DE3485921T2 - PLASMIDES FOR TRANSFORMING PLANT CELLS. - Google Patents

Description

Technical area

This invention lies in the field of genetic manipulation, plant biology and bacteriology.

Technical background

Over the last decade, the science of genetic engineering has developed rapidly. A number of methods are known for inserting a heterologous gene into bacteria, thereby making the bacteria capable of efficiently expressing the inserted genes. Such methods usually involve the use of plasmids that can be cleaved at one or more selected sites by restriction endonucleases. Typically, a gene of interest is obtained by cleaving a piece of DNA and the resulting DNA fragment is mixed with a fragment obtained by cleaving a vector, such as a plasmid. The different DNA strands are then joined together ("joined") to form a reconstituted plasmid; see, for example, U.S. Patents 4,237,224 (Cohen and Boyer, 1980), 4,264,731 (Shine, 1981), 4,273,875 (Manis, 1981), 4,322,499 (Baxter et al., 1982), and 4,336,336 (Silhavy et al., 1982).

A number of other references are available. Some of these describe the natural process by which DNA is transcribed into mRNA and mRNA is translated into protein, see L. Stryer, Biochemistry, 2nd ed., pp. 559 ff. (W.H. Freeman and Co., 1981); A.L. Lehninger, Biochemistry, 2nd ed., pp. 853 ff. (Worth Publ., 1975). Others describe the processes and products of genetic manipulation, see, for example, T. Maniatis et al., Molecular Cloning, A Laboratory Manual (Cold Spring Harbor Labs, 1982); J.K. Setlow and A. Hollaender, Genetic Engineering, Principles and Methods (Plenum Press, 1979). In the following, all references are cited in abbreviated form; a list of full citations appears after the examples.

Most of the genetic manipulation work currently being done involves the insertion of genes into various types of cells, mainly bacteria such as E. coli, various other microorganisms such as yeast, and mammalian cells. However, many of the techniques used for genetic manipulation of animal cells and Techniques and substances used to manipulate microorganisms are not directly applicable to genetic manipulation involving plants.

The term "plant" as used herein refers to a multicellular differentiated organism capable of photosynthesis, such as angiosperms and multicellular algae. This term does not include microorganisms such as bacteria, yeast and fungi. The term "plant cell" includes any cell derived from a plant; it includes undifferentiated tissue such as callus or crown gall tumor, as well as plant seeds, offshoots, pollen or plant embryos.

Ti and Ri plasmids

The tumor-inducing (Ti) plasmid of Agrobacterium tumefaciens has been proposed for use as a natural vector for introducing foreign genetic information into plant cells (Hernalsteen et al., 1980, Rorsch and Schilperoort, 1978). Certain species of A. tumefaciens can infect a wide variety of plant cells, causing crown gall disease. The infection process is not fully understood. At least part of the Ti plasmid enters the plant cell. Various metabolic changes occur and part of the Ti plasmid is inserted into the plant genome (probably into the chromosomes). The part of the Ti plasmid that enters the plant genome is called "transferred DNA" (T-DNA). T-DNA is stably maintained in plant DNA (Chilton et al., 1977; Yadev et al., 1980, Willmitzer et al., 1980; Otten et al., 1981).

Research conducted by various laboratories has led to the characterization of several structural genes (i.e., protein-coding genes) located in T-DNA (Garfinkel et al., 1981; Leemans et al., 1982), as well as other DNA sequences that appear to serve various other functions. For example, two sequences called the "left border" and "right border" appear to delinearize T-DNA, and may be involved in the process by which T-DNA is transferred into plant chromosomes (Zambryski et al., 1982).

Another species of Agrobacterium, A. rhizogenes, carries a "root-inducing" (Ri) plasmid similar to the Ti plasmid Infection of a plant cell by A. rhizogenes causes hairy root disease. Like the Ti plasmid, a segment of DNA called "T-DNA" (also called "R-DNA" by some researchers) is transferred into the plant genome of an infected cell.

There are also reports of several other bacteria capable of causing genetic transformation of plant cells, including A. rubi and certain bacteria of the genus Rhizobia that have been treated with a mutagenic agent; Hooykaas et al., at page 156 of Setlow and Hollaender, 1979.

The term "Ti plasmid" as used herein includes any plasmid (1) contained in a microorganism other than a virus that is capable of causing genetic transformation of one or more plant species or plant cells, and (2) that contains a DNA segment inserted into a plant genome. It includes Ri plasmids.

The term "T-DNA" as used herein refers to a DNA segment in or from a Ti plasmid (1) that has been inserted into the genome of one or more types of plant cells, or (2) that is contained in a DNA segment located between two base sequences that can serve as T-DNA boundaries. The terms "T-DNA boundary" and "boundary" as used herein are empirically determined and applied; these terms are intended to refer to a base sequence that appears at or near the end of a DNA segment that has been transferred from a Ti plasmid into a plant genome.

Despite the existing knowledge about T-DNA and Ti plasmids, before this invention, no one was able to use these vectors to introduce foreign genes that are expressed in genetically modified plants. Attempts to manipulate genes have encountered a number of obstacles to such use. Such obstacles are:

1) the large size (approximately 200,000 base pairs) and resulting complexity of Ti plasmids preclude the use of standard recombinant DNA techniques to genetically modify and/or insert foreign genes into specific sites in the T-DNA. For example, there are no known unique restriction endonuclease sites in a Ti plasmid (Leemans et al., 1982);

2) the T-DNA inserted into and expressed in plant cells contains genes involved in the production of large amounts of phytohormones in the transformed plant cells (Leemans et al., 1982). The large amounts of phytohormones impair the normal metabolic and regenerative processes of the cells and prevent the formation of phenotypically normal plants from the cells (Braun and Wood, 1976; Yang et al., 1980). Exceptions to this are rare cases where the T-DNA was subjected to extensive spontaneous deletions in planta to eliminate those genes involved in phytohormone production. It has been reported that under these conditions normal plants are obtainable at low frequencies (Otten et al., 1981). However, the T-DNA genes involved in phytohormone production could not be deleted before this invention because they were very important in the identification and/or selection of transformed plant cells (Marton et al., 1979).

As described above, simple recombinant DNA techniques for introducing foreign genes into plasmids are not applicable to the large Ti plasmid. As a result, several indirect methods have been developed, which are discussed below. The first reported use of the Ti plasmid as a vector was in model experiments in which bacterial transposons were inserted into T-DNA and subsequently introduced into plant cells. The bacterial transposons have been reported to be stably maintained in the plant genome (Hernalsteens et al., 1980, Garfinkel et al., 1981). However, in these cases, the transformed tumor tissues were found to be unable to regenerate into normal plants, and there is no reported evidence for the expression of bacterial genes in the plant cells. In addition, since the insertion of bacterial transposons is believed to be essentially random, much of the effort was required to identify and localize the position of the inserted DNA in these examples.

In later experiments, J. Leemans, H. de Greve et al. (1982) demonstrated the expression of dihydrofolate reductase, a characteristic product of the bacterial Tn 7 transposon, by crown gall tobacco cells transformed by mutant Ti plasmids containing the Tn 7 transposon sequence in the T-DNA. The authors speculate that it should be possible to insert behind the T-DNA promoters for opine synthesis the sequences of bacterial genes encoding functions that inactivate other antibiotics. In this way, they assume that it would be possible to achieve good expression and possibly sufficient antibiotic resistance to allow direct selection of the transformed plant cells.

Other investigators have reported the use of intermediate vectors that replicate in both E. coli and A. tumefaciens (Matzke and Chilton, 1981; Leemans et al., 1981; Garfinkel et al., 1981). The intermediate vectors contain relatively small subfragments of the Ti plasmid that can be manipulated using standard recombinant DNA techniques. The subfragments can be modified by the deletion of specific sequences or by the insertion of foreign genes at specific sites. The intermediate vectors containing the modified T-DNA subfragment are then introduced into A. tumefaciens by transformation or conjugation. Double recombination between the modified T-DNA fragment on the intermediate vector and its wild-type counterpart on the Ti plasmid results in the replacement of the wild-type copy by the modified fragment. Cells containing the recombined Ti plasmids can be selected using appropriate antibiotics.

Various foreign DNAs have been inserted at specific sites in the T-DNA by this method and were reported to be stably transferred into plant cells (Matzke and Chilton, 1981, Leemans et al., 1981, 1982). However, such foreign genes have not been reported to be capable of expression in plant cells. Possible reasons for this have been discussed by MD Chilton et al. (1981). For example, it is suggested that the length and sequence of the 5'-untranslated part of the transcript may affect the efficiency of capping, splicing, transport to the cytoplasm or initiation of translation. Furthermore, in the method described above, it is preferred that a double crossover takes place to replace the wild-type copy with the modified DNA fragment. A single crossover leads to the formation of a cointegral Plasmid containing two copies of the T-DNA subfragments. This duplication is undesirable in these methods because homologous recombination, which can occur in A. tumefaciens cells or in plant cells, can lead to the loss of the inserted foreign gene(s).

A major disadvantage of the above methods is that the frequency of double recombination is quite low, about 10-4 to 10-9 (Leemans et al., 1981), and that extensive effort is required to identify and isolate the rare double-crossover recombinants. As a result, the number and types of experiments that can be performed using existing methods for genetically manipulating the Ti plasmid are very limited.

Using a similar procedure, Choudary et al. (1982) reported the transfer of the phaseolin gene from bean to sunflower plant cells using recombinant A. tumefaciens cells as vectors. Transcription of the phaseolin gene in the transformed sunflower cells was reported, but phaseolin production could not be demonstrated.

Another report on the prospects of using Agrobacteria Ti plasmids as vectors for plant genetic manipulation is that of Ream and Gordon (1982).

EP-A-0 116 718, which is prior art under Article 54(3) EPC to the present application, describes modified acceptor Ti plasmids which allow the introduction of any gene(s) of interest contained in an intermediate cloning vector carrying a region of homology with a corresponding region in the acceptor Ti plasmid. This introduction is achieved in Agrobacterium as host by a single crossover taking place within the two homologous DNA segments. The resulting hybrid Ti plasmids in Agrobacterium can be used directly to infect plant cells which can subsequently be screened for expression of the product of the gene(s) of interest. It is possible to regenerate transformed reproductive plants from the transformed plant cells.

However, the hybrid Ti plasmids of EP-A-0 116 718, which produced by the single crossover, two copies of T-DNA subfragments. This duplication is undesirable for the reason given above.

A number of other methods for inserting DNA into plant cells have been reported. One such method involves the use of lipid vesicles, also called liposomes, to encapsulate one or more DNA molecules. The liposomes and their DNA content can be taken up by plant cells, see e.g. Lurquin, 1981. If the inserted DNA can be incorporated into the plant genome, replicated and inherited, the plant cells are transformed.

Other techniques involve contacting plant cells with DNA complexed with either (a) polycationic substances such as poly-L-ornithine (Davey et al., 1980) or (b) calcium phosphate (Krens et al., 1982). Using these techniques, all or part of the Ti plasmid is reportedly inserted into plant cells, causing tumorigenic alteration of the plant cells.

Another method has been developed that involves the fusion of bacteria containing desired plasmids with plant cells. Such methods involve the conversion of bacteria into spheroplasts and the conversion of plant cells into protoplasts. Both methods remove the cell wall barrier from the bacteria and plant cells using enzymatic digestion. The two cell types can then be fused together by exposing them to chemical agents such as polyethylene glycol, see Hasezawa et al., 1981.

However, all of the above techniques suffer from one or more of the following problems: (i) transformation efficiencies reported so far have been very low; (ii) only small DNA molecules can be inserted into plant cells; (iii) only small amounts of DNA molecules can be inserted into plant cells and/or (iv) a gene inserted into a plant cell will not be stably maintained by the plant cell unless it is incorporated into the genome of the plant cell, i.e. unless the gene is inserted into a chromosome or plasmid that replicates in the plant cell.

Prior to this invention, there was no satisfactory method for creating and identifying genetically transformed plant cells that could be routinely regenerated into morphologically normal plants.

Summary of the invention

This invention relates to various plasmids useful for creating transformed plant cells capable of subsequent regeneration into differentiated, morphologically normal plants. This invention also relates to microorganisms containing such plasmids and to methods for creating such plasmids and microorganisms.

This invention involves a first plasmid, such as pMON120, which has certain desired characteristics described below. A gene capable of being expressed in plant cells can be inserted into this plasmid to obtain a modified plasmid, such as pMON128. For example, plasmid pMON128 contains a chimeric gene that expresses neomycin phosphotransferase II (NPT II), an enzyme that inactivates certain antibiotics. The chimeric gene is capable of expression in plant cells.

The modified plasmid is inserted into a suitable microorganism, such as Agrobacterium tumefaciens cells, containing Ti plasmids. In the A. tumefaciens cells, some of the inserted plasmids recombine with Ti plasmids to form a cointegral plasmid due to a region of homology between the two plasmids. Only a single crossover is required to form the desired cointegral plasmid.

Due to the features of the inserted plasmid of this invention, the resulting cointegral Ti plasmid contains the chimeric gene and/or another inserted gene within the T-DNA region of the cointegral plasmid. The inserted gene(s) are surrounded by at least two T-DNA borders, at least one of which has been inserted into the Ti plasmid by the crossover. A. tumefaciens cells that do not have cointegral Ti plasmids with inserted genes are killed by appropriate antibiotics.

A. tumefaciens cells with cointegral plasmids are cocultured with plant cells, such as protoplasts, protoplast-derived cells, plant cuttings or intact plants, under conditions that allow the cointegral Ti plasmids or parts thereof to enter the plant cells. Once inside the plant cells, a portion of the Ti plasmid surrounded by the two T-DNA borders is inserted into the plant genome by natural processes. This DNA segment contains the chimeric gene and/or any other desired gene or genes. Preferably, the segment of vector DNA that is inserted into the plant genome does not contain any genes that would render the plant cell incapable of regenerating into a differentiated, morphologically normal plant. The transformed plant cell(s) can be regenerated into a morphologically normal plant that passes the inserted gene to its offspring.

There are various uses for plants transformed by the method of this invention. For example, a gene encoding an enzyme that inactivates a herbicide can be inserted into a plant. This would render the plant and its progeny resistant to the herbicide. On the other hand, a gene encoding a desired mammalian polypeptide, such as growth hormone, insulin, interferon or somatostatin, can be inserted into plants. The plants can be grown and harvested and the polypeptide could be extracted from the plant tissue.

Short description of the drawings

Figure 1 is a diagram showing the steps of this invention using pMON128 and the NOS-NPT II gene as an example.

Figure 2 shows the formation of pMON41, a plasmid used to construct pMON120.

Figure 3 shows the formation of M-4, an M13-derived DNA used to construct pMON109.

Figure 4 shows the formation of pMON54, a plasmid used to construct pMON109.

Figure 5 shows the formation of pMON109, a plasmid used to construct pMON120.

Figure 6 shows the formation of pMON113, a plasmid used to construct pMON120.

Figure 7 illustrates the construction of plasmid pMON120, an intermediate vector with three restriction endonuclease sites suitable for inserting a desired gene.

Figure 8 represents the formation of pMON128, an intermediate vector obtained by inserting a chimeric NOS-NPT 11 kanamycin resistance gene into pMON120.

Figure 9 illustrates the cointegration of pMON128 with a wild-type Ti plasmid through a single crossover, forming a cointegral plasmid with multiple boundaries.

Figure 10 shows the cointegration of pMON128 with a disarmed Ti plasmid, forming a non-tumorigenic cointegral plasmid.

Fig. 11 is a graph comparing the growth of transformed cells and non-transformed cells on kanamycin-containing medium.

Detailed description of the invention

In a preferred embodiment of this invention, a series of chimeric genes were inserted into plant cells using the steps summarized in the diagram of Figure 1. As shown in Figure 1, three pre-plasmids were prepared. These plasmids were designated as follows:

1. pMON41 containing a right border from a nopaline-type Ti plasmid and the 5' part of a nopaline synthase (NOS) gene. The construction of the plasmid pMON41 is described below and shown in Fig. 2.

2. pMON109 containing the 3' part of a NOS gene and a selection marker gene (spc/str) that allowed the selection of A. tumefaciens cells with cointegral Ti plasmids with chimeric genes. The construction of the plasmid pMON109 is described below and shown in Figures 3, 4 and 5.

3. pMON113, which contained a DNA region with a sequence identical to the sequence within the T-DNA portion of an octopine-type Ti plasmid. This region was designated the "left internal homology" (LIH) region. The construction of pMON113 is described below and shown in Fig. 6.

After assembly of these plasmids, each plasmid was digested by appropriate endonucleases to obtain a desired fragment. Three fragments (one from each of the three plasmids) were assembled in a triple ligation to obtain the intermediate vector, plasmid pMON120, as shown in Fig. 7.

Plasmid pMON120 plays a key role in the embodiment of this invention, which is described in detail below. This plasmid has the following features:

1. pMON120 has at least three unique restriction endonuclease sites (EcoRI, ClaI and HindIII) that allow convenient insertion of any desired gene.

2. pMON120 replicates within normal E. coli cells. However, it does not replicate within normal Agrobacterium cells unless it cointegrates with another plasmid, such as a Ti plasmid, that replicates in Agrobacterium cells.

3. pMON120 carries a marker gene encoding an enzyme that confers resistance to two antibiotics, spectinomycin (spc) and streptomycin (str). This gene, called the spc/str gene, is expressed in E. coli and in A. tumefaciens, but not in plant cells. pMON120 does not carry genes encoding resistance to ampicillin or tetracycline.

4. pMON120 carries a sequence homologous to a sequence within the T-DNA portion of an octopine-type A. tumefaciens Ti plasmid. This sequence is referred to as the "left inner homology" (LIH) region. This region of homology promotes crossover, whereby pMON120 or a derivative of pMON120, such as pMON128, forms a cointegrate with the Ti plasmid when the two plasmids exist within the same A. tumefaciens cell. By definition, the "cointegral" plasmid is formed by a single crossover. It contains all DNA sequences that previously existed in either the Ti plasmid or the plasmid derived from pMON120.

5. pMON120 carries a nopaline-type T-DNA "right border", i.e. a sequence capable of acting as one end (by convention referred to as the right border) of a T-DNA sequence transferred from a Ti plasmid and inserted into the chromosome of a plant cell during transformation of the cell by A.

tumefaciens is inserted.

6. pMON120 carries a gene (including a promoter) that encodes the expression of an enzyme, nopaline synthase (NOS). Once introduced into a plant cell, the NOS enzyme catalyzes the production of nopaline, a type of opine. In most plant species, opines are non-harmful compounds that accumulate in low amounts; the presence of nopaline can be easily detected in plant tissue (Otten and Schilperoort, 1978). Opine genes can serve as useful marker genes for confirming transformation, since opines usually do not exist in non-transformed plant cells. If desired, the NOS gene in pMON120 can be made non-functional by a variety of techniques known to those skilled in the art. For example, a BamHI site exists within the coding portion of the NOS gene; a stop codon or other suitable oligonucleotide sequence could be inserted into this site to prevent translation of NOS.

7. The relative location of the various genes, cleavage sites and other sequences in pMON120 is very important for the practice of this invention. The entire pMON120 plasmid or its modified plasmid, such as pMON128, is contained in the cointegral Ti plasmid. However, only a portion of the cointegral Ti plasmid (the modified T-DNA region) is inserted into the plant genome. Therefore, only a portion of the pMON120-derived plasmid is inserted into the plant genome. This portion begins at the T-DNA boundary and extends in one direction only to the homology region. In pMON120, the NOS count marker, the spc/str selection marker and the three insertion sites are within the portion of pMON120 that would be transferred into the plant genome. However, the pBR322-derived region next to the LIH and the PvuI site would likely not be transferred into the plant genome. Importantly, this arrangement of pMON120 and its derivatives prevents the transfer of more than one region of homology into the plant genome, as discussed below.

8. pMON120 is approximately 8 kilobases long. This is sufficiently small to enable all the aims of this invention to be achieved. However, if desired, it can be made smaller by deleting one or more nucleotide sequences that are not essential, using techniques known to those skilled in the art. Such a reduction in size could improve the efficiency or other characteristics of the plasmid when used in this invention or for other purposes, as can be determined by those skilled in the art.

It is understood that a variety of intermediate vectors that differ from pMON120 in one or more respects can be made and used by those skilled in the art. For example, the NOS marker used to count transformed plant cells could be deleted or replaced or supplemented by another counting or selection marker. Such a marker gene could include an antibiotic resistance gene such as the NOS-NPT II-NOS chimeric gene described below. As another example, the spc/str marker gene used to select A. tumefaciens cells could be deleted with cointegral plasmids or replaced or supplemented by another counting or selection marker expressed in Agrobacteria. As another example, a series of T-DNA boundaries (such as a "left" nopaline-type boundary or a left or right octopine-type boundary) could be used. Similarly, more than one boundary (such as two or more right nopaline boundaries or a right nopaline boundary and a right octopine boundary) could be inserted into the intermediate vector in the desired orientation; this can increase the frequency of insertion of the T-DNA into the plant genome, as can be determined by those skilled in the art. It is also possible to insert both left and right boundaries (of either type) into an intermediate vector. It is also possible to increase the length of the homology region; this is probably to increase the frequency of the desired single crossover (Leemans et al., 1981). It is also possible to select an appropriate homology region from any type of desired plasmid, such as a nopaline or agropine Ti plasmid or a Ri plasmid; such regions enable the intermediate vector to form a cointegrate with any desired plasmid.

Method for forming pMON120

Plasmid pMON120 was constructed from fragments derived from three other plasmids. These three plasmids were designated pMON41, pMON109 and pMON113. The construction of each of these three plasmids is summarized below; further information is given in the Examples.

Plasmid pMON41 contributed a nopaline-type T-DNA right border and the 5' portion of a nopaline synthase (NOS)S gene to pMON120. It was constructed by the following method.

A nopaline-type Ti plasmid, designated pTiT37 plasmid, can be digested with HindIII endonuclease to generate a series of fragments, including a 3.4 kb fragment designated HindIII-23 fragment. This fragment contains the entire NOS gene and the T-DNA right border. Applicant inserted a HindIII-23 fragment into a plasmid, pBR327 (Soberon et al., 1980), which had been digested with HindIII. The resulting plasmid, designated pMON38, was digested with both HindIII and BamHI. This resulted in a 2.3 kb fragment containing the nopaline-type right border and the 5' part of a NOS gene (including the promoter region, the 5' untranslated region and part of the structural sequence). This 2.3 kb fragment was inserted into a pBR327 plasmid digested with HindIII and BamHI. The resulting plasmid was designated pMON41 as shown in Fig. 2.

A number of strains of A. tumefaciens are publicly available from the American Type Culture Collection (Rockville, MD); assignment numbers are provided in each ATCC catalog. Each strain contains a Ti plasmid likely to be suitable for use in this invention, as can be determined by one of skill in the art through routine experimentation.

Plasmid pMON109 contributed a spc/str selection marker gene and the 3' part of a NOS gene to pMON120. It was constructed by the following method.

Plasmid pMON38 (described above and shown in Figure 2) was digested with RsaI to give blunt ends as shown:

5'-GTAC

CATG

A 1.1 kb fragment was isolated and digested with BamHI to yield fragments of 720 bp and 400 bp, each of which had a blunt Rsa end and a cohesive BamHI end. These fragments were annealed to double-stranded DNA from a phage M13 mp8 (Messing and Vieira, 1982) that had been digested with SmaI (thus yielding blunt ends) and BamHI. The mixture was ligated, transformed into cells and plated for recombinant phage. Recombinant phage DNAs containing the inserted 720 bp fragment were identified by the size of the BamHI-SmI insert. One of these phage DNAs was designated M-4, as shown in Figure 3. The 720 bp insert contained the 3' untranslated region (including the polyadenylation signal indicated by a thick dot in the figures) of the NOS gene as well as the 3' part of the structural sequence of the NOS gene. The 720 bp insert is surrounded in M-4 by EcoRI and PstI sites that were present in the M13 mp8 DNA.

A bacterial transposon, designated Tn7, is known to contain the above-mentioned spc/str gene. The Tn7 transposon also contains a gene that causes the host cell to be resistant to the antibiotic trimethoprim. The exact location and orientation of the spc/str gene and the trimethoprim resistance gene in Tn7 are not known. The Tn7 transposon can be obtained from a number of cell strains that are publicly available. A strain of A. tumefaciens was isolated in which the Tn7 transposon had been inserted into the HindIII-23 region of a pTiT37 plasmid. The modified pTiT37 plasmid was designated pGV3106 (Hernalsteens et al., 1980).

Plasmid pGV3106 was digested with HindIII and the fragments were shotgun cloned into pBR327 plasmids digested with HindIII. These plasmids were inserted into E. coli cells and cells that were ampicillin resistant (due to a pBR327 gene) and trimethoprim resistant (due to a Tn7 gene) were selected. The plasmid obtained from one colony was designated pMON31. This plasmid contained a 6 kb HindIII insert. The insert contained the spc/str resistance gene and the trimethoprim resistance gene from Tn7 and the 3' part of a NOS gene (which came from the pTiT37 plasmid).

Plasmid pMON31 was reduced in size twice. The first reduction was performed by digesting the plasmid with EcoRI, diluting the mixture to remove an 850 bp fragment and rejoining the large fragment. The resulting plasmid, designated pMON53, was obtained from transformed cells selected for resistance to ampicillin and streptomycin. Resistance to trimethoprim was not observed.

Plasmid pMON53 was further reduced in size by digesting the plasmid with ClaI, diluting the mixture to remove a 2 kb fragment, and rejoining the large fragment. The resulting 5.2 kb plasmid was designated pMON54, as shown in Fig. 4. This plasmid contains the spc/str gene.

Plasmid pMON54 was digested with EcoRI and PstI and a 4.8 kb fragment containing the spc/str gene was isolated. M-4 DNA was digested with EcoRI and PstI and a 740 bp fragment containing the NOS 3' untranslated region was isolated. These fragments were joined together to form pMON64. To obtain the NOS 3' portion and the spc/str gene on a single EcoRI-BamHI fragment, the orientation of the spc/str gene was reversed by digesting pMON64 with ClaI and rejoining the mixture. Plasmids with the desired orientation were identified by digestion using EcoRI and BamHI. These plasmids were designated pMON109 as shown in Figure 5.

Plasmid pMON113 contributed a homology region to pMON120, which allows pMON120 to form a cointegral plasmid when present in A. tumefaciens with a Ti plasmid. The homology region was taken from an octopine-type Ti plasmid. In the Ti plasmid, it is located near the left T-DNA border within the T-DNA portion of the Ti plasmid. This homology region is referred to as the "left inner homology" (LIH) region.

A region of homology can be derived from any type of plasmid capable of transforming plant cells, such as a Ti plasmid or a Ri plasmid. An intermediate vector can be constructed that contains a cointegral plasmid with any type of plasmid from which the region of homology is derived. was created.

Furthermore, the homology region does not necessarily have to be located within the T-DNA. For example, it may be possible for a homology region to be derived from a segment of a Ti plasmid that contains a T-DNA border and a sequence of bases outside the T-DNA region. In fact, if the intermediate vector contains two suitable T-DNA borders, it may be possible for the homology region to be located entirely outside the T-DNA region.

Applicant obtained an E. coli culture with a pBR-derived plasmid containing the Bam-8 fragment of an octopine-type Ti plasmid. The Bam-8 fragment, which is approximately 7.5 kb, contains the left border and LIH region of the Ti plasmid (Willmitzer et al., 1982; DeGreve et al., 1981). The Bam-8 fragment was inserted into the plasmid pBR327 which had been digested with BamHI. The resulting plasmid was designated pMON90, as shown in Figure 6.

Plasmid pMON90 was digested with BglII and a 2.6 kb fragment containing the LIH region but not the left border was purified. The 2.6 kb fragment was treated with Klenow polymerase to convert the cohesive ends to blunt ends, and the fragment was digested with HindIII to obtain a 1.6 kb fragment (the desired fragment) and a 1 kb fragment. Both fragments were mixed with a pBR322 plasmid digested with PvuII and HindIII. The mixture was ligated and inserted into E. coli cells. The cells were selected for ampicillin resistance and scored for the presence of a SmaL site, which exists on the 1.6 kb fragment but not on the 1 kb fragment. A colony with the desired plasmid was identified and the plasmid from this colony was designated pMON113, as shown by Fig. 6.

To assemble pMON120, three fragments had to be isolated. Plasmid pMON41 was digested with Pvul and BamI and a 1.5 kb fragment containing a nopaline-type right border and the 5' part of a NOS gene was isolated. Plasmid pMON109 was digested with BamHI and EcoRI and a 3.4 kb fragment containing a spc/str gene and the 3' part of a NOS gene was isolated. Plasmid pMON113 was digested with PvuI and EcoRI and a 3.1 kb fragment containing the LIH region was isolated.

The three fragments were mixed together and ligated to form pMON120, as shown in Fig. 2. A culture of E. coli containing pMON120 had been deposited with the American Type Culture Collection. This culture was assigned the number 39263.

It is clear that a number of different methods could be used to form pMON120 or a similar intermediate vector. For example, instead of three-way linkage, it would be possible to assemble two of the desired fragments in a plasmid and insert the third fragment into the plasmid.

Method of using pMON120

As mentioned above, pMON120 has three unique cleavage sites (EcoRI, ClaI and HindIII) suitable for insertion of any desired gene. These cleavage sites are located in the part of pMON120 that is inserted into a plant genome, so that the inserted gene is also inserted into the plant genome.

A series of chimeric genes capable of expressing bacterial and mammalian polypeptides in plant cells have been constructed by the applicant. These chimeric genes are described in detail in a separate application entitled "Chimeric Genes Suitable for Expression in Plant Cells", U.S. Application Ser. No. 458,414, filed January 17, 1983.

These chimeric genes are suitable for use in this invention. They can be inserted into pMON120 to form a modified plasmid which can be used as described below.

In a preferred embodiment of this invention, a chimeric gene was created comprising the following DNA sequences:

1. a promoter region and a 5'-untranslated region derived from a nopaline synthase (NOS) gene;

2. a structural sequence derived from a neomycin phosphotransferase II (NPT II) gene and

3. a 3'-untranslated region derived from a NOS gene.

This chimeric NOS-NPT II-NOS gene was isolated on a DNA fragment with EcoRI ends. This fragment was inserted into the EcoRI site of pMON120 and the resulting plasmids (with chimeric gene inserts with opposite orientations) were designated pMON128 and pMON129, as shown in Fig. 8. Plasmid pMON129 has two copies of the chimeric gene; this may be a useful feature in certain types of work. Each plasmid can be used to transform plant cells in the following manner. A culture of E. coli containing pMON128 has been deposited with the American Type Culture Collection. This culture has been assigned the number 39264.

Plasmid pMON128 (or another plasmid obtained by inserting a desired gene into pMON120) is inserted into a microorganism containing an octopine-type Ti plasmid (or another suitable plasmid). Suitable microorganisms are A. tumefaciens and A. rhizogenes carrying Ti or Ri plasmids. Other microorganisms that may also be suitable for use in this invention are other species of Agrobacterium as well as bacteria of the order Rhizobia. The suitability of these cells or any other cells now known or later discovered or created for use in this invention can be determined by those skilled in the art by routine experimentation.

The plasmid can be inserted into the microorganism by any desired method, such as transformation (i.e., contacting plasmids with cells treated to increase their DNA uptake) or conjugation with cells containing the pMON128 or other plasmids.

The inserted plasmid (such as pMON128) has a region that is homologous to a sequence within the Ti plasmid. This "LIH" region of homology allows a single crossover whereby pMON128 and an octopine-type Ti plasmid combine to form a cointegral plasmid; see, e.g., Stryer, supra, pp. 752-754. Usually this takes place within the A. tumefaciens cell after pMON128 has been inserted into the cell. Alternatively, the cointegral plasmid can be formed in another cell type or in vitro and then inserted into an A. tumefaciens cell or other type of cell. which can transfer the cointegral plasmid into plant cells.

The inserted plasmid, such as pMONM128, combines with the Ti plasmid in the manner shown in Figures 9 or 10, depending on the type of Ti plasmid involved.

In Fig. 2, 2 represents the T-DNA part of an octopine-type Ti plasmid. 4 represents the inserted plasmid, such as pMON128. When these two plasmids coexist in the same cell, crossover can occur, leading to the formation of the cointegral plasmid 6.

The cointegral plasmid 6 has one left border 8 and two right borders 10 and 12. The two right borders are referred to here as the "proximal" right border 10 (the right border closest to the left border 8) and the "distal" right border 12 (the right border further from the left border 8). The proximal right border 10 was carried by plasmid 4; the distal right border was contained on Ti plasmid 2 before cointegration.

A culture of A. tumefaciens GC3111 containing a cointegral plasmid formed by pMON128 and wild-type Ti plasmid pTiB653 was deposited at the American Type Culture Center. This culture was assigned the number 39266.

When the cointegral Ti plasmid 6, shown in Fig. 9, is inserted into a plant cell, either of two DNA regions, T-DNA region 14 or T-DNA region 16, can enter the plant genome.

T-DNA region 14 is delimited by left border 8 and proximal right border 10. Region 14 contains the chimeric gene and all other genes contained in plasmid 4, such as the spc/str selection marker and the NOS count marker. However, region 14 does not contain any of the T-DNA genes that cause crown gall disease or would otherwise disrupt the metabolism or regenerative capacity of the plant cell.

T-DNA region 16 contains left border 8 and both right borders 10 and 12. This segment of T-DNA contains the chimeric gene and other genes contained in plasmid 4. However, T-DNA region 16 also contains the T-DNA genes believed to cause crown gall disease.

Either of the above T-DNA segments, region 14 or region 16, could be transferred into the plant DNA. This is assumed to occur with a probability of 50:50 for any given T-DNA transfer. This is likely to result in a mixture of transformed cells, some tumorous and some non-tumorous. It is possible to isolate non-tumorous cells from the mixture and culture them as described in the Examples.

An alternative approach has also been developed that avoids the need to isolate tumorous cells from non-tumorous cells. Several mutant strains of A. tumefaciens have been isolated that are incapable of causing crown gall disease. Such strains are commonly referred to as "disarmed" Ti plasmids. A Ti or Ri plasmid can be disarmed by one or more of the following types of mutations:

1. Removal or inactivation of one of the border regions. One such disarmed octopine plasmid, which has a left border but no right border, is called pAL4421; this plasmid is contained in A. tumefaciens strain LBA4421 (Oooms et al., 1982; Garfinkel et al., 1981).

2. Removal or inactivation of one or more of the "tumor morphology" genes, referred to as tmr and tms genes, see e.g. Leeman et al., 1982.

Various other types of disarmed Ti plasmids can be prepared using methods known to those skilled in the art; see Matzke and Chilton, 1981; Leeman et al., 1981; Koekman et al., 1979.

Fig. 10 represents an octopine-type Ti plasmid with a T-DNA region 22 that undergoes mutation to delete the tms and tmr genes and the right border. This results in a disarmed Ti plasmid with partial T-DNA region 24. When plasmid 26 (such as pMON128) is inserted into a cell carrying the disarmed Ti plasmid 24, a crossover occurs, forming a cointegral Ti plasmid with disarmed T-DNA region 28. The LIH homology region is repeated in this Ti plasmid, but the disarmed Ti plasmid does not contain any oncogenic genes. On the other hand, if only the right border had been deleted from the T-DNA region 22, the tms and tmr genes and the Octopine synthase (OCS) gene contained in the cointegral disarmed Ti plasmid; however, they would be outside the T-DNA boundaries.

The disarmed cointegral Ti plasmid is used to infect plant cells and the T-DNA region 28 enters the plant genome as shown by transformed DNA 30. Plant cells transformed by disarmed T-DNA 28 exhibit normal phytohormone metabolism and normal ability to be regenerated into differentiated plants.

After inserting pMON128 into A. tumefaciens cells, the desired crossover occurs in a certain fraction of the cells. Cells containing cointegral plasmids (whether virulent or disarmed) can be easily selected from other cells in which the crossover did not occur in the following way. Plasmid pMON120 and its derivatives contain a marker gene (spc/str) that is expressed in A. tumefaciens. However, these plasmids do not replicate in A. tumefaciens. Therefore, the spc/str marker gene is not replicated or stably inherited by A. tumefaciens unless the inserted plasmid combines with another plasmid that can replicate in A. tumefaciens. The most likely such combination based on the region of homology is the cointegrate formed with the Ti plasmid. A. tumefaciens cells containing this cointegral plasmid can be easily identified and selected by growing the cells on medium containing either spc or str or both.

The Ti plasmid 28 shown in Fig. 10 contains two LIH regions. It is possible that cointegral Ti plasmids undergo a subsequent crossover in which the two LIH regions recombine. This case is undesirable because it may result in deletion of the DNA between the LIH regions, including the chimeric gene. However, this is unlikely to cause serious difficulties for two reasons. First, this case is likely to occur with a relatively low probability, such as 10:2. Second, the plasmid pMON120 and its derivatives were constructed so that the selection marker gene (spc/str) is located in the DNA region that would be deleted by the crossover. Therefore, the selective conditions required to identify and grow of Agrobacteria cells containing cointegral plasmids, also to kill the progeny of cells undergoing a subsequent crossover that eliminates the chimeric gene from the Ti plasmid.

Only one of the LIH regions in the cointegral Ti plasmid is inserted into the plant genome, as shown in Fig. 10. This important feature results from the construction of pMON120 and it distinguishes this cointegral plasmid from previously formed undesirable cointegral plasmids. The LIli region, which lies outside the T-DNA boundaries, is not inserted into the plant genome. This leads to at least two important advantages. First, the presence of two LIH regions inserted into the plant genome could result in crossover cases that would lead to the loss of the inserted genes in the transformed plant cells and their progeny. Second, the presence of two DNA homology regions can significantly complicate efforts to analyze the DNA inserted into the plant genome (Matzke and Chilton, 1981).

After A. tumefaciens cells containing the cointegral Ti plasmids with the chimeric genes have been identified and isolated, the cointegral plasmids (or portions thereof) must be inserted into the plant cells. Finally, methods can be developed to perform this step directly. In the meantime, a method has been developed that can be used conveniently and with good results. This method is described in two separate pending U.S. patent applications entitled "Transformation of Plant Cells by Extended Bacterial Co-Cultivation," Ser. No. 458,413, and "Rapid Culture of Plant Protoplasts," Ser. No. 458,412, both filed January 17, 1983. The method described in these applications can be briefly summarized as follows.

The plant cells to be transformed are exposed to enzymes that remove the cell walls. This converts the plant cells into protoplasts, which are viable cells surrounded by membranes. The enzymes are removed and the protoplasts begin to regenerate cell wall material. At an appropriate time, the A. tumefaciens cells (containing the cointegral Ti plasmids with chimeric genes) are treated with the plant protoplasts. The cells are co-cultured for a period of time that allows infection of the plant cells by A. tumefaciens. After an appropriate co-culture period, the A. tumefaciens cells are killed and the plant cells are propagated.

Plant cells that have been transformed (i.e., cells that have received DNA from the cointegral Ti plasmids) and their progeny can be selected by a variety of methods, depending on the type of gene(s) that have been inserted into the plant genome. For example, certain genes can cause different antibiotics to be inactivated; such genes include the chimeric NOS-NPT II-NOS gene carried by pMON128. Such genes can serve as selection markers; a group of cells can be grown on medium containing the antibiotic that is inactivated by the chimeric gene product, and only those cells that contain the selection marker gene survive.

A number of genes can serve as counting markers in plant cells. For example, pMON120 and its modified plasmids, such as pMON128, carry a nopaline synthase (NOS) gene that is expressed in plant cells. This gene encodes an enzyme that catalyzes the formation of nopaline. Nopaline is a non-harmful compound that is usually accumulated in low amounts in most plant species; it can be easily detected by electrophoretic or chromatographic methods.

When a plant is transformed by a gene that produces a polypeptide that is difficult to detect, the presence of a selection marker gene (such as the chimeric NOS-NPT II-NOS gene) or a counting marker gene (such as the NOS gene) in the transformation vector can help to identify and isolate transformed cells.

In fact, any desired gene can be inserted into pMON120 or other plasmids designed to cointegrate with Ti or similar plasmids. For example, the applicant has created a number of chimeric genes discussed in the cited application "Chimeric Genes Suitable for Expression in Plant Cells".

The suitability of a gene for use in this invention can be determined by those skilled in the art through routine experimentation. Such use is not limited to chimeric genes; for example, this invention can be used to insert multiple copies of a natural gene into plant cells.

This invention is suitable for use with a variety of plants, as can be determined by those skilled in the art through routine experimentation. For example, this invention is likely to be useful for transforming cells from any plant species that can be infected by bacteria from the genus Agtrobacterium. It is believed that virtually all dicotyledonous plants and certain monocotyledons can be infected by one or more strains of Agrobacterium. In addition, microorganisms from the genus Rhizobia are likely to be useful for carrying cointegral plasmids of this invention, as can be determined by those skilled in the art. Such bacteria may be preferred for certain transformation types or plant species.

Certain types of plant cells can be grown in vitro and regenerated into differentiated plants using techniques known to those skilled in the art. Such plant species include potatoes, tomatoes, carrots, alfalfa and sunflower. Research into in vitro plant culture techniques is progressing rapidly and it is likely that methods will be developed that allow the regeneration of a much wider range of plants from cells grown in vitro. Cells from such a plant with regenerative capacity are likely to be transformable by the in vitro co-cultivation process discussed above, as can be determined by routine experimentation by those skilled in the art. Such transformed plant cells can be regenerated into differentiated plants using the methods described in the examples.

The in vitro co-cultivation method offers certain advantages in the transformation of plants that are sensitive to in vitro cultivation and regeneration. However, this invention is not limited to in vitro cell culture methods. For example, a number of plant shoots and cuttings (including soybeans, carrots and sunflowers) by contact with A. tumefaciens cells carrying the cointegral plasmids of this invention. It is also possible to regenerate virtually any type of plant from a cutting or shoot. Therefore, it may be possible to develop methods for transforming shoots or cuttings using virulent or preferably disarmed cointegral plasmids of this invention or mixtures thereof and then regenerating the transformed shoots or cuttings into differentiated plants that pass the inserted genes to their progeny.

As mentioned above, it is not essential to this invention that cocultivation be used to transfer the cointegral plasmids of this invention into plant cells. A number of other methods are used to insert DNA into cells. Such methods include encapsulation of DNA in liposomes, complexation of DNA with chemicals such as polycationic substances or calcium phosphate, fusion of bacterial spheroplasts with plant protoplasts, microinjection of DNA into a cell, and induction of DNA uptake by electric current. Although such methods for inserting DNA into plant cells have not yet been used with satisfactory efficiency, they are being actively investigated and may be useful for inserting foreign genes into plant cells using the intermediate vectors and cointegral plasmids of this invention or plasmids derived therefrom.

This invention may be useful for a variety of purposes. For example, certain bacterial enzymes, such as 5-enolpyruvylshikimate-3-phosphoric acid synthase (EPSP synthase), are inactivated by certain herbicides; other enzymes, such as glutathione S-transferase (GST), inactivate certain herbicides. It may be possible to insert chimeric genes into plants that cause expression of such enzymes in the plants, thereby causing the plants to become more resistant to one or more herbicides. This would allow the herbicide that normally kills the non-transformed plants to be applied to a field of transformed plants. The herbicide would act as a weed killer and leave the transformed plants unharmed.

On the other hand, it may be possible to insert chimeric genes into plants that cause the plants to produce mammalian polypeptides such as insulin, interferon, growth hormone, and the like. At an appropriate time, the plants (or cultured plant tissue) would be harvested. Using a variety of techniques known to those skilled in the art, the desired protein can be extracted from the harvested plant tissue.

Another use of this invention is to create plants with high levels of desired substances, such as storage proteins or other proteins. For example, a plant could contain one or more copies of a gene encoding a desired protein. Additional copies of this gene can be inserted into the plant by this invention. Alternatively, the structural sequence of the gene could be inserted into a chimeric gene under the control of a different promoter, causing prolific transcription of the structural sequence.

The methods of this invention can be used to identify, isolate, and examine DNA sequences to determine whether they are capable of promoting or otherwise regulating the expression of genes within plant cells. For example, DNA from any type of cell can be fragmented using partial endonuclease digestion or other methods. The DNA fragments can be randomly inserted into plasmids similar to pMON128. These plasmids, rather than having a complete chimeric gene such as NOS-NPT II-NOS, have a partial chimeric gene with a site for the insertion of DNA fragments in place of a NOS promoter or other promoter. The plasmids with inserted DNA are then inserted into A. tumefaciens where they can rejoin the Ti plasmids. Cells with cointegral plasmids are selected using the spc/str or other marker gene. The cointegral plasmids are then inserted into the plant cells by bacterial cocultivation or by other means. The plant cells contain a selectable marker structural sequence, such as the NPT II structural sequence, but this structural sequence is not transcribed unless the inserted DNA fragment serves as a promoter for the structural sequence. The plant cells can be selected by growing them on medium containing kanamycin or other suitable antibiotics.

Using this technique, it is possible to examine the promoter regions of bacteria, yeast, fungi, algae, other microorganisms and animal cells to determine whether they function as gene promoters in different types of plant cells. It is also possible to examine promoters from one plant species in other types of plant cells. By using similar techniques and varying the cleavage site in the starting plasmid, it is possible to examine the performance of any DNA sequence as a 5' untranslated region, 3' untranslated region or any other type of regulatory sequence.

As used herein, the term "a piece of DNA" includes plasmids, phages, DNA fragments and polynucleotides, whether natural or synthetic.

The term "chimeric piece of DNA" as used herein is limited to a piece of DNA that contains at least two parts (i.e., two nucleotide sequences) derived from other and distinct pieces of DNA. For example, a chimeric piece of DNA cannot be formed merely by deleting one or more parts of a naturally existing plasmid. A chimeric piece of DNA can be formed by a variety of methods, such as joining two fragments from different plasmids together or by synthesizing a polynucleotide in which the base sequence has been determined by analyzing the base sequences from two different plasmids.

The term "a chimeric piece of DNA" as used herein is limited to DNA that has been assembled, synthesized, or otherwise produced as a result of human effort, and any piece of DNA that is replicated or otherwise derived therefrom. "Human effort" includes enzymatic, cellular, and other biological processes when such processes occur under conditions induced, enhanced, or regulated by human effort or intervention; this excludes plasmids, phages, and polynucleotides formed solely by natural processes. The term "derived from" as used herein is intended to be broadly construed. Whenever used in a claim, it is intended to include the term "chimeric" may be a material limitation.

As used here, "foreign DNA" includes any DNA that is inserted into a pre-existing plant cell. A "foreign gene" is a gene that is inserted into the pre-existing plant cell.

As used herein, a "marker gene" is a gene that confers a phenotypically identifiable characteristic on the host cell, allowing transformed host cells to be distinguished from non-transformed cells. It includes screening, counting and selection markers.

As used herein, a "region of homology" refers to a sequence of bases in one plasmid that has sufficient correlation with a sequence of bases in another plasmid to cause recombination of the plasmid to occur at a statistically determinable frequency. Preferably, such recombination should occur at a frequency that allows for the convenient selection of cells with combined plasmids, e.g., greater than 1 per 106 cells. This term is described in more detail in a number of publications, e.g., Leemans et al., 1981.

By "tumor-producing genes" are meant those genes of the Ti plasmid whose presence in the transformed plant leads to a morphologically abnormal plant, i.e. induces crown gall formation.

Those skilled in the art will know, or be able to find, numerous equivalents to the specific embodiments of the invention discussed herein using no more than routine experimentation. Such equivalents are within the scope of this invention.

Example 1: Formation of plasmid pMON41

A culture of E. coli carrying a pBR325 plasmid (Bolivan, 1978) with the HindIII-23 fragment of pTiT37 (Hernalsteens et al., 1980) inserted at the HindIII site was obtained from Drs. M. Bevan and MD Chilton, Washington University, St. Louis, MO. 10 µg of the plasmid from this clone was treated with 10 units of HindIII (unless otherwise stated, all restriction endonucleases used in these constructions were purchased from New England Biolabs, Beverly, MA, and treated with buffers according to the supplier's instructions) for 1 h at 37ºC. The 3.4 kb HindIII-23 fragment was purified by adsorption onto glass beads (Vogelstein and Gillespie, 1979) after separation from the other HindIII fragments by electrophoresis on a 0.8 96 agarose gel. The purified 3.4 kb HindIII fragment (1.0 µg) was mixed with 1.0 µg of plasmid pBR327 DNA (Soberon et al., 1980) digested with both HindIII (2 units, 1 h, 37°C) and calf alkaline phosphatase (CAP, 0.2 units, 1 h, 37°C), deproteinized with phenol, precipitated with ethanol, and resuspended in 10 µl of TE (10 mM Tris HCl, pH 8, 1 mM EDTA). 1 unit of T4 DNA ligase (prepared by the method of Murray et al., 1979) was added to the fragment mixture. 1 unit is defined as the concentration sufficient to obtain greater than 90% circularization of 1 µg HindIII-linearized pBR327 plasmid in 5 min at 22°C. The mixed fragments were contained in a total volume of 15 µl of 25 mM Tris-LiCl pH 8, 10 mM MgCl2, 1 mM dithiothreitol, 200 µM spermidine-HCl and 0.75 mM ATP (ligase buffer).

The mixture was incubated for 3 h at 22°C and then mixed with E. coli C600 A56 cells prepared for transformation by treatment with CaCl2 (Maniatis et al., 1982). After a period for expression of the ampicillin resistance determinant carried by the pBR327 vector, cells were spread on LB solid medium plates (Miller, 1972) containing ampicillin at 200 µg/ml. After incubation at 37°C for 16 h, several hundred clones were obtained. Plasmid minipreps (Ish-Horowicz and Burke, 1981) were performed on 24 of these colonies and aliquots of the resulting plasmid DNAs (0.1 µg) were digested with HindIII to demonstrate the presence of the 3.4 kb HindIII fragment. One plasmid showed the expected structure and was designated pMON38. pMON38 DNA was prepared using Triton X-100 lysis and CsCl gradient methods (Davis et al., 1980).

50 µg of pMON38 DNA was digested with HindIII and BamHI (50 units each, 2 h, 37ºC) and the 2.3 kb HindIII-BamHI fragment was purified as described above. The purified fragment (1 µg) was mixed with 1 µg of the 2.9 kb HindIII-BamHI fragment of the pBR327 vector purified as described above. After linking (T4 DNA ligase, 2 units) and transformation of E. coli cells as described above, 50 ampicillin-resistant colonies were obtained. DNAs from 12 plasmid minipreps were digested with HindIII and BamHI to detect the presence of the 2.3 kb fragment. A plasmid of the correct structure was chosen and designated pMON41 as shown in Fig. 2. A quantity of this DNA was prepared as described above.

Example 2: Creation of M13 clone M-4

30 µg of plasmid pMON38 (described in Example 1) was digested with RasI (30 units, 2 h, 37°C) and the 1100 bp RasI fragment was purified after separation by agarose gel electrophoresis using the glass bead method described in the previous example. The purified 1100 bp RasI-RsaI fragment (1 µg) was digested with 2 units of BamHI and the BamHI was inactivated by heating. This DNA was mixed with 0.2 µg of phage M13mp8RF DNA previously digested with SmaI and BamHI (2 units each, 1 h, 37°C) and 0.2 units of calf alkaline phosphatase (CAP). After linking with 100 units of T4 DNA ligase and transforming E. coli JM101 cells as described in the previous example, the transformed cells were mixed with soft agar and plated under conditions that allow identification of recombinant phage (Messing and Vieira, 1982). 12 recombinant phage producing cells were taken and RF plasmid minipreps were obtained as described in the previous example. The RF DNAs were digested with BamHI and SmaI to demonstrate the presence of the 720 bp RsaI-BamHI fragment. One of the recombinant RF DNAs carrying the correct fragment was designated M13 mp8 M-4. This procedure is shown in Fig. 3. M-4 RF-DNA was prepared using the methods of Ish-Horowicz and Burke, 1981, and Colman et al., 1978.

Example 3: Construction of pMON109

20 µg of plasmid pGV3106 (Hernalsteens et al., 1980, prepared according to the method of Currier and Nester, 1976) was digested with HindIII (20 units, 2 h, 37ºC) and mixed with 2 µg of HindIII-digested pBR327. After ligation (T4 DNA ligase, 2 units) and transformation of E. coli cells as described above, a culture medium containing trimethoprim (100 µg/ml) and ampicillin was resistant colony was obtained. Digestion of plasmid DNA from this cell showed the presence of a 6 kb HindIII fragment. This plasmid was designated pMON31.

Plasmid pMON31 from a miniprep (0.5 µg) was digested with EcoRI (1 unit, 1 h, 37ºC) and the endonuclease was inactivated by heating (10 min, 70ºC). The 8.5 kb plasmid fragment was recircularized in a 100 µl ligation reaction (T4 DNA ligase, 1 unit) and used to transform E. coli cells with selection for ampicillin and streptomycin (25 µg/ml) resistant colonies. Plasmid miniprep DNAs from 6 clones were digested with EcoRI to detect loss of the 850 bp fragment. A plasmid lacking the 859 bp EcoRI fragment was designated pMON53. This plasmid was introduced by transformation into E. coli Gm42 dam cells (Bale et al., 1979) as described.

Plasmid pMON53 (0.5 µg) from a miniprep prepared from dam cells was digested with ClaI and recircularized in dilute solution as described above. After transformation of E. coli GM42 cells and selection for clones resistant to ampicillin and spectinomycin (50 µg/ml), 50 colonies were obtained. Digestion of plasmid miniprep DNAs from 6 colonies showed that all lacked the 2 kb ClaI fragment. One of these plasmids was designated pMON54 as shown in Figure 4. Plasmid DNA was prepared as described in Example 1.

Plasmid pMON54 DNA (20 μg) was digested with EcoRI and PstI (20 units each, 2 h, 37°C) and the 4.8 kg fragment was purified from agarose gels using Na-45 membrane (Schleicher and Schuell, Keene, N.li.).

The purified 4.8 kb fragment (0.5 µg) was mixed with 0.3 µg of a 740 bp EcoRI-PstI fragment obtained from M13mp8 M-4 RF DNA (described in Example 2) purified using NA-45 membrane. After linking (T4 DNA ligase, 2 units), transformation of E. coli GM42 dam cells and selection for spectinomycin resistant cells, 20 colonies were obtained. Plasmid miniprep DNAs prepared from 12 of these clones were digested with PstI and EcoRI to show the presence of the 740 bp fragment. A plasmid carrying this fragment was designated pMON64. A quantity of this plasmid DNA was prepared as described in Example 1.

DNA (0.5 µg) from pMON64 was digested with ClaI (1 unit, 1 h, 37°C), ClaI heat inactivated and the fragments were rejoined with T4 DNA ligase (1 unit). After transformation and selection for spectinomycin-resistant cells, plasmid minipreps were made from 12 colonies. The DNAs were digested with BamHI and EcoRI to determine the orientation of the 2 kb ClaI fragment. Half of the clones contained the ClaI fragment in the reverse orientation of that in pMON64. One of these plasmids was designated pMON109 as shown in Figure 5. DNA was prepared as described in Example 1.

Example 4: Creation of plasmid pMON113

Plasmid pNW31C-8,29C (Thomashow et al., 1980) was obtained from Dr. S. Gelvin of Purdue University, West Lafayette, IN. This plasmid carries the pTiA6 7.5 kb Bam-8 fragment. The Bam-8 fragment was purified using the Na-45 membrane from 50 µg BamHI-digested pNW31C-8,29C as described in the previous examples. The purified 7.5 kb Bam fragment (1.0 u) was mixed with 0.5 µg of pBR327 vector DNA that had previously been digested with both BamHI (2 units) and 0.2 units of calf alkaline phosphatase (CAP) for 1 h at 37°C; the mixture was deproteinized and resuspended as described in the previous examples. The mixed fragments were treated with T4 ligase (2 units), used to transform E. coli C600 A cells, and ampicillin-resistant colonies were selected as described above. Minipreps were performed on 12 of these clones to obtain plasmid DNA. The DNA was digested with BamHI to show the presence of the pBR327 vector and 7.5 kb Bam-8 fragments. A plasmid containing both fragments was designated pMON90. DNA was prepared as described in Example 1.

25 µg of pMON90 DNA was digested with BglII (25 units, 2 h, 37°C) and the 2.6 kb BglII fragment was purified using the NA-45 membrane. To create blunt ends, the fragment (2 µg) was resuspended in 10 µl of 50 mM HaCl, 6.6 mM Tris-HCl pH 8, 6.6 mM MgCl₂ and 0.5 mM dithiothreitol (Klenow buffer). The 4 deoxynucleoside triphosphates (dATP, dTTP dCTP and dGTP) were added to a final concentration of 1 mM and 1 unit of E. coli large Klenow fragment of DNA polymerase I (New England Biolabs, Beverly, MA) was added. After incubation for 20 min at 22°C, the Klenow polymerase was heat inactivated and 10 units of HindIII were added. HindIII digestion was carried out for 1 h at 37°C and then the enzyme was heat inactivated. The HindIII-BglII (blunt) fragments (1 µg) were added to 0.25 µg of the 2.2 k HindIII-PvuII fragment of pBR322 (Bolivar et al., 1977) generated by HindIII and PvuII digestion and then treated with calf alkaline phosphatase as described in the previous examples. After linkage using 100 units of T4 DNA ligase, transformation of E. coli LE392 cells and selection of ampicillin-resistant colonies as described in the previous examples, 19 colonies were obtained. Plasmid minipreps were prepared from 12 colonies and digested with HindIII to determine the size of the recombinant plasmid and with SmaI to determine that the correct fragment had been inserted. A plasmid with the correct structure was designated pMON113 as shown in Figure 6. Plasmid DNA was prepared as described in Example 1.

Example 5: Creation of plasmid pMON120

20 µg of plasmid pMON109 (described in Example 3) was digested with EcoRI and BamHI (20 units each, 2 h, 37°C) and the 3.4 kb BamHI-EcoRI fragment was purified using the NA-45 membrane as described in the previous examples. 20 µg of plasmid pMON41 (described in Example 1) was digested with BamHI and PvuI (20 units each, 2 h, 37°C) and the 1.5 kb BamHI-PvuI fragment was purified using the NA-45 membrane as described in the previous examples.

20 µg of pMON113 DNA (described in Example 4) was digested with PvuI and EcoRI (2 units each, 2 h, 37°C) and the 3.1 kb PvuI-EcoRI fragment was purified using the NA-45 membrane as above. To assemble plasmid pMON120, the 3.1 kb EcoRI-PvuI pMON113 fragment (1.5 µg) was mixed with 1.5 µg of the 3.4 kb EcoRI-BamHI fragment of pMON109. After treatment with T4 ligase (3 units) for 16 h at 10°C, the Ligase was inactivated by heating (10 min, 70°C) and 5 units of BamHI were added. Digestion continued for 30 min at 37°C, at which time the BamHI endonuclease was inactivated by heating as above. Thereafter, 0.75 µg of the 1.5 kb PvuI-BamHI fragment of pMON41 was added together with T4 DNA ligase (2 units) and fresh ATP to a final concentration of 0.75 mM. The final ligase reaction was carried out for 4 h at 22°C, at which time the mixture was used to transform E. coli LE 392 cells with subsequent selection for spectinomycin-resistant cells as described above. Plasmid minipreps from 12 of several thousand colonies were checked for plasmids approximately 8 kb in size containing unique sites for BamHI and EcoRI. A plasmid showing the correct structure was designated pMON120, which is shown in Figure 7 using an alternative construction method. pMON120 DNA was prepared as described in Example 1.

A culture of E. coli containing pMON120 was deposited with the American Type Culture Collection. This culture was assigned the number 39263.

Example 6: Creation of plasmids pMON128 and pMON129

Plasmid pMON75 (described in detail in a separate application entitled "Chimeric Genes Suitable for Expression in Plant Cells," cited above) contains a chimeric NOS-NPT II-NOS gene. This plasmid (and pMON128, described below) can be digested by EcoRI and a 1.5 kb fragment can be purified containing the NOS-NPT II-NOS gene.

Plasmid pMON120 was digested with EcoRI and treated with calf alkaline phosphatase. After phenol deproteinization and ethanol precipitation, the EcoRI-digested linear pMON120 DNA was mixed with 0.5 µg of the 1.5 kb EcoRI chimeric gene fragment from pMON75 or 76. The mixture was treated with 2 units of T4 DNA ligase for 1 h at 22°C. After transformation of E. coli cells and selection of spectinomycin-resistant colonies (50 µg/ml), several thousand colonies appeared. Six of these were picked and grown and plasmid minipreps were made. The plasmid DNAs were digested with EcoRI to check for the 1.5 kb chimeric gene insert and with BamHI to determine the orientation of the insert. BamHI digestion showed that in pMON128 the chimeric gene was transcribed in the same direction as the intact nopaline synthase gene of pMON120. A culture of E. coli containing pMON128 was deposited with the American Type Culture Collection. This culture was assigned the number 39264. The orientation of the insert in pMON129 was opposite that in pMON128; the appearance of an additional 1.5 kb BamHI fragment in digests of pMON129 indicated that plasmid pMON129 carried a tandem duplication of the chimeric NOS-NPT II-NOS gene, as shown in Fig. 8.

Example 7: Creation of co-integral plasmid pMON128::pTIB6S3TraC

Plasmid pMON128 (described in Example 6) was transferred into chloramphenicol-resistant Agrobacterium tumefaciens strain GV3111=C58Cl carrying Ti plasmid pTiB6S3traC (Leemans et al., 1982) using a triparental plate mating procedure as follows. 0.2 ml of a culture of E. coli carrying pMON128 was mixed with 0.2 ml of a culture of E. coli strain HB101 carrying a pRK2013 plasmid (Ditta et al., 1980) and 0.2 ml of GV3111 cells. The cell mixture was grown in Luria broth (LB), spread on an LB plate and incubated for 16 to 24 h at 30°C to allow plasmid transfer and formation of cointegral plasmids. The cells were resuspended in 3 ml of 10 mM MgSO4 and 0.2 ml aliquots were then plated on an LB plate containing 25 µg/ml chloramphenicol and 100 µg/ml each of spectinomycin and streptomycin. After incubation for 48 h at 30°C, approximately 10 colonies were obtained. A colony was chosen and grown at 30°C in LB medium containing chloramphenicol, spectinomycin and streptomycin at the concentrations indicated above.

A separate type of co-integral plasmid for use in control experiments was prepared by inserting pMON120 into A. tumefaciens cells and selecting for cells with co-integral plasmids using spectinomycin and streptomycin as described above. Like pMON120, these plasmids do not contain the chimeric NOS-NPT II-NOS gene. Example 8: Solution used in plant cell culture The following solutions were used by the Applicant: per litre of enzyme mixture: Cellulysin Macerozyme Ampicillin KH₂PO₄ KNO₃ CaCl₂ MgSO₄·7H₂O KI CuSO₄·5H₂O Mannitol MS9: MS salts (see below) Sucrose B5 vitamins (see below) Mannitol Phytohormones: Benzyladenine (BA) 2,4-D MS-ES: MOS: Feeding plate medium: MS salts Sucrose B5 vitamins Mannitol Phytohormones: BA Ms2C: MS104: MS11: B5 vitamin supply: Rinse fluid: Chlorophenoxyacetic acid NAA Zeatin Myoinositol Thiamine.HCl Nicotinic acid Pyrodoxine.HCl PVP-40

MS salts are purchased premixed as a dry powder from Gibco Laboratories, Grand Island, N.Y.

Example 9: Production of protoplasts

Mitchell petunia plants were grown in growth chambers with 2 or 3 rows of fluorescent lamps and two rows of incandescent lamps (about 5000 lux). The temperature was kept constant at 21ºC and the lights were on for 12 h per day. The plants were grown in a 50/50 mixture of vermiculite and Promix BX (Premier Brands Inc., Canada). The plants were watered once a day with Hoagland nutrient solution. Tissue was taken from dark green plants with compact bushy growth. The leaves were sterilized in a solution of 10% commercial bleach and a small amount of detergent or Tween 20 for 20 min with occasional agitation. The leaves were rinsed two or three times with sterile distilled water. Thin strips (about 1 mm) were cut from the leaves perpendicular to the main rib. The strips were added to the enzyme mixture at a ratio of approximately 1 g tissue to 10 ml enzyme. The dishes were sealed with parafilm and incubated in the dark or under low indirect light with continuous gentle agitation (e.g., 40 rpm on a double rotating shaker). Enzyme incubations were generally carried out overnight for approximately 16 to 20 hours.

The digestion mixture was sieved through a 68, 74, or 88 µm sieve to remove large debris and leaf material. The filtrate was spun at 70 to 100 g for 5 min to pellet the protoplasts. The supernatant was decanted and the pellet was slightly resuspended in rinse solution. This suspension was poured into Babcock bottles. The bottles were filled to within 2 or 3 cm of the base of the neck. 1 ml of growth medium MS9 was carefully added to the top of the rinse solution.

The Babcock flasks were equilibrated and centrifuged at 500 to 1000 rpm for 10 to 20 min. The protoplasts formed a compact band in the neck at the interface. The band was removed with a pipette, taking care not to pick up excess rinsing fluid. The protoplasts were diluted in MS9. At this point, the protoplasts were washed with MS9 or diluted without washing for plating.

Protoplasts were suspended in MS9 medium at 5 x 10⁴ per ml and plated in T-75 flasks at 6 ml per flask. The flasks were incubated on a horizontal surface with low indirect light or in the dark at 26 to 28°C. On the third day after removal of enzymes from leaf tissue, MSO (medium containing no mannitol) was added to each flask using an amount equal to half the original volume. The same amount of MSO is added again on day 4. This reduces the mannitol concentration to about 0.33 M after the first dilution and about 0.25 M after the second dilution.

Example 10: Co-cultivation with bacteria

On day 5 after protoplast isolation, 5- to 7-day-old tobacco suspension cultures (TXD cells) were diluted (if necessary) with MS2C medium to such an extent that 1 ml would easily spread over the surface of agar medium in a 100 x 15 mm Petri dish [i.e., a 10 to 15% suspension (w/v)]. The agar medium was obtained by mixing 0.8% agar with MS-ES medium, autoclaving the mixture, and cooling the mixture until it solidified in the plate. 1 ml of the TXD suspension was spread over 25 ml of feeding plate medium. An 8.5 cm disk of Whatman No. 1 filter paper was placed over the TXD feeding cells and smoothed. A 7 cm disk of the same paper was placed in the center of the larger one.

Separately, aliquots of a culture of A. tumefaciens cells (grown in fruit extract-peptone medium) were added to the flasks containing the plant cells. One set of aliquots contained cells with the cointegral pMON128::Ti plasmids with chimeric NOS-NPT II-NOS genes. The other set of aliquots contained cells with the cointegral pMON120::Ti plasmids that lacked chimeric NOS-NPT II-NOS genes.

Bacteria were added to the flasks to a density of 108 cells/ml. 0.5 ml of the cell mixture was spread in a thin layer on the surface of the 7 cm filter paper disk. Plates were wrapped in parafilm or plastic bags and incubated under direct fluorescent illumination, no more than 5 plates in a stack.

Colonies were observed within 7 days. Within 14 days, the 7 cm disks with attached colonies were transferred to new MSO agar medium (without feeder cells) containing 500 ug/ml carbenicillin and 50 ug/ml kanamycin sulfate (Sigma, St. Louis, MO). Within 2 weeks, vigorously growing green colonies were observed on plates containing plant cells co-cultured with A. tumefaciens strains containing the pMON128 cointegral NOS-NPT II plasmid. No transformed colonies were observed on plates containing plant cells co-cultured with A. tumefaciens strains containing the pMON120 cointegral plasmid. The kanamycin-resistant transformants are capable of uninterrupted growth in the kanamycin-containing culture medium. Southern blotting experiments (as described in E. Southern, J. Mol. Biol. , 503 (1975),) confirmed that these cells contain the chimeric NOS-NPT II gene.

Both sets of transformed cells (and a third set of cells similarly transformed by a chimeric gene encoding the NPT type I enzyme) were tested for kanamycin resistance. The results are shown in Fig. 11.

Example 11: Regeneration of transformed plants

The transformed kanamycin-resistant colonies described in Example 10 contained both tumorous and non-tumorous cells as described in Figure 9 and the related text. The following procedure was used to isolate non-tumorous transformed cells from tumorous transformed cells and to regenerate differentiated plant tissue from the non-tumorous cells.

Colonies were grown on MS104 agar medium containing 30 u/ml kanamycin sulfate and 500 ug/ml carbenicillin until they reached a diameter of approximately 1 cm. Predominantly tumorous colonies show a slightly paler shade of green and are more loosely organized than predominantly non-tumorous colonies. Non-tumorous colonies were removed from the MS104 medium with forceps and placed on MS11 medium containing 30 ug/ml kanamycin and 500 ug/ml carbenicillin. When the colonies continued to grew, colonies that appeared pale green and were loosely organized were removed and discarded.

MS11 medium contains zeatin, a phytohormone that induces shoot formation in nontumorous colonies. Some shoots were eventually found to sprout from kanamycin-resistant colonies. These shoots can be grown to the desired size, cut with a sharp blade, and placed in agar medium without phytohormones such as MSO so that they can form roots. If desired, the medium can be supplemented with naphthaleneacetic acid to induce root formation. Plants can be grown in agar medium to the desired size and then transferred to soil. With appropriate cultivation, such plants will grow to maturity and form seeds. The offspring will inherit the acquired trait according to classical Mendelian laws.

Literature:

A. Bale et al, Mut. Res. 59: 157 (1979)

F. Bolivar, Gene 4: 121 (1978)

A. Braun and H. Wood, Proc. Natl. Acad. Sci. USA 73: 496 (1976)

M.D. Chilton et al, Cell 11: 263 (1977)

M. D. Chilton et al, Stadler Symposia, Volume 13, 39-51 (1981), University of Missouri

P. V. Choudary et al, Miami Winter Symposia, Volume 19, (1982), Page 514, Academic Press

A. Colman et al, Eur. J. Biochem 91: 303-310 (1978)

T. Currier and E. Nester, J. Bact. 126: 157 (1976)

M. Davey et al, Plant Sci. Lett. 18: 307 (1980)

R. W. Davis et al, Advanced Bacterial Genetics, Cold Spring liarbor Laboratory, New York, (1980)

H. De Greve et al, Plasmid 6: 235 (1981)

G. Ditta et al., Proc. Natl. Acad. Sci. US 77: 7347 (1980)

D. Garfinkel et al, Cell 27: 143 (1981)

S. Hasezawa et al, Mol. Gen. Genet. 182: 206 (1981)

J. liernalsteens et al, Nature 287: 654 (1980)

D. Ish-Horowicz and J. F. Burke, Nucleic Acids Res. 9: 2989-2998 (1981)

B. Koekman et al, J. Bacteriol. 141: 129 (1979)

F. Krens et al, Nature 296: 72 (1982)

J. Leemans et al, J. Mol. Appl. Genetics 1: 149 (1981)

J. Leemans, H. de Greve et al, Molec. Biology of Plant Tumors, Academic Press (1982), 537-545

J. Leemans et al, The EMBO J. 1: 147 (1982)

A. L. Lehninger, Biochemistry, 2nd ed., (Worth Publ., 1975)

P. Lurquin, Nucleic Acids. Res. 6: 3773 (1979)

T. Maniatis et al, Molecular Cloning, A Laboratory Manual (Cold Spring Harbor Labs., 1982)

L. Marton et al, Nature 277: 129 (1979)

T. Matzke and M.D. Chilton, J. Mol. Appl. Genet. 1: 39 (1981)

J. Messing et al, Nucleic Acids Res 9: 309 (1981)

J. Messing and J. Vieira, Gene 19: 269-276 (1982)

J. Miller, Experiments in Molecular Genetics, Cold Spring liarbor Laboratory, N.Y. (1972)

N. Murray et al, J. Mol. Biol. 132: 493 (1979)

G. Ooms et al, Plasmid 7: 15 (1982)

L. Otten and R. Schilperoort, Biochim. Biophys. Acta 527: 497 (1978)

L. Otten, Mol. Gen. Genet. 183: 209 (1981)

L. W. Ream and M. P. Gordon, Science, 218, 854-859, (1982)

R. Roberts, Nucleic Acids Res. 10: r117 (1982)

A. Rorsch and R. Schilperoort, Genetic Engineering, 189, Elsevier/North Holland, N.Y. (1978)

J. K. Setlow and A. Hollaender, Genetic Engineering, Principles and Methods (Plenum Press 1979)

X. Soberon et al., Gene 9: 287 (1980)

L. Stryer, Biochemistry, 2nd ed. (W. H. Freeman and Co., 1981)

M. Thomashow et al, Cell 19: 729 (1980)

B. Vogelstein and D. Gillespie, Proc. Natl. Acad. Sci.: 615-619 (1979)

L. Willmitzer et al, Nature 287: 359 (1980)

L. Willmitzer et al, The EMBO J 1: 139 (1982)

N. Yadev et al, Nature 287: 1 (1980)

F. Yang et al, Mol. Gen. Genet. 177: 707 (1980)

P. Zambryski et al, J. Mol. Appl. Genetics 1: 361 (1982)

Claims (18)

1. Cointegral plasmid for use in transforming plant cells and formed by recombination through a single crossover (A) a chimeric plasmid comprising a gene that functions in plants to express an encoded polypeptide, which plasmid comprises in sequence: (a) a DNA region homologous to T-DNA located near the left T-DNA border of a tumor-inducing plasmid of Agrobacterium and capable of causing in vivo recombination of the chimeric plasmid with the tumor-inducing plasmid of < Agrobacterium by a single crossover, b) a gene comprising a promoter acting in plant cells, a coding structure sequence encoding a polypeptide expressible in plant cells, and a 3'-untranslated region encoding a polyadenylation signal, wherein the promoter and the polyadenylation signal are operably linked to the coding structure sequence, and c) an Agrobacterium plasmid T-DNA right border sequence that enables the transfer and introduction of T-DNA into the genome of a plant cell, which chimeric plasmid does not contain any plant tumorigenic genes between and including region (a), gene (b) and border sequence (c), and (B) a Ti plasmid capable of transferring the T-DNA region into the genome of a plant cell, such that the cointegral plasmid has a T-DNA region with the following elements in sequence: (i) a left T-DNA border sequence, (ii) the gene (b) of the chimeric plasmid, (iii) at least one right T-DNA border sequence, wherein the T-DNA border sequences enable the transfer of the T-DNA region into the genome of a plant cell and wherein there is a boundary between and including the left T-DNA border sequence, gene (b) and right T-DNA border sequence or (if there is more than one right T-DNA border sequence) between gene (b) and the right T-DNA border sequence which is closest to the left T-DNA border sequence, (1) there are no genes which would render a transformed plant cell tumorous or incapable of regeneration into a morphologically normal plant, and (2) there is no duplication of DNA sequences which duplication by homologous recombination could result in loss of gene (b). 2. Plasmid according to claim 1, in which the polypeptide acts as a selection marker in transformed plant cells. 3. Plasmid according to claim 2, in which the polypeptide causes the plant cells to be antibiotic resistant. 4. A plasmid according to claim 3, wherein the polypeptide is neomycin phosphotransferase. 5. Plasmid according to one of claims 1 to 4, in which the plant tumor-inducing plasmid is a wild-type Ti plasmid 6. A plasmid according to any one of claims 1 to 4, in which the plant tumor-inducing plasmid is a disarmed Ti plasmid in which the tumorigenic genes and the right T-DNA border have been removed. 7. Cointegral Ti plasmid according to claim 1, having a first gene that causes expression of a polypeptide that acts as a selection marker in bacteria, and a T-DNA region comprising a second gene that causes expression of a polypeptide that acts as a selection marker gene in the transformed plant cells, and a third gene that causes expression of a third polypeptide in plant cells. 8. Plasmid according to claim 7, in which the third gene acts as a counting marker gene. 9. A plasmid according to claim 8, in which the third gene is a chimeric gene. 10. Plasmid according to one of claims 8 and 9, formed by a single crossover between a chimeric plasmid comprising a gene which functions in plants to express an antibiotic resistance marker, which plasmid comprises in sequence: (a) a DNA region homologous to T-DNA located near the left T-DNA border of an Agrobacterium tumor-inducing plasmid and capable of causing in vivo recombination of the chimeric plasmid with the Agrobacterium tumor-inducing plasmid by a single crossover, b) a gene comprising a promoter acting in plant cells, a coding structure sequence encoding an antibiotic resistance marker expressible in plant cells, and a 3'-untranslated region encoding a polyadenylation signal, wherein the promoter and the polyadenylation signal are operably linked to the coding structure sequence, and c) an Agrobacterium plasmid T-DNA right border sequence which enables the transfer and introduction of T-DNA into the genome of a plant cell, which chimeric plasmid does not contain any plant tumorigenic genes between and including region (a), gene (b) and border sequence (c), and a plant tumor-inducing plasmid from Agrobacterium. 11. A plasmid according to claim 7, wherein the second gene confers antibiotic resistance to the plant cells. 12. A plasmid according to claim 11, in which the second gene expresses neomycin phosphotransferase. 13. Plasmid according to claim 7, which does not contain tumorigenic genes. 14. A microorganism containing a chimeric plasmid according to claim 1. 15. A microorganism containing a chimeric plasmid according to claim 7. 16. A microorganism according to claim 15, which contains a chimeric plasmid according to claim 12. 17. Culture of microorganisms of claim 14. 18. The culture of claim 17, identified by ATCC No. 39266.