DANIEL McALPINE MEMORIAL LECTURE 1887

Dr Allen Kerr, Waite Agricultural Research Institute, Glen Osmond, SA 

Agrobacterium: pathogen, genetic engineer and biological control agent

I have some affinity with Daniel McAlpine. We were both born in Scotland and both came to Australia where we studied plant diseases. He arrived in Melbourne in 1884 and I in Adelaide in 1951 - 67 years later. In fact before coming to Australia, I knew of Daniel McAlpine through my teacher and mentor Malcolm Wilson who, like McAlpine, was a rust enthusiast. I feel very honoured to be asked to give the Daniel McAlpine Memorial Lecture.
Plant pathology is entering a very exciting phase. In a few years, plant genes for disease resistance will have been located and cloned and will be transferred from plant to plant by genetic engineering. Not only will this have immense practical implications for agriculture, it will also enable the elucidation of the plant pathogen interaction, of which we are abysmally ignorant, despite many books on the topic. Much of the excitement stems from studies on Agrobacterium, a soil-inhabiting, single-celled prokaryotic organism. I'd like to tell you its story, as I see it.

Pathogen

In 1907, Erwin F. Smith and C. 0. Townsend established that crown gall, a disease affecting a wide range of plants, was caused by a bacterium which they named Bacterium tumefaciens later changed to Agrobacterium tumefaciens by Conn in 1942. Also in 1942, White and Braun showed that although A. tumefaciens was necessary for the induction of crown gall, once the gall was initiated, the bacteria could be eliminated without affecting gall growth. The next significant development was the discovery of opines in the late 1960s by Morel and his colleagues in France. Opines are unusual compounds present in crown galls. Examples are octopine, nopaline, agropine and agrocinopine. All consist of two common substances joined together in a very uncommon manner. The kind of opine synthesized depends not on the plant but on the inciting bacterium which can catabolize the synthesized opine and use it as a source of energy, carbon and usually nitrogen. Strains of A. tumefaciens are frequently classified according to opines synthesized.
Through the work of White and Braun and Morel and his colleagues, the idea of genetic information being transferred from bacterium to plant was born but the next major breakthrough was not until 1975 when workers in Gent, Leiden and Seattle showed that genes for pathogenicity are located on a large plasmid called the tumour-inducing (Ti) plasmid. It is a conjugative plasmid but the transfer genes are normally repressed; some opines de-repress these genes and promote conjugation.
An important development in crown gall research was the demonstration by Mary-Dell Chilton and her colleagues that part of the Ti plasmid is transferred to plant cells; it is called transfer DNA or T-DNA. Several genes on the T-DNA have been identified, the most important for crown gall induction being iaam and laah which are concerned with indole acetic acid (IAA) synthesis and ipt which encodes the synthesis of a cytokinin. The transfer of these genes from bacterium to plant induces cell division and explains the hormone independence of crown gall tissue, a fact known for many years. As expected, the genes encoding synthesis of opines are also located within the T-DNA. As a result of TDNA transfer, the plant is directed to synthesize compounds which it cannot utilize; only the inciting bacterium benefits. There are still some genes in the T-DNA whose functions are not known.
Insertion mutations in the T-DNA may result in abnormal galls but all mutants are pathogenic. However, mutations in another region of the Ti plasmid may cause complete loss of pathogenicity. This region is known as the Vir region, in which six genes (vir A-E, G) have been located. Recent work by Stachel and Zambryski has done much to elucidate the functions of the vir genes. Most vir genes are normally repressed but are de-repressed by plant exudates containing acetosyringone and OH-acetosyringone. These substances are involved in the biosynthesis of lignin and are peculiar to plants; presumably they signal the presence of a wounded plant to the bacterium. Genes vir A and vir G are regulatory genes and control the expression of vir B-E. These latter genes are concerned with the processing of T-DNA and presumably with its transfer to the plant cell although the details of transfer and integration into the plant genome have not yet been elucidated. The process resembles conjugation and plasmid transfer in bacteria; because the conjugation is between a bacterium and a plant it has been described as "unnatural" and a recent article in Nature had the caption "a bizarre vegetal bestiality". Perhaps "a bizarre bacterial vegetality" would be more accurate but not so colourful.
Two chromosal genes are also necessary for crown gall induction. They are called chv A and B and control the attachment of bacterial cells to plant cells; chv 8 encodes the synthesis of a 01,2-giucan which is presumably required for attachment. The function of chv A is not known. Shortly after attachment, bacteria can be washed off plant cells but soon they become anchored by means of bacterial cellulose.
To summarise, pathogenicity by A. tumefaciens involves: (a) attachment of A. tumefaciens to cells of a wounded plant; (b) anchoring of bacterial cells by a bacterial cellulose; (c) induction of vir genes by plant exudates; (d) processing of T-DNA which is transferred to a plant cell; (e) integration of T-DNA into the plant genome; (@ synthesis of plant hormones; (g) rapid cell division to form galls; (h) synthesis of opines; (i) preferential growth of A. tumefaciens because of opine nutrition; and 0) transfer of pTi to other bacterial cells through conjugation induced by opines. We now know more about the pathogenicity of A. tumefaciens than of any other plant pathogen. Unfortunately, except for the closely related Agrobacterium rhizogenes, it appears to be a unique system and has not helped to elucidate pathogenesis in other organisms. However, the clear demonstration that "signal" molecules from plants play an important role in pathogenesis of A. tumefaciens may stimulate similar studies on other pathogens.
Genetic engineer

A. tumefaciens can be considered a natural genetic engineer; it inserts DNA into the genome of a plant cell which, as a result, is directed to synthesise plant hormones and opines. However, it is not useful to an agriculturalist to engineer plants that are galled and diseased. The solution is to remove the genes that produce the disease symptoms. In particular, the genes that encode the synthesis of plant hormones must be eliminated. It is usually necessary to insert a selectable marker to distinguish transformed cells from normal cells. Frequently, an antibiotic resistance gene derived from a bacterium is used but because bacterial genes are not expressed in plants, a chimaeric gene is constructed consisting of a promoter active in plants and a structural gene for antibiotic resistance e.g. kanamycin. Other genes of interest can be inserted into the TDNA and they also are transported into the plant. The Vir region, of course, must be present but Schilperoort and his colleagues showed that the vir genes can act in trans i.e. the vir genes and the T-DNA do not have to be on the same plasmid. This has formed the basis of binary vectors in which the vector containing T-DNA can be manipulated in vitro, transformed into E. coli and then transferred to a strain of Agrobacterium containing a Vir region. The vector plasmid and the Vir region plasmid remain separate. When such strains are inoculated into plants, the T-DNA is transferred, as previously described.
What can be achieved by this mechanism? There have been several spectacular successes. A gene for resistance to the herbicide glyphosate has been located in the bacterium Salmonella typhimurium, cloned in E. coal, inserted into T-DNA on a vector, transferred to Agrobacterium and then to tobacco and tomato. Because a suitable promoter was used, glyphosate resistance was expressed in t he transformed plants, rendering them resistant to the herbicide. Another spectacular achievement has been to introduce the toxin gene from Bacillus thuringiensis into tobacco. Such plants are protected from the ravages of many caterpillars. Most effort has been put into studying fundamental aspects of plant gene regulation. Some genes are active in seeds, other in roots or leaves, some in young plants and some in old. A gene for leghaemoglobin synthesis has been transferred from Medicago sativa to Lotus corniculatus vi@Agrobacterium. As might be expected, the gene is expressed only in root nodules of both plants. It can now be manipulated in vitro before transfer and the base sequences in the promoter responsible for specific expression in nodules should be identified.
One of the most exciting developments relates to cross protection in plant viruses. When a plant is infected with a virus, it is usually protected from infection by another, closely related virus. With a few notable exceptions cross-protection has not been used in the field. One argument against crossprotection has been that it is undesirable to spread a virus, even though it appears to be innocuous. The mild strain may mutate to virulence, or when combined with another unrelated virus, it may cause serious damage. These objections do not apply to the new approach to biological control through cross-protection because only a small piece of the viral genome is used. For example, protection of plants by inserting genes encoding virus coat protein has been reported. Other approaches to the control of virus diseases, using the Agrobacterium system, are also being taken.

In my opinion, the most important development in molecular plant pathology will be the cloning of a gene for disease resistance. We are attempting to clone a gene from flax for resistance to flax rust. This was the system used by Flor in his classical studies from which the gene-for-gene hypotheses developed. Disease resistance is clearly defined genetically and the genome of flax is relatively small. This gives a better chance of finding the needle in the haystack. Briefly, the procedure is to construct a genomic library of disease-resistant flax in E. coli, transfer it to Agrobacterium and thence to a susceptible cultivar of flax which must then be regenerated. The regenerated plants will be inoculated with rust and a change from susceptibility to resistance will indicate that a resistance gene has been transferred.
We calculate that more than 20 000 regenerants will have to be screened to give 95% probability of success. This is still an immense task.

Biological control agent

Since 1973, crown gall on stone fruit and roses has been controlled commercially by dipping planting material (seeds, cuttings or roots of young plants) in a bacterial suspension of Agrobacterium radiobacter strain K84, a non-pathogenic relative of A. tumefaciens. This method of biological control has generally proved very effective and is now practised throughout the world. Control is achieved through production of an antibiotic called agrocin 84 by strain K84. Max Tate and his students at the Waite Institute have determined the structure of agrocin 84. The biosynthesis of agrocin 84 is encoded by a 48kb plasmid called pAgK84. A potential threat to the continued success of the biological control of crown gall is that pAgK84 is a conjugative plasmid which can transfer to other agrobacteria, including pathogens. Following plasmid transfer, such pathogens are no longer subject to biological control. The transfer genes on pAgK84 have been located and a deletion mutant constructed by recombinant DNA technology. The new strain K1026 is identical with strain K84 except that it is transfer deficient (Tra-). It will soon be tested as a new biocontrol agent.*
I have now been working on Agrobacterium for 20 years. It has provided me with endless fascination and great intellectual stimulation. Some of the recent developments, particularly with the genetic engineering of plants were inconceivable just ten years ago. I wonder what the next ten years will bring.

*K1026 was the first genetically engineered organism to be released in Australia for testing with no containment facilities. The experiment took place on 16 June 1987.