1IGER, Aberystwyth, Dyfed SY23 3EB, UK; 2Laboratory of Plant Pathology, Mie University, Tsu 514, Japan

Background and objectives
Epidermal cells in barley leaves have a general penetration resistance to Erysiphe graminis that halts primary infection before a haustorium is produced. It is associated with deposition of papillae beneath appressoria, and is independent of the further compatibility or incompatibility of the interaction. In the field, barley is exposed to polycyclic attacks from E. graminis hordei as well as other pathogens and non-pathogens. These repeated infection attempts affect host resistance and may therefore have great importance for the level of partial resistance and/or adult plant resistance that barley plants express. Here we report data showing that prior attack on a barley leaf epidermal cell by a compatible isolate of E. graminis hordei dramatically affects its accessibility (susceptibility) or inaccessibility (resistance) to subsequent attack.

Materials and methods
Barley seedling leaves were fixed horizontally with the adaxial side upwards and inoculated with a compatible E. graminis hordei isolate as 'inducer' inoculum. After 48 h, incubation fungal spores and hyphae were removed from the leaf surface using cellulose acetate. Only haustoria from successful infections and papillae from failed penetration attempts were left within epidermal cells. Uninoculated leaves were treated with cellulose acetate and used as controls. Immediately after removing the inducer inoculum, leaves were inoculated again with 'challenge' inoculum of the same isolate. After an additional 48 h, incubation leaf segments were fixed and cleared. Fungal structures and host-cell responses were observed on short epidermal cells using differential interference contrast and fluorescence microscopy.

Results and conclusions
When the inducer attack failed, a papilla remained in the attacked cell. Almost all subsequent challenge attacks (91%) on such cells failed to penetrate. Thus failure of inducer attack rendered cells 'inaccessible' to subsequent attack. Conversely, if the inducer attack succeeded in forming a haustorium, most subsequent attacks (97%) also succeeded. Thus success of inducer attack rendered cells 'accessible' to subsequent attack. Induction towards inaccessibility or accessibility was not restricted to the challenged cell. Cells adjacent to the challenged cell were induced towards the same state, although to a slightly lesser degree. However, two cells away no induction was observed.

An important host-cell response related to resistance against penetration is local accumulation of autofluorescent (phenolic) compounds beneath fungal appressoria. Induction towards increased inaccessibility or accessibility dramatically affected the intensities of autofluorescence associated with attempted penetration in the challenged cells. When the inducer failed to penetrate, autofluorescence responses to the challenger were strongly increased. In adjacent cells the increase was moderate, and two cells away autofluorescence responses were similar to controls. Conversely, when the inducer succeeded (haustorium formed), autofluorescence responses to the challenger were strongly suppressed. In adjacent cells autofluorescence responses were moderately suppressed, and two cells away autofluorescence responses were again similar to controls. Papilla diameters beneath challenge appressoria which had failed to penetrate were measured, and compared to diameters of papillae made by the host when the inducer failed to penetrate. There were no significant differences between these papillae. The average diameters of papillae produced in response to the inducer was 8 Ám, and to the challenger 7 Ám.

The phenomena of induced inaccessibility and accessibility are important for understanding disease development in the field. Further evaluation of the physiological, biochemical and molecular basis of these phenomena will increase understanding of host-pathogen relationships, disease epidemiology and the basis of durable resistance.