) incorporated into triglycerides as a result of the acyltransferase activities. The fatty acid composition of an oil determines its physical and chemical properties, and thus its uses. Plants, especially plant species which synthesize large amounts of oils in plant seeds, are an important source of oils both for edible and industrial uses. Various combinations of fatty acids in the different positions in the triglyceride will alter the properties of triglyceride. For example, if the fatty acyl groups are mostly saturated fatty acids, then the triglyceride will be solid at room temperature. In general, however, vegetable oils tend to be mixtures of different triglycerides. The triglyceride oil properties are therefore a result of the combination of triglycerides which make up the oil, which are in turn influenced by their respective fatty acid compositions.

For example, cocoa-butter has certain desirable qualities (mouth feel, sharp melting point, etc.) which are a function of its triglyceride composition. Cocoa-butter contains approximately 24.4% palmitate (16:0), 34.5% stearate (18:0), 39.1% oleate (18:1) and 2% linoleate (18:2). Thus, in cocoa butter, palmitate-oleate-stearate (POS) comprises almost 50% of triglyceride composition, with stearate-oleate-stearate (SOS) and palmitate-oleate-palmitate (POP) comprising the major portion of the balance at 39% and 16%, respectively, of the triglyceride composition. Other novel oils compositions of interest might include trierucin (three erucic) or a triglyceride with medium chain fatty acids in each position of the triglyceride molecule.

Of particular interest also are triglyceride molecules in which stearate is esterified at the sn-1 and sn-3 positions of a triglyceride molecule and oleate is at sn-2. Vegetable oils rich in such SOS (Stearate-Oleate-Stearate) molecules share certain desirable qualities with cocoa butter yet have a degree of additional hardness when blended with chocolate- based fats. SOS - containing vegetable oils are currently extracted from relatively expensive oilseeds from certain trees grown in tropical areas such as Sal, Shea, and Illipe trees from India, Africa, and Indonesia respectively. Cheaper and more conveniently grown sources for SOS -type vegetable oils are desirable.

In addition, vegetable oils rich in stearate fatty acid content tend to be solid at room temperature. Such vegetable fats can be used directly in shortenings, margarine and other food "spread" products, obviating the need for chemical hydrogenation. Hydrogenation is a process to convert unsaturated fatty acids in liquid oils to a saturated form which in turn converts the oil into a solid fat useful in margarine and shortening applications. The cost and any other factors associated with chemical hydrogenation, such as the production of trans fatty acids, can be avoided if the vegetable oil is engineered to be stearate rich in the plant seed.

Moreover, some plant tissues use 18 carbon fatty acids as precursors to make other compounds. These include saturated long chain fatty acids longer than 18 carbons in length. Since very little stearate typically accumulates in common oilseed crops, it may be necessary to increase stearate accumulation if one wants to increase production of compounds which depend upon supply of stearate fatty acids for synthesis.

Thus, a variety of plant oils modifications are desired, including alternative crop sources for certain oils products and/or means to provide novel fatty acid compositions for plant seed.

DESCRIPTION OF THE FIGORES Figure 1. An amino acid sequence alignment of representative Class I (FatA) and Class II (FatB) thioesterases is provided. UcFatBl (SEQ ID NO:l) is a California bay C12 thioesterase. CcFatBl (SEQ ID NO:2) is a camphor C14 thioesterase. CpFatBl (SEQ ID NO:3) is a Cuphea palustris C8 and CIO thioesterase. CpFatB2 (SEQ ID NO:4) is a Cuphea palustris C14 thioesterase. GarmFatAl (SEQ ID NO:5) is a mangosteen 18:1 thioesterase which also has considerable activity on C18:0 acyl-ACP substrates. BrFatAl (SEQ ID NO:6) is an 18:1 thioesterase from Brassica rapa (aka Brassica campestris) . Amino acid sequences which are identical in all of the represented thioesterases are indicated by bold shading.

Figure 2. The nucleic acid sequence and translated amino acid sequence of a mangosteen Class I acyl-ACP thioesterase cDNA clone, GARM FatAl (SEQ ID NO:7), are provided.

Figure 3. The nucleic acid sequence and translated amino acid sequence of a mangosteen Class I acyl-ACP thioesterase cDNA clone, GARM FatA2 (SEQ ID NO:8), are provided. Figure 4. Data from fatty acid composition (weight percent) analyses of T2 mature pooled seed from Brassica plants transformed with napin/mangosteen TE construct pCGN5266 are provided.

Figure 5. Fatty acid composition data (weight percent) from analyses of T2 half seeds from events 5266-LP004-2 and 5266-SP30021-29 are provided.

Figure 6. The nucleic acid sequence and translated amino acid sequence of a Brassica napus BND11 stearoyl-ACP desaturase cDNA clone are provided. Figure 7. The nucleic acid sequence and translated amino acid sequence of a Brassica napus BND9 stearoyl-ACP desaturase cDNA clone are provided.

SUMMARY OF THE INVENTION

This invention relates to plant thioesterases, specifically plant acyl-ACP thioesterases having substantial activity on 18:0-ACP substrates such that the C18:0 content in a target plant seed oil can be dramatically increased upon expression of the plant acyl-ACP thioesterase in the seeds of the target plant.

DNA constructs useful for the expression in a plant seed cell of a plant acyl-ACP thioesterase having substantial activity on 18:0-ACP substrates are described herein. Such constructs will contain a DNA sequence encoding the plant acyl-ACP thioesterase under the control of regulatory elements capable of preferentially directing the expression of the plant acyl-ACP thioesterase in seed tissue, as compared with other plant tissues. At least one element of the DNA construct will be heterologous to at least one other of the elements, or when found in a plant cell, at least one element will be heterologous to the plant cell. Of particular interest for use in the constructs of the present invention is a Class I acyl-ACP thioesterase, Gar FatAl, obtained from mangosteen ( Garcinia mangifera) .

In yet a different embodiment, host plant cells containing a first DNA construct capable of expressing a plant acyl-ACP thioesterase having substantial activity on 18:0-ACP substrates and a second DNA construct capable of expressing an anti-sense stearoyl-acyl ACP desaturase sequence are desired. Such a first DNA construct will contain a DNA sequence encoding the plant acyl-ACP thioesterase of interest under the control of regulatory elements capable of preferentially directing the expression of the plant acyl-ACP thioesterase in seed tissue as compared with other plant tissues when such a construct is expressed in a transgenic plant. The second DNA construct will contain a DNA sequence encoding a plant stearoyl-acyl ACP desaturase element positioned in an anti- sense orientation under the control of regulatory elements capable of directing the transcription of the plant stearoyl- acyl ACP desaturase in the plant host cell.

In a different aspect, this invention relates to methods of using a DNA sequence encoding a plant acyl-ACP thioesterase for modifying the composition of triglycerides, i.e., plant oils, produced by a plant seed. Such modified oil compositions may be obtained either by expression of the acyl- ACP thioesterase alone, or by expression of the acyl-ACP thioesterase in combination with a second construct which provides for reduction of the level of the native stearoyl-ACP desaturase of the target plant species. Plants and plant parts, particularly seeds and oils extracted from such seeds, having such a modified fatty acid composition are contemplated herein. Also provided in the present invention is a novel antisense desaturase plant expression construct which provides regions from two different classes of desaturase genes from a multigene family. Expression of this construct in transgenic plant seeds provides an improved method for increasing stearate levels in plant seeds by antisense methods. Higher levels of stearate are obtained as a result of this improved method, as well as higher percentages of first generation transformants which display the desired increased stearate phenotype.

DETAILED DESCRIPTION OF THE INVENTION

By this invention, a mechanism for the increased accumulation of stearate (C18:0) in plants is provided. As described herein, plant acyl-ACP thioesterases having substantial activity toward 18:0-ACP substrates are involved in the accumulation of stearate in at least some plant species and may be expressed in transgenic plant seeds to provide an increase in the levels of stearate fatty acids. Furthermore, it is demonstrated that such 18:0-ACP thioesterases are members of the Class I group of plant acyl-ACP thioesterases.

Class I type thioesterases have been found in essentially all plant sources examined to date, and are suggested to be essential "housekeeping" enzymes (Jones et al . The Plant Cell (1995) 7:359-371) required for membrane biosynthesis. Class I type thioesterases have previously been shown to have activity primarily on 18:1 acyl-ACP substrates, with some lesser amount of activity on 16:0 substrates, and only little or no activity on 18:0 substrates. Examples of Class I thioesterases from safflower, Cuphea hookeriana and Brassica rapa ( campestris) , which have activity primarily on 18:1-ACP substrate, have been described (WO 92/20236 and WO 94/10288) . Other 18:1 thioesterases have been reported in AraJidopsis thaliana (Dormann et al . (1995) Arch. Biochem. Biophys . 316: 612-618 ) , Brassica napus (Loader et al . (1993) Plant Mol . Biol . 23 : 169- 778) and coriander (Dormann et al . (1994) Biochem. Biophys . Acta 2212:134-136). A Class I thioesterase from soybean (WO 92/11373) was reported to provide 10- and 96-fold increases in 16:0-ACP and 18:1-ACP activity upon expression in E. coli , and a smaller (3-4 fold) increase in 18:0-ACP activity.

A second class of plant thioesterases, Class II (or FatB) thioesterases, include enzymes that primarily utilize fatty acids with shorter chain-lengths, from C8:0 to C14:0 (medium chain fatty acids) as well as C16:0. Class II thioesterases preferably catalyze the hydrolysis of substrates containing saturated fatty acids. Class II (or FatB) thioesterase genes have been isolated from California Bay, elm, Cuphea hookeriana, Cuphea palustris, Cuphea lanceolata, nutmeg, Arabidopsis thaliana, mango, leek and camphor. A FatB thioesterase gene was also identified in mangosteen in the gene isolation experiments described herein. Expression of the FatB gene in E. coli demonstrated hydrolysis activity primarily on 16:0 substrates. In the following examples, isolation of genes encoding Class I acyl-ACP thioesterases from mangosteen is described. Two different types of Class I thioesterase genes were discovered. One mangosteen thioesterase gene, (GarmFatA2), is shown herein to be an 18:1-ACP specific Class I thioesterase similar to those discovered previously in other plant tissues. However, a second type of mangosteen Class I thioesterase gene, represented by clone GarmFatAl, was discovered which demonstrates 18:1-ACP thioesterase activity (100-fold increase upon expression in E. coli ) , but also demonstrates substantial activity on 18:0-ACP substrates. The 18:0 activity of

GarmFatAl is approximately 25% of the 18:1 activity, whereas in most Class I thioesterases analyzed to date, the 18:1 activity is highly predominant, with activity on 16:0 and 18:0 substrates detectable at less than 5% of the 18:1 activity levels. Additionally, most plant Class I thioesterases demonstrate approximately equal activity on 16:0 and 18:0 substrates, whereas the GarmFatAl mangosteen thioesterase of the present invention demonstrates a clear preference for hydrolysis of 18:0 substrates over 16:0 substrates. The novel mangosteen GarmFatAl thioesterase clone will thus be useful for production of the 18:1/18:0 thioesterase in host cells, and particularly for expression in plant seed cells for modification of TAG fatty acid composition to provide increased levels of C18:0 fatty acyl groups. Furthermore, the mangosteen GarmFatAl clone will be useful in plant genetic engineering application in conjunction with plants containing elevated levels of C18:0 (stearate) fatty acids. Such plants may be obtained by antisense gene regulation of stearoyl-ACP desaturase in Brassica seeds as described by Knutzon et al . ( Proc. Nat . Acad. Sci . (1992) 55:2624-2628), and may also be obtained by co-suppression using sense expression constructs of the stearoyl-ACP desaturase gene, or by conventional mutation and plant breeding programs.

The determination that Class I type plant thioesterases are active in the in vivo accumulation of 18:0 fatty acids suggests several possibilities for additional plant sources of genes which encode thioesterase proteins having substantial activity on 18:0 substrates. Stearate is found in some natural plant species, particularly tropical plant species, in abundance. For example, other species in the genus Garcinia accumulate triglycerides containing stearate in their seeds, e.g., kokum. Other natural plant source of C18:0 fatty acids include plants of the Mangifera family: e.g., mango,

Butyrospermuπ. (shea) , Pentadesma (tallow tree) , Illipe (illipe butter) , Theobroma (cocoa) , Si arouJba (tree of paradise) and Shorea (sal) .

A plant acyl-ACP thioesterase DNA sequence useful for alteration of stearate levels as described herein encodes for amino acids, in the form of a protein, polypeptide or peptide fragment, which amino acids demonstrate substantial activity on 18:0 acyl-ACP substrates to form 18:0 free fatty acid (i.e., stearate) under plant enzyme reactive conditions. By "enzyme reactive conditions" is meant that any necessary conditions are available in an environment (i.e., such factors as temperature, pH, lack of inhibiting substances) which will permit the enzyme to function.

One skilled in the art will readily recognize that antibody preparations, nucleic acid probes (DNA and RNA) and the like may be prepared and used to screen and recover "homologous" or "related" Class I acyl-ACP thioesterases from a variety of plant sources. Typically, nucleic acid probes are labeled to allow detection, preferably with radioactivity although enzymes or other methods may also be used. As plant thioesterase genes are known to contain extensive sequence homology, various DNA screening methods such as PCR or DNA hybridization methods may be used to identify related Class I thioesterases. Plant thioesterase genes show at least approximately 50% sequence identity at the nucleic acid level. Between members of Class I thioesterases the percentage of sequence identity is as high as 70-80%, and Class II thioesterases typically demonstrate sequence identity of at least 60%.

Thus, in order to obtain additional stearoyl-ACP thioesterases, a genomic or other appropriate library prepared from the candidate plant source of interest is probed with conserved sequences from one or more Class I plant thioesterases to identify homologously related sequences. Positive clones are then analyzed by restriction enzyme digestion and/or sequencing. When a genomic library is used, one or more sequences may be identified providing both the coding region, as well as the transcriptional regulatory elements of the thioesterase gene from such plant source. Probes can also be considerably shorter than the entire sequence. Oligonucletides may be used, for example, but should be at least about 10, preferably at least about 15, more preferably at least 20 nucleotides in length. When shorter length regions are used for comparison, a higher degree of sequence identity is required than for longer sequences. Shorter probes are often particularly useful for polymerase chain reactions (PCR) , especially when highly conserved sequences can be identified. (See, Gould, et al . , PNAS USA (1989) 85:1934-1938.)

When longer nucleic acid fragments are employed (>100 bp) as probes, especially when using complete or large cDNA sequences, one can still screen with moderately high stringencies (for example using 50% formamide at 37°C with minimal washing) in order to obtain signal from the target sample with 20-50% deviation, i.e., homologous sequences. (For additional information regarding screening techniques see Beltz, et al . Methods in Enzymology (1983) 200:266-285.). Once the related acyl-ACP thioesterase sequence is obtained, expression of the plant acyl-ACP thioesterase in a host cell may be obtained to further characterize the activity of the thioesterase. In this manner, additional plant acyl- ACP thioesterases having substantial activity on stearoyl-ACP may be identified for use in the methods of the present invention. As demonstrated herein by increasing the amount of stearoyl-ACP preferring thioesterase available to the plant FAS complex, an increased percentage of stearate may be provided. Additionally, by decreasing the amount of stearoyl- ACP desaturase available to the plant FAS complex in conjunction with an increase of the amount of stearoyl-ACP thioesterase available, a more marked increased percentage of stearate may be obtained.

The nucleic acid sequences which encode plant stearoyl- ACP thioesterases may be used in various constructs, for example, as probes to obtain further sequences. Alternatively, these sequences may be used in conjunction with appropriate regulatory sequences to increase levels of the respective thioesterase of interest in a host cell for recovery or study of the enzyme in vi tro or in vivo or to decrease levels of the respective thioesterase of interest for some applications when the host cell is a plant entity, including plant cells, plant parts (including but not limited to seeds, cuttings or tissues) and plants. A nucleic acid sequence encoding a plant acyl-ACP thioesterase of this invention which has substantial activity on 18:0 acyl groups may include genomic, cDNA or mRNA sequence. A cDNA sequence may or may not contain pre¬ processing sequences, such as transit peptide sequences. Transit peptide sequences facilitate the delivery of the protein to a given organelle and are cleaved from the amino acid moiety upon entry into the organelle, releasing the "mature" sequence. The use of the precursor plant acyl-ACP thioesterase DNA sequence containing the transit peptide and mature protein encoding sequences is preferred in plant cell expression cassettes. Other plastid transit peptide sequences, such as a transit peptide of seed ACP, may be employed to translocate the plant stearoyl-ACP thioesterase of this invention to various organelles of interest. Furthermore, as discussed above the complete genomic sequence of the plant stearoyl-ACP thioesterase may be obtained by the screening of a genomic library with a probe, such as a cDNA probe, and isolating those sequences which regulate expression in seed tissue. In this manner, the transcription and translation initiation regions, introns, and/or transcript termination regions of the plant stearoyl- ACP thioesterase may be obtained for use in a variety of DNA constructs, with or without the thioesterase structural gene. Once the desired plant stearoyl-ACP thioesterase nucleic acid sequence is obtained, it may be manipulated in a variety of ways. Where the sequence involves non-coding flanking regions, the flanking regions may be subjected to resection, mutagenesis, etc. Thus, transitions, transversions, deletions, and insertions may be performed on the naturally occurring sequence. In addition, all or part of the sequence may be synthesized. In the structural gene, one or more codons may be modified to provide for a modified amino acid sequence, or one or more codon mutations may be introduced to provide for a convenient restriction site or other purpose involved with construction or expression. The structural gene may be further modified by employing synthetic adapters, linkers to introduce one or more convenient restriction sites, or the like. The nucleic acid or amino acid sequences encoding a plant stearoyl-ACP thioesterase of this invention may be combined with other non-native, or "heterologous", sequences in a variety of ways. By "heterologous" sequences is meant any sequence which is not naturally found joined to the plant stearoyl-ACP thioesterase, including, for example, combinations of nucleic acid sequences from the same plant which are not naturally found joined together.

The DNA sequence encoding a plant stearoyl-ACP thioesterase of this invention may be employed in conjunction with all or part of the gene sequences normally associated with the thioesterase. In its component parts, a DNA thioesterase encoding sequence is combined in a DNA construct having, in the 5' to 3 ' direction of transcription, a transcription initiation control region capable of promoting transcription and translation in a host cell, the DNA sequence encoding plant stearoyl-ACP thioesterase and a transcription and translation termination region.

Potential host cells include both prokaryotic and eukaryotic cells. A host cell may be unicellular or found in a multicellular differentiated or undifferentiated organism depending upon the intended use. Cells of this invention may be distinguished by having a plant stearoyl-ACP thioesterase foreign to the wild-type cell present therein, for example, by having a recombinant nucleic acid construct encoding a plant stearoyl-ACP thioesterase therein.

Depending upon the host, the regulatory regions will vary, including regions from viral, plasmid or chromosomal genes, or the like. For expression in prokaryotic or eukaryotic microorganisms, particularly unicellular hosts, a wide variety of constitutive or regulatable promoters may be employed. Expression in a microorganism can provide a ready source of the plant enzyme.

For the most part, the constructs will involve regulatory regions functional in plants which provide for modified production of plant stearoyl-ACP thioesterase, and modification of the fatty acid composition. The open reading frame, coding for the plant stearoyl-ACP thioesterase or functional fragment thereof will be joined at its 5' end to a transcription initiation regulatory region such as the wild- type sequence naturally found 5 ' upstream to the thioesterase structural gene, or to a heterologous regulatory region from a gene naturally expressed in plant tissues. Examples of useful plant regulatory gene regions include those from T-DNA genes, such as nopaline or octopine synthase, plant virus genes, such as CaMV 35S, or from native plant genes.

For such applications when 5 ' upstream non-coding regions are obtained from other genes regulated during seed maturation, those preferentially expressed in plant embryo tissue, such as ACP and napin-derived transcription initiation control regions, are desired. Such "seed-specific promoters" may be obtained and used in accordance with the teachings of USPN 5,420,034 having a title "Seed-Specific Transcriptional Regulation" . Transcription initiation regions which are preferentially expressed in seed tissue, i.e., which are undetectable in other plant parts, are considered desirable for fatty acid modifications in order to minimize any disruptive or adverse effects of the gene product. Regulatory transcript termination regions may be provided in DNA constructs of this invention as well. Transcript termination regions may be provided by the DNA sequence encoding the plant stearoyl-ACP thioesterase or a convenient transcription termination region derived from a different gene source, for example, the transcript termination region which is naturally associated with the transcript initiation region. Where the transcript termination region is from a different gene source, it will contain at least about 0.5 kb, preferably about 1-3 kb of sequence 3 ' to the structural gene from which the termination region is derived.

Plant expression or transcription constructs having a plant stearoyl-ACP thioesterase as the DNA sequence of interest for increased or decreased expression thereof may be employed with a wide variety of plant life, particularly, plant life involved in the production of vegetable oils for edible and industrial uses. Most especially preferred are temperate oilseed crops. Plants of interest include, but are not limited to, rapeseed (Canola and High Erucic Acid varieties) , sunflower, safflower, cotton, Cuphea, soybean, peanut, coconut and oil palms, and corn. Depending on the method for introducing the recombinant constructs into the host cell, other DNA sequences may be required. Importantly, this invention is applicable to dicotyledons and monocotyledons species alike and will be readily applicable to new and/or improved transformation and regulation techniques.

The method of transformation is not critical to the instant invention; various methods of plant transformation are currently available. As newer methods are available to transform crops, they may be directly applied hereunder. For example, many plant species naturally susceptible to Agrobacterium infection may be successfully transformed via tripartite or binary vector methods of Agrobacterium mediated transformation. In addition, techniques of microinjection, DNA particle bombardment, eiectroporation have been developed which allow for the transformation of various monocot and dicot plant species.

In developing the DNA construct, the various components of the construct or fragments thereof will normally be inserted into a convenient cloning vector which is capable of replication in a bacterial host, e.g., E. coli . Numerous vectors exist that have been described in the literature. After each cloning, the plasmid may be isolated and subjected to further manipulation, such as restriction, insertion of new fragments, ligation, deletion, insertion, resection, etc., so as to tailor the components of the desired sequence. Once the construct has been completed, it may then be transferred to an appropriate vector for further manipulation in accordance with the manner of transformation of the host cell.

Normally, included with the DNA construct will be a structural gene having the necessary regulatory regions for expression in a host and providing for selection of transformant cells. The gene may provide for resistance to a cytotoxic agent, e.g. antibiotic, heavy metal, toxin, etc., complementation providing prototrophy to an auxotrophic host, viral immunity or the like. Depending upon the number of different host species the expression construct or components thereof are introduced, one or more markers may be employed, where different conditions for selection are used for the different hosts. A number of markers have been developed for use for selection of transformed plant cells, such as those which provide resistance to various antibiotics, herbicides, or the like. The particular marker employed is not essential to this invention, one or another marker being preferred depending on the particular host and the manner of construction.

As mentioned above, the manner in which the DNA construct is introduced into the plant host is not critical to this invention. Any method which provides for efficient transformation may be employed. Various methods for plant cell transformation include the use of Ti- or Ri-plasmids, microinjection, eiectroporation, DNA particle bombardment, liposome fusion, DNA bombardment or the like. In many instances, it will be desirable to have the construct bordered on one or both sides by T-DNA, particularly having the left and right borders, more particularly the right border. This is particularly useful when the construct uses A . tumefaciens or A . rhizogenes as a mode for transformation, although the T- DNA borders may find use with other modes of transformation. Once a transgenic plant is obtained which is capable of producing seed having a modified fatty acid composition, traditional plant breeding techniques, including methods of mutagensis, may be employed to further manipulate the fatty acid composition. Alternatively, additional foreign fatty acid modifying DNA sequence may be introduced via genetic engineering to further manipulate the fatty acid composition. One may choose to provide for the transcription or transcription and translation of one or more other sequences of interest in concert with the expression of a plant stearoyl-ACP thioesterase in a plant host cell. In particular, the reduced expression of stearoyl-ACP desaturase in combination with expression of a plant stearoyl-ACP thioesterase may be preferred in some applications. Stearoyl-ACP thioesterases may also be used in combination with other thioesterase genes with differing specificities. For example, a transgenic oilseed crop expressing both a lauroyl-ACP thioesterase (WO 92/20236) and a stearoyl-ACP thioesterase during seed development will produce an oil enriched in both lauric and stearic acids. Similarly, a transgenic oilseed crop expressing both a palmitoyl-ACP thioesterase (WO 95/13390) and a stearoyl-ACP thioesterase during seed development will produce an oil enriched in both palmitic and stearic acids. These thioesterase gene constructs may be linked to each other in the genome of the transgenic plant or may be unlinked. Conversely the thioesterase genes may be combined in the same transgenic plant by generating plants with one or the other thioesterase gene and subsequently crossing two plants, one of each type. By selecting parents for such crosses, it is possible to further manipulate the relative ratios of desired fatty acids in a seed oil. When one wishes to provide a plant transformed for the combined effect of more than one nucleic acid sequence of interest, typically a separate nucleic acid construct will be provided for each. The constructs, as described above contain transcriptional or transcriptional and translational regulatory control regions. The constructs may be introduced into the host cells by the same or different methods, including the introduction of such a trait by crossing transgenic plants via traditional plant breeding methods, so long as the resulting product is a plant having both characteristics integrated into its genome.

By decreasing the amount of stearoyl-ACP desaturase, an increased percentage of saturated fatty acids is provided. Using anti-sense, transwitch, ribozyme or some other stearoyl- ACP desaturase reducing technology, a decrease in the amount of stearoyl-ACP desaturase available to the plant cell is produced, resulting in a higher percentage of saturates such as one or more of laurate (C12:0), myristate (C14.-0), palmitate (C16:0), stearate (C18:0), arachidate (C20:0), behenate (C22:0) and lignocerate (C24:0) . In rapeseed reduced stearoyl-ACP desaturase results in increased stearate levels and total saturates (Knutzon et al . (1992) Proc. Nat . Acad. Sci 8.9:2264-2628) . A novel construct is also provided herein which may be used in antisense reduction of stearoyl-ACP desaturase to increase both the percentage of stearate which may be obtained and the percentage of primary transformants expressing the increased stearate trait.

Of special interest is the production of triglycerides having increased levels of stearate or palmitate and stearate. In addition, the production of a variety of ranges of such saturates is desired. Thus, plant cells having lower and higher levels of stearate fatty acids are contemplated. For example, fatty acid compositions, including oils, having a 10% level of stearate as well as compositions designed to have up to an appropriate 60% level of stearate or other such modified fatty acid(s) composition are contemplated.

Oils with increased percentages of stearate, are desired. Increased stearate percentages (by weight) ranging from native levels to increases of up to 25 fold are described. By manipulation of various aspects of the DNA constructs (e.g., choice of promoters, number of copies, etc.) and traditional breeding methods, one skilled in the art may achieve even greater levels of stearate. By combination of the plant stearoyl-ACP desaturase sequence in combination with the expression of a plant stearoyl-ACP thioesterase in seed tissue, an increased percentage of stearate can be achieved in rapeseed and other plant species. DNA sequence of C. tinctorius stearoyl-ACP desaturase gene, as well as DNA sequences of stearoyl-ACP desaturase genes from a Ricinus, a Brassica and a Simmondsia plant are found in WO 91/13972.

The invention now being generally described, it will be more readily understood by reference to the following examples which are included for purposes of illustration only and are not intended to limit the present invention.

EXAMPLES

Example 1 Mangosteen Thioesterase Gene Sequences

A cDNA bank is prepared from seeds extracted from mature mangosteen fruit using the methods as described in Stratagene Zap cDNA synthesis kit (Stratagene; La Jolla, CA) . Oil analysis of the mangosteen tissues used for RNA isolation reveals 18:0 levels of approximately 50%. Oil analysis of seeds from less mature mangosteen fruit reveals 18:0 levels of

20-40%. Total RNA is isolated from the mangosteen seeds by modifying the CTAB DNA isolation method of Webb and Knapp

( Plant Mol . Biol . Reporter (1990) 8:180-195) . Buffers include:

REC: 50 mM TrisCl pH , 0.7 M NaCl, 10 mM EDTA pH8,

0.5% CTAB. REC+: Add B-mercaptoethanol to 1% immediately prior to use. RECP: 50 mM TrisCl pH9, 10 mM EDTA pH8, and 0.5%

CTAB. RECP+: Add B-mercaptoethanol to 1% immediately prior to use.

For extraction of 1 g of tissue, 10ml of REC+ and 0.5 g of PVPP is added to tissue that has been ground in liquid nitrogen and homogenized. The homogenized material is centrifuged for 10 min at 12000 rpm. The supernatant is poured through miracloth onto 3ml cold chloroform and homogenized again. After centrifugation, 12,000 RPM for 10 min, the upper phase is taken and its volume determined. An equal volume of RECP+ is added and the mixture is allowed to stand for 20 min. at room temperature. The material is centrifuged for 20 min. at 10,000 rpm twice and the supernatant is discarded after each spin. The pellet is dissolved in 0.4 ml of 1 M NaCl (DEPC) and extracted with an equal volume of phenol/chloroform. Following ethanol precipitation, the pellet is dissolved in 1 ml of DEPC water. Briefly, the cloning method for cDNA synthesis is as follows. First strand cDNA synthesis is according to Stratagene Instruction Manual with some modifications according to Robinson, et al . (Methods in Molecular and Cellular Biology (1992) 3:118-127) . In particular, approximately 57μg of LiCl precipitated total RNA was used instead of 5μg of poly(A)+ RNA and the reaction was incubated at 45°C rather than 37°C for 1 hour.

Probes for library screening are prepared by PCR from mangosteen cDNA using oligonucleotides to conserved plant acyl-ACP thioesterase regions. Probe Garm 2 and Garm 106 are prepared using the following oligonucleotides. The nucleotide base codes for the below oligonucleotides are as follows:

A = adenine C = cytosine T = thymine U = uracil

G = guanine S = guanine or cytosine

K = guanine or thymine W = adenine or thymine

M = adenine or cytosine R = adenine or guanine

Y = cytosine or thymine B = guanine, cytosine or thymine

H = adenine, cytosine or thymine

N = adenine, cytosine, guanine or thymine

Garm 2 4874: 5' CUACUACUACUASYNTVNGYNATGATGAA 3' (SEQ ID NO:9)

4875: 5' CAUCAUCAUCAURCAYTCNCKNCKRTANTC 3' (SEQ ID NO:10) Primer 4874 is a sense primer designed to correspond to possible encoding sequences for conserved peptide V/L/A W/S/Y V/A M M N, where the one letter amino acid code is used and a slash between amino acids indicates more than one amino acid is possible for that position. Primer 4875 is an antisense primer designed to correspond to possible encoding sequences for peptide D/E Y R R E C.

Qarπ 106

5424: 5' AUGGAGAUCUCUGAWCRBTAYCCTAMHTGGGGWGA 3' (SEQ ID NO:11)

5577: 5' ACGCGUACUAGUTTNKKNCKCCAYTCNGT 3' (SEQ ID NO:12) Primer 5424 is a sense primer designed to correspond to possible encoding sequences for peptide E/D H/R Y P K/T W G D. Primer 5577 is an antisense primer designed to correspond to possible encoding sequences for peptide T E W R K/P K.

The DNA fragments resulting from the above reactions are amplified for use as probes by cloning or by further PCR and radiolabeled by random or specific priming.

Approximately 800,000 plaques are plated according to manufacturer's directions. For screening, plaque filters are prehybridized at room temperature in 50% formamide, 5X SSC, 10X Denhardt's, 0.1% (w/v) SDS, 5mM Na₂EDTA, 0.lmg/ml denatured salmon sperm DNA. Hybridization with a mixture of the Garm 2 and Garm 106 probes is conducted at room temperature in the same buffer as above with added 10% (w/v) dextran sulfate and probe. Plaque purification and phagemid excision were conducted as described in Stratagene Zap cDNA Synthesis Kit instructions.

Approximately 90 acyl-ACP thioesterase clones were identified and sorted as to thioesterase type by DNA sequencing and/or PCR analysis. Of the analyzed clones, at least 28 were Class I (FatA) types, and 59 were Class II (FatB) types. Two subclasses of FatA type clones were observed, the most prominent type is termed GarmFatAl and the single clone of the second subclass is termed GarmFatA2. DNA and translated amino acid sequence of GarmFatAl clone C14-4 (pCGN5252) (SEQ ID NO:7) is presented in Figure 2. DNA sequence and translated amino acid sequence of the FatA2 clone C14-3 (SEQ ID NO:8) is presented in Figure 3.

Constructs for expression of the Figure 2 Garm FatAl clone in E. coli are prepared as follows. Restriction sites are inserted by PCR mutagenesis at amino acid 49 (SacI) , which is near the presumed mature protein amino terminus, and following the stop codon for the protein encoding region (BamHl) . The mature protein encoding region is inserted as a SacI/BamHl fragment into pBC SK (Stratagene; La Jolla, CA) resulting in pCGN5247, which may be used to provide for expression of the mangosteen thioesterase as a lacZ fusion protein.

Results of thioesterase activity assays on mangosteen Class I thioesterase clone GarmFatAl using 16:0, 18:0 and 18:1 acyl-ACP substrates are shown below.

Acyl-ACP Thioesterase activity (cpm/min)

16:0 18:0 18:1

Control 1400 3100 1733 GarmFatAl 4366 23916 87366

The GarmFatAl cone demonstrates preferential activity on C18:l acyl-ACP substrate, and also demonstrates substantial activity (approximately 25% of the 18:1 activity) on C18.-0 acyl-ACP substrates. Only a small increase in C16:0 activity over activity in control cells is observed, and the 16:0 activity represents only approximately 3% of the 18:1 activity. Expression of GarmFatA2 thioesterase in E. coli and assay of the resultant thioesterase activity demonstrates that C18:l is highly preferred as the acyl-ACP substrate. The thioesterase activity on 16:0 and 18:0 acyl-ACP substrates are approximately equal and represent less than 5% of the observed 18:1 activity.

Expression of Class II type mangosteen thioesterase clones in E. coli demonstrates that 16:0 is highly preferred over other acyl-ACP substrates.

Example 2 Plant Transformation methods

A. AgroJbacterium-mediated Transformation

Methods which may be used for AgroJbacterium-mediated transformation of Brassica are described by Radke et al . ( Theor. Appl . Genet . (1988) 75:685-694; Plant Cell Reports (1992) 22:499-505).

Transgenic Arabidopsis thaliana plants may be obtained by AgroJbacterium-mediated transformation as described by Valverkens et al . , ( Proc. Nat . Acad. Sci . (1988) 85:5536- 5540) . B. Particle Bombardment

DNA sequences of interest may be introduced as expression cassettes, comprising at least a promoter region, a gene of interest, and a termination region, into a plant genome via particle bombardment as described for example in European Patent Application 332 855 and in co-pending application USSN 07/225,332, filed July 27, 1988.

Briefly, tungsten or gold particles of a size ranging from 0.5mM-3mM are coated with DNA of an expression cassette. This DNA may be in the form of an aqueous mixture or a dry DNA/particle precipitate. Tissue used as the target for bombardment may be from cotyledonary explants, shoot meristems, immature leaflets, or anthers. The bombardment of the tissue with the DNA-coated particles is carried out using a Biolisticsfe particle gun (Dupont; Wilmington, DE) . The particles are placed in the barrel at variable distances ranging from lcm-14cm from the barrel mouth. The tissue to be bombarded is placed beneath the stopping plate; testing is performed on the tissue at distances up to 20cm. At the moment of discharge, the tissue is protected by a nylon net or a combination of nylon nets with mesh ranging from lOmM to 300mM.

Following bombardment, plants may be regenerated following the method of Atreya, et al . , (Plant Science Letters (1984) 34:379-383). Briefly, embryo axis tissue or cotyledon segments are placed on MS medium (Murashige and Skoog, Physio . Plant . (1962) 15:473) (MS plus 2.0 mg/1 6-benzyladenine (BA) for the cotyledon segments) and incubated in the dark for 1 week at 25 ± 2°C and are subsequently transferred to continuous cool white fluorescent light (6.8 W/m²) . On the 10th day of culture, the plantlets are transferred to pots containing sterile soil, are kept in the shade for 3-5 days are and finally moved to greenhouse. The putative transgenic shoots are rooted. Integration of exogenous DNA into the plant genome may be confirmed by various methods know to those skilled in the art.

Example 3 Construct for Expression of Garm FatAl Thioesterase in Plant Seeds

The napin expression cassette, pCGN1808, is described in

USPN 5,420,034 which is incorporated herein by reference. pCGNl808 is modified to contain flanking restriction sites to allow movement of only the expression sequences and not the antibiotic resistance marker to binary vectors such as pCGN1557 (McBride and Summerfelt (1990) Plant Mol . Biol . 14:269-276) . Synthetic oligonucleotides containing Xpnl, NotI and Hindlll restriction sites are annealed and ligated at the unique Hmdlll site of pCGΝ1808, such that only one Hindlll site is recovered. The resulting plasmid, pCGN3200 contains unique Hindlll, NotI and Xp.nl restriction sites at the 3 ' -end of the napin 3 ' -regulatory sequences as confirmed by sequence analysis. The majority of the napin expression cassette is subcloned from pCGΝ3200 by digestion with Hindlll and SacI and ligation to Hindlll and SacI digested pIC19R (Marsh, et al . (1984) Gene 32:481-485) to make pCGN3212. The extreme 5'- sequences of the napin promoter region are reconstructed by PCR using pCGN3200 as a template and two primers flanking the SacI site and the junction of the napin 5 ' -promoter and the pUC backbone of pCGN3200 from the pCGNl808 construct. The forward primer contains Clal, Hindlll, NotI , and Xpnl restriction sites as well as nucleotides 408-423 of the napin 5'-sequence (from the EcoRV site) and the reverse primer contains the complement to napin sequences 718-739 which include the unique SacI site in the 5" -promoter. The PCR was performed using in a Perkin Elmer/Cetus thermocycler according to manufacturer's specifications. The PCR fragment is subcloned as a blunt-ended fragment into pUC8 (Vieira and

Messing (1982) Gene 2.9:259-268) digested with Hindi to give pCGΝ3217. Sequence of pCGN3217 across the napin insert verifies that no improper nucleotides were introduced by PCR. The napin 5-sequences in pCGN3217 are ligated to the remainder of the napin expression cassette by digestion with Clal and SacI and ligation to pCGN3212 digested with Clal and SacI . The resulting expression cassette pCGN3221, is digested with Hindlll and the napin expression sequences are gel purified away and ligated to pIC20H (Marsh, supra) digested with Hindlll. The final expression cassette is pCGN3223, which contains in an ampicillin resistant background, essentially identical 1.725 napin 5' and 1.265 3' regulatory sequences as found in pCGNl808. The regulatory regions are flanked with Hindlll, NotI and Xpnl restriction sites and unique Sail ,

Bglll , Pstl , and Xhol cloning sites are located between the 5' and 3' noncoding regions. pCGΝ3223 is also described in WO 92/20236 which is incorporated herein by reference.

Mangosteen acyl-ACP thioesterase clone pCGN5252 (Garm FatAl) is digested with Pstl and Xhol and cloned into

Pstl/Xhol digested pCGN3223 resulting in pCGN5253 containing the mangosteen thioesterase encoding region positioned for transcriptional control from the napin promoter. The Pstl site in pCGN5252 is located in the cloning vector at the 5 ' end of the cDNA clone, and the Xhol site is located in the 3' untranslated region at nucleotides 1233-1238.

For Agrobacterium-mediated plant transformation, pCGN5253 is digested with Asp718 and cloned into Asp718 digested pCGN1578, a binary vector for plant transformation, resulting in clone pCGN5255. To insert a repeat of the napin 5' /garm FatAl/napin 3 ' into the binary vector, pCGN5253 is digested with Asp718 and the napin 5 ' /Garm FatAl/napin 3' region gel purified. The gel purified fragment is then ligated into Asp718 digested pCGN5255, resulting in clone pCGN5266 pCGN5255 is transformed into Agrobacterium tumefaciens strain EHA101 and used to transform a high oleic acid line of Brassica napus . pCGN5266 is transformed into Agrobacterium tumefaciens strain EHA101 and used to transform B. napus variety Quantum (SP30021) and a low linolenic B. napus line, Q04 (LP004) . Example 4 Analysis of Transgenic Plants Expressing Mangosteen Thioesterase

A. Analysis of 5255-Transformants

Pooled segregating mature seeds from pCGN5255 transformants are analyzed to determine fatty acid composition by GC as described in WO 92/20236. In the 22 plants analyzed, 18:0 (stearate) levels ranged from 2.8 to 14.2 weight percent, compared to background levels in a nontransformed plant of 2.2 percent. Analyses of pooled seed samples from additional transformants were conducted, and 18:0 levels of 16.6 weight percent were detected in transformant 5255-29. Levels of 16:0 fatty acids are not affected by expression of the Garm FatAl thioesterase.

Half seeds from transgenic plants 5255-20 and 5255-3 were similarly analyzed to determine the levels of 18:0 fatty acids obtained in individual seeds. For event 5255-20, half seeds with 18:0 fatty acid contents of up to 22.6 weight percent were obtained, and from event 5255-3, half seeds with 18:0 fatty acid contents of up to 10.0 weight percent were obtained. Similar data were obtained with other 5255 transgenic events and revealed 18:0 fatty acid levels of up to 26.6% in half seeds of 5255-29 and up to 15% in half seeds of 5255-19.

The highest half-seed plants were grown to maturity and T3 pooled seed was analyzed for fatty acid composition.

Stearate levels up to 26% in pooled seed were observed in selections from transformant 5255-29. Half-seeds from the T3 pools of event 5255-20 showed individuals with stearate contents up to 39%. Selections from event 5255-3, which contains the transgene in only one locus, had T3 pooled seed with the same stearate content (11%) as the half-seeds from which they were selected, as would be expected if the insertion were homozygous. The trait is being inherited in a Mendelian fashion and the stearate levels are being maintained.

B. Analysis of 5266-Transformants

Pooled segregating mature seeds from pCGN5266 transformants are analyzed to determine fatty acid composition by GC as described in WO 92/20236. Results of these assays for the 10 transformants in each variety with the highest stearate levels are presented in Figure 4 as weight percent fatty acid composition. In plants analyzed to date, 18:0 levels range from approximately 4 to 22 weight percent, compared to background levels in nontransformed plants of 1.4 (Q04) and 1.8 (Quantum) percent.

Half seeds from transgenic plants 5266-LP004-2 and 5266- SP30021-29 were similarly analyzed to determine the levels of 18:0 fatty acids obtained in individual seeds. Fatty acid composition data from the ten seeds of those analyzed which contained the highest levels of stearate are presented in Figure 5. For event 5266-LP004-2, half seeds with 18:0 fatty acid contents of up to 33.13 weight percent were obtained, and from event 5266-SP30021-29, half seeds with 18:0 fatty acid contents of up to 38.86 weight percent were obtained. Similar data were obtained with other 5266 transgenic events.

Example 5 Construct for Antisense Regulation of Stearoyl-ACP Desaturase in Brassica

Brassica napus stearoyl-ACP desaturase clones were isolated as follows. A cDNA library was constructed from RNA isolated from mid maturation (30 days after pollination) seeds of Brassica napus cultivar 212/86. The library was constructed using the lambda Uni-ZAP (Stratagene) vector kit according to the manufacturers directions with the following modification: lOOμg of total RNA was used for cDNA synthesis, and first strand cDNA synthesis was carried out at 42^*C. The cDNA library was screened with the coding region of a delta-9 desaturase gene isolated from Brassica campestris (Knutzon et al . (1992) Proc. Natl Acad. Sci . 85:2624-2628) . Partial DNA sequence was obtained from 42 clones that hybridized with the probe. The clones fell into 5 classes. DNA sequence was obtained from the largest clone from each class. BND9 and BND16 were most closely related and 78.6% of the cDNA clones were of this class. BND11 and BND53 were related, and 19% of the cDNA clones were of this class.

An antisense gene was constructed to generate antisense RNA homologous to both of the major classes of desaturase genes, BND9 and BND11. A fragment containing the majority of the coding region of BND9 was excised using the enzymes Hindlll and Pvull. The ends were filled in with the klenow fragment of DNA Polymerase I and deoxynucleotides. Plasmid pCGN3223, containing a napin expression cassette, was linearized with Bglll and the ends were filled in with the klenow fragment of DNA Polymerase I and deoxynucleotides. Ligation of the pCGN3223 vector and the BND9 fragment resulted in pCGN7826. pCGN7826 was digested with Xhol and the ends were filled in with the klenow fragment of DNA Polymerase I and deoxynucleotides. The filled in fragment isolated from BNDll after digestion with Bglll and Pvull was ligated to PCGN7826 to yield pCGN7690. pCGN7690 contains a napin promoter positioned for expression of an antisense transcript homologous to both the BND9 and BNDll mRNAs. The antisense gene was excised from pCGN7690 using Asp718 and cloned into Asp718 digested binary plant transformation vector PCGN1559PASS to yield pCGN7696. pCGNl559PASS is a binary vector such as those described by McBride et al . ( supra) and is prepared from pCGN1559 by substitution of the pCGN1559 linker region with a linker region containing the following restriction digestion sites: Asp718/Asd/ Pad /Xbal/ BamHl /Swal/Sse8387 (Pstl) /Hindlll. Ascl, Pad , Swa I and Sse8387 have 8-base restriction recognition sites and are available from New England BioLabs: Ascl, Pa ; Boehringer Manheim: Swal and Takara (Japan) : Sse8387.

Example 6 Analysis of Transgenic Plants Expressing an Antisense Desaturase Construct

Approximately 70 transgenic plants containing pCGN7696 were generated. Pools of 50 seeds were analyzed using gas chromatography to determine the fatty acid composition of the seed oils. The average level of stearate in pooled seeds from these transgenic plants was 7.7%. Pooled seeds from plant

7696-31 contained 14.7% (by weight) stearate, pooled seeds from plant 7696-54 contained 14.1% stearate, pooled seeds from plant 7696-69 contained 11.4% stearate, pooled seeds from plant 7696-36 contained 11.3% stearate, and pooled seeds from plant 7696-45 contained 9.5% stearate. By comparison, pooled seeds from untransformed control plants contain less than 3.0% stearate. Also, in pooled seeds from 20 transgenic plants containing an antisense construct containing only the BND9 desaturase sequence under control of the napin promoter, an average of 3.6% stearate in the pooled seeds was observed. This data demonstrates the dual antisense desaturase gene construct, pCGN7696, provides for improved stearate production as compared to a single gene construct. Additional data confirm these results as the highest stearate level observed in pooled seed from a BND9 transformant was 7.6 % as compared to 14.7% with 7696 transformants. Similarly, the highest single seed level of stearate observed with BND9 transformants was 10.1% in comparison to levels of up to 29% obtained in 7696 transformants.

Segregation analysis of the oil composition of single seeds suggested that 7696-54 and 7696-31 contained two T-DNA loci, while 7696-36, 7696-69 and 7696-45 contained single T- DNA loci. Individual seeds were analyzed to determine stearate ranges for individual transformants. Stearate levels in single seeds from 7696-54 and 7696-31 ranged from 2 to 29 weight percent while single seeds from the single locus plants contained stearate contents ranging between 2 and 20 percent by weight.

Example 7 Crosses and Analysis of Plants Containing Both Antisense Desaturase and Mangosteen Thioesterase

Constructs

Half-seed analysis of T2 seeds from 5255 and 7696 transformants were used to select individual transformants to be crossed for combination of the stearoyl-ACP thioesterase and antisense desaturase traits. Transgenic plants were grown from the remaining half-seed where desirable levels of 18:0 and/or numbers of GarmFatAl gene inserts^' were observed. 18:0 contents of representative half seeds for 5255-3, 5255-19, 5255-29 and 5255-20 events and 7696-31, 7696-36, 7696-45 and 7696-54 events selected for crossing experiments are provided in Table 1 below. TABLE 1

1/2 Seed 1/2 Seed

5255 Event 18:0 % 7696 Event 18:0 %

5255-29-7 21.31 7696-54-11 29.30

5255-20-11 20.33 *7696-45-l 17.65

*5255-3-29 8.11 *7696-36-l 29.62

*5255-19-l 14.95 7696-31-1 28.12

* Single insertions by segregation analysis

Data from various Fl crosses of the transgenic events indicated above are provided in Table 2. The highest stearate levels observed by half-seed analysis of seeds from selected Fl crosses are shown as weight percent of total fatty acids.

TABLE 2

Parent 5255-29-7 5255-20-11 5255-3-29 5255-19-1

7696-54-11 47.14 42.40 35.46 40.28

7696-45-1 18.77 - 7.61 13.39

7696-36-1 34.03 - 21.33 -

7696-31-1 31.84 - - -

The above results demonstrate the existence of numerous half seeds with stearate contents above 30% (as high as 47%) from crosses of multiple loci plants from each transgene as well as from single locus events. The combination of antisense desaturase and mangosteen TE gives higher stearate than either hemizygous parent. In some cases the Fl 's have higher stearate than seen so far in either homozygous parent suggesting the genes are acting synergistically. Example 8 Transformation and Transσenic Plant Analysis with a Construct Containing Antisense Desaturase and Manσosteen Thioesterase Constructs

As an alternative to separate transformations and plant breeding as described above to obtain transgenic plants containing both antisense stearoyl-ACP desaturase and stearoyl thioesterase genes, such transgenic plants may be obtained by transformation with a single construct containing both antisense stearoyl-ACP desaturase and stearoyl thioesterase genes. One such construct is the binary transformation vector pCGN7748. pCGN7748 contains 2 copies of napin expression cassettes with antisense desaturase BND9 and BNDll genes. It also contains one copy of an expression cassette containing antisense desaturase BND9 and BNDll genes under the regulatory control of a stearoyl-ACP desaturase gene promoter, and one copy of a napin expression cassette for expression of the mangosteen thioesterase gene. pCGN7748 was constructed by cloning two copies of the napin/antisense BND9/ antisense BNDll gene from pCGN7690 (see Example 5) as Asp718 fragments into Asp718 digested binary vector pCGN1559PASS to yield pCGN7859. The napin/mangosteen thioesterase gene was excised from pCGN5253 (see Example 3) using Asp718 and the ends filled in with the klenow fragment of DNA polymerase one and all 4 dNTPs. The resulting blunt ended DNA fragment was cloned into Swal digested pCGN7859 to yield plasmid pCGN7743. The DNA fragment containing BND9/BNDll from pCGN7690 was excised with Sail and Xhol, and treated with the klenow fragment of DNA polymerase one and the nucleotides dCTP and TTP and ligated to the desaturase expression cassette of pCGN5207, which had been digested with BamHl and treated with klenow fragment of DNA polymerase one and the nucleotides dGTP and dATP. [pCGN5207 contains approximately 1.5Kb of the 5' regulatory region and 1.3 kB of the 3 ' regulatory region of the Brassica rapa stearoyl-ACP desaturase gene (Knutzon et al . (1992) Proc. Nat . Acad. Sci . 85:2624-2628) . A polylinker containing BamHl-Pstl -NotI-Xbal- Nael-JScoRI-Clal restriction sites separates the 5" and 3' regulatory regions in pCGΝ5207.] The clone resulting from insertion of the BND9/BNDll fragment in the desaturase expression cassette was pCGN7745. The Smal fragment from pCGN7745 was cloned into pCGN7743 digested with Sse8387 (after treating with the klenow fragment of DNA polymerase one and all 4 dNTPs) to yield pCGN7748. pCGN7748 is transformed into Agrobacterium tumefaciens strain EHA101 and used to transform B. napus variety Quantum (SP30021) .

T2 pooled seed from a plant transformed with pCGN7748 was analyzed to determine fatty acid composition and demonstrated the following fatty acid composition profile (weight percent fatty acids) :

7748-SP30021-1

%16 : 0 = 4 . 8 %18 : 0 = 43 . 8 %20 : 0 = 6 .4

%16 : 1 = 0 . 1 %18 : 1 = 16 . 6 %20 : 1 = 0 .2

%18:2 = 19.5 %22:0 = 1.4

%18:3 = 7.2

These results demonstrate that mangosteen thioesterase clone GarmFatAl may be used to increase the 18:0 content of seed oils from transgenic plants, and that improved vegetable oils having a stearic acid content of greater than 30 weight percent may be obtained from such seeds following crushing and fatty acid extraction procedures. Furthermore, in plants expressing the mangosteen thioesterase in combination with an antisense desaturase construct, levels of 18:0 can be further increased and may exceed 50% of the total percentage of fatty acids in the plant seed oil. All publications and patent applications mentioned in this specification are indicative of the level of skill of those skilled in the art to which this invention pertains. All publications and patent applications are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be obvious that certain changes and modifications may be practiced within the scope of the appended claims.

Claims

What is claimed is:

1. A method of increasing the levels of 18:0 fatty acids in plant seed triglycerides, wherein said method comprises: growing a plant having integrated into its genome a DNA construct, said construct comprising in the 5' to 3' direction of transcription, a promoter functional in a plant seed cell, a DNA sequence encoding an acyl-ACP thioesterase protein having substantial activity on C18:0 acyl-ACP substrates and little or no activity on C16:0 acyl-ACP substrates, and a transcription termination region functional in a plant cell.

2. The method of Claim 1 wherein said plant is an oilseed crop plant.

3. The method of Claim 2 wherein said oilseed crop plant is a Brassica plant.

4. The method of Claim 1 wherein said thioesterase encoding sequence is from a mangosteen plant.

5. The method of Claim 4, wherein said DNA sequence encodes the mangosteen GarmFatAl thioesterase presented in Figure 2.

6. The method of Claim 1, wherein said promoter is from a gene preferentially expressed in plant seed tissue.

7. The method of Claim 1, wherein said plant seed triglycerides comprise at least 20 mole percent C18:0 fatty acyl groups.

8. The method of Claim 1, wherein said plant seed triglycerides comprise at least 30 mole percent C18:0 fatty acyl groups.

9. The method of Claim 1, wherein said plant has integrated into its genome a second DNA construct, wherein said second DNA construct provides for antisense expression of a stearoyl-ACP desaturase gene native to said plant.

10. The method of Claim 9, wherein said plant seed triglycerides comprise at least 40 mole percent C18:0 fatty acyl groups.

11. The method of Claim 9, wherein said plant is an oilseed Brassica plant, said thioesterase encoding sequence is from a mangosteen plant and said second DNA construct is PCGN7696.

12. A plant seed comprising a minimum of 20 mole percent stearate in total fatty acids, wherein said stearate is incorporated into at least one position of a triglyceride molecule and wherein wild-type seed of said plant contains less than 5.0 mole percent stearate in fatty acids.

13. The seed of Claim 12 comprising a minimum of about 30 mole percent stearate in fatty acids.

14. The seed of Claim 12 comprising a minimum of about 40 mole percent stearate in fatty acids.

15. Plant seed oil, wherein a minimum of 20 mole percent of the acyl groups of said oil are stearate acyl groups, and wherein said oil is derived from a seed of Claim 12.

16. Plant seed oil, wherein a minimum of 40 mole percent of the acyl groups of said oil are stearate acyl groups, and wherein said oil is derived from a seed of Claim 12.

17. A Brassica seed comprising a minimum of 20 mole percent stearate in total fatty acids.

18. A Brassica seed comprising a minimum of 40 mole percent stearate in total fatty acids.

19. Plant seed oil, wherein a minimum of 20 mole percent of the acyl groups of said oil are stearate acyl groups, and wherein said oil is derived from a Brassica seed of Claim 17.

20. Plant seed oil- wherein a minimum of 40 mole percent of the acyl groups of said oil are stearate acyl groups, and wherein said oil is derived from a Brassica seed of Claim 18.

21. A construct for antisense regulation of a plant stearoyl-ACP desaturase gene, wherein said construct comprises a promoter providing for expression in a plant seed of an antisense transcript comprising DNA sequence from at least two plant stearoyl-ACP desaturase genes.

22. The construct of Claim 21, wherein said two plant stearoyl-ACP desaturase DNA sequences are from two different classes of genes from a stearoyl-ACP desaturase gene family.

23. The construct of Claim 21 represented by pCGN7696.

24. A method for increasing the stearate content in transgenic plant seeds by antisense regulation of a native stearoyl-ACP desaturase gene, the improvement of which comprises: growing a plant comprising a construct for expression in plant seed tissues, wherein said construct provides for transcription of an antisense DNA sequence comprising portions of at least two plant stearoyl-ACP desaturase genes.