Genome-wide microarray analysis of tomato roots showed defined responses to iron deficiency

Abstract

Background

Plants react to iron deficiency stress adopting different kind of adaptive responses. Tomato, a Strategy I plant, improves iron uptake through acidification of rhizosphere, reduction of Fe³⁺ to Fe²⁺ and transport of Fe²⁺ into the cells. Large-scale transcriptional analyses of roots under iron deficiency are only available for a very limited number of plant species with particular emphasis for Arabidopsis thaliana. Regarding tomato, an interesting model species for Strategy I plants and an economically important crop, physiological responses to Fe-deficiency have been thoroughly described and molecular analyses have provided evidence for genes involved in iron uptake mechanisms and their regulation. However, no detailed transcriptome analysis has been described so far.

Results

A genome-wide transcriptional analysis, performed with a chip that allows to monitor the expression of more than 25,000 tomato transcripts, identified 97 differentially expressed transcripts by comparing roots of Fe-deficient and Fe-sufficient tomato plants. These transcripts are related to the physiological responses of tomato roots to the nutrient stress resulting in an improved iron uptake, including regulatory aspects, translocation, root morphological modification and adaptation in primary metabolic pathways, such as glycolysis and TCA cycle. Other genes play a role in flavonoid biosynthesis and hormonal metabolism.

Conclusions

The transcriptional characterization confirmed the presence of the previously described mechanisms to adapt to iron starvation in tomato, but also allowed to identify other genes potentially playing a role in this process, thus opening new research perspectives to improve the knowledge on the tomato root response to the nutrient deficiency.

Background

Iron (Fe) deficiency is a yield-limiting factor for a variety of field crops all around the world and generally results from the interaction of limited soil Fe bioavailability and susceptible genotype cultivation [1]. Iron is an important microelement for plant life due to its involvement as redox-active metal in photosynthesis, mitochondrial respiration, nitrogen assimilation, hormone biosynthesis, production and scavenging of reactive oxygen species, osmoprotection and pathogen defence [2].

Under aerated conditions at neutral alkaline pH, the soluble Fe concentration in soil solution is very low. To cope with Fe shortage plants have developed two strategies for its acquisition. The Strategy I (all higher plants except grasses) relies on improvement of Fe uptake through acidification of soil solution by excretion of protons via a plasmalemma P-type ATPase resulting in an increased Fe solubility, reduction of Fe³⁺ to the more soluble Fe²⁺ by a Fe^III-chelate reductase and plasmalemma transport of Fe²⁺ by the activity of a Fe transporter [3]. Some model plants used to study Strategy I are dicots such as Arabidopsis thaliana, Solanum lycopersicum and Pisum sativum[4].

Plant responses to Fe deficiency have been recently analyzed on the basis of large-scale changes not only in transcriptome [5–14], but also in proteome [15–20] and metabolome [17]. Results of transcriptome analysis are influenced by differences in experimental plans, plant species and microarray platforms, and thus difficult to compare and be generalized. Notwithstanding this drawback, recently, a set of 92 transcripts that robustly reflect the transcriptional response of Arabidopsis to Fe deficiency [21], has been described as the "ferrome" by Schmidt and Buckhout [21]. The "ferrome" consists of a list of transcripts considered to be involved in the basic response to iron deficiency. The ferrome is particularly enriched in genes related to heavy metal cation transport and metal homeostasis. Focusing on tomato, a plant often used as a model to study Fe deficiency (Strategy I) and a crop of economic importance, no information is available at genome-wide transcriptional level. Two proteomic characterizations of tomato roots in response to 1-week of Fe deprivation showed 23 [15] and 15 [16] differentially expressed protein spots respectively. Modifications in proteome suggest changes in energy metabolism, sulfur metabolism, response to oxidative stress and signal transduction.

In the present work a genome-wide transcriptional characterization of tomato roots in response to Fe deficiency is presented. This approach allowed indentifying 97 differentially expressed transcripts involved in the responses to the nutritional stress. Transcriptional changes, mainly related to positive modulation of glycolysis, TCA and methionine cycle, suggest that tomato roots behave similarly to Arabidopsis under Fe deficiency. Furthermore, flavonoid biosynthesis and root morphological changes are revealed as specific tomato responses to Fe shortage.

Results and discussion

Responses to Fe-deficiency

Typical responses of Fe-deficiency [22] were observed in tomato plants grown for 14 d in the presence of a low amount of Fe and thereafter subjected to 7 d of Fe deprivation. The chlorophyll content (SPAD index value) was reduced in Fe-deficient plants (Table 1). A concomitant increase in root Fe^III-chelate reductase activity (Table 1) was also observed with values similar to those commonly found in roots of Fe-deficient tomato plants [23]. Furthermore, Fe-deprived tomato plants developed more lateral roots and showed an abundant production of root-hairs (Figures 1 and 2).

Table 1 Leaf SPAD index values and root Fe^III-chelate reductase activity

Figure 1

Shoot and root apparatus of tomato plants grown under different Fe-supply conditions. Leaf detail of Fe-deficient (A) and Fe-sufficient (B) plants. Shoot (C) and roots (E) of Fe-deficient plants and shoot (D) and roots (E) of Fe-sufficient plants.

Figure 2

Root apparatus of tomato plants grown under different Fe-supply condition. Detail of root apparatus of A) Fe-deficient and B) Fe-sufficient plants.

Comparison of root gene expression profiles in Fe-deficient and Fe-sufficient tomato plants

Differences in root gene expression between Fe-sufficient and Fe-deficient tomato plants were obtained by a genome-wide gene expression analysis using a tomato chip developed through Combimatrix technology [24]. This chip allows monitoring simultaneously the expression of more than 25,000 tomato transcripts. Ninety-seven differentially expressed transcripts between Fe-deficient and Fe-sufficient tomato roots (75 up-regulated and 22 down-regulated) were identified by Linear Models for MicroArray (LIMMA) [25] (adjusted p-value ≤ 0.05; |FC| ≥ 2). This result obtained using a large-scale chip reinforce the idea that plant transcriptional response to Fe shortage is based on the modulation of a relative small set of genes as previously observed for the Arabidopsis "ferrome" [21].

Manually curated annotation of the 97 differentially expressed transcripts was based on results of BlastP analysis against UniProt [26] database (Figure 3; Additional file 1) using terms of biological process of Gene Ontology (GO) [27]. Sequence grouping in functional categories according to the GO terms revealed that the most abundant functional category was "metabolic process" both for up-regulated and down-regulated transcripts (35% and 45% respectively; Figure 3). Other up-regulated transcripts belonged to "establishment of localization" (12%) and "cell wall organization and biogenesis" (8%), while for the down-regulated transcripts "response to stimulus" (9%) was one of the most representative main functional categories (Figure 3). Only up-regulated transcripts are present in the "secondary metabolic process" category (Table 2). Transcripts encoding proteins with no sequence homology to known proteins were defined as "no hits found" (12% and 5% for up-regulated and down-regulated transcripts respectively), while a similar percentage of transcripts showed homology to proteins involved in "unknown" biological process (24% and 23% respectively). A selection of differentially expressed and discussed in relation to Fe deficiency is reported in Table 2. The up-regulation and down-regulation of six differentially expressed transcripts in response to Fe deficiency were confirmed through Real-time RT-PCR experiments (Table 3).

Figure 3

Functional categories distribution of differentially expressed transcripts. Distribution in main functional categories according to the GO "biological process" terms of the 75 up-regulated (A) and 22 down-regulated (B) transcripts in roots of Fe-deficient relative to Fe-sufficient plants. Percentage of transcripts is reported for each functional category.

Table 2 List of transcripts modulated in response to Fe-deficiency and reported in the Discussion

Table 3 Real-time RT-PCR validation of a set of genes differentially expressed in microarray analysis

TC ID, description, Real-time RT-PCR relative expression value (Fe-deficient vs. Fe-sufficient) and microarray fold change value (Fe-deficient vs. Fe-sufficient). Real time data were normalized on the EF1a gene and were performed 4 times on 3 independent experiments. Real-time RT-PCR data are expressed as mean ± SD.

Sixty-one of 97 transcripts are ascribable to adaptive responses to Fe deficiency involving Fe homeostasis, metabolic process, oxidative stress responses, root morphological modification, transport processes, hormone metabolism and signaling. The others are hardly related to specific role showing homology to protein without a specific biological process or lacking homology to known protein ("no hits found").

Fe homestasis

Our transcriptional analysis confirmed that roots of Fe-deficient tomato plants overexpressed genes involved in Fe uptake and reduction, including the transcripts encoding IRON-REGULATED TRANSPORTER (IRT) [28–30] and Fe^III-chelate reductase (FRO) [31]. The tomato bHLH protein (encoded by LeFER) plays a role in Fe-deficiency responses through the expression of these two tomato Fe mobilization genes belonging to the Fe uptake systems of the Strategy I plants [29, 32, 33]. Our data show the up-regulation of the FER transcript (#24) in Fe-deficient roots, which is in agreement with the positive modulation of Fe-uptake-related transcripts such as LeFRO1 (#68), LeIRT1 (#39) and a transcript encoding a NATURAL RESISTANCE-ASSOCIATED MACROPHAGE PROTEIN1 (LeNRAMP1; #38) [29, 32]. A positive modulation of FER and Fe mobilization proteins (IRT1, NRAMP1 and FRO1) was not found in two proteomics studies performed in tomato roots grown in conditions similar to those used in the present work [15, 16]. Authors justified these results as related to the features of proteomic approach, which was not sensitive enough to detect FER and not well suited for membrane-bound proteins. However, functional characterizations of fer mutant proved that FER controls the expression of the iron-uptake genes [29, 32, 33]. A transcriptional behaviour similar to that described in the present work was observed for Arabidopsis orthologous genes in Fe-starved roots [5, 8, 11, 34].

Our data show a strong down-regulation of another ferric chelate reductase (#70). Previous results indicated that the same transcript (TC194139 of the Release 12.0, corresponding to TC124302 of the Release 9.0) specific to the Solanum lycopersicum genome, is only slightly regulated by Fe and that its function is not essential for Fe uptake [34]. Our results showing a negative regulation of this FRO transcript in response to Fe-deficiency in roots, also quantified by Real-time RT-PCR experiment, confirm that this gene does not play a crucial role in deficiency-induced Fe uptake and suggest the involvement in other biological process.

Together with the positive modulation of LeFER, LeFRO1 and LeIRT1, we observed for the first time a high up-regulation of another bHLH transcript (#26); this result suggests that like in Arabidopsis[35] also in tomato plants the response to Fe deficiency through FER activity may need the interaction with another bHLH protein. BlastP analysis against TAIR database [36] using protein sequence obtained from the predicted coding sequence of the TC194645 showed the highest sequence homology with the protein encoded by AtbHLH38 (score: 120; Evalue: 9 E-28; identity: 35%; positives: 57%) known to interact with FIT, the Arabidopsis orthologous of tomato FER [35].

The down-regulation of two ferritin transcripts (#25 and 30) is in line with the negative regulation of ferritin genes observed in roots of Fe-depleted Arabidopsis plants [5, 8, 11]. It has been suggested that ferritins can be involved in Fe homeostasis [37] with a main role of plastidial Fe [11]. Arabidopsis nodulin-like genes have been recently described to be putatively involved in Fe transport and storage under metal cation sufficiency [38]. Transcriptional levels of Arabidopsis nodulin-like genes were down-regulated at least until 72-h of Fe-deficiency, while two other nodulin-like genes were not modulated in response to different Fe conditions [38]. On the other hand in tomato roots we recorded a positive modulation of two transcripts encoding protein with a nodulin-like domain (#9 and 22). BlastP analysis against TAIR database [36] using protein sequence obtained from the predicted coding sequence of the TC214599 (#22) showed the highest sequence homology with the protein encoded by At3g43660 (score: 325; Evalue: 2 E-29; identity: 54%; positives: 68%) which is one of the two Arabidopsis genes not modulated by iron [38].

Metabolic processes

Carbohydrate metabolism

As observed in Arabidopsis root microarray analyses [5, 8, 11] glycolysis-related genes are positively modulated in Fe-deficient roots. These transcriptional data fit well with the increased activity of glyceraldehyde-3-phosphate dehydrogenase (GADPH), pyruvate kinase (PK), and phosphofructokinase (PFK) recorded in response to Fe starvation in cucumber roots [39]. Furthermore, increased levels of protein related to glycolysis under Fe shortage were recently reported in sugar beat [17] and Medicago truncatula[20] roots. All together these evidences are consistent with the idea of a shift from anabolic to catabolic metabolism. In the present work, a fructose-bisphosphate aldolase (FBP aldolase) (#59) and a PFK (#58) transcripts were up-regulated under Fe deficiency, further confirming changes in primary metabolism in response to Fe starvation. The positive modulation of transcript encoding a PFK, an enzyme catalysing a protogenic reaction, supports the role of glycolysis in different process such as production of ATP and H⁺ for H⁺-ATPase, reducing equivalents for ferric chelate reductase and of phosphoenolpyruvate (PEP) [40]. Up-regulation of three phosphoenolpyruvate carboxylase (PEPC) transcripts (#48, 49 and 63) agrees with results of several proteomic and physiological studies on the response to Fe deficiency in tomato [15, 16, 41], sugar beat [17] and Medicago truncatula roots [20], showing a positive modulation of proteins involved both in glycolysis and TCA cycle. PEPC activity, through pyruvate consumption, can keep active the glycolytic pathway and give a contribution to the control of cytosolic pH [40, 41]. In Fe deficiency, starch catabolism was reported to be enhanced both at transcriptional [5] and protein level [16]. Our analysis revealed a down-regulation of a transcript encoding a protein showing homology to a potato chloroplastic ß-amylase (#52) involved into starch degradation in plastids [42]. Since starch catabolism mediated by this enzyme occurs in plastids, the negative modulation of ß-amylase could be ascribed to other causes than an accelerated glycolysis.

The positive modulation of a cytochrome C oxidase (#76) is in line with the transient induction of electron-transport-chain genes observed in Arabidopsis[5] and consistent with the enhancement of respiration rate observed in cumber [39] and sugar beat [43] Fe-deficient roots. This behaviour was interpreted as an attempt to increase energy production through oxidative phosphorylation. However, more recently it has been suggested that the increased respiration rate in root segments of Fe-starved cucumber plants should not be interpreted as an increase in mitochondrial activity but rather as the result of an increase in the number of less efficient mitochondria and of the induction of different O₂-consuming reaction [44].

Methionine cycle

Nicotianamine (NA) is considered to be a key molecule for long-distance transport of Fe in plants [45]. Proteomic analysis of Fe-starved tomato roots [16] showed a positive modulation of proteins related to metabolism of methionine (e.g. methionine synthase), a precursor of nicotianamine. An up-regulation of a transcript encoding a methylthioribose kinase (MTK, #55), another enzyme of the methionine cycle, was observed in our transcriptional analysis. A positive modulation of MTK transcripts was also recorded in roots of rice, a Strategy II plant species, under Fe deficiency [46, 47], although in this case, related to the necessity to increase the production of mugineic acid.

A down-regulation of a sulfate adenylyltransferase (ATPS; ATP sulfurylase) gene (#54) was observed in this work. However, the opposite was found in Arabidopsis[11]. The connection between the sulphur nutritional status and capability to respond to Fe shortage has been demonstrated in both Strategy I and Strategy II plants [30, 48]. Recently an increase in methionine content related to phytosiderophore synthesis without significant changes in ATPS activity was described in Fe-deficient barley roots [48]. Interestingly, in a recent study of Medicago truncatula root proteome [20] two enzymes related to biosynthesis of cysteine and methionine were negatively affected by Fe deficiency.

Protein turnover

Response to Fe starvation induced the accumulation of gene transcripts related to protein turn-over, including protease (#53) and peptidase (#50 and 51) involved in protein degradation, and a gene encoding a heterogeneous nuclear ribonucleoprotein (#28) that can act in the pre-mRNA metabolism preceding protein synthesis.

The activity of these genes can be related to molecular events controlling plant responses to abiotic stresses such as nutritional deficiencies [49]. A general increase in protein synthesis was reported as a response to Fe-deficiency in cucumber roots [50]; furthermore, protein recycling in response to Fe starvation was suggested by analysis of expression profiles of soybean [13] and proteomic changes in cucumber [18] and Medicago troncatula[20] roots. Transcriptomic data presented here are in line with the idea that, under Fe deficiency, N recycling reactions take place, possibly related to the necessity of additional anaplerotic source of C and N [18, 20].

Secondary metabolism

Phenolic compounds are reported as components of root exudates in Fe-deficient Strategy I plants. These molecules are involved in chelation and/or reduction of rhizospheric insoluble Fe [51, 52]. Recently, phenolics have been proposed to selectively influence rhizospheric microorganisms and be involved in the reutilization of apoplastic Fe [53]. Our results showed an up-regulation of a flavonoid-3-hydroxylase (#81) gene and of three genes putatively involved into flavonoid glycosylation (#82, 83 and 84). Two out of the last three genes showed sequence homology to a Catharanthus roseus flavonoid glucoside 1,6-glucosyltransferase catalysing 1,6-glucosylation of flavonol and flavone glucosides [54]. A positive modulation of genes related to general phenylpropanoid pathway (e.g. genes encoding PAL and 4CL) was reported in Fe-deficient Arabidopsis roots [11]. Data of the present work underline the up-regulation of transcripts involved in a more specific branch of phenolic pathway supporting the idea that Fe-starved roots might operate flavonoid secretion into the rhizosphere in order to promote Fe acquisition [55].

Oxidative stress responses

Many proteins involved in antioxidative defence response contain Fe in heme group or coordinated to the thiol group of cysteine. The modulation of transcripts and proteins related to oxidative stress response in roots of Fe-starved plants seems to depend on the species and the experimental conditions [5, 13, 17, 19]. However, catalase (CAT) and peroxidase (POX) activities are known to be depressed under Fe deficiency conditions [56] in tomato leaves. Our data showed a main down-regulation of transcripts encoding thioredoxin (TRX) (#29) and detoxifying enzymes catalase (CAT; #62), superoxide dismutase (#69 and 71), peroxidase (POX; #73, 78, 85 and 87). Two other POX transcripts (#66 and 77) and a thioredoxin peroxidase gene (#27) were, conversely, up-regulated. At protein level, a decrease of a CAT was reported while some POXs showed higher levels in response to Fe deficiency in tomato roots [15]; a different response of peroxidase isoforms has also been reported in sunflower [57]. Taken together, data of the present work suggest a different role in response to nutrient stress condition between the thioredoxin (ABB) and POX isogenes and are in agreement with previous results obtained in Medicago truncatula[20] and sugar beat [17] Fe-deficient roots.

A germin-like transcript (#15) was up-regulated under Fe deficiency, similarly to what has been observed in the tomato root proteome analysis [15]. The positive modulation of a germin protein reported in this proteomic study was justified hypothesizing its role in producing hydrogen peroxide for apoplastic Fe reduction or in other stress response on the basis of sequence similarity to a Nicotiana attenuata germin protein [15, 58].

Root morphological adaptation

Morphological modifications in roots of Fe-deficient plants are well documented [59, 60]. Root hairs proliferation and development of transfer cells were described in Fe-starved tomato plants [29, 61–63]. Enhanced formation of lateral roots and root hairs was also recorded in our experiment (Figure 1). Extensin proteins seem to be involved in this latter phenomenon; in fact, we observed a strong positive modulation of LeExt1 (#32, 33 and 35), LeDif10 (#36) and LeDif54 (#31 and 34). It was reported that these three genes encode extensin, a structural protein putatively conferring physical characteristics of the cell wall [64], and act during root hair formation in tomato due their predominant expression in root hair cells [64, 65].

The overexpression of an endo-1,4-β-glucanase (#80) and a xyloglucan endotransglucosylase/hydrolase protein (#61) transcripts could be involved into the cell wall loosening [66] associated to root morphological adaptation to Fe deficiency, as previously observed in a tomato root proteomic analysis [16]. These changes in root morphology are also supported by the up-regulation of a transcript showing homology to a tobacco actin-depolymerizing factor (#37) related to the pollen tube elongation [67]. Here, we report a negative modulation of a transcript (#45) showing homology to the tobacco aquaporin PIP2 [68] suggesting a role of the PIP2 tomato protein in Fe-related morphological root changes. It has been suggested that PIP aquaporins can play a role not only in root water uptake but also in root development [68]. Transgenic plants exhibiting RNAi of PIP2 aquaporins showed a significant increase in the length of primary roots [68].

Transport processes

Among the positively modulated transcripts belonging to "transport" functional category, a stronger modulation (more than 17 times) of an oligopeptide transporter (OTP) gene (#44) was observed under Fe shortage. BlastP analysis against TAIR database [36] using protein sequence obtained from the predicted coding sequence of the TC208376 (#44) showed the highest sequence homology with the protein encoded by AtOPT3 (score: 541; Evalue: E-154; identity: 83%; positives: 90%). Positive modulation of AtOPT2 and AtOPT3 has been recorded in Fe-deficient Arabidopsis roots [8, 11, 21, 69]. The plant members of OTP family have been described to have different functions in transport physiology such as long-distance sulphur distribution, nitrogen mobilization, metal homeostasis, and heavy metal sequestration. OPTs can transport glutathione, metal-chelates and peptides [70]. It was hypothesized that some plants OTPs are able to transport Fe-chelates and Fe-NA suggesting a role of these proteins in long-distance metal transport in planta[71, 72]. A similar function can be hypothesized for the tomato up-regulated OPT transcript, due to the previously described positive modulation of genes related to NA synthesis (see methionine cycle paragraph).

Another up-regulated gene involved in transport phenomena in response to Fe-deficiency is LeAMT1 (#41). This tomato gene, firstly isolated from a root hair cDNA library [73] was root-specifically expressed and positively regulated by ammonium (NH₄⁺) in root hairs [73, 74]. This result suggests the presence of a linkage between NH₄⁺ uptake and Fe shortage. It has been demonstrated that NH₄⁺-dependent rhizosphere acidification can improve Fe availability in the rhizosphere [75]. Interestingly, nitrate acquisition is limited under Fe deficiency [20, 76].

Furthermore, favouring ammonium uptake with respect to nitrate could reduce competition for reducing equivalent between nitrogen and Fe acquisition. This is also in agreement with the hypothesized N-recycling in Fe-deficient plant roots [20].

Fe-deficient tomato roots strongly overexpressed two transcripts (#40 and 47) encoding a protein sharing features of major facilitator superfamily (MSF) with putative transport activities. However, on the basis of their sequence homology it was not possible to hypothesize an involvement in transport of a specific metabolite or mineral nutrient.

A positive modulation of a gene encoding a metal tolerance protein (MTP; #42) was also recorded. A similar behaviour was described in Fe-deficient Arabidopsis roots [8, 11] and interpreted in the light of low specificity of IRT1 transporters, which can transport different metals into the Fe-deficient plants. MTP genes might therefore play a role in the detoxification of zinc ions taken up absorbed under Fe deficiency conditions [11].

Fe deficiency in tomato induced the expression of a transcript (#43) encoding a protein showing sequence homology to Arabidopsis vacuolar aluminium-activated malate transporter (ALMT) 9 [77]. An increase in organic acids concentration mainly citric and malic acids in plant roots under Fe starvation has been reported for many plant species [78]. However, a decrease of malate levels was observed in tomato root tips after 15-d of Fe deficiency while higher contents were recorded in leaves and xylem sap [41]. The tomato malate transporter gene might be involved in malate fluxes at intracellular level and/or in long-distance transport in planta.

Hormone metabolism and signaling

The role of plant hormones in the regulation of Fe deficiency responses has been extensively studied [79, 80]. As for Arabidopsis Col-0 accession, we identified the positive modulation of a methionine cycle gene, MTK gene, that could be related not only to NA synthesis (see above) but also to the recycling of methylthioadenosine during ethylene production [11]. This is in agreement with the hypothesized involvement of ethylene and/or auxin in control of hair root production under Fe deficiency [79]. Furthermore, data of the present work suggest the involvement of other hormones (such as gibberellin and cytokinin) in Fe deficiency response in tomato roots. Concerning gibberellin, we recorded the up-regulation of two gibberellin oxidase 20 (GA20OX) transcripts (#72 and 74). They can regulate root morphological changes through the synthesis of active GA1. Indeed, it has been reported that GAs are related to tomato root growth [81]. In addition, it was shown that the expression of a tomato SlGA20ox1:GUS construct in Arabidopsis localized also in columella of secondary roots [82], thus suggesting a role in secondary root formation. The same authors also reported that the expression of the construct was positively affected in cotyledons, hypocotyls and roots by benziladenine, a synthetic cytokinin [82]. However cytokinins were described as negative regulators of root Fe uptake mechanism in Arabidopsis through a root-growth dependent pathway [83]. Cytokinins could play a similar role also in tomato roots. This hypothesis is supported by the up-regulation of a cytokinin oxidase/dehydrogenase (CKX) gene (#65) under Fe starvation. CKX genes are involved in regulation of plant growth and development through the control of the cytokinin concentration [84].

Focusing on signal transduction, we observed the up-regulation of a transcript (#60) encoding a protein with sequence homology to the MCP5, a protein phosphatase-2C (PP2C) of Mesembryanthemum crystallinum[85]. McMCP5 is expressed in roots and is induced in response to salt and drought stresses [85]. The tomato PP2C gene could play a role in signal transduction of stress nutrient condition as Fe starvation. The positive modulation of a gene encoding a SRC2 protein containing a C2 domain (#1) suggests a putative involvement of Ca as secondary messenger. Focusing on Ca, in our experiment we also identified a down-regulated gene (#57) in Fe-starved roots showing homology to an Arabidopsis gene (AtSOS2) encoding a CBL-interacting serine/threonine-protein kinase 11. Arabidopsis SOS2 interacts with the Ca binding protein SOS3 (SALT OVERLY SENSITIVE 3), thus controlling K and Na homeostasis and the response to salt stress [86, 87]. In addition a negative modulation of a gene (#67) showing homology to a tobacco Avr/Cf-9 rapidly elicited (ACRE) transcript encoding a PP2C was observed [88]. Our data confirm the results obtained with proteome analysis of tomato roots in response to Fe-deficiency, where changes in the levels of proteins involved in signal transduction were reported [16]. The observed transcriptional changes in tomato roots can be the result of the perception of nutrient stress condition and the following signal transduction.

Taken together, these results suggest that Fe deficiency responses are, at least in part, dependent on hormonal balance modifications possibly resulting from signal perception and transduction.

Conclusion

Ninety-seven differentially expressed transcripts were identified comparing root transcriptional profiles of Fe-deficient and Fe-sufficient tomato plants. Tomato roots respond to Fe deficiency by modulating the expression of a number of transcripts similar to the model plant Arabidopsis. The comparison of tomato Fe-responsive transcript set with the Arabidopsis "ferrome" [21], encompassing 92 transcripts that robustly represent the response to Fe shortage, confirms the involvement of the well know homologous key regulatory elements (e.g. bHLHs) controlling the expression of transcripts related to Fe uptake and translocation (e.g. IRT and FRO). As showed by Arabidopsis "ferrome" [21], tomato roots modulate transcripts involved in homeostasis of Fe and heavy metal cations (e.g. IRT, NRAMP, MTP, ferritin) and others cation (e.g. AMT). Both plant species require the up-regulation of transcripts related to glycolysis (e.g. PFK) and methionine cycle (e.g. MTK), the latter pathway being putatively linked to NA biosynthesis in response to Fe deficiency. Fe-NA complexes could be transported both in tomato and Arabidopsis plants through OPTs during the response to Fe shortage. Here we describe, for the first time, the modulation of a specific branch of phenolic (flavonoids) biosynthesis in response to Fe deficiency. In addition, tomato roots seem to be more characterized by root morphological adaptation, mainly linked to hair root production, as suggested by the strong up-regulation of extensin transcripts.

Therefore, this transcriptional study, while confirming evidence coming from proteomic studies, allowed identifying new putative targets for further functional investigations on the response to Fe deficiency in tomato roots.

Methods

Plant material, growth conditions and RNA extractions

Tomato seedling (Solanum lycopersicum L., cv. 'Marmande superprecoce' from DOTTO Spa, Italy), germinated for 6 days on filter paper moistened with 1 mM CaSO₄, were grown for 14 days in a continuously aerated nutrient solution (pH adjusted at 6.0 with 1 M KOH) as reported by Tomasi at al. [22] with 5 μM Fe (Fe-EDTA); thereafter, most of the plants were transferred for a further week to a Fe-free nutrient solution (Fe-deficient) and some tomato plants were transferred for a week to a nutrient solution containing 100 μM Fe-EDTA (Fe-sufficient plants) as control. Nutrient solutions were renewed every three days. The controlled climatic conditions were the following: day/night photoperiod, 16/8 h; light intensity, 220 μE m^-2s^-1; temperature (day/night) 25/20°C; RH 70 to 80%.

At the end of the growing period (27 days), Fe-deficient tomato plants clearly showed visible symptoms of Fe deficiency yellowing of the fully expanded apical leaves and proliferation of lateral roots and root hairs and increase in the diameter of the sub-apical root zone. 24 hours before harvesting, all nutrient solutions (both for Fe-deficient and Fe-sufficient plants) were renewed and the pH was adjusted to 7.5 with 10 mM Hepes-KOH. The pH of the growing medium was adjusted to this value to mimic as close as possible the conditions that are occurring in Fe-deficiency-inducing soil. However, in order to favour an equilibrate development of tomato plants growing in the Fe-free nutrient solution, the exposure to the pH of 7.5 was limited to the last two days.

Roots of Fe-deficient and Fe-sufficient tomato plants (27 d-old) were harvested five hours after the beginning of light phase. The collected roots were immediately frozen in liquid nitrogen and stored until further processing at -80°C. The collection was repeated in three independent cultivations and the roots from six plants were pooled for each treatment.

Ferric-chelate reduction

To determine the root capacity to reduce Fe(III)-EDTA, accordingly to Pinton et al. [89] roots of a single intact (Fe-sufficient or Fe-deficient) tomato plants were incubated in the dark at 25°C for 60 min in 50 mL of an aerated solution containing CaSO₄ 0.5 mM, BPDS 0.5 mM, Hepes-KOH 10 mM (pH 5.5) and 0.25 mM of Fe(III)-EDTA. Thereafter, the absorbance of the solutions at 535 nm was measured at intervals of 15 min and the amount of Fe(III) reduced calculated by the concentration of the Fe(II)-BPDS₃ complex formed, using an extinction coefficient of 22.1 mM^-1 cm^-1.

Microarray analysis

Transcriptional analysis was carried out using a Combimatrix [24], produced by the Plant Functional Genomics Center, University of Verona [90]. The chip (TomatoArray2.0) carries 25,789 nonredundant probes (23,282 unique probes and 2,507 probes with more than one target) randomly distributed in triplicate across the array, each comprising a 35-40-mer oligonucleotide designed using the program oligoarray 2.1 [91]. The source of sequence information included tentative consensus sequences (TCs) derived from the DFCI Tomato Gene Index [92] Release 12.0 and expressed sequence tags. Eight bacterial oligonucleotide sequences provided by CombiMatrix, 8 probes designed on 8 Ambion spikes and 40 probes based on Bacillus anthracis, Haemophilus ducreyi and Alteromonas phage sequences were used as negative controls. Complete description of chip is available at the Gene Expression Omnibus [93] under the series entry (GPL13934).

Total RNA was isolated using the Spectrum™ Plant Total RNA kit (Sigma-Aldrich) and quantified by spectrophotometry using NanoDrop™ 1000 Tem Scientific). RNA quality was evaluated using Agilent 2100 Bioanalyzer (Agilent). Total RNA (1 μg) was amplified and labelled using the RNA ampULSe kit (Kreatech). After checking the quantity and quality of aRNA by spectrophotometry using NanoDrop™ 1000 (Thermo Scientific) and the quality subsequent labelling, 4 μg of labelled aRNA was hybridized to the array according to the manufacturer's recommendations [24]. Pre-hybridization, hybridization, washing and imaging were performed according to the manufacture's protocols. The array was scanned with an Axon GenePix^® 4400A scanner (MDS Analytical Technologies).

Analysis of raw data was performed using the open source software of the Bioconductor project [94, 95] with the statistical R programming language [96, 97]. Background adjustment, summarization and quantile normalization were performed using limma package [25]. Differentially expressed probes were identified by linear models analysis [25] using limma package and applying Bayesian correction, adjusted p-value of 0.05 and a |FC| ≥ 2. All microarray expression data are available at the Gene Expression Omnibus [93] under the series entry (GSE31112). Genes were grouped in main functional categories according to the "biological" terms of the Gene Ontology [27] assigned to each tomato TC or EST (Release 12.0) on the basis of the results of BlastP analysis [98] against the UniProt database [26] (Additional file 1). Genes without significant BlastP results were classified as "no hits found" (Evalue < 1e-8; identity > 40%).

Real-time RT-PCR experiments

0.5 μg of total RNA (checked for quality and quantity using a spectrophotometer NanoDrop™ 1000 (Thermo Scientific), followed by a migration in an agarose gel) of each sample was retrotranscribed using 1 pmol of Oligo d(T)23VN (New England Biolabs, Beverly, USA) and 10 U M-MulV RNase H for 1 h at 42°C (Finnzymes, Helsinki, Finland) following the application protocol of the manufacturers. After RNA digestion with 1 U RNase A (USB, Cleveland, USA) for 1 h at 37°C, gene expression analyses were performed by adding 0.16 μl of the cDNA to the realtime PCR complete mix, FluoCycleTM sybr green (20 μl final volume; Euroclone, Pero, Italy), in a DNA Engine Opticon Real-Time PCR Detection (Biorad, Hercules, USA). Specific primers (Tm = 58°C) were designed to generate 80-140 bp PCR products (Additional file 2). Three genes were used as housekeeping to normalized the data: elongation factor 1-alpha EF1a (X14449; TC203463; forward: 5'- TGGATATGCTCCAGTGCTTG-3'; reverse: 5'-TTCCTTACCTGAACGCCTGT-3'), histone H1 (AJ224933; TC192148; forward: 5'- CAAAGGCCAAAACTGCTACC-3'; reverse: 5'-AGGCTTTACAGCTGCTTTCG-3') and ubiquitin Ubi3 (X58253; TC196208; forward: 5'-AGCCAAAGAAGATCAAGCACA-3'; reverse: 5'-GCCTCTGAACCTTTCCAGTG-3'). Each Real-Time RT-PCR was performed 4 times on 3 independent experiments; analyses of real-time result were performed using Opticon Monitor 2 software (Biorad, Hercules, USA) and R [93] with the qpcR package [99]. Efficiencies of amplification were calculated following the authors' indications [100]: PCR efficiencies were 99.15%, 89.16% and 87.25%, for EF1a, H1 and Ubi3 genes, respectively. The efficiencies for TC191581, TC192292, TC194139, TC216882, TC205660, TC208376 and NP796451 were respectively 94.75, 85.03, 98.92, 91.55, 92.70, 96.61 and 92.43%. The reported Real time data were normalized on the EF1a gene. Gene expression data were illustrated considering the differences in the amplification efficiency of PCR and using the gene expression levels in roots of Fe-sufficient plants as reference; applying the following formula:

$$gene\mathsf{\text{exp}}ression{x}_{-Fe}=\frac{{\left(2\times E_x\right)}^{\left[c_t\left(x+Fe\right)-c_t\left(x-Fe\right)\right]}}{{\left(2\times E_y\right)}^{\left[c_t\left(y+Fe\right)-c_t\left(y-Fe\right)\right]}}$$

Where: E_{x or y} is the percentage value of PCR efficiency for the amplification of the gene x or y, respectively; C_t(x_+Fe) Ct for the control treatment (+Fe) and the considered gene (x); C _t (x_-Fe); Ct for the treated roots (-Fe) and the considered gene (x); C _t (y_+Fe) Ct for the control treatment (+Fe) and the housekeeping gene (y); C _t (y_-Fe) Ct for the treated roots (-Fe) and the housekeeping gene (y).