The use of Agrobacterium for plant improvement.

Abstract

Agrobacterium species have potential as tools for plant improvement, in that they can be used as biological control agents against crown gall disease, vectors for gene insertion into plants and plant root-inducing bacteria. For a bacterium to be successful as a biocontrol agent against crown gall disease, it must be able to produce an effective agrocin specific for A. tumefaciens pathogens and be able to colonise host plants effectively. Successful biological control of crown gall disease has been achieved on a limited range of host plants by application of Agrobacterium radiobacter. Strain K84 is non-pathogenic, produces an antibiotic type substance, agrocin 84, which kills a specific spectrum of pathogenic Agrobacterium strains, and is a good plant coloniser. The colonisation abilities of this bacterium and a number of potential biocontrol bacteria, D286, 173 and its derivatives, H8, and H6 and its derivatives, were compared in vitro and in vivo. In addition the ability of these strains to control crown gall pathogens in vitro and in vivo was assessed. Although D286 and H8 were good long term colonisers they were subsequently eliminated from the programme because D286 was unable to control grapevine crown gall disease, and H8 was a pathogen, with a narrow host range for biotype 3 pathogens, which had not been cured. Good long term colonisation ability was shown to be critical to the success of the biological control strain. Both K84 and J73 strains produced fibrils which attached them to tomato root surfaces and have similar colonisation efficiencies up to 14 days after inoculation. However, the ability of J73 to colonise plants for longer periods was significantly less than that of K84. Thus, the presence of fibrils is not sufficient to ensure colonisation. No correlation was found between hydrophobicity and colonisation. Poor in vivo biological control of tomatoes and grapevines by J73 and its derivatives has been shown to be due to poor long term colonisation. The additions of the agrocin 84 plasmid to J73 improved in vitro biological control by J73 but was not effective in vivo. Thus although J73 produced a broad host range agrocin it was ineffective as a biocontrol agent due to its poor long term colonisation ability. H6 and its derivative proved to be good long term colonisers of the grape vine rhizosphere and were able to effectively control crown gall disease on grapevines. Eucalyptus grandis was propagated in vitro from axillary buds. The effect of the gelling agents gelrite, agarose and agar on propagation was determined. Shoot multiplication and elongation on gelrite-containing media was found to be superior to that obtained on agarose- and agar-containing media. Rooting was enhanced with gelrite as the gelling agent. In vitro clone development from seed of selected E. grandis genotypes and the integration of this into vegetative clonal propagation programmes is proposed. Not all clones developed in this manner rooted well under in vitro conditions. Thus the potential use of A. rhizogenes to improve rooting was investigated. A. rhizogenes is a bacterial pathogen which causes 'hairy root' disease on certain plants. Eucalyptus species do not fall into the natural host range of this organism, however, it is able to infect these plant species. In an attempt to improve rooting and root quality of in vitro and in vivo propagated Eucalyptus species and clones, the root-inducing genes on the Ri plasmids of a number of A. rhizogenes strains were inserted into Eucalyptus by inoculating in vitro and in vivo stem cuttings with the selected A. rhizogenes strains. The resultant chimeric plants have transformed roots and normal shoots. Root development was monitored in vitro and after the plantlets had hardened off, and in vivo. Only transformed roots grew as root cultures in hormone-free liquid medium. The potential use of this procedure for improving rooting of clonal material is discussed. Under in vitro conditions for example, one of the broad host range A. rhizogenes strains, LBA9402, was able to induce up to 80 % rooting on E. grandis, E. nitens and E. dunnii explants. While under nursery conditions for example, one of the two E. globulus clones tested, HM15, developed up to ten times as many roots in response to two strains of the A. rhizogenes bacterium (LBA9402 and TR8,3) while the other clone failed to respond. Not only the inherent rooting abilities of the numerous Eucalyptus genotypes and clones tested, but also hormone induced-rooting and Agrobacterium-mediated rooting were found to vary from clone to clone and genotype to genotype. A. rhizogenes strains were however, able to overcome the genetic control on rooting of certain clones but this was found to be dependent on both the genotype of the bacterium and the plant. The novel concept of using bacterial cocktails to improve rooting was investigated, with the aim of overcoming the limited host range of individual A. rhizogenes strains. Both in vitro and nursery trials were established to test the potential use of bacterial cocktails as a means of improving rooting of a range of Eucalyptus genotypes and clones. Agrobacterium cocktails not only improved the ability of certain clones and genotypes to root but also the quality of the roots. The anatomy of transformed roots from chimeric plants was compared to that of non-transformed roots, revealing no difference in root anatomy of these roots. Nutrient uptake studies using a radioisotope nutrient uptake bioassay showed no difference in phosphorus, potassium and nitrogen uptake by transformed and nontransformed roots. The potential of A. rhizogenes strains to improve rooting of two other species, Anacardum occidentale (cashews) and Pinus (a number of pine hybrids) was also shown in preliminary trials.