rosea conidia and allowed to interact for 5 days. Water inoculated roots were used as control. After surface sterilization, colonization levels were determined by counting colony forming units (cfus). No significant differences in root colonization ability were recorded between WT and the ΔHyd1 strain. In contrast, root colonization by the ΔHyd3 strain was significantly (P < 0.001) reduced (Figure 8). Interestingly,

the double deletion ΔHyd1ΔHyd3 strain showed increased (P < 0.001) colonization ability compared to WT or single deletion strains (Figure 8). Figure 8 A. thaliana root colonization by C. rosea strains. A. thaliana roots were detached 5 days post inoculation and washed. After sterilization in 2% NaOCl for 1 min, the roots were homogenized in water and serial dilutions were plated on PDA plates under sterile BGJ398 condition at 25°C. Different letters find more indicate statistically

significant differences (P ≤ 0.05) based on the Tukey-Kramer test. Discussion Filamentous fungi generally contain multiple hydrophobin genes, which play important roles in fungal growth, development and environmental communication [1, 2, 6, 7]. We identified only 3 class II hydrophobin genes in the genome of the mycoparasite C. rosea. This is in strong contrast with the closely related mycoparasites T. atroviride and T. virens that contain high numbers (10 and 9 respectively) and diversity of class II hydrophobins [29]. This indicate important ecological differences between C. rosea and Trichoderma spp., and emphasize that different mycoparasites may rely on different mechanisms of interaction. The expansion of the hydrophobin gene family in Trichoderma spp. is hypothesized to help the fungus to attach Wilson disease protein to the hyphae of a broad range of asco- and basidiomycetes [29]. The high expression of Hyd1 in conidiating mycelia in comparison with germinating conidia indicates that Hyd1 may have a role during conidiophore development. This is consistent with the expression pattern of hyd1 in M. anisoplia where expression is low in germinating conidia and high in mycelium

with conidiophores [35]. The expression, but lack of regulation, of Hyd1, Hyd2 and Hyd3 on different nutrient regimes, and between developmental stages of Hyd2 and Hyd3, indicate a constitutive role of the corresponding proteins in C. rosea. Constitutive roles of hydrophobins in fungal growth and development are reported in many species [6, 7, 36]. However, certain hydrophobins from Trichoderma spp. and M. brunneum are regulated by nutritional conditions and between different life cycle stages [5, 11, 28, 37]. Expression levels of Hyd1, Hyd2 and Hyd3 are repressed in C. rosea during interactions with B. cinerea and F. graminearum, which is consistent with the expression pattern of T. atroviride hydrophobin genes hfb-1b, hfb-2c and hfb-6a[37]. This may suggest that Hyd1, Hyd2 and Hyd3 are not involved in protecting hyphae from recognition by other organisms [6, 7].