The aim of this study was to investigate how plants can sense the presence of a fungus able to effect plant disease control in the rhizosphere. As experimental system we used plant cell cultures which were challenged with Trichoderma metabolite mixtures. Moreover, possible modifications in the pattern of the secreted molecules were analysed by testing the effects of culture filtrates of Trichoderma growing in direct antagonism with the phytopathogenic fungus Botrytis. The use of suspension cultured plant cells as a simplified approach, although not closely mimicking the natural situation, offers several advantages in the dissection of complex cellular responses at a molecular level. These in vitro studies may represent a valuable starting point for future experiments to be carried out in planta.
The molecular nature of the elicitors produced by Trichoderma strains has been at least partially unravelled (see [3,4] for reviews). Some of these compounds have been tested in purity for their ability to induce expression of plant defense genes and disease resistance [16,7]. However, the most direct way to have an overview of the complex reactions and effects caused by the fungal metabolites in plants is to assay the natural mixtures. Plants have been found to be actually penetrated and colonized by the fungus at the root level [5,6], and thus the secreted fungal molecules directly interact with living plant cells. It is at the plant-fungus interface that the first steps of the molecular interaction occur and generate the multiple effects observed both in vitro and in agriculture conditions.
Our results highlight the induction of Ca2+-mediated signal perception as an early step during the interaction of soybean cells with Trichoderma metabolites. Although the involvement of Ca2+ in pathogen sensing by plants has been frequently claimed , to our knowledge a Ca2+-mediated perception by plant cells of a fungal biocontrol agent has not yet been reported. Since Ca2+ has been recently demonstrated to be involved also in the molecular communication between plant cells and mycorrhizal fungi , a transient variation in [Ca2+]cyt proves once again to be the most general way for plants to open a dialogue with their fungal partners. The specificity of the Ca2+ changes that we recorded in the single and two-fungal partner interactions (pathogenic and antagonist fungus alone and in combination) guarantees that this intracellular messenger delivers to cells different messages, which are progressively decoded into definite downstream responses. A specificity of the perception mechanism by plant cells is confirmed by the fact that different patterns of intracellular ROS accumulation and cell death induction were determined by the application of the various fungal mixtures. This does not necessarily imply that the cascade of events leading to the physiological responses follows separate, independent pathways, but rather that a network of overlapping pathways may be activated [19,20]. In view of the complexity of signalling crosstalks, firm causal links among Ca2+, ROS and cell death are not easy to assess.
The fungal growth conditions used in this work are well-known to induce the accumulation in the Trichoderma culture filtrates of specific compounds including enzymes, oligosaccharides and secondary metabolites. Separation of fungal culture filtrates by a 3 kDa cut-off let us discriminate differential cell responses to the active molecules recovered in the two fractions.
The larger MW (>3 kDa) fraction is known to contain a battery of hydrolytic enzymes, released by Trichoderma [21-23] as well as Botrytis [24,25], capable of digesting the plant cell wall. Increasing evidence indicates that the elicitor function of fungal lytic enzymes is unrelated to their enzymatic activity, but instead due to the direct perception by plant cells of the protein per se, rather than just through their hydrolysis products [26-28].
The Trichoderma fraction includes, as major secondary metabolites, peptaibols such as trichorzianines A1 and B1 , which are known to form oligomeric voltage-dependent ion channels in the plasma membrane of fungal hosts and plant cells, thus affecting membrane permeability [29,30]; cell death may occur as a consequence of cytoplasmic leakage through these ion channels . Oligosaccharides, gradually released by the action of Trichoderma hydrolytic enzymes on the fungal host cell wall, may also be active components of the small MW fraction generated in the two-way interaction (Trichoderma-Botrytis). They are perceived by both the biocontrol agent as mycoparasitism/antagonism inducers  and by plant cells as elicitors [33,7]. Chitooligomers have been demonstrated to activate in plant cells an increase in [Ca2+]cyt [34,35] and defense responses [33,36]. All the 2+ changes and physiological parameters more remarkable than the higher MW mixtures. The enhancement of the cellular responses that we recorded upon cell treatment with the Trichoderma-Botrytis coculture filtrates cannot be attributed to the mere co-presence of elicitors released by the two fungi alone. Instead, qualitative/quantitative differences in the secreted compound mixture may arise when the biocontrol agent and the pathogen are grown together in direct antagonism, due to competition for nutrients and/or a direct effect of mycoparasitism [7,38]. Isolates of T. harzianum have been previously shown to reduce the activity of hydrolytic enzymes produced by B. cinerea . Furthermore, the activity of Trichoderma hydrolytic enzymes gradually releases oligosaccharides from the Botrytis cell wall, which accumulate in the coculture medium. Our results indicate that plant cells are able to sense different elicitors that Trichoderma mainly addresses to its fungal host and consequently activate their own signal transduction pathway. This notion is also supported by recent findings concerning the activation in plants of a specific pattern of gene expression by the same Trichoderma strain .
Reduction in cell viability was recorded with all the metabolite mixtures of either Trichoderma or Botrytis. This result is expected for the pathogen, for which a Ca2+- and caspase-mediated hypersensitive response (HR) has already been reported [39-41], but also not surprising for the biocontrol agent. It is a common finding that plant roots and seeds treated with Trichoderma are dotted with small necrotic spots probably caused by a HR, which results in localized callose deposition that limits the colonization of plant tissues by the fungus to the first few layers of cells [5,3].
The lack of the 42 kDa endochitinase in the high MW fraction of the deletion mutant culture medium  determined a Ca2+ transient clearly different from that of the wild type. This result indicates that the knock-out of the ech42 gene and its effect on the molecules secreted in the culture medium deeply modifies the fungal signal which is perceived by plant cells through Ca2+. Furthermore, the inactivation of the ech42 gene seems to produce, during the two-way interaction (Δech42-Botrytis), a Trichoderma-Botrytis. This fraction is also able to induce a PCD pathway instead of a necrotic cell death, probably because of a reduced secretion of toxic secondary metabolites by the biocontrol agent. It has been demonstrated that the knock-out of the 42 kDa endochitinase alters the biocontrol ability of Trichoderma virens  and reduces the mycoparasitic and disease control efficacy in vivo of T. atroviride strain P1 . In addition, the ISR-inducing ability of the Δech42 mutant is also lower than the wild type in assays where bean roots treated with Trichoderma are leaf-inoculated with B. cinerea (M. Lorito, unpublished). It has been previously demonstrated that complementation of the mutant culture filtrate with the purified 42 kDa endochitinase fully recovers the Trichoderma biocontrol activity . Application to plant cells of the unfractionated Trichoderma culture filtrate complemented with CHIT 42 may provide a useful validation of the results obtained in this paper. Nevertheless, in view of possible pleiotropic effects of the endochitinase deletion on the fungal metabolism, this experimental approach may lead to an oversimplification of the complex network of synergistic interactions between the Trichoderma bioactive molecules .
The changes in [Ca2+]cyt triggered by Trichoderma metabolites in plant cells, although monitored only in vitro, provide new insight into the mechanisms by which these beneficial fungi affect plant physiology and resistance to stress. Our findings suggest the chance of using the Trichoderma secreted molecules, in mixtures or purified, as elicitor treatments against phytopathogens. This possibility is particularly intriguing, since a recognition of the biocontrol agent metabolites would allow the plant to perceive the presence of Trichoderma, thus pre-activating defense mechanisms against different pathogens , and also inducing a variety of other beneficial effects (i.e. promotion of plant growth, nutrient uptake, seed germination, resistance to abiotic stresses).