Introduction

Faba bean (Vicia faba L.) is one of the most important legume crops providing protein for both human and animal consumption in many regions of the world (Duc et al. 2010; Rubiales 2010). However, the production of this crop is limited by various factors including diseases, which plays an essential role (Stoddard et al. 2010). Chocolate spot is one of the most important foliar diseases affecting the crop; under propitious conditions (i.e. mild temperatures around 15–22 °C and relative humidity higher than 80 %), this disease may cut down yields drastically (Villegas-Fernández et al. 2009). The foliar disease is caused by Botrytis fabae Sard., which is related to the necrotrophic fungus Botrytis cinerea Pers.: Fr. B. cinerea is a polyphagous pathogen with a wide range of host (more than 200 species), including strawberries, grapes and tomatoes (Staats et al. 2005). This pathogen may also infect faba bean plants producing chocolate spot symptoms, but without serious consequences (Harrison 1988).

Spraying crops with fungicides is the only available measure to effectively control chocolate spot. However, fungicides are often too expensive and not environmentally adequate (Stoddard et al. 2010). A number of agronomic practices such as intercropping can reduce this infection (Fernandez-Aparicio et al. 2011) but needs to be further adjusted. Growing resistant varieties becomes a key strategy when dealing with this disease since they could replace chemical measures. Although various researches have been carried out to find the sources of resistance and incomplete-resistant varieties have been introduced in some countries, further work is needed (Sillero et al. 2010). It is important to detect plants with a higher degree of resistance and look into the mechanisms accounting for such resistance. This will help to understand plant protection against infection. Characterising mechanisms of resistance are comprehensive; it covers from the macroscopic level to the molecular level. Unfortunately, there has been no important development of infrastructure, both in terms of knowledge and physical capacity, to carry out molecular studies of V. faba (Varshney et al. 2009).

Medicago truncatula Gaertn is a well-established model plant for legumes, and a considerable amount of molecular and physiological information is available (Rose 2008; Young and Udvardi 2009). M. truncatula has been described as a promising system for studying plant–pathogen interactions and some studies show interesting results (Rispail et al. 2010; Tivoli et al. 2006b). With regard to Botrytis spp., to our knowledge, these pathogens have not been reported to damage M. truncatula in nature, so it presents as a non-host resistance to Botrytis spp. Non-host resistance has been widely researched in different pathosystems (some examples: Hadwiger (2008); Die et al. (2009); Lipka et al. (2008)) and this type of resistance is largely based on the response by the plant at the molecular level (Holub and Cooper 2004; Nurnberger and Lipka 2005). Nevertheless, it is possible to bring about infection on M. truncatula by Botrytis spp. under controlled conditions in the laboratory (Ellwood et al. 2007); so, breaking the natural non-host resistance of M. truncatula to Botrytis spp. The study of M. truncatula/Botrytis spp. interactions might provide useful information for the better understanding of the V. faba/B. fabae pathosystem.

Transcription factors (TFs) are crucial in plant development and differentiation, they regulate the expression of complex traits in plants (Udvardi et al. 2007), such as resistant responses to biotic stresses (Singh et al. 2007). Identifying the TFs involved in these responses would be an important step to unravel the intricate mechanisms responsible for resistance, and could open the door to manipulation for the future of the metabolic pathways implicated (Iwase et al. 2009). TFs present low expression levels compared to those of other genes; hence, a very sensitive technique such as quantitative real-time PCR (qPCR), which is estimated to be at least 100-fold more sensitive than DNA arrays in transcripts detection (Czechowski et al. 2004), is required for the study of TFs. In the case of M. truncatula, a genome guide qPCR-based platform for a quantitative measurement of 1,084 TF genes is available (Kakar et al. 2008) and has been employed in various studies (Gao et al. 2010; Madrid et al. 2010; Verdier et al. 2008).

The objective of this work was to carry out a gene expression profile of M. truncatula transcription factors and identify those involved in the interaction with pathogens B. fabae and B. cinerea. The evaluation of the differential responses of two M. truncatula genotypes to the pathogens was first established under laboratory-controlled conditions.

Materials and methods

Evaluation of Medicago truncatula genotypes

The responses of two accessions of M. truncatula to separate inoculation with two isolates of B. fabae and B. cinerea were assessed at the seedling stage under laboratory-controlled conditions. Two consecutive series of experiment were performed, with four replications per genotype in the first experiment and ten replications in the second experiment (each replication consisting of a single plant). A third experiment was carried out to perform microscopic observations of the first stages of the infection at two different times (with three replications per time). Management of plant and pathogen material and inoculation procedures were the same for the three experiments.

Plant material

Two genotypes of M. truncatula supplied by Dr. J. M. Prosperi (INRA, Montpellier, France), were employed for this study, namely, line A17 (cultivar ‘Jemalong’) and line Esp162. These accessions had presented the most extreme reactions to the disease in a previous evaluation of a germplasm collection of M. truncatula (data not shown). Plants were grown in a growth chamber at 20 °C with a photoperiod of 14 h of visible light (150 μmol m−2 s−1 photon flux density) and 10 h of darkness.

Pathogen

A monosporic isolate of B. cinerea (Bc-CA-06) and another one of B. fabae (Bf-CO-05) were used in these tests. The B. cinerea isolate was multiplied on 9-cm Petri dishes with potato dextrose agar medium (Oxoid) and kept at 19 °C with a cycle of 12 h of visible light + near-ultraviolet radiation and 12 h of darkness for 10 days, after which time, a carpet of sporulating mycelium covered the surface of the plate. The B. fabae isolate was grown in modified V8 medium as described by Villegas-Fernández et al. (2012) until sporulating mycelium was attained.

Inoculation

A spore suspension of each isolate was obtained by adding glucose solution (1.2 % w/v) to the plates where the fungi had sporulated (10–15 ml per dish) and then scraping the surface of the medium with a needle to dislodge the spores. The suspension was filtered through a sieve and the spore concentration was adjusted to 1,150,000 spores/ml with a hematocymeter slide. Tween-20 was finally added (0.03 % v/v).

Plants were sprayed with approximately 2.5 ml of the spore suspension, taken to an incubator and kept in the darkness for 48 h at 20 °C with approximately 100 % continuous relative humidity. The plants were transferred to a growth chamber under the same conditions; the pots in the trays were covered with plastics whose inner sides were sprinkled daily with water in order to maintain a relative humidity greater than 95 %.

Disease-response evaluation

Evaluation was made at two time points: 2 and 8 days after inoculation (2 and 8 dai, respectively). These points were chosen on the basis of previous studies, in which the appearance timing of symptoms had been established. Sporulation would start 8 days after inoculation, giving rise to a second cycle of infection that would interfere with the evaluation, so no more evaluations were made beyond that time point. The recorded parameter was disease severity (DS), expressed as the percentage of foliar surface of each plant covered with lesions, and a correction to increase by 50 % the weight of the sporulating lesions, when present:

$$ \mathrm{Disease}\ \mathrm{severity}=\left(\%\ \mathrm{non}-\mathrm{sporulating}\ \mathrm{lesions}\right)+1.5\kern0.5em \times \kern0.5em \left(\%\ \mathrm{sporulating}\ \mathrm{lesions}\right) $$

Microscopy observations

Pathogens spore germination was determined at the microscopic level. Plant samples were taken twice (24 and 48 h) after inoculation. One complete leaf was detached from each plant, its three leaflets separated and processed according to Rubiales and Moral (2004): they were laid with the adaxial surface up on filter paper dipped in fixative (1:1, absolute ethanol/glacial acetic acid, v/v). After several changes of the fixative, the leaflets were transferred to filter paper, moistened with tap water for at least 2 h, then to lactoglycerol (1:1:1, lactic acid/glycerol/water, v/v/v) for at least, another 2 h. A drop of trypan blue in lactoglycerol (0.1 %, w/v) was placed on a cover glass; the sample was carefully laid with the adaxial surface toward the cover glass and then mounted in lactoglycerol on a microscope slide. The percentage of germinated spores was determined under ×200 magnification (a minimum of 100 spores per leaflet was observed; those with a germ tube and those without were counted, and the proportion was subsequently calculated).

Data analysis

The statistical software employed was Statistix 8 (Analytical Software, Tallahassee, FL, USA). The factorial analyses of variance (factors being genotype and pathogen) were carried out for the studied parameters in the evaluation experiments (i.e., severity and percentage of germination) at different times. Data were arcsine transformed as required. The means were separated using least significant difference (LSD) test (p = 0.05).

Gene expression profiling of transcription factors

A large-scale profiling of transcription factors involved in the interaction of M. truncatula–Botrytis spp. was carried out. The tested M. truncatula genotypes were A17 and Esp162. Three different treatments were applied to six plants of each genotype: (1) inoculation with a B. fabae isolate, (2) inoculation with a B. cinerea isolate and (3) no inoculation. Three biological replications were performed.

Plant material, inoculation and sampling

Plant and pathogen management and inoculation were carried out in the same way as described in the evaluation experiment. Non-inoculated plants were sprayed with the glucose and Tween-20 solution. Sampling was made 48 h after the inoculation (after taking plants out of the incubator). All the leaves of the six plants subjected to the same treatment were detached and pooled, thus making up an individual sample, immediately frozen in liquid nitrogen for storage at −80 °C.

Total RNA extraction and cDNA synthesis

Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Its integrity was checked via denaturing agarose gel electrophoresis (2 % w/v) and the RNA concentration and purity were spectrophotometrically determined. To remove genomic DNA contamination, 200 μg of total RNA were treated with RNAse-free DNAse RQ ITM (Promega, Madison, USA); the RNA integrity, concentration and purity were again determined. PCR with primers specific for M. truncatula ubiquitin gene was employed to confirm that no genomic DNA remained (Kakar et al. 2008). First-strand cDNA was synthesised with the aid of SuperScriptTM III reverse transcriptase (Invitrogen, Carlsbad, USA) according to the manufacturer’s protocol.

Real-time PCR

A collection of primer pairs corresponding to 1,084 putative transcription factors (TFs) in M. truncatula was generated using available genomic databases (Kakar et al. 2008). Polymerase chain reactions were carried out as described by Kakar et al. (2008) in optical 384-well plates with an ABI PRISM® 7900HT sequence detection system (Applied Biosystems, Foster City, CA, USA) using SYBR® Green to monitor double-stranded DNA (dsDNA) synthesis. Four housekeeping genes (MtUbi, MtGapdh, MtPdf2, MtEf1) were replicated twice in each plate.

Data analysis

SDS 2.0 software (Applied Biosystems) was used to analyse the data. Cq values were those corresponding to a threshold of 0.3. Amplification reactions with more than one melting curve peak were not further considered. Posterior analysis included the determination of PCR efficiency (E) with the help of LinReg software (Ramakers et al. 2003). Only reactions whose efficiencies were greater than 1.6 were chosen for this analysis. Software application BestKeeper (Pfaffl et al. 2004) was employed to identify the most stable housekeeping genes for use as reference genes.

Gene expression value was calculated as ECq. The normalisation index in each plate was defined as the geometric mean of the expression values of the reference genes in the plate; expression level for every gene was subsequently normalised by dividing the values by that index. Finally, factorial analyses of variance were carried out for these normalised expression values; factors were genotype and pathogen (with three levels: inoculation with B. fabae, inoculation with B. cinerea and no inoculation). Means were separated using LSD test (p = 0.05). These statistical analyses were performed using Statistix 8 (Analytical Software, Tallahassee, FL, USA).

The TFs showing significant differences only for pathogen were clustered using a k-means clustering analysis. A data matrix with the log2 of the ratios of the inoculated/non-inoculated expression values was employed to generate a Euclidean distance matrix. Cluster v3.0 (Eisen et al. 1998) was used for clustering and TreeView v1.05 software was used for visualising.

Results

Evaluation of Medicago truncatula genotypes

Before carrying out the expression profiling of TFs of M. truncatula, it was necessary to identify genotypes of M. truncatula showing differential reaction to infection by Botrytis spp. It was equally important to characterise the differential aggressiveness of both pathogenic species (B. cinerea and B. fabae) against M. truncatula. Those were the objectives of the first experiment under controlled conditions, confronting two genotypes of M. truncatula against two isolates of B. cinerea and B. fabae. The genotypes were chosen on the basis of previous tests (data not shown).

First symptoms of infection appeared as soon as 2 dai. These were small necrotic spots covering leaf tissue. They were not abundant, so severity was still low, being the highest values just around 7 %. Nevertheless, at this early stage, it was already possible to discriminate between genotypes, though still not between isolates: genotype A17 presented very low severity compared to that of genotype Esp162, while both B. fabae and B. cinerea showed similar aggressiveness (Fig. 1a shows the results for each genotype/pathogen combination). The factorial analysis of the variance for severity confirmed this, showing significant genotypic effects, but not detecting differences between isolates. No interaction between genotype and pathogen was detected either (Supplementary Table 1).

Fig. 1
figure 1

Severity values for two genotypes of M. truncatula (A17 and Esp162) 2 days (a) and 8 days (b) after inoculation with either B. cinerea or B. fabae. Vertical bars represent means of standard deviations (LSD test for arcsin √x transformed data; p < 0.05)

Full size image

Symptoms had increased at 8 dai. Not only did they consist of necrotic spots, but also whole necrotic areas had developed in some leaves, resembling aggressive lesions typical of infection by B. fabae on faba bean (Harrison 1988). Some of these lesions had even started sporulating. Again, these symptoms were more frequent in genotype Esp162 (Fig. 2), reaching very high levels in some cases (∼60 %); on the contrary, severity values for genotype A17 ranged from 6 to 12 % (Fig. 1b). On this occasion, however, it was possible to discriminate between the pathogens: the isolate of B. fabae was more aggressive than the isolate of B. cinerea. All this was confirmed by the factorial analysis of variance (Supplementary Table 2). Actually, interaction between genotype and pathogen was detected. This interaction is explained because the response of A17 to both pathogens was quite similar; whereas the differences in aggressiveness between B. cinerea and B. fabae were found on genotype Esp162.

Fig. 2
figure 2

Leaves of genotype Esp162 of Medicago truncatula infected by Botrytis fabae (8 days after inoculation), showing typical necrotic symptoms (a) and complete necrosis with sporulation (b)

Full size image

This first experiment, hence, permitted the identification of two genotypes of M. truncatula showing differential responses to Botrytis spp.: genotype A17, presented a reaction that can be defined as resistant, while genotype Esp162 was susceptible to the pathogens. Additionally, the superior aggressiveness of B. fabae over B. cinerea was established.

Once this had been done, the next step was the identification of the best time point for the collection of samples for the profiling of TFs. The microscopic observation of the early stages of the interaction plant/pathogen was deemed useful for this purpose.

The spores of Botrytis spp. germinate producing hyphae that grow on the leaf surface until they penetrate the plant epidermis. Then, they grow in the mesophyll, branching in different directions and killing cells to feed on them. We wanted to focus on the initial stages of infection, so the aim of the microscopy experiment was to evaluate spore germination after inoculation. We tested both genotypes of M. truncatula, in case any difference in the behaviour of the pathogens on them could be detected.

The rate of spore germination of B. fabae was below 30 % in both A17 and Esp162 at 24 h after inoculation, lower than that of B. cinerea (∼80 % for both genotypes of M. truncatula). At 48 h after inoculation, the rate of spore germination for B. fabae had risen up to 60 %, still lower than the rate for B. cinerea, which had only increased slightly (Fig. 3). The analysis of variance confirmed these differences for factor pathogen at both time points, whereas it did not detect any difference for genotype or the interaction genotype–pathogen (Supplementary Tables 3 and 4). Therefore, the differential reaction to infection by Botrytis spp. of these genotypes of M. truncatula is not related to the process of germination of the spores of the pathogens.

Fig. 3
figure 3

Germination values for spores of B. fabae and B. cinerea isolates on two genotypes of M. truncatula (A17 and Esp162) 24 and 48 h after inoculation. Vertical bars represent means of standard deviations

Full size image

Considering the results of these two experiments, we chose the time point for the collection of samples for the profiling of transcription factors at 48 h after inoculation. At this time, a considerable number of pathogen spores are germinated and some symptoms have become visible, showing that the infection is already under way, but still in its initial stages.

Gene expression profiling of transcription factors

The objective of this experiment was to identify those transcription factors of M. truncatula involved in the interaction with Botrytis spp. The two chosen genotypes of M. truncatula present differential responses to Botrytis spp., and the pathogens (B. fabae and B. cinerea) present differential aggressiveness; hence, our study had to consider both the effect of plant genotype and of pathogen.

A total of 874 out of the 1,084 TF-encoding transcripts (i.e. 80.6 %) could be amplified. The analysis with Bestkeeper allowed choosing MtGapdh, MtPdf2 and MtEf1 as the best reference genes, since they were the most stable across the different biological situations tested. These three genes were used to calculate the index for normalisation. The ensuing analyses of variance showed that 170 amplified transcripts showed significant differences only for genotype (Supplementary Table 5), 126 amplified transcripts showed significant differences just for pathogen (Supplementary Table 6) and, finally, 37 amplified transcripts showed significant differences for factors genotype and pathogen (Table 1). In Table 2, information about the proportions of the families of TFs of M. truncatula present in these three groups is provided.

Table 1 Transcription factors that have presented significant differences in expression for both genotype and pathogen after inoculation with either B. fabae or B. cinerea. Their ascription to the different families of transcription factors of Medicago truncatula is provided. Different ratios of expression levels are shown
Full size table
Table 2 Percentages of the genes that have shown significant differences for genotype, pathogen or both genotype and pathogen after inoculation with Botrytis spp., belonging to the different families of transcription factors of Medicago truncatula
Full size table

A k-means clustering was performed on the 126 genes showing significant differences for pathogen in order to facilitate their analysis. This resulted in their classification into seven different clusters (Fig. 4). These clusters correspond with the different patterns of reaction to inoculation by the pathogens. Clusters 1 and 5 comprise TFs which are similarly down-regulated on infection by either B. fabae or B. cinerea (down-regulation were higher in cluster 1 than in cluster 5); in cluster 2, on the contrary, up-regulated TFs after infection by both pathogens are found; induced TFs only after inoculation with B. fabae are present in cluster 3; cluster 4 includes those TFs which undergo repression after infection, being this repression higher for B. fabae than for B. cinerea; finally, cluster 6 presents the repressed TFs after infection with B. cinerea and cluster 7 shows the repressed TFs after infection with B. fabae. These various behaviours are shown in Fig. 4. As for the family distribution, cluster 2 includes three genes belonging to the Zn-Finger (C2H2 type) family, two genes belonging to the WRKY family, one gene belonging to the AP2/EREBP family, one gene to the BHLH family and one gene to the HD-like family. In cluster 3, five AP2/EREBP, four MYB, two Zn-finger (one C2H2 type and one CO-like type), one GRAS and one NAC are found, among others. In cluster 4, five AP2/EREBP, two Myb, two BHLH, one Zn-Finger (YABBY type), one bZIP, one PHD and one TCP are found. Clusters 1, 5, 6 and 7 are made up of TFs belonging to different families, no family standing out in particular.

Fig. 4
figure 4

Heat maps of the expression profiles (log2 of the ratio inoculated/non-inoculated) of the TFs presenting significant differences for pathogen, separated in the clusters obtained by K-means clustering (colour scale is shown: green for down-regulation, red for up-regulation and black for values close to 0); the number of genes included in each cluster is indicated on its left. The average of the ratios for each cluster is represented to its right, blue bar being for B. fabae and red bar for B. cinerea

Full size image

As for the 37 genes showing significant differences for genotype and pathogen, most of them presented down-regulation, its degree depending on the genotype. Only three genes showed up-regulation for resistant genotype A17 after inoculation.

These results show that nearly a fifth of the 874 amplified TFs (126 for pathogen and 37 for both genotype and pathogen) are involved in the interaction M. truncatula/Botrytis spp.

Discussion

In this work, we have studied the interaction between M. truncatula and two important causal agents of diseases in legumes at the level of the regulation of transcription. Given that M. truncatula is not a natural host of these pathogens, the first step was to set up a system that reproduces an effective infection process. Although the procedures have been described before (Ellwood et al. 2007), no differential genotypic responses to Botrytis spp. had been previously reported. We have identified two genotypes of M. truncatula that may be defined as partially resistant (A17) and susceptible (Esp162) to Botrytis spp. under test conditions. The severity values of their differential responses are similar to those obtained by genotypes of V. faba in experiments under controlled conditions (Villegas et al. 2012). It is remarkable that B. fabae has turned out to be more aggressive than B. cinerea. The former is a specialist pathogen that infects a few species of the Fabaceae, mainly V. faba (Staats et al. 2005), while the latter is a generalist pathogen responsible for diseases in legumes such as Lens culinaris and Medicago sativa (Tivoli et al. 2006a; Williamson et al. 2007), not posing a serious challenge for V. faba. The behaviour of M. truncatula resembles more of V. faba than that of other legumes in the case of these pathogens, thus making it a good model plant for the study of the V. faba/Botrytis spp. pathosystem.

The microscopy studies have shown that B. cinerea spores reached the maximum level of germination at 24 h after inoculation; the spores of B. fabae need longer time to germinate, always maintaining a lower rate of germination than B. cinerea. This might account for similar severities caused by both pathogens only 48 h after inoculation; later, however, after the number of germinated spores of B. fabae has increased, this pathogen proves to be more aggressive.

The availability of this biological material has allowed us to investigate in two different directions: firstly, through the identification of genes of M. truncatula involved in the process of infection by Botrytis spp., focusing on specific reactions to each pathogen, different in aggressiveness as they are; and secondly, through the identification of genes that characterise the differential response of each genotype when confronted with a given pathogen. Timing for sampling was chosen on the basis of this macroscopical and microscopical information, considering that 48 h after the inoculation, infection had just started to take place, and the plant would be reacting to it.

Nearly 20 % of the 874 amplified genes changed their expression levels 48 h after inoculation with either B. fabae or B. cinerea (37 genes showed significant differences for genotype and pathogen, while 136 genes showed for only pathogen, regardless of the genotype). This proportion is higher than that (∼13 %) found for Uromyces striatus (Madrid et al. 2010) in the only other study reported so far to have used the M. truncatula TF qPCR platform to assess the response to fungal diseases. We can hypothesise that this difference is because both B. fabae and B. cinerea are necrotrophic pathogens that quickly penetrate plant tissue, grow inside it and feed on plant cells after killing them, while U. striatus is a biotrophic fungus that takes a longer period to expand inside and damage the plant, therefore causing higher physiological alteration at the beginning of infection by Botrytis spp.

Cluster analysis allowed the grouping of the TFs that differentially responded to infection. Most of these TFs were down-regulated (99 out of 126, 57 independently of the pathogen); they are of various types, belonging to different families, and though their repression could be due to the hampering of biological processes by the pathogen, it is also true that down-regulation may be an integral part of a defensive response by plants (Bustos et al. 2010). Further work would be necessary to elucidate this issue, and establish the role of each of these TFs under these circumstances. As for the nine genes up-regulated after infection with both pathogens, six of them belong to families related to stress responses: AP2/EREBP, which is a unique family to plants, characterised by the presence of the AP2 DNA-binding domain, and is known to be linked to the response to abiotic and biotic stresses (Dietz et al. 2010); C2H2 or Cys2/His2-type zinc finger proteins, which have been implicated in the response to different stresses, both biotic and abiotic (Kielbowicz-Matuk 2012); and WRKY, also an exclusive family to plants, whose members constitute what has been called a ‘web of plant defence regulators’ against pathogens (Eulgem 2006). These genes might constitute the core of the defensive reaction of M. truncatula to Botrytis infection.

Of special interest is the differential response observed with the two pathogens. Apart from the nine genes previously mentioned, 18 TFs are also up-regulated when infection is caused by B. fabae. Again, among them, we find genes belonging to the families related to reaction to stresses: AP2/EREBP and C2H2, as previously seen; GRAS, a family of proteins that, though not known to bind to a specific DNA sequence, are classified as transcriptional regulators, and some of its members are described as essential for the activation of stress-inducible promoters (Fode et al. 2008); NAC, which has been reported as key in regulating stress perception (Jensen et al. 2010); and Myb (one homeodomain-like), a family widely represented in plants whose members have been assigned many roles, in some cases linked to defence regulation (Dubos et al. 2010). All of these points to an increase in the defence responses in M. truncatula when confronted with B. fabae higher than with B. cinerea. Our results show that B. fabae is more aggressive than B. cinerea; therefore, we may assume that the plant unchains additional defensive mechanisms to tackle a more serious challenger. It is more remarkable considering that 24 h after inoculation, there are more germinated spores of B. cinerea than of B. fabae, therefore, the number of infection sites is, in principle, higher for the former than for the latter; however, it seems that the plant recognises B. fabae as a more dangerous foe even at this early stage and reacts accordingly. In the case of the V. faba/Botrytis spp. pathosystem, it has been suggested that B. fabae is more aggressive than B. cinerea because it is capable of metabolising the phytoalexin (wyerone acid) produced by the plant (Harrison 1988). It has also been reported that B. fabae produces a phytotoxin (regiolone), which is more toxic against V. faba than the equivalent toxin in B. cinerea (isosclerone) (Evidente et al. 2011). Our hypothesis suggest that something similar could happen in the case of M. truncatula/Botrytis spp., therefore leading the plant to resort to extra defence resources in an attempt to stop infection. Finally, it is also noticeable that there are different genes down-regulated depending on the type of pathogen (20 genes for B. cinerea and 9 genes for B. fabae). As in the previous case, it is difficult to know if this is part of a defence response or if it is a reflection of the hindrance of the different physiological processes by each pathogen. Nevertheless, it would not be likely to expect that B. cinerea would hamper more physiological processes than B. fabae, given that the former has turned out to be less aggressive a pathogen than the latter. So, it is possible to assume that most of those 20 genes, down-regulated after infection by B. cinerea, are somehow related to defence mechanisms. All these results reveal that 48 h after inoculation, although no difference is detected at the macroscopic level; however, at the molecular level, the plant–pathogen interaction is specific for each Botrytis species, anticipating the diverse reaction clearly obvious after some days with the difference in aggressiveness of each pathogen.

As for the differential response of the tested genotypes to infection, it is noteworthy that most of the genes presenting different behaviours for both the genotype and the pathogen are down-regulated. Although it is difficult to determine whether this repression is part of the defence reaction, there are some cases pointing to this direction: TFs 1443.m00022, 1699.m00013 and 939.m00019 are all down-regulated in resistant genotype A17 and up-regulated in susceptible Esp162. The same could be said of other genes as 980.m00009 or 989.m00004, which show lower levels of expression in non-inoculated A17 than in non-inoculated Esp162, and experience further repression in A17 after inoculation with both pathogens; the low expression may be related to a more effective defensive response. On the contrary, there are some genes that, in spite of undergoing down-regulation, present higher expression levels in infected A17 than in infected Esp162. We could assume that the down-regulation is, on this occasion, a consequence of malfunction due to the presence of the pathogen, and that the resistant genotype has managed to alleviate this effect. On the other hand, only three TFs undergo up-regulation after inoculation, two of these are common to the infection by B. fabae and B. cinerea (1103.m00014 and 1323.m00024) and the third TF is exclusively up-regulated after B. fabae inoculation (1524.m00006). All of them reach higher levels of expression in A17 than in Esp162 after inoculation, therefore, they may be suggested as taking part in the defensive reaction against Botrytis spp.

Apart from these induced responses to infection, constitutive defences should not be disregarded. M. truncatula presents, in principle, non-host resistance to Botrytis spp.; although we have managed to overcome this resistance and achieve infection, it is likely that the resistant reaction shown by A17 is largely based on non-host resistance. It is generally accepted that non-host resistance consists of two components: constitutive and induced resistance. The constitutive resistance is based on a series of different elements, ranging from structural barriers to inhibitory metabolites or proteins (Ferreira et al. 2006). We found a great number of TFs (170) showing constitutive differences between the two genotypes regardless of inoculation with the pathogens. This result is in contrast with that of Madrid et al. (2010), who only reported nine TFs presenting constitutive differences between their tested genotypes; this suggests that the genotypes selected for our work differs greatly one from the other. It is probable that some of these 170 TFs are involved in constitutive resistance to infection. Among these TFs, there are members of families related to plant resistance, such as C2H2, AP2/EREBP, WRKY or NAC; NAC has been reported to be involved in non-host resistance (Uma et al. 2011).

Finally, it is of interest to notice that only 2 of the 13 defence-related genes pointed out by Madrid et al. (2010) appeared as influenced by infection in our results. TF 1082.m00015 is up-regulated in the resistant response to U. striatus, while 1063.m00010 is repressed. In our case, the behaviour is quite similar, since 1082.m00015 is up-regulated and 1063.m00010 is down-regulated after inoculation with both B. fabae and B. cinerea (with no difference between genotypes). In any case, it is not surprising that most of the TFs involved in the reactions to each of the diseases are not the same: the studied systems are very different, both for the type of pathogen and for the genotypes of M. truncatula chosen (it is remarkable that A17 is susceptible to U. striatus and presents resistance to Botrytis spp.).

In conclusion, this work has established a system to study M. truncatula/Botrytis spp. interactions that may be a useful model for the V. faba/Botrytis spp. pathosystem. The profiling of transcription factors has permitted the identification of, on the one hand, those involved on the differential responses to B. fabae and B. cinerea and, on the other hand, those participating in the resistance to these pathogens. Future work should focus on assigning a clearer role to these TFs, describing and integrating the pathways in which they may take part, and complementing these results with additional molecular and pathological studies. This would increase the functional knowledge of the M. truncatula genome and help in breeding programmes of V. faba for resistance to chocolate spot.