1School of Biological and Chemical Sciences, Deakin University, Geelong, Victoria, 3217 Australia; 2School of Botany, University of Melbourne, Parkville, Victoria, 3052 Australia

Background and objectives
Signals that are sent between host and pathogen are critical to the establishment of an incompatible interaction where the result is activation of elaborate defence-specific responses. Perception of a signal molecule by a host cell triggers a cascade of events that may include phosphorylation and dephosphorylation, secondary messengers such as Ca2+ [1] and possibly cytoskeletal proteins [2]. We have little knowledge, however, of either the signal molecules involved or their subcellular action. Among those that appear to function in a signalling role is abscisic acid (ABA). Through its regulation of specific genes, ABA can control numerous cellular processes. Our research has investigated two possibly interlinked components in the early cellular events in the gene-for-gene interaction of Phytophthora sojae with soybean: (i) the role of ABA as a signal molecule and (ii) the involvement of the cytoskeleton.

Materials and methods
Pathogen, plants, and inoculation. The race 1 isolate of P. sojae and seeds of soybean cultivars Harosoy (rps Rps7, compatible) and Haro1272 (Rps1a Rps7, incompatible) were the gift of Dr Malcolm Ryley, Department of Primary Industries, Queensland. Plants were grown in the dark for experiments with hypocotyls or in the light for leaf experiments. Cell suspension cultures were initiated from callus cultures. Hypocotyls, leaves and suspension cells were inoculated with a drop of a zoospore suspension (1 x 105/ml).
Treatments. Hypocotyls and/or leaves were treated with an abiotic elicitor, were wounded, or treated with ABA or the carotenoid biosynthesis inhibitor norflurazon, or treated with known inhibitors of signal transduction. In some experiments stomatal pore areas were measured.
PAL activity and gene expression. Activity of PAL was measured in treated leaves and gene expression was determined in total RNA with a PAL cDNA probe (gift of Dr MK Bhattacharrya).
Fixation and labelling protocols. Hypocotyl pieces were fixed in paraformaldehyde and sectioned using a cryostat. Microtubules and actin (in suspension cells) were visualized with a b-tubulin monoclonal antibody and Phalloidin-FITC, respectively, using fluorescence microscopy.

Results and conclusions
ABA, signalling and stomatal closure. ABA and norflurazon had opposing effects on the interaction, thus PAL gene expression and PAL activity increased in normally compatible interactions after treatment with norflurazon, but decreased in normally incompatible interactions treated with ABA. There was systemic closure of stomata in the incompatible, but not the compatible interaction, but not after wounding or treatment with the abiotic elicitor indicating the presence of an induced mobile factor. The effects of treatment with signal transduction inhibitors were complex, but in general their application altered the phenotype of the interaction.
Cytoskeletal changes. There were clear differences in the response of the cytoskeleton to inoculation in compatible and incompatible interactions. In hypocotyl and suspension-cultured cells disorganization and loss of microtubules and actin filaments were characteristics of the incompatible interaction at very early stages after inoculation. The proximity of the cytoskeleton to the plasma membrane and its close links with the nucleus mean that it may function as an early sensor of pathogen invasion.

There are many factors that interact to determine the outcome of a host-pathogen interaction, and ABA must now be considered as a key signalling molecule in these interactions. Rearrangements of cytoskeletal proteins (that are not necessarily linked to ABA's action) are also important for the establishment of incompatibility.

1. Mackintosh C, Lyon GD, Mackintosh RW, 1994. Plant Journal 5, 137-147.
2. Kobayashi I, Kobayashi Y, Hardham AR, 1994. Planta 195, 237-247.