1 Introduction

Tomatoes (Solanum lycopersicum L.) are considered one of the commercially significant crops in the world [1]. According to The Observatory of Economic Complexity records, S. lycopersicum had a total trade of 9.81 billion dollars in 2020 (https://oec.world/en/profile/hs/tomatoes). It has grown 8.01% from 2019, making it one of the most crucial vegetables in the market. It is a great source of vitamins, minerals, and antioxidants. The demand increases exponentially with the global need for more variation, quality improvement, taste enhancement, health benefits, and longer shelf life. Although tomatoes are considered tropical plants, they are grown almost everywhere in the world and are reportedly susceptible to various biotic as well as abiotic stress.

Drought is one among the greatest predominant abiotic stresses affecting tomatoes (Plant Nutrition and Food Security in the Era of Climate Change. Academic Press; 2022). In tomatoes, the drought response mechanisms are reported to affect seed germination, water and mineral uptake, photosynthesis frequency, reactive oxygen species (ROS) generation, stomatal closure and root-shoot growth [2]. Due to the increased transpiration rate during drought, a decrease in relative water content (RWC) is observed, leading to dehydration and osmotic stress. The generation of stress-induced reactive oxygen species causes impairment to cellular structures and leads to cell death [3]. Thus, a drastic decrease in yield is observed in tomato plants due to drought stress.

Despite the availability of drought-resistant wild varieties of Solanum, the mechanisms and genes conferring resistance are obscure. Exploring drought-resistant wild types like Solanum pimpinellifolium, Solanum pennellii, Solanum habrochaites, etc., would provide better insights into the drought-response strategies in tomatoes. Identifying differentially expressed genes (DEGs) in drought stress conditions could be an important step in better understanding drought tolerance and its mechanisms. Secondly, in plants, a sizable share of genes in the genome (up to 10%) possibly code for transcription factors (TFs) [4]. Complex regulatory networks in plants control the drought stress response. TFs are essential managers in this network, which play a vital role by activating or hindering the expression of stress-linked target genes [5]. One of the major transcription factors reportedly intricate in drought response are AP2/ERF (APETALA2/Ethylene Response Factor) [6]. AP2/ERF TFs are a large superfamily of transcription factors found in plant systems. They are known to be indulged in the management of stress-responsive genes, osmotic adjustment, ROS scavenging, and stomatal regulation [7]. Several studies [8, 9] compile investigations on the role of AP2/ERF on plant drought. Other major transcription factors known to be involved in drought stress include NAC (NAM, ATAF1/2 and CUC2) [10], MYB (Myeloblastosis) [11], bZIP (Basic Leucine Zipper) [12], WRKY [13], and bHLH [14]. These transcription factor families, among others, interact with promoters of target genes to fine-tune gene expression and coordinate the plant's counter-response to drought stress [15]. Studying these transcription factors’ functions and regulatory networks provides valuable acumens into the molecular machineries underlying drought endurance in plants and offers potential targets for crop enhancement. Identification of transcription factors can therefore be vital in deciphering plant stress response mechanisms.

Recently, researchers have been focusing on gene/protein interaction networks. Network analysis can provide a holistic view of gene interactions as they do not function as a single entity but in interaction with others. Network analysis helps interpret molecular interactions and provides meaningful insights using graph theoretical analysis of biological systems [16]. By applying network analysis to biological data, researchers can uncover hidden relationships, identify key regulatory elements, and understand complex biological processes holistically. This systems-level perspective facilitates the discovery of novel biomarkers, drug targets, and therapeutic interventions. Similar network analysis studies are reported in rice [17, 18], wheat [19], Barley [20], Arabidopsis thaliana [21] and Centella asiatica [22]. Overall, network analysis has revolutionized the field of biology by providing a powerful framework to integrate, analyze, and interpret biological data in a comprehensive and context-specific manner. Therefore, reconstructing the interconnected network of DEGs can aid in detecting the crucial genes which can be potential targets for crop betterment.

In the present study, we attempted to identify drought-responsive genes in tomatoes by analysing microarray data of two tomato species- the wild-type Solanum pimpinellifolium and the cultivated Solanum lycopersicum with differential drought responses. To obtain significant TFs among the DEGs, we did a transcription factor prediction using the Transcription Factor Database (TFDB). Subsequently, protein–protein interaction network creation, GO annotation and hub analysis were carried out to get more insights into the drought mechanism. Our findings suggest a regulatory role for the jasmonic acid/ethylene pathway, antioxidant response, ROS scavenging, and heat stress response proteins in conferring tolerance to wild-type S. pimpinellifolium. The potential role of TFs and other significant enzymes is also explored in this study.

2 Materials and methods

2.1 Gene expression datasets

The datasets taken for the present study are from GSE39894 (Affymetrix Tomato Genome Array), containing transcriptome profiles of root tissues of three natural varieties of S. lycopersicum and S. pimpinellifolium as four biological replicates (a total of 46) under well-watered and water-deficit treatments. The data was submitted by Takuya et al.  [23]. As per the submitters, drought plants were irrigated until the soil was fully hydrated and left un-watered until all leaves showed signs of wilting. Plants were grown in commercial garden soil with natural lighting at a temperature range of 22–30 degrees Celsius until they formed flower buds. Total RNA was taken from tomato root tissue by grinding to a powder in liquid nitrogen using the modified protocol of Griffiths et al. [24] and used for data generation. The platform used for the experiment was GPL4741 (Affymetrix tomato genome array), which consisted of ~ 10,000 probe sets to interrogate over 9200 transcripts.

The microarray expression data were retrieved from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). GEO database is an open database of high-throughput resources, including microarrays, gene expression data, and chips [25]

2.2 Identification of differentially expressed genes (DEGs)

The data retrieved from the GEO database were then analysed to find DEGs involved in responses to drought stress using GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/). GEO2R is a web-based tool to detect DEGs from the GEO series. This tool can be used to differentiate DEGs between drought and control samples. Probe sets without matching gene symbols or genes with more than one probe set were discarded. Differentially expressed genes with a log2FC of ≥ 0.5 and ≤ − 0.5 and adjusted p ≤ 0.05 were identified and considered for further study. Fold change (FC) is a measure used to quantify the variations in expression levels of a particular gene between two conditions or experimental groups. It represents the ratio of expression levels in one condition relative to another. Log2FC (log2 fold change) refers to the transformation of fold change values obtained by taking the logarithm base 2 of the fold change. It is commonly used in gene expression analysis to stabilize variance across expression levels and to make fold changes symmetrical around zero. The Benjamini and Hochberg  [26] method was used as the FDR control.

2.3 Identification of transcription factors from DEGs

In the current research, to detect transcription factors among the differentially expressed genes, the PlantTFDB [27]transcription factor prediction tool was used. This tool utilizes domain-specific bit-score to identify transcription factors.

2.4 Gene ontology (GO) and pathway enrichment analysis of DEGs

Gene ontology (GO) enrichment was analysed using STRING database 11.5 (https://string-db.org/). Annotations were performed for GO terms based on biological process (BP), cellular components (CC) and molecular function (MF). The significance of GO terms was verified using Fisher's exact test (p < 0.05). Fisher's exact test in GO enrichment analysis helps to identify biologically meaningful associations between gene sets and functional categories represented by GO terms, providing insights into the underlying biological processes or pathways associated with a particular gene set.

2.5 Protein–protein interaction network construction

Functional associations among DEGs during drought stress were figured out by STRING database version 11.5 (https://string-db.org/). Functional associations in STRING dB are made based on co-expression data in the NCBI-GEO database, co-occurrence of the genes in the similar organisms and text mining. It also contains information mined from the summaries of scientific literature and important protein interaction datasets from other researches [28]. The confidence score for interactions was set as medium (≥ 0.4)[29]. The PPI network obtained from STRING v.11.5 was imported into Cytoscape version 3.9.1 for additional analysis.

2.6 Hub analysis

Hub analysis was performed out using Cytoscape to identify vital genes in the drought response network (version 3.9.1). Cytoscape is an open-source software for visualising molecular interaction networks and biological pathways and integrating these networks with annotations and gene expression profiles. Network analyzer tool available in the Cytoscape was used to find the network's highly connected nodes (hubs) [30]. Three network analyzer parameters, namely degree (DE), betweenness centrality (BC), and closeness centrality (CC), were used for sleuthing hub genes.

3 Results and discussion

In this work, the identification of differential gene expression signatures and transcription factors in drought-tolerant and drought-resistant type of Solanum is analysed and discussed. The methodology implemented in this study is represented in Fig. 1.

Fig. 1
figure 1

Methodology overview detailing the experimental approach utilized in the study

Full size image

3.1 DEGs under drought stress

This current study was done to achieve in-depth information about the fluctuations in the expression of genes in the roots of S. lycopersicum and S. pimpinellifolium in response to drought stress. Through a comprehensive evaluation, we identified a set of 397 DEGs in S. lycopersicum, of which 53% were upregulated and 46.7% downregulated. Conversely, the percentage of up and downregulated genes in S. pimpinellifolium showed the inverse pattern with 43.9% upregulated and 56% downregulated genes out of 353 DEGs (Supplementary data 1). Top 10 upregulated and downregulated DEGs of S. lycopersicum and S. pimpinellifolium  (Fig.2 a, b) are discussed in the following section.

Solanum pimpinellifolium (wild type) exhibited distinct upregulation of genes involved in ROS scavenging (iron superoxide dismutase), cell wall modification (glucan endo-1,3-beta-d-glucosidase, chitinase), ABA signaling (SNF4 protein, nine-cis-epoxy carotenoid dioxygenase), and stress response (small heat shock protein, TAS14 protein, alternative oxidase 1c, jasmonic acid 2). Conversely, downregulated genes played roles in stress management (proteinase inhibitor, peroxidase, expansin) and osmolyte accumulation (extensin-like protein Dif54, st3 protein) (Fig. 2b).

Fig. 2
figure 2

a Top 10 Differentially expressed genes identified in Solanum lycopersicum. b Top 10 Differentially expressed genes identified in Solanum pimpinellifolium

Full size image

Solanum lycopersicum (cultivated) displayed upregulation of genes associated with stress response signaling (ethylene-forming enzyme, LeCBF1 protein), cell wall modification (xyloglucan endotransglucosylase-hydrolase XTH9), and nutrient uptake (NRT2 protein, inorganic phosphate transporter). Downregulated genes were linked to ROS scavenging (iron superoxide dismutase), energy metabolism (SNF4 protein), and osmolyte accumulation (heat shock protein, dehydration-responsive element-binding protein) (Fig. 2a).

Despite the shared expression patterns like up and downregulation of various antioxidant defense, cell wall remodelling and ROS scavenging indicative of a general retort to drought stress in both species, specific differences were detected. Interesting observations include that of differential response of superoxide dismutase, jasmonic acid, ethylene biosynthesis and heat shock proteins. These results provide interesting perspective of the differences in drought-response strategies adopted by both species.

A sum of 398 and 352 DEGs were identified for cultivated and wild varieties of Solanum, in which 194 (48%) and 188 (53%) encode for transcription factors.

3.2 Transcription factors expressed under drought stress

Seventy-eight differentially expressed genes were found to contain corresponding TFs in TFDB. We identified 44 and 34 DEGs that belonged to 13 and 14 transcription factor families in S. lycopersicum and S. pimpinellifolium (Fig. 3), respectively. The highest percentage of TFs observed in both species was that of the ethylene response factor (ERF) family. S. lycopersicum showed double the quantity of ERF genes compared to S. pimpinellifolium. The transcription factors of basic helix–loop–helix (bHLH), basic region leucine zipper (bZIP), double B-box zinc finger (DBB), GARP superfamily of transcription factors (G2-Like), Gibberellic acid insensitive (GAI), Repressor of GAI, Scarecrow (GRAS), and Nuclear Transcription Factor Y Subunit Gamma (NF-YC) were found in equal numbers in both species. The homeodomain-leucine zipper (HD-ZIP), myeloblastosis (MYB), Three-amino-acid-loop-extension (TALE) and NAM, ATAF1/2 and CUC2 domain (NAC) TFs were expressed more in S. lycopersicum. The expression of WRKY TFs was observed to be more in S. pimpinellifolium. On the other hand, TFs ethylene insensitive3‐like (EIL) and MCM1, AGAMOUS, DEFICIENS, and SRF serum response factor (MIKC_MADS) were expressed exclusively in S. pimpinellifolium, and the zinc finger homeodomain (ZF-HD) was expressed only in S. lycopersicum. The list of DEGs, containing these TFs, is provided in Supplementary data 2.

Fig. 3
figure 3

Comparison of differentially expressed Transcription factors in wild and cultivated Solanum

Full size image

Different transcription factors stand reportedly associated with various stress responses. The role of AP2/ERF in drought response of Oryza sativa [31], Arabidopsis thaliana [32], Vigna radiate [33] and Triticum aestivum [34] are well documented. Similarly, NAC family of TFs are related to drought response in Oryza sativa [35] Capsicum annum L. [36], Glycine max [37], Triticum aestivum [38,39,40], Zea mays [41, 42], S. lycopersicum [43, 44] and Hordeum vulgare [45]. Also, in Solanum tuberosum [46, 47], Arabidopsis thaliana [48] and soybean [49], the bZIP family of TFs are stated to be intricate in drought response.

As per Li et al. SlERF84 conferred improved tolerance to drought in tomatoes [9]. SlJUB1, belonging to the NAC transcription factor family, showed binding affinity to the promoter region of genes encrypting SlDREB and DELLA to control ROS homeostasis and stress-associated gene modulation concerning drought stress [50]. Jian et al. stated the involvement of SlNAC6, a NAC TF, in tomato drought stress [51]. Orellana et al. reported the role of transcription factor SlAREB in conferring drought and salt stress [52]. Zhu et al. revealed the part of the SlbZIP1 TF mediating drought stress [53]. Pan et al. reported the downregulation of SlbZIP38 in tomatoes during drought [54].

Interestingly, the species-specific expression of two TFs in each species was noted. Out of these, the cultivated species exhibited ZF-HD, which per previous reports are involved in drought in species like Arabidopsis [55], cotton [56] and tomato [57]. Abscisic acid-responsive element-binding protein (AREB)/ABF (ABRE binding factor) regulon is an ABA-dependent gene expression during drought stress [58]. Transgenic plants overexpressing ABF3, AREB1/ABF2 or AREB2/ABF4 show better drought tolerance and improved ABA sensitivity [59].

On the contrary, the expression of TFs—EIL and MIKC_MADS were detected in the drought-tolerant tomato, S. pimpinellifolium. EIL is a vital gene family which plays a critical role in the ethylene signalling pathway, which controls a wide-ranging variety of plant growth development, and defences against various biological and abiotic stresses. EIL shows a differential response to low temperatures during kiwifruit aging [60]. Evidence supporting the role of EIL in the controlling of salt tolerance and wound signalling in rice [61,62,– 63] is also reported. MIKC_MADS, belonging to the MADS-domain TFs, are vital members of regulatory systems underlying numerous developmental pathways in plants, including abiotic stress response [64]. In rice, MADS26 is a negative controller of drought stress [65]. But tomato MBP11 is required to give more salinity tolerance to the plant [66]. Likewise, in transgenic tobacco plants, the increased expression of MADS51 [67] improved plant growth under phosphorous. Although the role of MIKC_MADS is not widely understood in the case of drought, the possibility of cross-talk among salinity and drought stress in different plants highlights the potential of this TF in conferring drought tolerance in wild S. pimpinellifolium. These studies specified that precise TFs may be crucial in regulating the downstream stress response gene expression during drought. These findings also extrapolate the possibility of an ethylene/jasmonic acid-dependent signalling pathway in S. pimpinellifolium roots.

3.3 Gene ontology of DEGs

STRING database was utilised for exploring the GO enrichment for the total DEGs obtained from drought-tolerant and drought-sensitive Solanum varieties and discussed based on the strength and FDR value. In the biological process, several terms were detected as enriched terms in both species like carbohydrate metabolic process and response to stimulus.

In wild type, DEGs categorised under BP were that of carbohydrate metabolic process (Eg: XTH5, XTH1, XTH16, XET2, tXET-B2, XET4, XTH6, Cel3, XYL2, PMEU1, PGcat GolS-1, IMP2, chMDH, XOPG1, ARF1, CEVI-1, AUREA, CIP2b, ACI25, sbt3, ER5, EIL3, JA1, JA3, CDkB2, ETR6, systemin). The most significant molecular function terms were hydrolase activity ( Xyloglucan endotransglucosylase/hydrolase (XTH) or related enzymes (XTH5, XTH1, XTH16, XET2, tXET-B2, XET4, XTH6) along with endo-beta-1,4-d-glucanases (Cel3, Cel1, Cel5, Cel2), cellulase, xylanases [XYL2 (Xylanase 2), pectin methylesterase (PMEU1), polygalacturonase (PGcat)] and stress response (PR-P2, GolS-1, IMP2, chMDH, ARF1, TAPG4) As majority of the protein expression involved cell wall modification enzymes, the major cellular component ontology was enriched as cellular anatomical entity which comprises the structural components of the cell (Fig. 4). These observations strengthen the hypothesis of better cell wall modification/recycling strategies in wild species to combat drought.

Fig. 4
figure 4

Visualization of enriched Gene Ontology terms, biological process (BP), cellular component (CC) and molecular function (MF) in Solanum lycopersicum and Solanum pimpinellifolium

Full size image

Similarly, in cultivated species too, the most enriched BP was that of carbohydrate metabolic process with the expression of certain different proteins like AGP-S1 (ADP-glucose pyrophosphorylase large subunit), XTH3 (Xyloglucan endotransglucosylase/hydrolase), FUT13 (alpha-1,4-fucosyltransferase), Cel8 (endo-beta-1,4-d-glucanase), TBG2 (beta-glucosidase), chMDH (cytoplasmic malate dehydrogenase), SIP (plasma membrane intrinsic protein), PGcat (polygalacturonase), mdh (malate dehydrogenase), cel7 (endo-1,4-beta-d-glucanase)and tEG3 (ss-galactosidase). In MF, the most enriched term was that of hydrolase activity with an additional expression of few proteins like cel7 (endo-1,4-beta-d-glucanase). Similar to wild type, the most enriched CC was of cellular anatomical entity with few specific expression of cell wall remodelling proteins like CCD1-2, sbt3, AGP-S1, AGP-S2, FUT13, TBG2 along with signaling regulation (WRKYIIe-1) and stress response (EXPA3, JA1). The DEGs can be found in the Supplementary Data 3.

These differences provide a holistic view of the similarities and variations in gene expression during water scarcity in both species and will be of use for further experimental studies.

3.4 Enriched pathways among the DEGs

Most of the enriched pathways were common in both species like the metabolic pathways, biosynthesis of secondary metabolites, plant hormone signal transduction and MAPK signalling. Other pathways related to drought included arginine and proline metabolism [68] and carotenoid biosynthesis [69], arginine biosynthesis [70], cysteine and methionine metabolism and glutathione metabolism [71]. However, in wild type, enrichment on additional pathways like protein processing in endoplasmic reticulum [72], nitrogen metabolism [73] and pentose and glucuronate interconversions (Fig. 5) was observed. The proteins involved in the protein processing in ER were mainly molecular chaperones which assist in protein folding, assembly and in preventing misfolding under stress conditions like drought. There is also enrichment of heat shock proteins (HSP21, HSP17 and er-sHSP) which aids in stabilizing existing proteins and prevent them from misfolding/aggregating due to stress.

Fig. 5
figure 5

Enriched pathways among S. lycopersicum and S. pimpinellifolium DEGs

Full size image

Nitrogen metabolism comprised of carbonic anhydrase (CA1 and CA2) proteins which catalyzes reversible conversion of carbon dioxide and water to bicarbonate. The role of these proteins in combating drought is not very clear but they might be involved in stomatal regulation. Glutathione S-transferase 1 is involved in ROS detoxification which is crucial during drought stress. Another protein, glutamate dehydrogenase 1 which aids in interconversion of glutamate and other amino acids can be connected to the production of glutamate which is the precursor of proline synthesis. Proline is an osmoprotectant widely reported during drought stress.

Another wild-specific enrichment was that of the pentose and glucuronate interconversions which has polygalacturonates, its precursor, pectin methyl esterase and abscission polygalacturonase expression. These are pivotal for the breakdown of pectin which are a major cell wall component. The involvement of these enzymes can be that of cell wall modification during drought response. Abscission polygalacturonase is a specific type of polygalacturonase involved in shedding of leaves which might be a response strategy in tolerant variety for conserving water.

3.5 Protein–protein interaction network analysis

STRING analysis of DEGs obtained from GEO2R resulted in the reconstruction of gene networks (Fig. 6a, b) with 301 nodes and 591 edges in S. pimpinellifolium and 336 nodes and 740 edges in S. lycopersicum, with a PPI enrichment value < 1.0 e−16. The hub analysis using network analyzer parameters like degree centrality, betweenness centrality, and closeness centrality led to identifying 7 common (DE, CC, BC) hub genes (Fig. 7) in S. lycopersicum. They were identified to be glutathione synthetase (GSH2), zeaxanthin epoxidase (LOC544162/ZEP), ascorbate peroxidase 2 (APX2), allene oxide synthase (AOS), Delta-1-pyrroline-5-carboxylate synthase; (P5CS/PRO2), 1-deoxyglucose5-phosphate synthase (DXS) and mitogen-activated protein kinase 3 (MPK3). The AREB gene was common in degree and CC centralities in S. lycopersicum. Other significant hubs from the network analyzer are listed in Fig. 7. On the other hand, GSH2, Catalase 2 (CAT2), pathogen-related-protein 2 (PR-P2), AOS, c-type cyclin-dependent kinase (CDKB2), and MPK3 in S. pimpinellifolium, were found to be common in three network parameters. ACO5, ACS2 and tEG3 genes were common among degree and CC centralities in S. pimpinellifolium. GSH2, MPK3 and AOS genes were shared in both the species in all three centralities (Degree, CC, and BC), whereas the lipoxygenase D (LoxD) gene was present in degree and betweenness centrality in both species.

Fig. 6
figure 6figure 6

a Protein–protein interaction network of Solanum pimpinellifolium  highlighting key interactions. b Protein–protein interaction network of Solanum lycopersicum  highlighting key interactions. Blue colour denotes down, and red colour denotes upregulation. Size of the circle indicates the fold change

Full size image
Fig. 7
figure 7

Summary of network analysis—Degree, Betweenness and Closeness Centralities, the common genes among them along with fold changes

Full size image

3.6 Gene expression pattern of common hubs among S. lycopersicum and S. pimpinellifolium

In this study, four genes, MPK3, AOS, LOXD and GSH2, were shared in both varieties. According to previous reports, the first 3 genes are acknowledged to take part in the initial drought response of plants. MPK3 activation is a crucial response of plants towards different biotic/abiotic stressors. MPK3 activates many downstream essential biological processes like activating ABA/JA (jasmonic acid)/SA (salicylic acid) signalling and stress-related transcription factors essential for the drought response. The role of MPK kinases in promoting drought resistance is reported in various plants like tobacco [74]. Preceding reports have revealed that MPK3 regulated ethylene biosynthesis, phosphate acquisition and leaf senescence [75, 76] in various plants. The role of MAPK cascade in Arabidopsis leaf senescence is reported in 2009 [77] and its role in metal stress was reported in 2012 [78]. Overexpression of the tomato SlMAPK3 heightened endurance to cadmium and drought stress by refining various structural and physiological characteristics [79]. SlMAPK-knockout mutants showed more tolerance to heat stress in tomatoes [80]. In accordance with these reported studies, both species in our study showed an upregulation of MPK3 with a two-fold increase in drought-susceptible and a one-fold increase in drought-tolerant variety.

The second common gene, AOS, is involved in the biosynthesis of jasmonic acid (JA) and its precursor 12-oxo-phytodienoic acid (12-OPDA) from lipoxygenase-derived hydroperoxides. It catalyses the synthesis of unstable allene oxide, which is further converted spontaneously by hydrolysis or cyclisation. Both JA and 12-OPDA are testified to be associated in drought response in plants like Poncirus trifoliata [81]. The role of AOS in tomato drought and salt response were reported by Upadhyay et al. and Pedranzani et al. [82, 83]. Expression of AOS in our study correlated with the previous reports and showed an upregulation in both varieties explaining the participation of JA in the drought response in this study.

Another common gene, LoxD, is also involved in the oxylipin biosynthesis responsible to produce JA. The role of LoxD in drought is a less studied area requiring more focus. Although few studies exploring the part of LoxD in biotic and drought are available [84, 85], more research on the protein is needed for a deeper understanding. An overexpression of LoxD is associated with activating the H2O2 signal in guard cells, thereby arresting stomatal opening. This result complemented our results which showed an uptrend in LoxD expression in both species.

Lastly, GSH2, which encodes the chloroplast protein involved in the sub pathway synthesizing glutathione from l-cysteine and l-glutamate, was common in both species. However, the expression pattern of this enzyme showed a differing countenance. This protein comes under glutathione biosynthesis in the sulphur metabolism pathway. Glutathione, along with Glutathione S-transferases, is involved in reducing oxidative stress caused due to abiotic/biotic stress. It also involves morphological changes like root hair development as a stress response. In response to drought, differential glutathione levels are found in many plants. The upregulation of GSH2 in Arabidopsis and rice is related to improved drought tolerance, as Nguyen et al. and Siddiqui et al. reported [86, 87]. Similarly, Xia et al. stated the upregulation of GSH2 in increasing drought tolerance in transgenic tobacco [88]. Elevated levels of GSH2 were dependable with the foliar buildup of GSH in drought-treated Populus simonii saplings [89]. Mateusz et al. studied the role of Glutathione-dependent responses of different plants to drought and stated it as an essential protein in conferring drought tolerance [90]. In Solanum lycopersicum, exogenous abscisic acid (ABA) has been shown to escalate the ratio of GSH/GSSG levels [91]. During drought, maize showed a differential acclimation pattern of GSH in leaves and roots [92]. Our results showed a differential expression of GSH2 with upregulation in drought-susceptible and downregulation in the tolerant varieties. The fold change of GSH2 was minimal (− 0.6 in the wild and + 0.66 in the cultivated type) and, therefore, might be too small to bring in a significant change. The observed disparate expression in Solanum varieties can be attributed to a combination of compensatory responses, resource allocation strategies, feedback regulation, genetic variation, and temporal dynamics of gene expression in response to drought stress. The slight upregulation of GSH2 in the susceptible variety may indicate a compensatory mechanism to mitigate oxidative stress produced by drought, while downregulation in the tolerant variety could suggest more efficient or alternative stress-responsive pathways. Resource allocation may influence the expression patterns, with susceptible varieties reallocating resources towards antioxidant defense mechanisms at the expense of other genes. Feedback regulation and genetic variation further contribute to the observed expression patterns, which may fluctuate dynamically over period in response to changing stress levels. Further experimental work is needed to elucidate the species-specific roles of GSH2 in stress tolerance.

Taken together, the expression pattern of MPK3, AOS and LoxD implied the involvement of protein kinase cascade (MPK3) and JA in drought response in both varieties. A differential expression pattern was observed in the case of GSH2, which is involved in oxidative damage control but the fold change was found to be minimal. Despite the upregulation of cysteine and methionine biosynthesis, GSH2 showed a downtrend in the wild type. This pattern may be due to the downregulation of galactose metabolism in wild type.

3.7 Drought-susceptible S. lycopersicum-specific hubs

Our study identified the genes APX2, PRO2, DXS and AREB among the top 14 hubs using network analysis parameters. APX2, a gene involved in the H2O2- detoxification system in chloroplast and cytosol, was upregulated in accordance with other studies implying activation of basal level responses to drought. A similar pattern of APX2 in drought resistance is reported in other plants like Citrus rootstocks [93] and apples [94]

The other three genes- PRO-2, DXS, and AREB were downregulated in S. lycopersicum. In plants, the PRO-2/P5C5 gene family is a key rate-controlling enzyme in proline biosynthesis, leading to osmoregulation. The relationship amongst proline accumulation and drought tolerance is unclear. Several studies stated its role in different stress tolerance mechanisms in species like potato [95], transgenic tobacco [96], Cenchrus species [97], soybean [98] and wheat [99]. Generally, proline accretion is associated with increased stress tolerance [100]. In our study, a decreased drought tolerance in cultivated S. lycopersicum may be associated with the downregulation of proline. However, there are studies which challenge this postulation. Lv et al. showed no relationship between proline content and tolerance in 25 rice cultivars [101]. Also, a negative correlation between proline with salt stress was reported by Kishor et al. [100]. Hence, additional research is required to comprehend the expression pattern of proline in our research. DXS is the gene coding the first regulatory enzyme in the Rohmer pathway, also known as the 2-C-methyl-d-erythritol 4-phosphate (MEP) pathway. It catalyses the translation of pyruvate and d-glyceraldehyde 3-phosphate into 1-deoxy-d-xylulose 5-phosphate (DXP). In coordination with the Psy1 gene, it is also suggested to control carotenoid biosynthesis during fruit ripening in tomatoes. Carotenoids play a remarkable role in RO scavenging, and the downregulation of DXS can be deduced as a reason for decreased tolerance in cultivated types. A study by Loyola et al. showed a comparable reduction pattern in DXS expression in drought-stressed S. lycopersicum and S. chilense varieties [102].

The chief ABA-dependent gene activation pathway is facilitated by a precise subfamily of basic leucine zipper (bZIP) transcription factors that identify abscisic acid-responsive elements (ABREs) known as abscisic acid-responsive element-binding proteins (AREBs). AREB proteins are characterised widely for their active involvement in modulating drought stress responses. The promoters of drought-responsive genes contain ABRE: PyACGTGG/TC, to which the AREB TF binds and leads to the activation of drought response genes [103]. Arabidopsis plants expressing TF and other genes with ABRE motif-containing promoters, such as ANAC002/ATAF1, ANAC019, ANAC055 [104], and RD29B [105], also show increased drought tolerance. The significance of AREB in drought is reported in many plants like rice [106], grapes [107], and soybean [108].

3.8 Drought-tolerant S. pimpinellifolium-specific hubs

The identified hubs upregulated in the drought-tolerant S. pimpinellifolium variety were PR-P2, tEG3, ACS2 and ZEP. On the other hand, CAT2, CDKB2 and ACO5 showed a downregulating pattern.

Pathogenesis-related protein2 (PR-P2) is a defence gene identified as a characteristic marker gene of ethylene/JA and SA-mediated defence signalling pathways. Moreover, it is reportedly directly/indirectly regulated by ERF2 of the ERF family of transcription factors [109]. Therefore, a two-fold increase of PR-P2 gene countenance and the increased presence of transcription factors like ERF and EIL in Solanum pimpinellifolium suggest the possibility of an ethylene/JA and SA mediated signaling pathway during drought stress. Although few studies stating the role of PR-P2 in biotic pressure response in tomatoes are reported [110], the role of PR-P2 in drought is not well elucidated. Due to its presence in the extracellular and vacuole, it can be presumed to play a part in detoxification and preserving cell homeostasis. However, more investigational evidence is essential to confirm the role of this protein in tomato drought response.

Beta-galactosidase 3 or tEG3 belongs to the glycosyl hydrolase family, which is associated in hydrolysing galactosyl residues from cell wall polysaccharides by endo-cleavage of beta-1,4-galactan. It can liberate galactose from carbohydrates, galactolipids and glycoproteins. Evidence supporting its role in cell wall expansion, senescence and fruit ripening are also explicitly studied. Increased beta-galactosidase activity in drought-tolerant S. pimpinellifolium suggests its potential role in releasing/providing galactose, which is needed for synthesising raffinose. Raffinose is mainly involved in ROS scavenging during drought stress.

During drought, the MPK3 cascade activates the ACS2 enzyme and catalyses the translation of S-adenosyl methionine (SAM) to aminocyclopropane-1-carboxylic acid (ACC). ACC is the predecessor of ethylene. The transformation of ACC to ethylene requires oxygen and is thus regulated at the ACS enzyme level. An upregulation of ACS and the presence of TF EIL in S. pimpinellifolium strengthen the possibility of involvement of ethylene signalling in drought response.

Another upregulated gene to be identified as a hub is the chloroplastic ZEP. ZEP is involved in converting zeaxanthin into anthexanthin and subsequently to ABA. ABA is the most crucial phytohormone mediating abiotic/biotic stress-related downstream signalling. Increased ZEP concentration correlated with drought in tomato roots [111] and seedlings [112]. On the contrary, a decrease in ZEP was observed in tomato leaves during drought [113] indicating a tissue-specific expression.

As mentioned above, the downregulated genes in S. pimpinellifolium like CAT2, CDKB2 and ACO5 belong to antioxidant, cell cycle progression and ethylene biosynthetic processes, respectively. In contradiction to the general trend of increased catalase activity during drought tolerance, in this research work, the expression of CAT2 was downregulated in the drought-tolerant S. pimpinellifolium. The lack of correlation between CAT2 and ROS production suggests the involvement of other alternate genes that function in oxidative stress scavenging.

A downregulation of ACO5 and ACO2, ACO4, ACS10 and ACS11 are reported as drought response mechanisms in Arabidopsis. Correspondingly, a downtrend of ACO5 was observed in drought-stressed S. pimpinellifolium roots in our study. ACO5 is previously reported to harbour a leucine zipper domain, unlike other ACO genes, and its expression is accelerated during anaerobic stress like flooding [113]. The activity of ACO5 is reportedly associated with oxygen concentration [114, 115] and therefore, the presence of ROS during drought might cause the downregulation of ACO5 and aid in ethylene production.

The CDKB2 gene encrypts a cyclin-dependent protein kinase aiding in regulating of the G2/M transition of the mitotic cell cycle. In our study, downregulation of CDKB2 was observed, implying the possibility of endoreduplication as an energetically efficient alternative to cell proliferation. An earlier study has reported that endoreduplication in Arabidopsis leaf lessened the negative impact caused by drought stress [116].

The initial drought response in both species points towards an activation of MAPK signaling and the jasmonic acid biosynthesis pathway. When combined with the TFDB results, common transcription factors like ERFs, MYB, TALE, bHLH, and WRKY initiate the downstream gene regulation process. The presence of common TF ERF may suggest the JA/ABA-dependent pathways commonly observed in stress response pathways.

At the same time, the drought-tolerant Solanum pimpinellifolium exhibits better strategies to adapt to water deficits. Firstly, the expression of EIL and MIKC_MADS TFs was exclusively found in wild species, which further strengthens the possibility of early ethylene-based gene regulation. Also, presence of better ROS scavenging, heat shock proteins and nitrogen metabolism are also observed. The upregulation of ACS2, an essential gene from the ethylene biosynthetic pathway, along with the presence of transcription factor EIL and gene PR-P2, underlines the involvement of ethylene in drought in this species (Fig. 8). An additional presence of MIKC_MADS TF, reportedly involved in conferring salinity resistance, also improves its relevance as salinity and drought pathways often exhibit cross-talk in many plants. An increased expression of FeSOD corresponds to better antioxidant strategies adopted by the wild species. Along with these, the downregulation of CDKB2, which can be correlated to the strategy of endoreduplication in this species, underlines better stress adaptation strategies adopted by the wild species. Taken together, an increased stress resilience due to the regulation of jasmonic acid /ethylene-based signaling by inducing leaf senescence and stomatal closure to reduce transpiration can be elucidated. Also, protection of cellular machinery from drought stress damage by regulation of heat shock proteins along with active mitigation of oxidative damage by increased production of ROS might confer a better tolerance to wild species.

Fig. 8
figure 8

Proposed drought response mechanism in Solanum pimpinellifolium (Image is created using BioRender.com)

Full size image

However, the cultivated variety's gene expression patterns suggest a different drought response strategy. The absence of ethylene biosynthesis genes among the hubs indicates the plant might not be relying on ethylene-mediated stress response, thereby reducing the role of processes like leaf senescence and stomatal closure. Although GSH2 levels was upregulated in S. lycopersicum, indicating the presence of ROS scavenging mechanisms, the fold change is minimal and might be inadequate for a strong response The downregulation of hubs, PRO2 and DXS suggests a potential dampening of two key pathways involved in stress response by reducing proline biosynthesis and MEP pathway activity. These observations suggest a weaker defense system against drought stress in cultivated variety. (Fig. 9).

Fig. 9
figure 9

Proposed drought response mechanism in Solanum lycopersicum (Image is created using BioRender.com)

Full size image

The differences in drought response between wild and domesticated Solanum varieties can be attributed to the years of domestication. Domestication of tomato plants primarily focused on improving traits such as fruit size, quality, and shelf life. As a result, domesticated tomato varieties may have lost some of the drought resistance strategies in their wild relatives. Research has shown that the genetic diversity of cultivated tomato varieties is only a small fraction of that found in their wild relatives. This lack of genetic diversity can limit the potential of domesticated varieties to acclimatize to changing climatic environments, like drought. However, recent developments such as de-novo domestication strategies offer promising ways to address this issue. These strategies use genetic information from wild relatives to develop novel cultivars with required traits, including drought tolerance. By reintroducing genetic diversity from wild relatives, de-novo domestication can help improve cultivated tomatoes' ability to handle with drought and other environmental tensions.

4 Conclusion

Increasing incidents of drought and the limited knowledge of the mechanisms that preside over the responses to water stress in Solanum species pose a challenge to improving their drought tolerance strategies. To this end, our study attempts to look into the transcriptomic responses during drought using microarray data and relies on a multi-omics approach to analyze the expression differences between wild and cultivated tomatoes (Solanum pimpinellifolium L. and Solanum lycopersicum L.) with varying drought tolerance.

Several key observations, including the upregulation of genes involved in antioxidant defense, stress signaling, and cell wall modification in wild species, suggest a proactive response to drought stress aimed at enhancing stress tolerance and survival. Downregulation of lignin biosynthesis and nutrient uptake genes may reflect a prioritization of resources toward stress adaptation and resilience. On the other hand, the upregulation of defense-linked genes and stress signaling pathways indicates a similar response to stress, albeit with a focus on pathogen defense and regulatory mechanisms. Downregulation of antioxidant enzymes and stress-responsive genes suggests a potential trade-off between stress adaptation and resource conservation, possibly due to domestication-related factors or genetic selection for specific traits. Similarly, dominance of drought-responsive TFs like AP2/ERFs, which is a widely reported TF during stress have been observed in both the species with a species-specific expression of EIL and MIKC_MADS in the wild type and ZF-HD in the cultivated type. EIL, activated by ethylene signaling aids in stomatal closure and also in regulation root development. Osmolyte biosynthesis, ABA biosynthesis and senescence are regulated through MIKC_MADS family of TFs. Similarly, ZF-HDs regulate stomatal regulation and stress protection through specific pathways. Enrichment of gene ontologies correlated with the general stress response strategies adopted by plants. Another significant observation included the enrichment of protein processing in ER, nitrogen metabolism and pentose and glucuronate interconversions. The network and hub analysis suggest similar activation of stress response signaling pathways (MPK3), ABA-dependent pathway (PR-P2 and AREB), oxylipin biosynthesis (AOS, LOXD) antioxidant defense mechanisms (APX2, GSH2), and metabolic adaptations. Differences like the presence of ethylene biosynthetic genes (ACO5 and ACS2) and cell cycle regulation (CDKB2) were observed as hub genes in the wild variety and genes. The differences may be attributed to evolutionary adaptations, domestication effects, or other environmental and genetic factors.

Taken together, the drought tolerance in wild species of Solanum may be due to the species-specific regulation of jasmonic acid and ethylene pathways accompanied with the efficient response of ROS scavenging and heat shock proteins. The observations in this study can be used for experimental studies and can be used to develop drought-tolerant cultivated varieties using marker-assisted breeding, genome editing technologies like CRISPR-CAS, introgression of genomic regions /alleles associated with drought tolerance from wild to cultivated varieties or by including drought tolerant traits like antioxidant activity, hormone signaling into phenotypic selection criteria for breeding programs. Additionally, understanding the roles of TFs like ERFs and species-specific TFs can guide efforts to manipulate gene expression and regulatory networks to improve drought resilience in tomatoes.