Introduction
Whitefly Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) is a major pest of both greenhouse and open-field horticultural crops worldwide. Severe damages by B. tabaci are caused directly through phloem feeding and indirectly by the transmission of a number of different plant viruses to a wide range of plants in tropical, subtropical, and Mediterranean climate conditions. Among horticultural crops affected by B. tabaci, tomato (Solanum lycopersicum) is the most common host and the second most important vegetable crop next to potato (FAOSTAT, 2001). Besides its agricultural interest, tomato has several advantages as a model plant, such as small genome (950 Mb), a short generation time, availability of transformation protocols and genetic and genomic resources (Pascual et al., Reference Pascual, Blanca, Cañizares and Nuez2009), leading to the complete sequencing of tomato genome (Mueller et al., Reference Mueller, Solow, Taylor, Skwarecki, Buels, Binns, Lin, Wright, Ahrens, Wang, Herbst, Keyder, Menda, Zamir and Tanksley2005; Tomato Genome Consortium, 2012). Several tomato varieties are resistant to both B and Q biotypes of B. tabaci (Nombela et al., Reference Nombela, Beitia and Muñiz2000, Reference Nombela, Beitia and Muñiz2001) currently renamed as Middle East-Asia Minor 1 and Mediterranean species, respectively (De Barro et al., Reference De Barro, Liu, Boykin and Dinsdale2011). The resistant response to B. tabaci is mediated by the major resistance gene (R gene) Mi-1 (Nombela et al., Reference Nombela, Williamson and Muñiz2003), introduced into a cultivated tomato from its wild relative, S. peruvianum (Smith, Reference Smith1944). Mi-1 also confers resistance against other phloem feeders, such as three species of root-knot nematodes (RKN) Meloidogyne spp. (Roberts and Thomason, Reference Roberts and Thomason1986), the potato aphid Macrosiphum euphorbiae (Rossi et al., Reference Rossi, Goggin, Milligan, Kaloshian, Ullman and Williamson1998) and the tomato psyllid Bactericerca cockerelli (Casteel et al., Reference Casteel, Walling and Paine2006). Mi-1 was localized in a 52-Kb region of the short arm of chromosome 6 of tomato and subsequently cloned (Kaloshian et al., Reference Kaloshian, Yaghoobi, Liharska, Hontelez, Hanson, Hogan, Jesse, Wijbrandi, Simons, Vos, Zabel and Williamson1998; Milligan et al., Reference Milligan, Bodeau, Yaghoobi, Kaloshian, Zabel and Williamson1998). This gene codifies for a CC-NB-LRR protein with 1257 amino acids, similar to other R proteins (Milligan et al., Reference Milligan, Bodeau, Yaghoobi, Kaloshian, Zabel and Williamson1998; Williamson, Reference Williamson1998; Martin et al., Reference Martin, Bogdanove and Sessa2003) and it was the first cloned R gene conferring plant resistance to an insect pest. The Mi-1 gene is constitutively expressed very early in development in every tissue of resistant tomato (Martinez de Ilarduya and Kaloshian, Reference Martínez de Ilarduya and Kaloshian2001), but the Mi-1 protein is stored in an inactive conformation in the absence of an attacker organism (Hwang and Williamson, Reference Hwang and Williamson2003). Upon detection of effector molecules from a nematode or an insect, Mi-1 protein experiences a conformational change and activates different signals leading to the resistance response (Williamson and Roberts, Reference Williamson, Roberts, Perry, Moens and Starr2009). It is well known that, apart from an R gene, the presence and function of additional genes in certain signal transduction pathways leading to defense against the attacker organism are necessary for an effective pest resistance (Williamson and Roberts, Reference Williamson, Roberts, Perry, Moens and Starr2009). In addition to Mi-1, other genes have been identified in tomato that are required for the Mi-1-mediated resistance, such as Rme1 against aphids, nematodes, and whiteflies (Martínez de Ilarduya et al., Reference Martínez de Ilarduya, Moore and Kaloshian2001, Reference Martínez de Ilarduya, Xie and Kaloshian2003, Reference Martínez de Ilarduya, Nombela, Hwang, Williamson, Muñiz and Kaloshian2004) and Hsp90 and Sgt1 against nematodes and aphids (Bhattarai et al., Reference Bhattarai, Li, Liu, Dinesh-Kumar and Kaloshian2007). A model of interaction has been proposed between the proteins encoded by these genes with Mi-1 forming an R-signaling complex with HSP90 and SGT1, and this complex guards RME1 (Bhattarai et al., Reference Bhattarai, Li, Liu, Dinesh-Kumar and Kaloshian2007). Further research on molecular aspects of plant resistance is essential to identify new components of Mi-1-mediated resistance, particularly on the mechanisms regulating those processes related to resistance to insects and the genes that control and modulate the resistant response.
Global analysis of gene expression has been widely done by means of high-performance technologies such as microarrays, allowing the detection of changes in the expression of thousands of genes simultaneously (Berrar et al., Reference Berrar, Downes and Dubitzky2003). The development of this technology for the analysis of expression profiles, along with the availability of databases of genomic sequence and expressed sequence tag from many plants, has allowed the study of transcriptional reprogramming in many different physiological situations (Aharoni and Vorst, Reference Aharoni and Vorst2001; Rensink and Buell, Reference Rensink and Buell2005). This included changes in response to the infection with bacterial pathogens (Tao et al., Reference Tao, Xie, Chen, Glazebrook, Chang, Han, Zhu, Zou and Katagiri2003; Balaji et al., Reference Balaji, Mayrose, Sherf, Jacob-Hirsch, Eichenlaub, Iraki, Manulis-Sasson, Rechavi, Barash and Sessa2008), phytopathogenic nematodes (Puthoff et al., Reference Puthoff, Nettleton, Rodermel and Baum2003; Alkharouf et al., Reference Alkharouf, Klink, Chouikha, Beard, MacDonald, Meyer, Knap, Khan and Matthews2006; Barcala et al., Reference Barcala, García, Cabrera, Casson, Lindsey, Favery, García-Casado, Solano and Escobar2010; Uehara et al., Reference Uehara, Sugiyama, Matsuura, Arie and Masuta2010; Portillo et al., Reference Portillo, Cabrera, Lindsey, Topping, Andrés, Emiliozzi, Oliveros, García-Casado, Solano, Koltai, Resnick, Fenoll and Escobar2013), or insect feeding (Korth, Reference Korth2003; Thompson and Goggin, Reference Thompson and Goggin2006). A number of previous studies have used microarray analysis to identify changes in the plant transcriptomic profiles in response to RKN feeding during compatible and/or incompatible interactions with Arabidopsis (Hammes et al., Reference Hammes, Schachtman, Berg, Nielsen, Koch, McIntyre and Taylor2005; Jammes et al., Reference Jammes, Lecomte, Almeida-Engler, Bitton, Martin-Magniette, Renou, Abad and Favery2005; Barcala et al., Reference Barcala, García, Cabrera, Casson, Lindsey, Favery, García-Casado, Solano and Escobar2010; Portillo et al., Reference Portillo, Cabrera, Lindsey, Topping, Andrés, Emiliozzi, Oliveros, García-Casado, Solano, Koltai, Resnick, Fenoll and Escobar2013), soybean (Ibrahim et al., Reference Ibrahim, Hosseini, Alkharouf, Hussein, Gamal El-Din, Aly and Matthews2011), or tomato (Bar-Or et al., Reference Bar-Or, Kapulnik and Koltai2005; Bhattarai et al., Reference Bhattarai, Xie, Mantelin, Bishnoi, Girke, Navarre and Kaloshian2008). More specifically, resistance to RKN mediated by the Mi-1 gene was studied in tomato roots by means of cDNA microarrays (Schaff et al., Reference Schaff, Nielsen, Smith, Scholl and Bird2007; Bhattarai et al., Reference Bhattarai, Xie, Mantelin, Bishnoi, Girke, Navarre and Kaloshian2008, Reference Bhattarai, Atamian, Kaloshian and Eulgem2010).
Signaling pathways involved in plant–aphid susceptible interactions have been more frequently studied by comparative transcriptome analysis (de Vos et al., Reference De Vos, Jae and Jander2007; Kuśnierczyk et al., Reference Kuśnierczyk, Winge, Midelfart, Armbruster, Rossiter and Bones2007; Li et al., Reference Li, Zou, Li, Bilgin, Vodkin, Hartman and Clough2008). These studies suggest, broadly speaking, that aphid feeding causes activation of responses different to those caused by chewing herbivores, with changes in the expression of enzymes involved in the synthesis of secondary metabolites, as demonstrated in rice (Zhang et al., Reference Zhang, Zhu and He2004; Cho et al., Reference Cho, Jung, Jeung, Kang, Shim, You, Yoo, Ok and Shin2005). Additionally, responses induced by aphids in Arabidopsis, Nicotiana attenuata, certain gramineae, and tomato were different to changes produced by chewing insects, but similar to those triggered by bacterial and fungal pathogens (Kaloshian and Walling, Reference Kaloshian and Walling2005; Thompson and Goggin, Reference Thompson and Goggin2006). The fact that whiteflies have the same type of piercing–sucking mouthparts like that of aphids initially led to the assumption that changes provoked by aphids should be the same or very similar to those following whitefly feeding. However, a study of Affymetrix microarrays during whitefly feeding on Arabidopsis showed qualitative and quantitative differences with respect to the results obtained with aphids, not only chewing herbivores (Kempema et al., Reference Kempema, Cui, Holzer and Walling2007).
Despite all the aforementioned background, studies have been scarce using microarrays to analyze in the leaf tissues the mechanisms that regulate processes related to plant resistance to insect pests. Previous research on wheat resistance to aphids demonstrated a general activation of the oxidative stress pathway, similar to the resistant responses mediated by pathogen-induced R genes (Boyko et al., Reference Boyko, Smith, Thara, Bruno, Deng, Starkey and Klaahsen2006). Another relevant study used microarrays to compare susceptible and partially resistant lines of barley in response to aphids (Delp et al., Reference Delp, Gradin, Åhman and Jonsson2009). A similar methodology was used to analyze changes of expression in tomato induced by whitefly feeding throughout insect development, but only on susceptible plants (Estrada-Hernández et al., Reference Estrada-Hernández, Valenzuela-Soto, Ibarra-Laclette and Délano-Frier2009). However, insufficient use had been made so far of microarray technology to study Mi-1-mediated resistance to whiteflies in tomato, or to identify new components of this resistance. So, more than 200 genes differentially expressed in different plant organs were obtained by cDNA arrays in cherry tomato at 25 days of infestation with B. tabaci but, again, only on susceptible plants (McKenzie et al., Reference McKenzie, Bausher, Albano, Shatters, Sinisterra and Powell2005).
In the present work, the GeneChip™ Tomato Genome Array (Affymetrix®), with over 9200 transcripts, was used for the first time in an unbiased study to detect basal differences in the global gene expression of tomato associated with the presence/absence of the R gene Mi-1. With this goal, uninfested leaf tissues of adult tomato plants of a susceptible cultivar (Moneymaker) and a Mi-1-bearing (resistant) cultivar (Motelle) were analyzed and their transcriptional profiles were compared. In a later phase of this study, plants of the same resistant and susceptible cultivars were again compared by microarrays 2 days after being infested with B. tabaci adults, to investigate how whitefly infestation modifies the basal differences previously detected in the comparison of the uninfested Motelle and Moneymaker.
Materials and methods
Insects, plant material, and growth conditions
Adult females of the Mediterranean B. tabaci were used for plant infestation. A population of these whiteflies, originally collected from cropped tomato, was reared for several generations in our laboratory, free from any plant pathogen, on the susceptible tomato cv. Marmande.
Six uninfested plants of each tomato cultivar Motelle (Mi-1/Mi-1) and Moneymaker (mi-1/mi-1) were compared by microarrays. These cultivars are near-isogenic lines (Laterrot, Reference Laterrot1987) differing only in the presence of a 650-kb introgressed region from Lycopersicon peruvianum (currently Solanum peruvianum) containing the Mi-1 gene, in chromosome 6 of Motelle (Ho et al., Reference Ho, Weide, Ma, van Wordragen, Lambert, Koornneef, Zabel and Williamson1992).
Tomato seeds were germinated and the plants were raised inside a growth chamber at a constant temperature of 25°C, L16:D8 h photoperiod and 70% r.h. Plants were grown in 1-liter plastic pots filled with autoclaved vermiculite (number 3, Projar, Spain), irrigated every 15 days with a nutritive complex 20-20-20 (Nutrichem 60; Miller Chemical, Hanover, PA, USA) at a concentration of 3 g l−1, and with water when needed in the meantime.
All plants were 8-week-old, with 8–9 true leaves each, at the time of analysis.
Whitefly infestations
Simultaneously to the analysis of the uninfested plants, six Motelle and six Moneymaker whitefly-infested plants of the same age were compared. For plant infestation, 50 ml Falcon tubes were modified from the clip-cage system for whiteflies (Muñiz and Nombela, Reference Muñiz and Nombela2001). Each tube was cut transversally to remove the conical bottom and a very thin polypropylene tissue (anti-thrips mesh) was attached by paraffin wax to the end of the tube. In addition, a lateral hole was drilled in the tube to introduce the insects later. The selected leaflet was inserted through the other end of the tube. Three modified Falcon tubes were used per plant, and each tube was placed in a well-developed leaflet of a leaf located in the middle-high zone of the plant (fig. 1).
Tomato plant with three modified Falcon tubes used for whitefly infestation.
Thirty adult females of B. tabaci were selected from the whitefly breading population and deposited into each tube through the lateral hole which was closed by a sponge plug. To maintain the same conditions, empty tubes were placed in the non-infested plants. After 2 days, tubes and whiteflies were carefully removed from all plants.
Sample collection
The samples were collected immediately after removing the whiteflies. From each tomato cultivar (Motelle or Moneymaker) and treatment (infested or non-infested), three biological replicates were collected, each consisting of six leaflets, one from each plant. The collected samples were immediately frozen in liquid nitrogen and stored at −80°C until RNA extraction.
Microarray hybridization and analysis
Gene expression of tomato leaves was performed using the Affymetrix GeneChip™ Tomato Genome Array, which contains over 10,000 probe sets to interrogate over 9200 tomato transcripts (http://www.affymetrix.com/products_services/arrays/specific/tomato.affx). Total RNA was isolated from leaves of plants from three independent biological replicates using Trizol Reagent (Thermo Fisher Scientific, Waltham, MA, USA) and further purified with RNeasy mini kit (‘clean-up’ protocol, Qiagen, Hilden, Germany), following the manufacturers' recommendations, and assessed in a Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA). cDNA was synthesized from 4 µg of total RNA using one-cycle target labeling and control reagents (Affymetrix®, Santa Clara, CA, USA), to produce biotin-labeled cRNA. The cRNA preparation (15 µg) was fragmented at 94°C for 35 min into segments 35–200 bases in length. Labeled cRNAs were hybridized to Affymetrix® arrays in a hybridization solution containing 100 mM 2-(N-morpholino) ethanesulfonic acid, 1 M Na+, and 20 mM EDTA in the presence of 0.01% Tween 20 to a final cRNA concentration of 0.05 µg ml−1 for 16 h at 45°C. Each microarray was washed and stained with streptavidin-phycoerythrin in a Fluidics station 450 (Affymetrix®) and scanned at 1.56 µm resolution in a GeneChip Scanner 3000 7 G system (Affymetrix®).
Bioinformatic and statistical data analyses
The GeneChip intensities were background-corrected, normalized, and summarized by the robust multiarray average (RMA) method (Irizarry et al., Reference Irizarry, Bolstad, Collin, Cope, Hobbs and Speed2003) using the affy package from Bioconductor (https://www.bioconductor.org/). Differentially expressed transcripts were determined using the moderated t test as implemented in the limma package from Bioconductor (Smyth, Reference Smyth, Gentleman, Carey, Huber, Irizarry and Dudoit2005). Raw P values were adjusted for multiple hypotheses testing using the false discovery rate (FDR) method (Benjamini and Hochberg, Reference Benjamini and Hochberg1995). Genes with a fold-change in expression ≥ 2 or ≤−2 and FDR < 0.05 were considered as differentially expressed.
The VENNY program version 2.1 (Oliveros, Reference Oliveros2007) was used to compare the lists of previously selected genes and to identify the genes shared in the different gene lists.
Descriptions of the genes and target sequences corresponding to GeneChip probesets were obtained from Affymetrix, Tomato Annotations Release 36 (NetAffx Analysis Center). Target sequences were also used in BLAST searches of their corresponding tomato genes (version SL3.0 and Annotation ITAG3.10) in Sol Genomics database (Fernandez-Pozo et al., Reference Fernandez-Pozo, Menda, Edwards, Saha, Tecle, Strickler, Bombarely, Fisher-York, Pujar, Foerster, Yan and Mueller2015).
Validation of microarray data by real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
For qRT-PCR validation, total RNA was extracted as previously detailed and 1 µg was retrotranscribed with the High Capacity Reverse transcription Kit (Thermo Fisher Scientific) using random primers, and then amplified with the primers listed (table 4) using a Hot FIREPol EvaGreen Plus-based system. The relative quantity (2 − ΔΔCt) of each mRNA was calculated after normalization to the housekeeping gene Ubi3.
To analyze the correlation between the data obtained by microarray and qRT-PCR, the Pearson correlation coefficient (r) was calculated using the GraphPad Prism program (version 4.00 for Windows, GraphPad Software, San Diego, California, USA, www.graphpad.com). Data obtained by qRT-PCR were transformed to a logarithmic scale since the microarray data were expressed in log2 scale. The r values oscillate between 1 (total positive correlation between both variables) and −1 (total negative correlation). The program also calculates the coefficient of determination (r 2), which establishes a proportion of variability shared or explained for both variables, and the P-value to establish whether the correlation between both variables is statistically significant.
Results
Basal differences between uninfested Moneymaker and Motelle
When comparing Moneymaker and Motelle cultivars in the absence of any infestation, 18 differentially expressed transcripts were obtained (fig. 2a). Of them, ten transcripts were significantly more expressed (up-regulated) in Motelle than in Moneymaker, whereas eight transcripts were less expressed (down-regulated) in Motelle relative to Moneymaker (table 1). Approximately half of these 18 differentially expressed transcripts were expressed more than fivefold in one cultivar than in the other.
Gene expression levels of the differentially regulated transcripts in each tomato cultivar and treatment. (a) Plants in the absence of infestation. (b) Plants infested by B. tabaci. Each bar (gray color for Moneymaker; black color for Motelle) corresponds to the mean signal of three replicates (Log2 Mean ± SE).
Transcripts up-regulated at least double (fold-change ≥ 2) or down-regulated at least half (fold-change ≤ −2) in leaves of the tomato cv. Motelle compared to cv. Moneymaker, in the absence of infestation and considering only significant values (FDR < 0.05)
a Transcript Identifier in the Affymetrix Genechip™.
b GenBank (NCBI) Transcript Identifier, provided by Affymetrix (Release 36, January 2017).
c Functional description of transcript, provided by Affymetrix (Release 36, January 2017).
d Tomato locus, Genome version SL3.0 and Annotation ITAG3.20
e Description of the tomato locus (Annotation ITAG3.20)
f Relative Expression in Motelle compared to Moneymaker.
g FDR value (corrected P-value) of the Relative Expression.
Among the ten transcripts more expressed in Motelle than in Moneymker, a transcript stands out (FC = 80.19), with a sequence similar to the NTGP4 gene of Nicotiana tabacum, which participates in different processes of response to biotic stimuli. The next transcript, with 28.42-fold expression greater in Motelle than in Moneymaker, corresponds to a gene with participation in translation elongation. The Les.3272.1.S1_at transcript (FC = 11.30) is similar to an Arabidopsis thaliana gene involved in the processes of amino acid biosynthesis. It is also worth noting the overexpression of the gene encoding the H + -ATPase vacuolar subunit that is involved in transport and binding to ATP (FC = 4.42) and the VDAC gene related to response to biotic stimuli (FC = 2.50). The final two up-regulated transcripts related to the glycogen glucosyltransferase (FC = 2.08) and to the F-box family protein (FC = 2.04). It is important to mention that three probesets corresponding to the homologous genes Mi-1.1 and Mi-1.2 were detected in our analysis, with expression values approximately sevenfold higher in Motelle than in Moneymaker, as expected for the presence of the Mi-1 locus in Motelle.
Approximately half of the eight transcripts down-regulated in Motelle relative to Moneymaker had highly significant FDR (P < 0.0001). The most repressed one (FC = −9.62) encodes the DELLA GAI protein, a negative regulator of gibberellin (GA) signaling. This was followed by the Pys gene (FC = −8.77) involved in secondary metabolism, the ADK gene (FC = −8.64) which catalyzes AMP synthesis from adenosine and ATP, as well as other gene encoding a selT-like protein (FC = −5.65). Also observed was down-regulation of the gene MADS-box 15 (FC = −3.15), a transcription factor involved in different plant development processes. The three last genes with lower expression in Motelle than Moneymaker codified a short-chain dehydrogenase/reductase (FC = −2.60), the CAD enzyme (FC = −2.39), key in the synthesis of lignin, and a Glutaredoxin (FC = −2.13) whose function is to protect cells against oxidative stress, thus maintaining cellular homeostasis.
Differences between Moneymaker and Motelle after B. tabaci infestation
The analysis of the transcriptomic profiles after 2 days of infestation revealed 28 transcripts with differential expression between Motelle and Moneymaker cultivars. Of them, 14 transcripts were expressed significantly more in Motelle than in Moneymaker, and 14 transcripts were expressed less in Motelle when compared to Moneymaker (fig. 2b). The expression range in up-regulated transcripts was 2–38 times greater in Motelle than in Moneymaker, while the down-regulated transcripts ranged from two to nine times lower in Motelle (table 2).
Transcripts up-regulated at least double (fold-change ≥ 2) or down-regulated at least half (fold-change ≤ −2) in whitefly infested leaves of the tomato cv. Motelle compared to infested leaves of cv. Moneymaker, considering only significant values (FDR < 0.05)
a Transcript Identifier in the Affymetrix Genechip™.
b GenBank (NCBI) Transcript Identifier, provided by Affymetrix (Release 36, January 2017).
c Functional description of transcript, provided by Affymetrix (Release 36, January 2017).
d Tomato locus, Genome version SL3.0 and Annotation ITAG3.20
e Description of the tomato locus (Annotation ITAG3.20)
f Relative Expression in Motelle compared to Moneymaker.
g FDR value (corrected P-value) of the Relative Expression.
When these results were compared with those previously obtained from uninfested plants, it was observed that whitely infestation substantially modified the basal differences among Motelle and Moneymaker cultivars (fig. 3). Out of the 18 transcripts differentially expressed in the uninfested plants, 14 were also up-regulated or down-regulated in Motelle regarding Moneymaker after whitefly infestation. Moreover, 14 additional transcripts showed differential expression between cultivars only after infestation with B. tabaci.
Venn diagrams comparing the number of transcripts with differential expression between tomato cultivars, before (no infested) or after (infested) B. tabaci infestation. Up-regulated represent transcripts more expressed in Motelle than in Moneymaker. Down-regulated represent transcripts less expressed in Motelle than in Moneymaker. Only transcripts are included with statistically significant values (FDR < 0.05) of relative expression (fold-change or FC) ≥2 (up) or ≤−2 (down).
All differential transcripts common to both analyses (infested and non-infested plants) are listed in table 3 with their corresponding relative expression values before and after whitefly infestation. For most of these transcripts, expression differences between Motelle and Moneymaker were moderately or markedly reduced after infestation. Only in four cases did the differences increase or remain similar to those in uninfested plants.
Transcripts with differential expression between tomato cultivars detected in the analysis of both uninfested and whitefly-infested plants.
a Transcript Identifier in the Affymetrix Genechip™.
b Tomato locus and description.
c Relative Expression in Motelle compared to Moneymaker. ↓↓ and ↑↑ represent marked decreases and increases, respectively, in the expression differences between cultivars. ↓ and ↑ represent moderate decreases and increases.
Seven additional transcripts were up-regulated in infested Motelle with respect to infested Moneymaker (table 2), which had not been previously highlighted in the comparison between cultivars in the absence of infestation. One of them was expressed 12.16-fold higher in Motelle than in Moneymaker, corresponding to a WD-40 repeat family protein. Among the other transcripts, Les.5012.1.S1_at (FC = 3.57) corresponds to the acid phosphatase 1 enzyme which participates in defense response processes, specifically against insects. Moreover, two transcripts correspond to the isoflavone reductase (FC = 3.39) and the methionine sulfoxide reductase (MsrA) (FC = 2.89), which both participate in processes of response to oxidative stress. The sulfate adenyltransferase enzyme (FC = 2.53) participates in sulfur assimilation processes. The transcript LesAffx.6110.1.S1_at (FC = 2.11) corresponds to a pectinesterase, involved in cell wall reorganization processes in response to pathogen attack. Finally, the enzyme NADH dehydrogenase (FC = 2.00) which is the first enzyme in Complex I of the electron transport chain in mitochondria, is correlated with programmed cell death.
Among the seven transcripts down-regulated in Motelle exclusively after whitefly infestation, we can highlight the ELI3 protein related to defense processes (FC = −5.79), a HMG type nucleosome/chromatin assembly factor (FC = −4.03), a signal transduction response regulator (FC = −2.54), and the E3 ubiquitin protein ligase (FC = −2.13). Moreover, the enzymes cytosine-5 DNA methyltransferase (FC = −2.57), aldehyde oxidase (FC = −2.08), and UDP-glucuronate decarboxylase (FC = −2.05), involved in the regulation of gene expression during development, hormone biosynthesis, and membrane-associated metabolic processes, respectively.
Validation of microarray data by qRT-PCR
The relative expression values of the 12 transcripts analyzed by qRT-PCR are shown in table 4. A positive correlation between these data with those previously obtained by microarray analysis was obtained (fig. 4), with a value of the Pearson correlation coefficient (r) of 0.7475, statistically significant (P < 0.0001), therefore validating the results obtained by microarray analysis.
Correlation between gene expression values obtained from the microarray analysis (axis X) and from qRT-PCR (axis Y), with a statistically significant (P < 0.0001) value of the Pearson correlation coefficient (r = 0.7475).
Analysis of relative expression by qRT-PCR
a Transcript Identifier in the Affymetrix Genechip™.
b Relative Expression in Motelle compared to Moneymaker according to qRT-PCR.
Discussion
Considering the literature reviewed so far, this is the first time that transcriptional profiles of non-infested foliar tissues have been compared by microarray from fully developed tomato plants belonging to different genotypes that are differentiated by the presence/absence of the Mi-1 gene. Motelle and Moneymaker are quasi-isogenic cultivars as they differ only in a 650-kb fragment of chromosome 6 in which Mi-1 is included (Messeguer et al., Reference Messeguer, Ganal, de Vicente, Young, Bolkan and Tanksley1991; Ho et al., Reference Ho, Weide, Ma, van Wordragen, Lambert, Koornneef, Zabel and Williamson1992). However, the presence of Mi-1 also appears to be associated with baseline differences in the expression of other genes not necessarily localized near it. The present study revealed the existence of 18 transcripts differentially expressed in the uninfested leaves, ten of which were expressed at least double in Motelle than in Moneymaker, while the other eight were expressed half or less. In principle, the genes represented by these 18 transcripts could be considered as good candidates to participate in the resistance to piercing–sucking insects mediated by the Mi-1 gene, although the relevance of each of them must be analyzed individually. Moreover, infestation with whitefly B. tabaci produces important changes in the transcriptome of tomato leaves, substantially modifying the initial differences between Moneymaker and Motelle.
Expression of Mi-1 gene
Three probesets detected as up-regulated in non-infested Motelle corresponded to the homologous genes Mi-1.1 and Mi-1.2. This result, not as expected less interesting, reflects the main difference between both cultivars in the absence of any type of infestation. A similar comparison between Motelle and Moneymaker had been made using the microarray technique in uninfested roots (Schaff et al., Reference Schaff, Nielsen, Smith, Scholl and Bird2007), analyzing the expression of 1547 genes, among which the Mi-1 gene was not included. In contrast, the expression of approximately 9200 genes in foliar tissue was analyzed in the present work to obtain additional information on the basal differences between both cultivars associated with the presence/absence of Mi-1. Moreover, the plants analyzed were fully mature, while roots from younger plants (4 weeks old) were used in the previous work by Schaff et al. (Reference Schaff, Nielsen, Smith, Scholl and Bird2007). This is important as Mi-1-mediated resistance of tomato to B. tabaci is dependent on plant age, and this resistance has very limited effectiveness in 5-month-old or younger plants (Rodríguez-Álvarez et al., Reference Rodríguez-Álvarez, Muñiz and Nombela2017). Individual expression of Mi-1.2 was previously analyzed by RT-PCR in these same tomato genotypes in the absence of any infestation, with expression of this gene in different plant tissues of only Motelle plants (Martínez de Ilarduya and Kaloshian, Reference Martínez de Ilarduya and Kaloshian2001).
The differential expression of Mi-1.1/Mi.1.2 was also maintained after B. tabaci infestation, with expression only slightly lower than that observed prior to infestation. This indicates that whitefly infestation does not cause substantial changes in Mi-1 expression, which agrees with the results previously obtained by other authors on the attack of nematodes or aphids (Martínez de Ilarduya and Kaloshian, Reference Martínez de Ilarduya and Kaloshian2001; Goggin et al., Reference Goggin, Shah, Williamson and Ullman2004). The detection of this fundamental difference between Motelle and Moneymaker in the present survey through global gene expression analysis can be considered as a validity test of this methodology for this purpose. Thus, this finding reinforces the use of DNA microarrays for the identification of differentially expressed genes.
Other baseline differences and changes after whitefly infestation
Leaving aside the Mi-1 gene, the transcript with the highest difference in expression between non-infested cultivars (more than 80-fold higher in Motelle than in Moneymaker) encoded the NtGP4 (N. tabacum geranylgeranylated 4) protein. This basal difference was considerably reduced to 37.59-fold after infestation with B. tabaci. Expression of NtGP4 gene can be involved in the processes of response to biotic stimuli as it was previously demonstrated to be induced in the roots of Moneymaker after nematode infection (Bhattarai et al., Reference Bhattarai, Xie, Mantelin, Bishnoi, Girke, Navarre and Kaloshian2008). Moreover, NtGP4 had a higher basal expression in leaves of a tomato cultivar tolerant to saline stress compared to sensitive Moneymaker, although this gene was not related to the salt response in either tomato genotype (Sun et al., Reference Sun, Xu, Zhu, Liu, Liu, Li and Hua2010). NTGP4 protein is similar to the AIG1 protein from Arabidopsis (Biermann et al., Reference Biermann, Morehead, Tate, Price, Randall and Crowell1994; Dykema et al., Reference Dykema, Sipes, Marie, Biermann, Crowell and Randall1999), which is involved in resistance to pathogenic bacteria in plants containing the resistance gene RPS2 together with the avirulence gene avrRpt2 (Reuber and Ausubel, Reference Reuber and Ausubel1996). The AIG1 protein has also been related to the ABA signaling pathway (Kim and Kim, Reference Kim and Kim2006) whose role in the resistance to plant diseases has been reviewed (Ton et al., Reference Ton, Flors and Mauch-Mani2009).
Another transcript that was 28.42-fold more expressed in Motelle than in Moneymaker, represents the translation elongation factor 1-γ. This basal difference was only slightly reduced after infestation with B. tabaci (FC = 26.22). Members of the eukaryotic Elongation Factor 1 (eEF1) complex have been implicated in a wide variety of cellular and viral processes (Sasikumar et al., Reference Sasikumar, Perez and Kinzy2012). Upregulation of elongation factor 1-γ-like in leaves has been shown as a first hint at stressful conditions in plants subjected to biotic stress (Weiß and Winkelmann, Reference Weiß and Winkelmann2017).
The third gene up-regulated in Motelle regarding Moneymaker (FC = 11.30) encodes the enzyme Diaminopimelate (DAP) epimerase which catalyzes the lysine biosynthesis from aspartate. In addition, it is thought that this enzyme could be used as a component in antimicrobial agents (Hor et al., Reference Hor, Dobson, Downton, Wagner, Hutton and Perugini2013). The differential expression of this gene between Motelle and Moneymaker remained fairly stable after infestation by B. tabaci (FC = 10.49).
Also a gene encoding the vacuolar H + -ATPase A2 subunit showed more than fourfold greater expression in Motelle leaves than in Moneymaker's, and subsequent infestation with B. tabaci almost did not alter that difference. The activity of this subunit was described in resistance mediated by Cf-9 gene to the pathogen Cladosporium fulvum expressing the Avr9 avirulence gene (Piedras et al., Reference Piedras, Hammond-Kosack, Harrison and Jones1998). The changes in the permeability of the plasma membrane are of the first events that occur in the defensive responses of the plants after the recognition of pathogens or elicitors. These changes produce a depolarization due to the entry of Ca+2 and H+ and to the exit of K+ and Cl− (Scheel, Reference Scheel1998). These fluxes appear to be necessary for the induction of expression of defensive genes against pathogen attack or wounds (Fukuda, Reference Fukuda1996; Jabs et al., Reference Jabs, Tschope, Colling, Hahlbrock and Scheel1997; Schaller and Oecking, Reference Schaller and Oecking1999; Schaller and Frasson, Reference Schaller and Frasson2001).
In the absence of infestation, another gene up-regulated in Motelle with respect to Moneymaker (FC = 2.50) encodes selective channels for voltage dependent ions (VDACs), or pores formed from transmembrane channel proteins (porins) present in the outer membrane of the mitochondria. These channels were better studied in animal cells than in plants but in both cases they are involved in apoptosis (Voehringer et al., Reference Voehringer, Hirschberg, Xiao, Lu, Roederer, Lock, Herzenberg and Steinman2000; Okada et al., Reference Okada, O'Neal, Huang, Nicholas, Ostrowski, Craigen, Lazarowski and Boucher2004; Veenman et al., Reference Veenman, Shandalov and Gavish2008; Kusano et al., Reference Kusano, Tateda, Berberich and Takahashi2009; Tateda et al., Reference Tateda, Watanabe, Kusano and Takahashi2011). It has been demonstrated that VDAC protein is necessary for normal plant growth and for defense in Arabidopsis, regulating the generation of hydrogen peroxide (Tateda et al., Reference Tateda, Watanabe, Kusano and Takahashi2011). The involvement of hydrogen peroxide in the VDAC pathway was previously observed in the non-specific resistance of Nicotiana benthamiana to Pseudomonas cichorii (Tateda et al., Reference Tateda, Yamashita, Takahashi, Kusano and Takahashi2009). VDAC was used as a marker in Arabidopsis of the hypersensitive response (HR) to Xanthomonas campestris (Lacomme and Roby, Reference Lacomme and Roby1999) or plant programmed cell death (Swidzinski et al., Reference Swidzinski, Leaver and Sweetlove2004). However, the differential expression of this gene between Motelle and Moneymaker was not observed after infestation with B. tabaci, suggesting that the attack of this insect does not promote HR in tomato leaves. These data agree with previously obtained results in Arabidopsis with B. tabaci where cytological analysis of the leaves showed that no HR was produced after feeding of the whitefly nymphs (Kempema et al., Reference Kempema, Cui, Holzer and Walling2007). Similarly, HR was not observed in tomato during the compatible and incompatible interactions with aphids (Martínez de Ilarduya et al., Reference Martínez de Ilarduya, Xie and Kaloshian2003).
Other two transcripts that were expressed approximately double in Motelle than in Moneymaker prior to infestation were not differentially expressed after infestation with B. tabaci. The sequence of one of them (FC = 2.08) corresponds to a gene encoding a glycogen glycosyltransferase which was detected in a previous study of microarrays in non-infested roots of Motelle and Moneymaker (Schaff et al., Reference Schaff, Nielsen, Smith, Scholl and Bird2007). In the same study, induction of this gene was also detected during nematode incompatible interaction, demonstrating its participation in Mi-1-mediated resistance to nematodes. Other studies suggested the involvement of glycosyltransferases in processes of biotic and abiotic stress responses (Vogt and Jones, Reference Vogt and Jones2000; Dixon, Reference Dixon2001; Mazel and Levine, Reference Mazel and Levine2002; Langlois-Meurinne et al., Reference Langlois-Meurinne, Gachon and Saindrenan2005; Qi et al., Reference Qi, Kawano, Yamauchi, Ling, Li and Tanaka2005; Meissner et al., Reference Meissner, Albert, Böttcher, Strack and Milkowski2008; von Saint Paul et al., Reference von Saint Paul, Zhang, Kanawati, Geist, Faus-Keßler, Schmitt-Kopplin and Schäffner2011) and synthesis of the cell wall (Lao et al., Reference Lao, Long, Kiang, Coupland, Shoue, Carpita and Kavanagh2003; Egelund et al., Reference Egelund, Skjot, Geshi, Ulvskov and Petersen2004; Baumann et al., Reference Baumann, Eklöf, Michel, Kallas, Teeri, Czjzek and Brumer2007). The second transcript (FC = 2.04) is related to an F-box protein. Many proteins in this family are involved in plant vegetative and reproduction growth and development, as abscisic acid (ABA) signaling to affect the seed germination of Arabidopsis (Peng et al., Reference Peng, Yu, Wang, Xie, Yuan, Wang, Tang, Zhao and Liu2012) or regulation of cell death and defense after pathogen recognition in tobacco and tomato (Van Den Burg et al., Reference van den Burg, Tsitsigiannis, Rowland, Lo, Rallapalli, MacLean, Takken and Jones2008). To analyze the possible participation of these and other genes in whitefly resistance, it would be necessary to perform complementary studies to obtain their expression differences between infested and non-infested plants.
Among the genes down-regulated in uninfested Motelle compared to uninfested Moneymaker, the largest difference was in the gene encoding the GAI protein (FC = −9.6) that belongs to the GRAS family; these proteins fulfill regulatory functions in different aspects of signaling and plant development (Bolle, Reference Bolle2004; Achard et al., Reference Achard, Cheng, De Grauwe, Decat, Schoutteten, Moritz, Van Der Straeten, Peng and Harberd2006). The GAI protein contains an N-terminal domain DELLA (Silverstone et al., Reference Silverstone, Ciampaglio and Sun1998), and proteins sharing this motif are also known as DELLA proteins (Eckardt, Reference Eckardt2003). GAI was the second protein that was cloned from this family (Peng et al., Reference Peng, Carol, Richards, King, Cowling, Murphy and Harberd1997) after cloning the SCR protein (Di Laurenzio et al., Reference Di Laurenzio, Wysocka-Diller, Malamy, Pysh, Helariutta, Freshour, Hahn, Feldmann and Benfey1996). DELLAs restrict plant growth by suppressing the action of GAs (Bolle, Reference Bolle2004). Reciprocally, GAs regulate growth through the degradation of DELLA proteins (Harberd, Reference Harberd2003; Jiang and Fu Reference Jiang and Fu2007; Wang et al., Reference Wang, Zhu, Huang, Li, Gong, Yao, Fu, Fan and Deng2009). In Arabidopsis and tomato, these proteins control plant defense by modulating the responses dependent on SA and JA (Navarro et al., Reference Navarro, Bari, Achard, Lisón, Nemri, Harberd and Jones2008; Bari and Jones, Reference Bari and Jones2009; Ding et al., Reference Ding, Wei, Wu, Davis, Jiang, Lee, Hammond, Shen, Sheng and Zhao2013). The fact that lower expression of the GAI protein was obtained in Motelle than in Moneymaker could be associated with greater growth of plants containing the Mi-1 gene. However, no obvious differences were observed in this study between plants of both genotypes, thus suggesting that the lower expression in Motelle would not affect the development of these plants. Subsequent infestation with B. tabaci reduced the difference between Motelle and Moneymaker (FC = −5.74). This reduction could be explained by a lower GAI expression in Moneymaker that would result in plant growth promotion, although this fact was not observed during our work. Alternatively, the reduction in the difference between Motelle and Moneymaker after whitefly infestation can be also explained as an increase of GAI expression in Motelle that would lead to a decrease in GA. This might suggest that DELLAs can be important in the Mi-1-mediated resistance against whiteflies. In Arabidopsis, DELLAs repress SA signaling pathway during P. syringae infection (Navarro et al., Reference Navarro, Bari, Achard, Lisón, Nemri, Harberd and Jones2008). However, SA plays an important role during the Mi-1-mediated resistance in tomato against B. tabaci (Rodríguez-Álvarez et al., Reference Rodríguez-Álvarez, López-Climent, Gómez-Cadenas, Kaloshian and Nombela2015). The role of the GA signaling pathway in plant defense is ambiguous as antagonistic effects have been observed in several studies (De Bruyne et al., Reference De Bruyne, Höfte and De Vleesschauwer2014). It would be interesting to later complement the present microarray analysis with more specific studies to confirm the role of GA/DELLAs in the Mi-1-mediated resistance in tomato against B. tabaci.
Another gene down-regulated in Motelle compared with Moneymaker (FC = −8.77) encodes the enzyme Pys1 (phytoene synthase 1) involved in secondary metabolism and related to fruit ripening (Gady et al., Reference Gady, Vriezen, Van de Wal, Huang, Bovy, Visser and Bachem2012). The difference in gene expression between these two cultivars decreased (FC = −5.47) after infestation with B. tabaci. Phytoene synthase catalyzes carotenoid biosynthesis (von Lintig et al., Reference Von Lintig, Welsch, Bonk, Giuliano, Batschauer and Kleinig1997; Toledo-Ortiz et al., Reference Toledo-Ortiz, Huq and Rodríguez-Concepción2010) and its coding gene is induced in tomato in response to saline stress (Zhou et al., Reference Zhou, Wei, Boone and Levy2007), as well as in banana under abiotic stresses (Kaur et al., Reference Kaur, Pandey, Shivani Kumar, Pandey, Kesarwani, Mantri, Awasthi and Tiwari2017), but not in compatible or incompatible interactions to nematodes (Bhattarai et al., Reference Bhattarai, Xie, Mantelin, Bishnoi, Girke, Navarre and Kaloshian2008). Neither Pys1 has shown differential expression in response of Arabidopsis to whiteflies (Kempema et al., Reference Kempema, Cui, Holzer and Walling2007).
The expression of the enzyme ADK (adenosine kinase) that catalyzes the synthesis of AMP from adenosine and ATP was more than eight times lower in Motelle than in Moneymaker in the absence of infestation. This difference was slightly reduced to FC = −6.09 after infestation of both cultivars with B. tabaci. It was previously known that this gene is induced in N. benthamiana after virus infection (Wang et al., Reference Wang, Hao, Shung, Sunter and Bisaro2003) and also by salt stress in Beta vulgaris and Spinacia oleracea, but not in other related plant species such as N. tabacum and Brassica napus (Weretilnyk et al., Reference Weretilnyk, Alexander, Drebenstedt, Snider, Summers and Moffatt2001). A more recent study has shown that ADK plays a role in plant development and defense (Liu et al., Reference Liu, Pedersen, Schultz-Larsen, Aguilar, Madriz-Ordeñana, Hovmøller and Thordal-Christensen2016). Additional analyses would be necessary to determine with certainty if this enzyme plays any role in the Mi-1-mediated resistance to whiteflies.
Another gene with lower expression in uninfested Motelle than in Moneymaker (FC = −5.65) encoded a selT-like protein. This difference between cultivars increased to FC = −8.58 after whitefly infestation. SelT-like protein precursors have been related to selenite resistance, as a SELT gene was more induced in a selenite-resistant accession of Arabidopsis than in a selenite-sensitive accession after selenite treatment (Tamaoki et al., Reference Tamaoki, Freeman and Pilon-Smits2008).
MADS-box transcription factors are involved in the regulation of different processes of plant development including flowering, fruit development, and embryogenesis (Busi et al., Reference Busi, Bustamante, D'Angelo, Hidalgo-Cuevas, Boggio, Valle and Zabaleta2003). In the present study, the expression of MADS-box 15 in Motelle was found to be repressed by comparison with Moneymaker in the absence of whiteflies (FC = −3.15) and this difference was maintained and even slightly increased (FC = −4.07) after infestation with B. tabaci. The expression of MADS-box 15 had been increased in both compatible and incompatible interactions of tomato with nematodes (Bhattarai et al., Reference Bhattarai, Xie, Mantelin, Bishnoi, Girke, Navarre and Kaloshian2008). In addition, several tomato MADS-box genes were induced during incompatible interactions with X. campestris pv. Vesicatoria (Bonshtien et al., Reference Bonshtien, Lev, Gibly, Debbie, Avni and Sessa2005) or in response to saline stress (Zhou et al., Reference Zhou, Wei, Boone and Levy2007). Further analysis of the expression of these transcription factors during the compatible and incompatible tomato–whitefly interactions would make it possible to define their role in a possible negative regulation of the development of the plant in favor of defensive processes.
Among the repressed transcripts in Motelle relative to Moneymaker was a gene encoding a short-chain dehydrogenase/reductase, with FC = −2.60 in uninfested plants and FC = −4.83 in whitefly infested plants. The short-chain dehydrogenases/reductases (SDR) constitute one of the largest enzyme superfamilies with over 46,000 members (Persson et al., Reference Persson, Kallberg, Bray, Bruford, Dellaporta, Favia, Duarte, Jörnvall, Kavanagh, Kedishvili, Kisiela, Maser, Mindnich, Orchard, Penning, Thornton, Adamski and Oppermann2009). More specifically, the transcript Solyc11g071460.1.1 was recently involved in compatible plant–microbe interactions, as it was identified among differential expressed genes in tomato leaf tissue, down-regulated at 24 h post-inoculation with Bacillus cinerea (Rezzonico et al., Reference Rezzonico, Rupp and Fahrentrapp2017).
The expression of a cinnamyl alcohol dehydrogenase (CAD) was 2.39 times lower in uninfested Motelle than in uninfested Moneymaker, but following infestation with B. tabaci, this differential expression was no longer detected. However, ELI3 protein, which is also a type of CAD protein (Logemann et al., Reference Logemann, Reinold, Somssich and Hahlbrock1997), was expressed 5.9 times less in Motelle than in Moneymaker, both infested. CADs are key enzymes of lignin synthesis and they catalyze the reversible conversion of cinnamyl aldehyde to the monolignols that will give rise to lignin, which is why CAD activity is correlated with lignification in tomato (Roth et al., Reference Roth, Boudet and Pont-Lezica1997). In addition to its role in lignification, the increase in the expression of some genes encoding CAD enzymes has been associated with a number of defensive responses to pathogens in compatible and incompatible interactions (Kiedrowski et al., Reference Kiedrowski, Kawalleck, Hahlbrock, Somssich and Dangl1992; Mitchell et al., Reference Mitchell, Hall and Barber1994; Coelho et al., Reference Coelho, Horta, Neves and Cravador2006). Thus, in the interaction, Arabidopsis–whiteflies, the expression of two CAD isoforms increased (Kempema et al., Reference Kempema, Cui, Holzer and Walling2007). CAD levels were also overexpressed after infection of parsley with fungi and bacteria (Schmelzer et al., Reference Schmelzer, Kruger-Lebus and Hahlbrock1989; Somssich et al., Reference Somssich, Bollmann, Hahlbrock, Kombrink and Schulz1989; Van Gijsegem et al., Reference Van Gijsegem, Somssich and Scheel1995; Logemann et al., Reference Logemann, Reinold, Somssich and Hahlbrock1997). More recently, CAD has been shown to be important for the resistance against Rhizoctonia cerealis in wheat (Rong et al., Reference Rong, Luo, Shan, Wei, Du, Xu and Zhang2016). Similarly, during the interaction of tomato with Xanthomonas axonopodis pv. vesicatoria, the level of plant resistance to the pathogen positively correlated with the levels of CAD enzyme (Umesha and Kavitha, Reference Umesha and Kavitha2011).
In the absence of infestation, a glutarredoxin was expressed 2.13 times less in Motelle than in Moneymaker, and a similar differential expression was maintained after the infestation with B. tabaci (FC = −2.34). Glutarredoxins are antioxidant enzymes that play an important role in the control of oxidative stress (Kalinina et al., Reference Kalinina, Chernov and Saprin2008; Meyer et al., Reference Meyer, Siala, Bashandy, Riondet, Vignols and Reichheld2008).
Differential genes detected only after infestation with B. tabaci
Among the genes only up-regulated in Motelle compared to Moneymaker when plants were infested by B. tabaci, it is a remarkable one (FC = 12.16) corresponding to the E3 ubiquitin-protein ligase TRAF7 which belongs to the WD-40 repeat protein family. In eukaryotes, proteins of this family are involved in a variety of functions such as signal transduction, cell division, cytoskeleton assembly, chemotaxis, RNA processing, and apoptosis (Xu et al., Reference Xu, Li and Shu2004; Stirnimann et al., Reference Stirnimann, Petsalaki, Russell and Müller2010). The N termini of TRAFs 2–7 render them genuine E3 ubiquitin ligases which are required in the process of protein ubiquitination and determine the substrate specificity (Huang et al., Reference Huang, Chen, Zhong, Li, Ao, Huang and Li2016). Interestingly, SCF-TRAFasome formation mediated by TRAF proteins may represent a method used by plants to assemble SCF complexes upon pathogen infection (Huang et al., Reference Huang, Chen, Zhong, Li, Ao, Huang and Li2016).
An enzyme similar to acid phosphatase 1 (Aps-1) stands out among the transcripts that were more expressed in Motelle than in Moneymaker only after whitefly infestation (FC = 3.57). Although the function of acid phosphatases is not well known, tomato Aps-1 could participate in response to invader organisms, as its enzyme activity increased in the roots of both susceptible and resistant tomato plants after infection with RKN (Williamson and Colwell, Reference Williamson and Colwell1991). This was later confirmed in microarray studies (Bhattarai et al., Reference Bhattarai, Xie, Mantelin, Bishnoi, Girke, Navarre and Kaloshian2008). Moreover, the Aps-1 gene, closely linked to Mi-1, was cloned and has been employed as a molecular marker for the presence of Mi-1 (Aarts et al., Reference Aarts, Hontelez, Fischer, Verkerk, van Kammen and Zabel1991; Williamson and Colwell, Reference Williamson and Colwell1991).
Three other genes more expressed in Motelle than in Moneymaker, only after infestation with B. tabaci, were related to protection against oxidative stress: Firstly, the gene that encodes the enzyme isoflavone reductase (FC = 3.39) is one of the key enzymes in the isoflavonoid biosynthesis and whose antioxidant function has been observed in Arabidopsis (Babiychuk et al., Reference Babiychuk, Kushnir, Belles-Boix, Van Montagu and Inze1995) and rice (Kim et al., Reference Kim, Kim, Wang, Kim, Lee, Kim, Kim, Lee and Kang2010). The gene encoding the peptide methionine sulfoxide reductase (PMSR) enzyme (FC = 2.89), which may play an important role in cell protection against oxidative stress, as it has been observed with PMSR2 in Arabidopsis (Bechtold et al., Reference Bechtold, Murphy and Mullineaux2004). Also remarkable is a gene encoding the subunit 3 of the enzyme NADH dehydrogenase (nad3) (FC = 2.00), a subunit present in Complex I of the electron transport chain in mitochondria. Complex I acts as a proton pump toward the intermembrane space of the mitochondria, thus avoiding acidification of the matrix that can lead to oxidative stress (reviewed by Subrahmanian et al., Reference Subrahmanian, Remacle and Hamel2016) and, ultimately, a cellular damage manifested in an HR. The fact that these three genes were more expressed in Motelle than in Moneymaker after the infestation with B. tabaci aligns with results from a previous study where HR was not observed in the Mi-1-mediated response of Motelle after aphid attack (Martínez de Ilarduya et al., Reference Martínez de Ilarduya, Xie and Kaloshian2003). This HR was also absent in Arabidopsis after whitefly infestation (Kempema et al., Reference Kempema, Cui, Holzer and Walling2007). All these data indicate that whitefly infestation does not provoke HR in bearing-Mi-1 tomato leaves, unlike what happens when roots are attacked by nematodes (Dropkin, Reference Dropkin1969).
Two other transcripts over-expressed in Motelle compared to Moneymaker only after infestation with B. tabaci were identified. One of them (FC = 2.11) corresponded to a 21 kDa pectinesterase; these enzymes are involved in cell wall reorganization processes as well as in plant response to pathogen attack (McMillan et al., Reference McMillan, Hedley, Fyffe and Pérombelon1993; Wiethölter et al., Reference Wiethölter, Graeßner, Mierau, Mort and Moerschbacher2003; Raiola et al., Reference Raiola, Lionetti, Elmaghraby, Immerzeel, Mellerowicz, Salvi, Cervone and Bellincampi2011). The second gene encodes the enzyme ATP sulforylase 1 (FC = 2.53) belonging to the family of sulfate adenylyltransferase enzymes. These enzymes are involved in the sulfate assimilation pathway by catalyzing the activation of sulfate ions by ATP to form adenosine-5′-phosphosulfate (APS) and pyrophosphate (Marzluf, Reference Marzluf1997). This reaction is the first enzymatic step in the use of sulfate upon its uptake.
Among the genes that were less expressed in Motelle than in Moneymaker after infestation with B. tabaci, but without differential expression in non-infested plants, the ELI3 encoding enzyme (FC = −5.79) is a CAD protein (Logemann et al., Reference Logemann, Reinold, Somssich and Hahlbrock1997) which has been discussed above. Also a methyl transferase (FC = −2.57) is involved in different cellular processes among which is the regulation of gene expression during development (Finnegan et al., Reference Finnegan, Peacock and Dennis1996).
Included in this group are three proteins involved in transcription processes. Transcript LesAffx.21605.1.S1_at (FC = −4.03) corresponds to the high mobility group B protein. Differential expression of this HMG type nucleosome/chromatin assembly factor has been associated with plant leaf development (Rantong et al., Reference Rantong, Van Der Kelen, Van Breusegem and Gunawardena2016) and more recently with thermotolerance in perennial grass (Xu and Huang, Reference Xu and Huang2018). LesAffx.10016.1.A1_at (FC = −2.54) represents a gene of the pseudo response regulator (PRR) family, which are sequentially expressed over the course of the day. More specifically, this locus Solyc06g069690 has been identified in maize with the timing of cab expression1 (TOC1) gene (Bendix, Reference Bendix2015), one of the main contributors to the plant clock system (Farré and Liu, Reference Farré and Liu2013). The product of Les.5732.1.S1_at (FC = −2.13) is similar to E3 ubiquitin protein ligase DRIP2 that acts as a negative regulator of the response to water stress in Arabidopsis (Qin et al., Reference Qin, Sakuma, Tran, Maruyama, Kidokoro, Fujita, Fujita, Umezawa, Sawano, Miyazono, Tanokura, Shinozaki and Yamaguchi-Shinozaki2008).
The last two enzymes only differentially expressed in infested plants are an aldehyde oxidase (AO1) (FC = −2.08) and the UDP-glucuronate decarboxylase 2 (FC = −2.05). AO1 was identified in tomato along with other enzymes of the same family by Min et al. (Reference Min, Okada, Brockmann, Koshiba and Kamiya2000) who suggested that each AO could play a different role in the growth and development of this solanaceae. AOs are involved in hormone biosynthesis processes, in particular catalyze the last step of ABA biosynthesis (Min et al., Reference Min, Okada, Brockmann, Koshiba and Kamiya2000). The activity of the enzyme UDP-glucuronate decarboxylase, involved in membrane-associated metabolic processes, has been detected in several plants and the expression of genes encoding these enzymes in barley has been studied (Zhang et al., Reference Zhang, Shirley, Lahnstein and Fincher2005). This enzyme has an important role in cell wall biosynthesis (Seifert, Reference Seifert2004).
Conclusions
Genes highlighted in the first phase of this study represent the baseline differences between the transcriptomic profiles of the Motelle and Moneymaker tomato cultivars, associated with the presence of the Mi-1 gene in the first of them. The observed changes in the relative expression of these genes following whitefly infestation, as well as the emergence of other genes with differential expression, illustrate how the baseline differences between Motelle and Moneymaker are substantially altered by this insect. Taken together, these results provide us with valuable information on candidate genes to intervene in one way or another in the tomato resistance mediated by the Mi-1 gene to B. tabaci. However, to analyze the actual participation of these genes in such a resistance, it would be necessary to perform complementary studies to obtain expression differences between infested and non-infested plants of the same cultivar. Based on the results of the present study, further analyses are currently underway in our laboratory to define the role of these and other genes during the compatible and incompatible interactions of adult tomato plants with the whitefly B. tabaci.
Acknowledgements
The authors are grateful to Professor José Javier Pueyo (ICA-CSIC) for his support with the qRT-PCR analysis, and to Dr Michael D. Bell (National Park Service, USA) for helping to improve the English version of the manuscript. This research was funded by a Project (AGL2007-65854/AGR) from the Plan Nacional I+D+I, Spanish Ministry of Education and Science. Clara I. Rodríguez Álvarez was financially supported by a fellowship/contract (AP2006-1035) from the Spanish FPU Program.
