As a basis for our study on the ISR-effect of T. hamatum T382 we provided proof-of-principle that T. hamatum T382 is effectively able to induce ISR and subsequently to reduce the symptoms of B. cinerea infection in A. thaliana (Figure 1A). To this end, 3-weeks old A. thaliana plants were treated with T. hamatum T382 by pipetting a spore suspension directly onto the roots, and their susceptibility to subsequent B. cinerea leaf infection was evaluated as compared to mock-treated plants. A preliminary experiment, in which T. hamatum T382 was administered on different time points ranging from 4 to 20 days before B. cinerea inoculation, indicated that a T. hamatum T382 treatment 6 days before B. cinerea inoculation resulted in the highest reduction of disease symptoms (results not shown). These conditions were applied in an extended disease experiment consisting of 12 repeats comprising a total of 2000 plants in which B. cinerea infection was scored symptomatically on a daily basis. By comparing the average lesion diameters of T. hamatum T382 and mock-treated plants using a two-sample one-sided Student’s t-test (Figure 1B) it was shown that treatment of A. thaliana with T. hamatum T382 resulted in a statistically significant (p < 0.001) reduction of the disease symptoms of on average 60% on 2, 3, and 4 days after B. cinerea inoculation. Moreover, the results of these disease assays were highly reproducible making the model A. thaliana – T. hamatum T382 – B. cinerea suitable for further transcriptomic analysis of ISR.

To determine relevant time points for a transcriptomic analysis of ISR, known marker genes were selected for the main defense-related pathways and their expression was monitored by qRT-PCR. We chose PDF1.2a for the combined JA- and Et-mediated pathway and PR1 for the SA-mediated pathway. For ISR-prime analysis, pools of leaves were collected from both T. hamatum T382 and mock-treated A. thaliana Col-0 plants (At + T vs. At) and the expression of the marker genes was examined on six consecutive days after T. hamatum T382 treatment. For ISR-boost analysis, the same two sets of plants were subsequently inoculated with B. cinerea and pools of systemic leaves were analyzed for marker gene expression on four consecutive days after B. cinerea inoculation (At + T + B vs. At + B).

During ISR-prime, a statistically significant (p < 0.01) induction was shown for PR1 on the first 3 days post-T. hamatum T382 inoculation (dpTi; Figure 2A). Two other SA-markers, PR2 and PR5, showed a similar induction pattern as PR1 (results not shown), while the expression of the Et/JA-marker PDF1.2a was not affected (p < 0.01; Figure 2B). After 3 dpTi, expression levels decreased until basal levels were reached at 5 and 6 dpTi. Based on these results we decided to use the samples taken on 2 dpTi in the subsequent transcriptomic analysis of ISR-prime. Interestingly, the changes in the expression of defense-related genes during ISR-prime coincided with the explosive growth of T. hamatum T382 in the soil and on the roots of the plants, which was observed both by qPCR using specific primers and plating on a selective medium. Both techniques showed an increase from 4 × 10⁵ CFU/g to 10⁸ CFU/g from 1 h to 3 days after T. hamatum T382 application in the soil. On the roots, T. hamatum T382 increased from 4 × 10⁴ CFU/g to 4 × 10⁸ CFU/g in the same time period.

During ISR-boost, the expression of PR1 was unaltered (Figure 2C) as was the case for the two other SA-responsive markers PR2 and PR5 (results not shown). Interestingly, PDF1.2a showed a six-fold downregulation in T. hamatum T382 treated plants on 1 day post-B. cinerea inoculation (dpBi) whereas no effect was observed in mock-treated control plants (Figure 2D). On 2 dpBi, however, the PDF1.2a expression levels increased up to 60-fold in control plants while in T. hamatum T382 treated plants still a slight downregulation (2.86-fold) was observed. Thus, during ISR-boost, we could not observe the strong induction of PDF1.2a on 2 dpBi, which is characteristic during the defense response induced by B. cinerea in A. thaliana (Manners et al., 1998). Based on these findings 1 and 2 dpBi were chosen as relevant time points for ISR-boost.

For both ISR-prime and ISR-boost, plant samples were collected at the above-selected time points from plants that showed a clear ISR-effect (defined as 40–80% reduction in disease index in the subsequent disease assay) and from their corresponding control plants. Three independent biological replicates were performed, each with a dye swap. Furthermore, control samples of the ISR-boost experiment (systemic leaves from A. thaliana Col-0 plants that were inoculated with B. cinerea without pretreatment with T. hamatum T382) were compared to those from untreated and uninfected plants (At + B vs. At) to characterize the defense response induced by B. cinerea (BIDR).

From all selected plant samples high quality mRNA was isolated, labeled, and hybridized on Agilent 4-pack Arabidopsis microarrays. Normalized microarray data were used to identify genes that are differentially expressed during ISR (either prime or boost) and BIDR. Comparing plants treated with T. hamatum T382 with mock-treated controls (At + T vs. At), we identified 2075 genes that are differentially expressed during ISR-prime as summarized in Table 1 (for a complete list see Table S1 in Supplementary Material). Furthermore, the microarray analysis revealed 276 and 1135 annotated genes that were differentially expressed during ISR-boost (Table 1, Table S1 in Supplementary Material) on 1 and 2 dpBi, respectively. The analysis of ISR-boost consisted of a comparison of plants treated with T. hamatum T382 and consequently inoculated with B. cinerea and mock-treated controls consequently inoculated with B. cinerea (At + B + T vs. At + B). For BIDR, comparison of these mock-treated plants that were consequently inoculated with B. cinerea and mock-treated plants without subsequent B. cinerea inoculation (At + B vs. At) resulted in the identification of 1119 and 7317 annotated genes that were differentially expressed on 1 and 2 dpBi, respectively (Table 1, Table S1 in Supplementary Material).

The high number of differentially expressed (DE) genes allowed us to examine the underlying mechanism of ISR at the level of biological processes and pathways leading to a more holistic view on ISR. As shown in Table 2, classification of the DE genes based on the standard Gene Ontology annotation (Gene Ontology Consortium, 2000), led to the identification of biological processes that were significantly (p < 0.001) enriched in the sets of DE genes as compared to the complete A. thaliana genome, according to the method described by Tavazoie et al. (1999). The main observations are discussed below for both ISR-prime and ISR-boost and compared to BIDR. A concise overview of these results is shown in Table 2 while the complete list of up- and downregulated processes can be found Table S2 in Supplementary Material.

With respect to upregulated processes a striking analogy between ISR-prime and BIDR (on 2 dpBi; Table 2) was observed including both defense responses such as “defense response to fungus,” “plant-type hypersensitive response,” or “response to chitin” and defense-related plant hormone responses like “response to SA” and “response to abscisic acid” (ABA). Nevertheless, still a number of differences between ISR-prime and BIDR could be observed. For instance, ISR-prime was characterized by the induction of “negative regulation of defense response” while this was not the case for BIDR. On the other hand, ISR-prime was not characterized by the induction of the Et- and JA-signaling pathways. Furthermore, anthocyanins were the main secondary metabolites formed during ISR-prime while defense during BIDR relied mainly on the production of camalexin. Specific for ISR-prime is also the stimulation of the transport of a variety of compounds in the plant, e.g. phospholipids and ions.

Remarkably, when looking at the downregulated processes, the correspondence between ISR-prime and BIDR was very low (Table 2). Where BIDR negatively affected the general metabolism of the plant (e.g. photosynthesis, translation, lipid metabolism), this phenomenon was not observed during ISR-prime.

It is important to stress here that, in order to characterize ISR-boost, we compared plants treated with T. hamatum T382 and subsequently inoculated with B. cinerea with mock-treated control plants subsequently inoculated with B. cinerea (At + T + B vs. At + B). As such the results for ISR-boost represent the biological processes that are specifically affected by the pretreatment with T. hamatum T382 and that occur on top of the B. cinerea induced defense response (BIDR).

On the first day after B. cinerea inoculation, BIDR affected a limited number of defense processes (Table 2) like “response to chitin” and “response to wounding.” Several defense-related plant hormone responses were induced, e.g. responses to JA, SA, and ABA (Table 2) while general metabolic processes, like translation, were downregulated. In contrast to BIDR, defense-related processes and hormone pathways were activated more rapidly (primed) during ISR-boost, such as “biosynthesis of JA” and “response to microbial phytotoxin,” or more strongly, such as “response to JA” and “response to wounding” (Table 2). Production of secondary metabolites such as galactolipids and anthocyanins were specifically induced during ISR-boost. Interestingly, defense–related ROS-production as reflected by the biological process “respiratory burst involved in the defense response” is specifically downregulated during ISR-boost.

On the second day after B. cinerea inoculation, the strong induction of defense processes that characterized BIDR was not or to a less extent observed during ISR-boost, indicating that the T. hamatum T382 treatment restrained the Botrytis-induced defense response (Table 2). Many defense-related processes showed an opposite behavior in ISR-boost as compared to BIDR such as “defense response to fungus,” “regulation of defense response,” and “response to chitin.” Similar observations can be made for the defense-related hormone pathways such as signaling mediated by JA and Et. However, in line with the situation on the first day after B. cinerea inoculation, the phenylpropanoid pathway–based biosynthesis of anthocyanins, as well as flavonoids, were specifically induced while “respiratory burst involved in the defense response” was specifically downregulated during ISR-boost. Again, general metabolic processes, like photosynthesis, were downregulated during BIDR.

The above-mentioned microarray results were validated by qRT-PCR and GUS-staining.

The involvement of the above-mentioned pathways that were identified in this study as contributing to ISR was further confirmed by mutant evaluation and anthocyanin analysis.

In contrast to what was generally observed for BCOs of bacterial origin (Verhagen et al., 2004; Cartieaux et al., 2008) which are believed not to significantly alter gene expression upon treatment, addition of Trichoderma hamatum T382 to the roots of the plant triggers a clear and pronounced defense response in the leaves on the second day after the treatment. Our finding is supported by the recent observation that Bacillus cereus AR156 treatment to the roots of A. thaliana activates expression of defense-related genes PR1, PR2, PR5, and PDF1.2 in the leaves (Niu et al., 2011) thereby inducing ISR against Pseudomonas syringae pv. tomato DC3000. Furthermore, a striking analogy can be observed between the biological processes that are induced during ISR-prime and BIDR. Indeed, based on the results for the enriched biological processes (Table 2) and genes (Table S4 in Supplementary Material; Figure 6A) it can be concluded that treatment with T. hamatum T382 results in a microbe-associated molecular pattern (MAMP)-induced defense response in A. thaliana leaves. The importance of MAMP-triggered immunity in the plant response to Trichoderma spp. has been described by Lorito et al. (2010). According to this review, Trichoderma-induced ISR includes the increase of the plant’s basic immunity or MAMP-triggered immunity upon the detection of a variety of MAMPs that are produced by the Trichoderma spp. This suggests that MAMP-triggered immunity by Trichoderma spp. at the roots can affect MAMP-triggered immunity in the leaves.

The MAMPs by which the plant recognizes T. hamatum T382, could be chitin-related because we observe a clear induction of “response to chitin” (Table 2) and several chitinases (Table S4 in Supplementary Material). However, it is not clear whether recognition of chitin from T. hamatum T382 at the roots would also be reflected by similar alterations in gene expression in systemic tissues such as leaves. In our results on BIDR we also see a clear induction of “response to chitin” and the MAMP-defense pathway in systemic leaves (Table 2). Nevertheless, MAMP recognition is known to result in SID2-dependent accumulation of SA and concomitant activation of the SA-signaling pathway both in local and systemic leaves, leading to increased resistance against subsequent pathogen infections (Mishina and Zeier, 2007; Tsuda et al., 2008), as reflected in our data by the induction of “response to SA” (Table 2). Additionally, Figure 6B clearly shows that during ISR-prime SA is indeed produced via isochorismate (through the actions of SID2) instead of via phenylalanine (through the actions of PAL1–4), which is the main source of SA during BIDR. The MAMP-triggered SA-signal activates ER-localized proteins that are involved in the folding of defense-related proteins (Li et al., 2009; Christensen et al., 2010) through the actions of TF BZIP60 (Iwata et al., 2008), which are all also upregulated in our data (Figure 6A; Table S4 in Supplementary Material). The latter is also the case for the LysM domain containing proteins, which are MAMP recognition receptors (PRRs; Li et al., 2009; Table S4 in Supplementary Material). Activation of PRRs leads to the onset of (i) the phenylpropanoid pathway leading to the accumulation of anthocyanins as reflected by the induction of “biosynthesis of anthocyanins” and “metabolism of anthocyanins” (Table 2; Figure 6D), (ii) a MAPK signaling cascade indicated by the induction of, e.g. MKK4 and MPK3 (Asai et al., 2002; Wan et al., 2004; Figure 6A; Table S4 in Supplementary Material), (iii) cell wall reinforcement, as reflected by the induction of PEN3 (Table S4 in Supplementary Material), and (iv) an increased Ca²⁺ influx in the cell by ion channels in the plasma membrane that are regulated by glutamate binding (Ranf et al., 2011), as reflected by the induction of “calcium ion homeostasis” and several Ca²⁺ transporters (Table 2; Table S4 in Supplementary Material). Ultimately Ca²⁺ regulates the channel activity of cyclic nucleotide-gated channels (Kaupp and Seifert, 2002), of which several are upregulated in our experiments, leading to the induction of the hypersensitive response (Moeder and Yoshioka, 2008) and the concomitant programmed cell death. Both processes are included in the list of upregulated biological processes during ISR-prime. In addition to Trichoderma-derived MAMPs inducing the defense response during ISR-prime, pectin could be involved in priming too. Indeed, our data demonstrate also upregulation of wall-associated kinases, which are known to interact with pectin released from the plant cell wall and to subsequently activate the MAPK cascade transferring the signal to the nucleus of the cell (Ringli, 2010). Furthermore, it has been demonstrated that MAMP-perception also results in systemic acquired resistance (SAR) in A. thaliana (Mishina and Zeier, 2007), which is reflected in our data by the induction of “systemic acquired resistance” and “regulation of SAR” during ISR-prime (Table 2).

In our data we also observe an induction of genes involved in the “negative regulation of defense response” (Table 2). This might be explained by the fact that T. hamatum T382 is recognized by the plant as a “beneficial invader.” For instance, we observe a clear induction of both PYK10 and BGLU18, which encode proteins accumulating in ER bodies (Table S4 in Supplementary Material). The latter are recently discovered ER compartments reportedly linked to wounding and defense responses (Ogasawara et al., 2009). PYK10, for example, is proposed to be involved in maintaining the interaction between the beneficial endophytic fungus Piriformospora indica and A. thaliana by repressing the defense response (Sherameti et al., 2008). Additionally, both the gene encoding protein kinase OXI1 and the gene encoding monodehydroascorbate reductase MDHAR, two enzymes which are also linked to the interaction with this endophyte (Vadassery et al., 2009; Camehl et al., 2011), are upregulated by T. hamatum T382 (Table S4 in Supplementary Material). Interestingly, mutant analysis showed that the production of ascorbate by MDHAR keeps the interaction between plant and endophyte in a mutualistic state whereas the OXI1 pathway seems to control plant growth promotion by the endophyte. Our observation of PYK10, MDHAR, and OXI1 induction during ISR-prime could support a similar role. However, it should be noted again that, in contrast to the results from the P. indica – A. thaliana interaction, our study did not focus on gene expression in roots but in systemic leaves.

In the next sections we will zoom in on the different defense-related pathways and discuss their involvement in ISR in more detail.

On the first day after B. cinerea inoculation, ISR-boost was characterized by a transient activation of JA-biosynthesis which was not observed during BIDR and reinforcement of the BIDR-related induction of JA-response (Table 2). Additionally, the induction of the defense-related process “response to wounding,” which is linked to the JA-pathway, was also reinforced by the ISR. Furthermore, during ISR-boost, the defense-related process “response microbial phytotoxin” and the biosynthesis of secondary metabolites (e.g. anthocyanins, galactolipids) were induced, whereas during BIDR these defense responses were not yet activated on the first day after B. cinerea inoculation. These findings correspond to the current view that (i) ISR primes the plant to react faster to subsequent pathogen infections (Prime-A-Plant Group et al., 2006; Pozo et al., 2008; De Vleesschauwer et al., 2009) and that (ii) Trichoderma spp. are able to activate ISR that leads to such primed responses (Lorito et al., 2010).

On the second day after B. cinerea inoculation the production of secondary metabolites (e.g. anthocyanins and flavonoids; Figure 6D; Table 2), was still reinforced. With respect to anthocyanins and flavonoids many different in planta functions have been proposed including their role as antioxidants or protectants against different types of abiotic (e.g. UV) and biotic stress. Regarding the latter, it is suggested that both compounds accumulate around fungal infection sites to protect host cells from oxidative damage as a result of the defense-related ROS-production (Hipskind et al., 1996; Kangatharalingam et al., 2002; Treutter, 2006). Furthermore, anthocyanins have been reported to play a role in attenuating defense-related ROS-production (Figueroa-Balderas et al., 2006; Senthil-Kumar and Mysore, 2010), while some flavonoids have direct antifungal activity (Treutter, 2006; Buer et al., 2010). Like flavonoids, galactolipids are also known to display antifungal effects. For instance, esters formed between jasmonates and galactolipids can inhibit the growth of Botrytis cinerea (Kourtchenko et al., 2007). The enhanced (“primed”) production of these antifungal compounds during ISR-boost might give a first explanation for the observed increased resistance toward B. cinerea resulting from T382-induced ISR.

On the second day after B. cinerea inoculation, we observed a clear moderation of BIDR during ISR-boost (Table 2), possibly as a result of the priming and the subsequent increased inhibition of B. cinerea proliferation (and concomitant BIDR-triggering) in T. hamatum T382 treated plants. The major induction of the defense system during BIDR was thus restrained during ISR-boost, as was the case for various defense processes (e.g. “defense response to fungus,” “response to chitin,”…), JA-, Et-, and ABA-responses, SA- and JA-mediated signaling and the production of camalexin. Such moderation of the normal defense response against pathogens has been previously observed for other BCAs, from both fungal (Wen et al., 2005) and bacterial (Cartieaux et al., 2008) origin. When studying the respective pathways in more detail, we observed the same sets of genes being upregulated during BIDR and downregulated during ISR-boost, thereby confirming the restraining of the BIDR (Figure 6). It is possible that during ISR-prime defensive proteins or their respective transcripts have accumulated so that the levels of these compounds at the moment of Botrytis inoculation are already higher in T. hamatum T382-treated plants than in control plants. This would leave out the need for a further strong induction of the defense response at the level of gene expression and, hence, would result in the observed restraining of the defense response during ISR-boost.

The induction of compounds with an antioxidant activity (e.g. anthocyanins, flavonoids), together with the observed restraining of the ROS response during ISR-boost (Figure 6A), might give a reasonable second explanation for the T382-induced ISR efficient against B. cinerea. The importance of ROS response in ISR is also reflected in our results for the vtc1 and rbohD mutant studies (Figure 4). Both mutants display increased resistance to B. cinerea infection and additional T. hamatum T382 treatment did not further boost this enhanced resistance. However, it should be noted here that the absence of an ISR-effect in mutants with already increased resistance to pathogens (such as B. cinerea) can be related to the fact that an additional boost of defense-related compounds by the biocontrol organism might not be possible in the plant. Different findings support the general idea that infection by B. cinerea is favored by ROS-production in the plant. First, as for other necrotrophic pathogens, infection and colonization of the plant will be promoted by necrosis-inducing components such as ROS (Govrin and Levine, 2000; Glazebrook, 2005). It has indeed been shown that B. cinerea virulence correlates with the intensity by which the plant produces ROS as defense reaction (Temme and Tudzynski, 2009). Secondly, it is known that this pathogen is actively secreting compounds to elicit an oxidative burst and subsequent programmed cell death (Govrin et al., 2006). Additionally, it has been demonstrated that B. cinerea infection can be suppressed by spraying antioxidants on plants (Elad, 1992).

In view of the very recent report by Brotman et al. (2012), it is worth here to briefly compare our results with their findings. In their study analysis of gene expression in A. thaliana leaves was done on a restricted set of 137 genes during the ISR induced by another Trichoderma species, T. asperelloides T203, and effective against the bacterial pathogen Pseudomonas syringae. During ISR-prime induced by T. asperelloides T203 they identified 15 up- and 2 downregulated genes. Four of these upregulated genes are also induced during T. hamatum T382-induced ISR-prime, namely WRKY40, WRKY55, TAT3, and PR5. The first one is a gene that is induced by SA-signaling, encoding a transcription factor that interacts with WRKY18 to form a negative feedback loop during MAMP-triggered defense (Pandey et al., 2010). Remarkably, the WRKY18 gene is also induced during T. hamatum T382-ISR-prime. The TAT3 gene is controlled by the JA-signaling pathway in which NPR1 acts as a positive regulator. This induction is linked to the activation of ROS signaling by NPR1 (Brosché and Kangasjärvi, 2012) upon activation of SA-signaling during the MAMP-response. Overall the similarities in ISR-prime between our results and these of Brotman et al. (2012) are low. A possible explanation for this discrepancy might be the fact that both studies characterize ISR-prime at different time points after Trichoderma treatment (4 dpTi in Brotman et al., 2012 vs. 2 dpTi in our study). As shown in Figure 2A, the induced response during ISR-prime triggered by T. hamatum T382 is transient, peaking in the first few days after Trichoderma treatment and then decreasing again until basal expression levels are reached.

Comparing our results on ISR-boost with their data is even more complex. First of all, as ISR-boost in both studies is initiated by different types of pathogen (fungal vs. bacterial) with different infection strategies, comparison of induced responses are expected to also significantly differ. Furthermore their study is restricted to only a comparison between the tripartite interaction (A. thaliana – T. asperelloides T203 – P. syringae) with untreated and mock-inoculated control plants, one can not rule out the effects of the pathogen (P. syringae) infection, as is the case in our study (BIDR). Nevertheless, some overlap in differentially expressed genes can be identified in both studies such as the induction of LTP4 and LOX2, and a downregulation of WRKY40. The LTP4 gene encodes a lipid transfer protein, a member of a family 14 of pathogenesis-related peptides (PR14) with reported in vitro antimicrobial activity (Sels et al., 2008). The JA-responsive LOX2 gene encodes a lipoxygenase that is required for JA production (Bell et al., 1995) during pathogen infection (Spoel et al., 2003). Enhanced expression of this gene also takes place during rhizobacteria-induced ISR-boost (Pineda et al., 2012).

Cultivation and spore harvesting of Trichoderma hamatum strain T382 (kindly provided by Tom De Ceuster, DCM, Sint-Katelijne-Waver, Belgium) and Botrytis cinerea strain B05-10 (kindly provided by Rudi Aerts, Katholieke Hogeschool Kempen, Geel, Belgium) was performed as described previously (Broekaert et al., 1990). Arabidopsis thaliana Col-0 plants were obtained from the European Arabidopsis thaliana stock centre (NASC). The Arabidopsis thaliana mutants sid2, npr1, myc2, ein2, etr1, chs, f3h, rbohd, vtc1, and tt were kindly provided by Prof. U. Conrath (RWTH Aachen University, Aachen, Germany), Prof. X. Dong (Duke University, Durham, NC, USA), Prof. S. Berger (Institut für Pflanzenbiochemie, Halle/Saale, Germany), Prof. F. Ausubel (Massachusetts General Hospital, Boston, MA, USA), Prof. J. Glazebrook (University of Maryland, Maryland, MD, USA) and the Arabidopsis Biological Resource Center, Columbus, OH, USA (accession numbers CS237 for etr1, CS3071 for ein2, N3130 for tt, N653439 for f3h, N671192 for chs, N671557 for rbohD and N8326 for vtc1), respectively. The transgenic Arabidopsis thaliana line containing the NahG gene was obtained from J. Ryals (Ciba Geigy, Research Triangle Park, NC, USA). The transgenic lines carrying PDF-promoter-GUS constructs were developed in house (De Coninck et al., 2010); the constructs contain the GUS-gene linked to the promoter fragments (1.25 kb) upstream of the predicted start codon of the PDF1.2a (At5g44420) and PR1 (At2g14610) genes.

Arabidopsis thaliana Col-0 plants were grown in untreated and unsterilized soil (“DCM potgrond voor Zaaien and Stekken”, DCM, Sint-Katelijne-Waver, Belgium) in a growth chamber with 21°C daytime temperature, 18°C night-time temperature, 75% humidity and a 12-h day-light cycle with a light intensity of approximately 120 μmol/m² s. Three weeks old plants were treated with Trichoderma hamatum T382 by pipetting a 50 μl spore suspension (2 × 10⁷/ml) directly onto the roots. Six days later, these and an equal number of mock-treated plants were inoculated with Botrytis cinerea B05-10 as described previously (Thomma et al., 1998). Briefly, a 5 μl drop of a Botrytis cinerea spore suspension (5 × 10⁵/ml in 1/2 PDB) was inoculated onto two leaves per plant. Disease symptoms were scored by measuring the diameters of the necrotic lesions on various days after B. cinerea inoculation. The disease assay was repeated 12 times with a total of 2000 plants. Each day, lesions diameters were measured and the average lesion diameters of treated and untreated plants were compared using a two-sample one-sided Student’s t-test (alternative hypothesis average lesion diameter of treated plants < average lesion diameter of untreated plants) implemented in R (R Development Core Team, 2011).

Trichoderma hamatum T382 density was determined by (i) dilution plate enumeration on Trichoderma selective medium and (ii) qPCR using Trichoderma hamatum T382 specific primers as previously described (Lievens et al., 2007).

In short, 10 g of each root sample was washed intensely in 100 ml phosphate buffer and 10 g of each soil sample was mixed with 90 ml phosphate buffer in a blender for 30 s at high speed. Next, a 10-fold dilution series was prepared and 100 μl of each dilution was plated and spread on a Trichoderma selective medium (Chung and Hoitink, 1990) in triplicate. After 5 days of incubation at 25°C in the dark, the colonies were counted.

In parallel, genomic DNA was extracted from 0.5 ml washed root or 0.5 g soil sample using Mo Bio Ultraclean Soil DNA Isolation kit according to the manufacturer’s specifications (Mo Bio Laboratories, Solana Beach, CA, USA). DNA extracts were diluted 10-fold and stored at −20°C. To specifically detect Trichoderma hamatum T382, qPCR amplification was performed in a total volume of 25 ml using the intercalating dye SYBR1 Green I on a SmartCyclerII1 instrument (Cepheid, Sunnyvale, CA, USA). Each reaction mixture contained 2 ml of the target DNA extract, 12.5 ml of the QuantiTectTM SYBR1 Green PCR Master Mix (Qiagen, Inc., Valencia, CA, USA), 0.625 ml of each primer (20 mM), and 9.25 ml sterile distilled water. Sequences of the Trichoderma hamatum T382 specific primers are shown in Table S3 in Supplementary Material. Thermal cycling conditions consisted of 10 min at 95°C followed by 45 amplification cycles of 15 s at 95°C, 30 s at 60°C, and 30 s at 72°C with a final 2-min elongation step at 72°C. Fluorescence was detected at the end of the elongation phase of each cycle. To evaluate amplification specificity, melt curve analysis was performed at the end of the PCR run. A melt curve profile was obtained by slowly heating the mixture from 60 to 95°C at 0.2°C/s with continuous measurement of fluorescence. Standard curves were generated by plotting the threshold cycle (Ct) of a 10-fold dilution series of standard DNA against the logarithm of the concentration. The regression line was used to calculate the DNA concentration of Trichoderma hamatum T382 in the samples via the obtained Ct-values (Brouwer et al., 2003; Lievens et al., 2006).

Primers were developed using Primer3 software (Rozen and Skaletsky, 2000), primer sequences are shown in Table S3 in Supplementary Material. Before qRT-PCR the ideal annealing temperature of each primer was determined in a regular PCR. Six different sets of leaves from mock-treated and T. hamatum T382 treated were collected daily and used for qRT-PCR to study gene expression during ISR-prime. For ISR-boost systemic leaves were collected from five of the same sets of mock-treated and Trichoderma hamatum T382-treated plants after additional Botrytis cinerea inoculation. For qRT-PCR validation of the microarray results, an additional biological repeat was used, for which samples were collected in the same manner. RNA extraction, DNase treatment and reverse transcription were done as described previously (Mirouze et al., 2006). The qRT-PCR analysis was carried out using the StepOnePlus System and Power SYBR Green PCR Master Mix (Applied Biosystems). The PCR parameters were: 10 min at 95°C, 40 cycles of amplification (10 s at 95°C, 10 s at 58°C, 10 s at 72°C) and a melting curve stage (15 s at 95°C, 1 min at 60°C increased to 95°C with steps of 0.3°C). Melt curve analysis was performed as described in the previous section. Elongation factor 1α (EF1α; At5g60390) was used as a reference gene (Becher et al., 2004). Transcript levels were normalized to the respective transcript level of EF1α. Relative log₂ induction ratios of treated samples compared with the mock treatment were calculated based on the ΔΔCt method (Livak and Schmittgen, 2001).

Samples from three independent sets of both ISR-primed (At + T) and ISR-boosted (At + B + T) plants, and from the corresponding control plants (At and At + B) were used for microarray analysis. For all samples a dye swap was performed. Three biological replicates were included to assess technical and biological variation. RNA was extracted using a combination of Trizol^® Reagent (Invitrogen Life Technologies, Paisley, UK) and a Qiagen RNeasy kit (Qiagen, Hilden, Germany). RNA quality control, labeling, hybridizations and imaging were performed at the MicroArray Facility (VIB, Leuven) according to the protocols specified on the web site (http://www.microarrays.be). Agilent Arabidopsis 4-pack microarrays (Agilent Technologies, Palo Alto, CA, USA) were used, normalization was done with the accompanying software (Agilent Technologies, Palo Alto, CA, USA). The correlations among the replicates (0.98–0.99 among technical replicates and 0.92 among dye swaps) and the scatter plots confirmed the high quality and reproducibility of the microarray experiments. Data from different hybridizations were centered and scaled using quantile normalization, implemented in R (R Development Core Team, 2011). For ISR-prime, ISR-boost and BIDR differentially expressed genes, defined as genes of which the expression level is significantly modified (raised or reduced) in a specific condition as compared to a control treatment were selected using the (adapted) t-test developed by Tusher et al. (2001). Analysis of enrichment of gene ontology (GO) terms was performed on the five sets of differentially expressed (DE) genes that were produced (ISR-prime, BIDR on 1 and 2 dpBi, ISR-boost on 1 and 2 dpBi) as described previously (Tavazoie et al., 1999). In short, each gene in the DE gene sets was attributed its corresponding GO terms (Gene Ontology Consortium, 2000) from the biological process ontology using the GO annotation on the TAIR website (Lamesch et al., 2012). Next, the hypergeometric probability statistic was used to calculate the probability that at random each GO term would have the observed number of instances among the DE gene sets as follows:

where m is the total number of DE genes in a specific set, N is the total number of genes in the genome for which GO annotation is available, n is the total number of genes that are annotated with a specific GO term and k is the number of genes which belong to the set of DE genes and are annotated with that GO term. Hypergeometric probability calculations were implemented in R (R Development Core Team, 2011).

Histochemical GUS-staining was performed as described (De Bondt et al., 1994) except that tissues soaked in substrate buffer were vacuum infiltrated for 5 min. prior to overnight incubation at 37°C.

Anthocyanin content was determined using a procedure modified from that of Neff and Chory (1998). For six replicates of four plants each, leafs of both ISR-primed (At + T) and ISR-boosted (At + B + T) plants, and from the corresponding control plants (At and At + B) were weighed, powdered in liquid nitrogen, and total plant pigments were extracted 48 h at 4°C with 0.6 ml methanol containing 1% HCl (w/v). After centrifugation (14000 rpm, 5 min), anthocyanin was extracted with 0.6 ml of chloroform. The absorbance of the aqueous phase was measured at 535 and 657 nm (A₅₃₅–A₆₅₇) and corrected for weight.