In silico analysis of ESTs from roots of Rangpur lime ( Citrus limonia Osbeck ) under water stress

CitEST project resulted in the construction of cDNA libraries from different Citrus sp. tissues under various physiological conditions. Among them, plantlets of Rangpur lime were exposed to hydroponic conditions with and without water stress using PEG6000. RNA from roots was obtained and generated a total of 4,130 valid cDNA reads, with 2,020 from the non-stressed condition and 2,110 from the stressed set. Bioinformatic analyses measured the frequency of each read in the libraries and yielded an in silico transcriptional profile for each condition. A total of 40 contigs were differentially expressed and allowed to detect up-regulated homologue sequences to well known genes involved in stress response, such as aquaporins, dehydrin, sucrose synthase, and proline-related synthase. Some sequences could not be classified by using FunCat and remained with an unknown function. A large number of sequences presented high similarities to annotated genes involved with cell energy, protein synthesis and cellular transport, suggesting that Rangpur lime may sustain active cell growth under stressed condition. The presence of membrane transporters and cell signaling components could be an indication of a coordinated morphological adaptation and biochemical response during drought, helping to explain the higher tolerance of this rootstock to water stress.


Introduction
Water stress limits crop production worldwide and has major impact on plant growth and development.Responses to abiotic stresses are complex and involve multiple stages of plant responses, according to the event (Chinnusamy et al., 2004).For instance, less water available at the root zone implies additional regulation to balance photosynthesis, water transpiration, and overall metabolism.Secondary consequences may also happen, such as the rise of a nutrient concentration to a toxic level, eventual foliar depletion, and reduction of crop yield.
The advent of new technologies, especially available during the 'post-genomic' era, allowed foresights into the complexity of plant responses (Oono et al., 2003).Even be-fore that, abiotic stresses have received proper attention in different studies, comparing non-stressed to stressed conditions caused by low temperature (Nordin et al., 1993), salinity (Forsthoefel et al., 1995), and water deficit (Vartanian, 1996).Helped by emerging protein and model plant databases, many molecular mechanisms have been envisioned in non-model plants as well, and citrus has been one of them (Sanchez-Ballesta et al., 2004).
Many studies have been conducted to elucidate the mechanisms that underlie plant responses to abiotic stress.The use of model plants like Arabidopsis thaliana has permitted some improvement in different studies, and recent efforts have focused on molecular response to water-deficit stress (Bray, 2004).Some drought-inducible genes have been identified: those involved in osmoprotection (e.g.sucrose, trehalose, and proline), protective factors of macromolecules (chaperone, heat shock proteins, and late embryogenesis abundant proteins), membrane transporters (aquaporins) and regulatory genes (transcription factors) (Yamaguchi-Shinozaki et al., 2002).In addition, it has been proven that other pathway components (including protein kinases) are required in many signal transduction pathways for expression of drought-inducible genes (Mikami et al., 1998;Soderman et al., 1996;Urao et al., 1994).Nevertheless, functions of most of the responsive genes during abiotic stress remain unknown.
Water is normally an agricultural limiting factor for most Brazilian soils.Commercial citrus groves, for instance, rely on grafting a scion of economic interest onto a more drought tolerant rootstock, which may endure from 4 to 6 months of low levels of water in the soil, depending on the region (Carlos et al., 1998).The industry has made efforts to implement irrigation, but the cost of energy and the necessary investment have restrained further development.Thus, achieving drought tolerance in citrus is still extremely important.For half a century, this problem has been partially managed in São Paulo state using Rangpur lime (Citrus limonia Osbeck), a drought-tolerant rootstock.This rootstock also shows other attributes, making it possible to grow citrus under different limiting conditions, such as low nutritional and general grove maintenance conditions (Pompeu Jr., 2001).Nationally recognized by growers, Rangpur lime has been widely used after the industry collapsed due to the incidence of Tristeza during the 1940s (Grant, 1961).In the following decades, its reputation grew markedly, reaching preference levels of above 90% in nurseries for new plantings (Pompeu Jr, 2001).Thus, during this period, the orchards became supported almost completely by a single species, which led, as expected, to critical phytosanitary limitations for the Brazilian citrus industry.During the 1970s, citrus blight was reported (Rossetti et al., 1991) and more recently, citrus sudden death, a poorly characterized disease that causes high losses in the citrus groves (Bassanezi et al., 2003).Both problems have been extremely important and have stimulated rootstock changes.Unfortunately, among the alternative rootstocks, drought tolerance has been a limiting characteristic.Therefore, Rangpur lime is still the most preferred option for most growers in spite of the phytosanitary risks.
The development of new drought tolerant rootstocks in citrus, through classical plant breeding and/or innovative strategies that may help plants to survive longer periods under water deficient condition, is essential for the citrus industry to maintain its position in the competitive international market.Consequently, understanding the tolerance of plants to drought and exploiting its potential has been a priority in many recent studies.
Actually, strategies to investigate drought tolerance in citrus should include artificial systems to mimic water stress and the correspondent gene expression.To create such drought condition, a hydroponic system using polyethylene glycol (PEG) has been considered.PEG is a nonpenetrating osmotic agent that lowers water potential causing a water stress condition (Piro et al., 2003) and previous studies using this system to induce a stress condition have shown its efficacy (Villela et al., 1991;Furlani, 1997;Piro et al., 2003;Wu et al., 2005).
In addition, an alternative to the study of gene expression is to use a citrus transcriptome database such as the citrus EST project (CitEST) created by the 'Centro APTA Citros Sylvio Moreira'.CitEST bioinformatic analyses allow the study of in silico gene expression and represent a powerful tool for understanding mechanisms involved in water stress.The databases permit the study of genes under specific condition or different tissues and the expression patterns may reflect the transcript frequency under such condition.Several water stress responsive transcripts can be identified in silico, and hopefully they can provide valuable information for understanding the mechanisms of the drought tolerance in Rangpur lime.The present study aims to evaluate transcriptional responses of Rangpur lime rootstock to osmotic stress in hydroponic systems, identifying genes possibly involved in drought tolerance.

Stress induction
Two month-old Rangpur lime (C.limonia Osbeck) plantlets were transferred to pots (2 L) containing hydroponic system with Hoagland's nutritive solution (Hoagland and Arnon, 1950) and aeration.The experimental design consisted of four plants per pot and two replications.All plants were maintained in this system during 30 days.After this period, they were subjected to water-stress treatment induced by 372.524 g.L -1 of PEG6000 (Synth, São Paulo, Brazil) diluted in water to allow an osmotic potential of -1.5 MPa (Villela et al., 1991).Roots were collected periodically, after 30, 45, 180, and 360 min after stress induction, washed with sterile water, surface-dried and kept at -80 °C until mRNA extraction.Equal weights of roots from each of the four collection times were mixed in order to generate a composite sample, from which the RNA was extracted.Roots from Rangpur lime plantlets, maintained in Hoagland's solution without PEG treatment, were collected and processed as described above and used as a mock (nonstressed) control.

cDNA libraries
RNA was extracted from 1 g of a pool of roots collected in the above-mentioned periods (30, 45, 180, and 360 min after stress induction) by using Trizol reagent (Invitrogen, Carlsbad, CA).The RNA poly A was isolated from 1 mg of total RNA using the polyATract mRNA Isolation System (Promega, Madison, WI).Libraries were constructed using the SuperScript Plasmid System with Gateway Technology for cDNA Synthesis and Cloning (Invitrogen).One library of stressed roots (a mix of each collection time) and one library of the control Boscariol-Camargo et al.
(non-stressed roots) were constructed and used in the in silico analysis.

EST sequencing and data analysis
Sequencing was carried out using the Big Dye Terminator v.3.1 Kit as described by the manufacturer (Applied Biosystems, Foster City, CA).Products were separated by capillary electrophoresis using an ABI 3700 sequencer (Applied Biosystems, Foster City, CA).
An in silico 2x2 hybridization experiment was performed for the stressed x non-stressed libraries.The ESTs were grouped in clusters using CAP3, with default parameters.The composing reads for each assembled contig were counted by library, and the significance was given by using the methodology described in Audic and Claverie (1997) for two libraries.
The chosen threshold was a p-value ≤ 0.05.The resulting contigs were compared with public protein databases through BLASTX search (http://www.ncbi.nlm.nih.gov/BLAST/), and an automatic categorization of them was performed using the FunCat Functional Catalogue (Ruepp et al., 2004).In addition, a manual categorization was utilized, especially when the result of automatic categorization had contigs classified into the 98 (classification not yet clear cut) or 99 (unclassified protein) classes.

Results and Discussion
In silico analysis A total of 4,130 valid reads were sequenced (2,110 and 2,020 reads from the stressed and non-stressed roots libraries, respectively), resulting in 852 contigs.After in silico hybridization, 39 significant contigs had different frequencies in each library, of which 28 were up-regulated (higher frequency) and 11 down-regulated (lower frequency) in roots of Rangpur lime under stressed condition.Contigs representing differentially expressed transcripts were classified according to functional category based on similarities to Arabidopsis (FunCat 2.0; Ruepp et al., 2004; Figure 1) and their expression profiles are shown in Figure 2. Most of them belonged to the "unclassified protein" (FunCat category 99).Other up-regulated transcripts were classified in the energy (FunCat 02), protein synthesis (12), protein metabolism ( 14), cellular transport (20), stress response (32) and cellular sensing and response (34.11) categories.Two contigs, C507 and C429, corresponding respectively to Dehydrin RAB18 and HAT22, after manual annotation, were associated with genes related with the ABA-dependent (abscisic acid) response to water stress.Additional genes related with ABA pathway were not found in this analysis.
Of the 11 down-regulated contigs, only one (contig 313 = C313) was classified as involved in biogenesis of cellular components (category 42.26.03),while all of the oth-ers fell in the unclassified proteins or in a classification of not yet clear-cut categories.

Genes potentially associated with Rangpur lime tolerance to water stress
Transcripts identified in the in silico analysis of Rangpur lime roots under water stress had high sequence similarity with genes involved in well-known water stress responses in other species, like proline-related synthase (P5CS), sucrose synthase, membrane proteins (aquaporins), and dehydrin (Table 1).Other contigs revealed similarity with genes involved in responses to different abiotic stresses.However, yet other genes found to be differentially transcribed in this study are not normally associated with any kind of stress.The putative up-regulated genes found in the present analysis are discussed below within their potential role during water stress response.

Osmoprotectants and carbohydrate synthesis
According to previous studies, osmoprotectants such as proline, glycine betaine and sugars may help plants to overcome water deficit stress.Osmoprotectants were shown to function as osmolytes which protect cells from dehydration by turgor maintenance of roots and shoots in response to drought (Yamaguchi-Shinozaki et al., 2002).
Free proline (Pro) accumulation is common in many eubacteria, algae and higher plants in response to osmotic stresses (Schobert, 1977;Csonka and Hanson, 1991) and was first observed in wilted rye grass (Kemble and MacPherson, 1954).Under normal conditions, Pro is preferentially synthesized via ornitine (Orn), whereas it is made directly from glutamate (Glu) under stress conditions (Delauney et al., 1993).The first two steps of the proline biosynthesis from glutamate in Arabidopsis are catalyzed by Δ 1 -pyrroline-5-carboxylate synthetase (P5CS), a bifunctional enzyme that catalyzes the conversion of Glu to Δ 1 - pyrroline-5-carboxylate, which is then reduced to proline (Hu et al., 1992).
There are several studies showing biosynthesis of Pro from Glu under stress conditions (Stewart and Hanson, 1980;Delauney et al., 1993).In our study, P5CS was differentially transcribed (p-value = 0.0099) in Rangpur lime under osmotic stress induced by PEG (see Table 1).These data suggest rapid synthesis and accumulation of Pro as one of the potential mechanisms used from this rootstock for its already known tolerance against drought.
Osmoregulation under water stress can also involve sucrose synthase activation, the most important enzyme of the sucrose biosynthesis pathway.The allocation of sucrose in leaves, stems, roots and nodules of soybean during water stress has been well characterized as well as after plant recovery (Fellows et al., 1987).Concentration of sucrose was significantly greater in leaves, roots and nodules (up to 3 times) during the stress period.Upon rehydration, the elevated concentrations of sucrose in roots and nodules decreased within 1 day to values of control plants.In our experiments, Rangpur lime appears to respond to water stress similarly.A putative sucrose synthase was differentially transcribed (p-value = 0.0499), suggesting adjustments of the carbohydrate biosynthesis as an additional mechanism of response to water stress.
Within the carbohydrate biosynthesis pathway, the cytosolic glyceraldehyde-3-phosphate dehydrogenase (Gap3) was also differentially transcribed (p-value = 0.0382) in Rangpur lime under PEG induced stress.Indeed, as shown by Yang et al. (1993), heat shock stress (and some responses are very similar to water stress) increased expression of the GapC in Arabidopsis, which could be expected.Additionally, Yang and collaborators showed similar induction of GapC (up to 10-fold) in response to sucrose supplied for plants in vitro.This study has a different orientation, but there is a chance that GapC induction may be a response to both the stress itself and its consequent increase on the cellular sucrose concentration, which is to be expected from higher sucrose synthase expression.
Expression of cysteine proteinase has been studied in different stages of plant development (Beyene et al., 2006), in response to a variety of stresses and also in programmed cell death (Schaffer and Fischer, 1988;Koizumi et al., 1993;Harrak et al., 2001;Solomon et al., 1999).Distinct drought-inducible cysteine proteinases were induced by changes in the osmotic potential of plant cells elsewhere (Koizumi et al., 1993).In the present work, a droughtinducible cysteine (Table 1) which was probably involved in the maintenance of osmotic potential was found to be up-regulated in water deficit conditions.

Membrane transporters
Aquaporins are membrane proteins that are related to water translocation between cells.Members of this family make water-specific channels for ions and solutes (Bray, 1993).The channels facilitate the water flow under osmotic gradient.Aquaporins facilitate the uptake of water from the soil and mediate the regulation of root hydraulic conductivity [Lp(r)] in response to a large variety of environmental stresses (Boursiac et al., 2005).A study using antisense lines of PIP1 and PIP2 aquaporins indicates that PIPs play an important role in the recovery of Arabidopsis from water-deficient conditions (Martre et al., 2002).During water stress, these water-channel proteins accumulate in the tonoplasto (vacuole membrane) and the movement of water and solutes from vacuole to the cytoplasm is affected, changing the amount of water and the cytoplasmic osmotic potential (Maurel et al., 1993).Aquaporins are found mainly in cells involved in hydraulic flow, such as root epidermis, root tip tissues and tissues near the xylem cells in roots (Bohnert et al., 1995).The expression of aquaporins of the tonoplast and cell membrane has been correlated also with cell elongation (Yamaguchi-Shinozaki et al., 1992;Daniels et al., 1994), supporting the idea that cells of Rangpur lime sustain continuous growth of the root system.Under stress conditions, trees with longer and deeper roots are favored to reach underground water.The gene expression analysis of aquaporins in response to abiotic stress in Arabidopsis thaliana revealed the presence of 35 genes of aquaporins, and from those, 13 had identity with intrinsic proteins of the plasmatic membrane (PIP) (Jang et al., 2004).The subfamily PIP is divided into 2 subgroups: PIP1 and PIP2, according to the aminoacid sequence at the amino-and the carboxy-terminal.Jang et al. (2004) observed using Real Time PCR approach, that different genes of PIP are predominantly expressed either in roots or in aerial tissues (PIP1;3, PIP1;4, PIP2;1 and PIP2;5) under water stressed condition.The same work also reported that the expression of aquaporin can be either dependent or independent of ABA.Here, we found membrane proteins that showed homology with PIP2.1 class, but we did not evaluate whether or not it is dependent on ABA.

Protective factors of macromolecules
Cellular protectors refer to the group of genes that code for certain late embryogenesis abundant (LEA) proteins, more commonly called dehydrins (Close, 1996).The exact function of these proteins is still unclear but numerous studies have demonstrated the induction of dehydrins by drought stress (Maitra and Cushman, 1994;Giordani et al., 1999;Thompson and Corlett, 1995).Dehydrins are normally synthesized in maturating seeds during their desiccation, and in vegetative tissues of plants treated with ABA or exposed to environmental stress factors that result in cellular dehydration.Allagulova et al. (2003)  generally accepted classification of dehydrins is based on their structural features, such as the presence of conserved sequences, designated as Y-, S-, and K-segments; and the K-segment representing a highly conserved 15 amino acid motif (EKKGIMDKIKE KLPG), forming amphiphilic alpha-helix, has been found in all dehydrins.Therefore, finding a putative dehydrin in citrus root tissues, taken from a water deficient condition, was not a surprise.In field conditions, Rangpur lime rootstock has relatively good drought resistance, resulting in reasonable harvests even without irrigation during the 3 to 6 months dry winter period in São Paulo state.This resistance mechanism remains unclear, but it may have higher levels of protective factors than other rootstocks, such as dehydrin molecules.This contribution to drought resistance remains quantitatively unconfirmed.
Although not included in the present analysis, multiple putative heat shock proteins that mainly belonging to the HSP70 family were found very close to statistical significance, showing a possible role in the response to water stress in Rangpur lime.Perhaps under different experimental conditions, these proteins would present a more significative expression.Most of the HSPs work as chaperones that help to correct folding and/or to prevent protein denaturation.Since stress promotes denaturation and protein aggregation, the higher rate of HSP synthesis in Rangpur lime under stressed condition would help protect the osmotic stress proteins that occur after dehydration of the cell (Zhu et al., 1997).Some of the HSPs induced in Rangpur lime are named binding protein (BiP).Plant BiP expression has also been shown to respond to a variety of abiotic and biotic stress conditions (Jelitto-Van Dooren et al., 1999;Kumar et al., 2004).Transgenic tobacco (Nicotiana tabacum L. cv.Havana) plants constitutively expressing BiP genes were more resistant to water stress than control plants (Alvim et al., 2001).The expression of these genes in Rangpur lime is probably involved in drought resistance.
Differential expression of an EST corresponding to the chloroplast Cpn21 protein was also identified in stressinduced Rangpur lime roots (p-value = 0.0170).Cpn21 was characterized as a 21-kDa protein chaperonin 21 (Cpn21), which is a functional homolog of the chaperonin 10 (Cpn10) and consists of two Cpn 10-like domains fused together in tandem (Hirohashi et al., 1999).Also named Cpn20 (TAIR, The Arabidopsis Information Resource), Cpn21 seems to work independently of a chloroplast Cpn10 (both are homologues of the E. coli GroEs) on the regulation of specific Cpn60 subunit homologues (E. coli GroEl homologues) in plastids (Koumoto et al., 2001).In Arabidopsis, the expression of the chloroplast Cpn10 in mature plants is detected only in leaves and stems, while Cpn20 is expressed in leaves, stems and roots (Koumoto et al., 1999 and2001).These findings suggest Cpn20 (or Cpn21) as the specific chaperonin involved in regulation of Cpn60 subunits in roots, and so, the assembly of proteins inside chloroplasts.Moreover, differential expression of the Cpn21 in stressed Rangpur lime suggests possible changes in the regulation of protein assembly or needs for increased protein synthesis.

Signaling
Signal transduction pathways for expression of drought-inducible genes require different transcription factors and other components like protein kinases.Phosphorylations promoted by diverse families of kinases are normally switches of important biochemical processes.Kinases that can be turned on and off under processes related to water stress are known (Zhang et al., 2004), and therefore, a putative kinase was expected to be seen in our results.However, a valuable question considering its potential insertion and contribution in different signaling pathways and different rootstock.
Another kinase found in the present analysis was phosphoenolpyruvate carboxykinase, which is an enzyme used in gluconeogenesis to convert oxaloacetate and guanosine triphosphate into phosphoenolpyruvate, guanosine diphosphate and carbon dioxide.It seems to be somehow involved in the stress response in many plants (Saez-Vasquez et al., 1995).In alfalfa, a cDNA library from inoculated and non-inoculated roots identified a homolog of phosphoenolpyruvate carboxykinase expressed in those plants subjected to nematode infection (Potenza et al., 2001).In sorghum, phosphoenolpyruvate carboxylase-kinase (PEPCase-k) activity is highly enhanced by salt stress in leaves (García-Mauriño et al., 2003).Here, we observed a higher expression of this enzyme in roots under water stress, but the exact role during this process needs to be further investigated.
A third transcription factor, a homeobox-leucine zipper homolog to the Arabidopsis HAT22 gene, was differentially expressed in the Rangpur lime stressed roots.A type of homeobox-leucine zipper gene has been identified in Arabidopsis thaliana (Schena and Davis, 1994).The authors suggested that these genes mediate specific regulatory events.In another study, Deng et al. (2002) reported five novel dehydration-responsive homeodomain leucine zipper genes (HDZips) from the resurrection plant Craterostigma plantagineum.All families of these genes were modulated in their expression in response to dehydration in leaves and roots, supporting the role of HDZips in related regulatory pathways in C. plantagineum, which lead to desiccation tolerance.In a recent work, Dezar et al. (2005) transformed Arabidopsis plants with a Hahb-4 homeobox-leucine zipper transcription factor from Helianthus annuus (sunflower).Transgenic plants were more tolerant to water stress conditions, showing improved plant development, healthier appearance and higher survival rates than wild-type plants.Plants under either normal or drought conditions produce approximately the same seed weight per plant as wild-type plants under normal growth conditions.Since Hahb-4 is a transcription factor, it was proposed that it may participate in the regulation of the expression of genes involved in developmental responses of plants to desiccation.Thus, the expression of a HAT22 homologue in our experiment probably indicates a regulatory role in stress response in citrus.

Other proteins activated during stress response
Several other deduced proteins could also be related to stress response and among them, an orthologue of Baxinhibitor was found to be significantly transcribed in stressed, but not in non-stressed, Rangpur lime root cDNA libraries.The Bax-inhibitor 1 (BI-1) is a membrane protein that plays an important role in preventing PCD (programmed cell death) related to Bax in mammals (Hückelhoven et al., 2003).In plants, it has been shown that expression of a Bax-inhibitor homolog is rapidly up-regulated during wounding or pathogen challenge, suggesting its role in responses to biotic and abiotic stresses (Sanchez et al., 2000;Hückelhoven et al., 2003;Matsumura et al., 2003;Kamauchi et al., 2005).No Bax-inhibitor like sequences had been reported in Citrus sp.prior to this work.Interestingly, all of the six putative Bax-inhibitor sequences tags found in our citrus EST libraries came from the drought stressed Rangpur lime root library.Even though no clear correlation has not been shown between Bax-inhibitor like protein and drought stress in the literature so far, it is reasonable to presume that the Bax protein might play a role in such a situation as well, and could possibly be overexpressed in order to avoid cell collapse and death.
Earlier reports have shown that water deficit causes significant decreases in protein synthesis (Hsaio, 1970;Dhindsa and Cleland, 1975;Rhodes and Matsuda, 1976;Mason et al., 1988;Bensen et al., 1988;Valluri et al., 1989).Interestingly, in our experiment we observed an increase in the levels of two proteins that constitute the 40S subunit of the ribosomes, S28 e S9.De novo protein synthesis is necessary for cold response.Three genes encoding different ribosomal proteins were isolated from soybean and are induced by low-temperature stress (Kim et al., 2004).This could have happened to enhance the translational process or help the proper ribosome assembly and functioning under this condition.We can also speculate that under water stress, Rangpur lime shows a higher level of ribosomal proteins in order to maintain the physiological processes that lead to tolerance to drought.On the other hand, higher expression of the ubiquitination related genes, encoding ubiquitin-conjugating enzymes, was also observed, suggesting that stress responses were taking place, leading to increased degradation of proteins.The ubiquitin-dependent proteolytic pathway regulates many essential processes in eukaryotes by selective protein breakdown (Ciechanover and Schwartz, 1998) and the ubiquitin conjugating enzyme is an E2 protein that synthesizes multiubiquitin chains from free ubiquitin (Chen and Pickart, 1990;van Nocker and Vierstra, 1991;van Nocker et al., 1996).
Another protein, a germin-like protein (C815), and classified in the stress response categories -as well as in the DNA processing/ DNA synthesis and replication -was found to be significantly up-regulated in stressed Rangpur lime root library.Interestingly, these proteins have been involved in an array of functions, most of them related to plant development and cell wall biogenesis, but also to C-compound and carbohydrate catabolism, osmotic regulation, photoperiodic oscillation, and defense (Çaliskan, 2000;Patnaik and Khurana, 2001).
The importance of germin-like proteins to salt stress tolerance in both monocot and dicotyledoneous plants has been known for years (Hurkman et al., 1991;Michalowski and Bohnert, 1992).However, their main and most-studied role is in early plant development (Lane et al., 1993).Wissel et al. (2003) demonstrated that regulation of a germin-like gene of Aspen is strongly related to the tree developmental stage, with overexpression only in young tissues and during the growth phase of the tree, regardless of the environmental conditions.

Up-regulated proteins not directly related to stress response
A putative protein found only in the stressed Rangpur lime library presented similarity with members of the MO25 (early mouse development) protein family, involved in cell differentiation.Even though there is no information about the role of such proteins in plants, hypothetical MO25-like genes have been found in Arabidopsis thaliana (www.arabidopsis.org).
In mammals, it has been shown that the MO25a/b protein interacts with STRADa/b and increases their ability to bind, activate and localize LKB1 (a 50 kDa serine/threonine kinase involved in suppression of tumor) in the cytoplasm (Boudeau et al., 2003).The complexes between the LKB1, STRADa/b and MO25 a/b are upstream kinases in the AMP-activated protein kinase (AMPK) cascade.AMPK is a sensor of cellular energy charge that acts as a 'metabolic master switch' inhibiting cell proliferation and protecting cells against nutritional or environmental stresses (Hardie and Hawley, 2001;Hardie et al., 2003;Hawley et al., 2003;Baas et al., 2004).
In plants, kinases have been shown to be activated by various abiotic stresses (Ikeda et al., 1999;Kelner et al., 2004;Thelander et al., 2004).It remains unknown and to be addressed in further studies whether the MO25 homologue has any implication in such activation or in root cell differentiation in outgrowing the stressful condition.
Another putative up-regulated protein in water stressed roots was a nodulin.Nodulins modulate metabolizing carbon and nitrogen, and transport compounds within the nodule.Kuster et al. (1996) studied different modular structures and compared the timing of gene expression.Cit-912 Drought stress in Rangpur lime rus does not produce nodules, such as beans, but it does have a root system populated by diverse organisms.Therefore, it is possible that water stress affects root-microorganism interactions in the citrus ryzosphere, and thus, affects gene expression of related mediators.
There are different forms of stress conditions, but for a regular cell expansion and root hair tip growth there is a highly dynamic demand for polar actin cytoskeleton formation.Ketelaar et al. (2004) studied such conditions.Under drought conditions, root development can be impaired to avoid dehydration, or initially stimulated to explore yet untouchable soil areas.Since gene families encode actin, in higher eukaryotes, it is possible that this putative actin gene is responsive to drought conditions.
In addition, other genes had a much higher expression induced by water stress, such as glycogenin glucosyltransferase (C613) and PGPD14 protein (C459).In rice, the accumulation of a glycogenin glucosyltransferase (OsGGT) seems to be involved in response to stress, including environmental stresses, such as drought and salt treatment (Qi et al., 2005).PGPD14 protein (pollen germination related protein) does not appear to be involved directly with stress because of lack of further reports.
Other unclassified proteins that also displayed differentially transcribed patterns under stressed condition can be targets for future investigation and perhaps they can help to find new answers about the mechanism involved in Rangpur lime drought resistance mechanism not yet described.

Concluding Remarks
Rangpur lime is a rootstock considered highly tolerant to drought stress.However, its mechanism of tolerance is not fully understood.There is little biological evidence suggesting that this tolerance is mostly related to a capacity to physically grow more and longer roots rather than only a biochemical ability to withstand water stress.Our EST data analysis showed that several of the proteins often associated with drought stress in other plant species, such as proline-related synthase, aquaporins, dehydrins, some sugars and others associated with different types of abiotic stresses, had transcripts that were significantly induced in the Rangpur lime library constructed under such condition.In addition, other proteins not normally related to abiotic stresses were also induced, but their roles remain unclear.
The Rangpur lime tolerance mechanism could be based on the activation of proteins involved in osmoprotection together with other responses that adjust cell turgor and an increase of expression of genes related to sustaining tissue development, which might keep root growth active.The responses of genes activated by abscisic acid (ABA) were infrequently observed.Only one transcription factor, a homologue of the Arabidopsis HAT22 gene known to be activated by ABA, was detected in our analysis.This gene might have a potential role in controlling the response to stress, similar to that occurring in other plants.Additionally, most of the up-regulated transcripts found in this study are related to protein synthesis and some to other functions.Future studies should test this hypothesis in vivo and validate the specificity and expression levels of the genes found here, considering related conditions.Thus, it is possible that the genetic response of Rangpur lime to drought leads to an efficient capacity of the root system to absorb water, potentially in a better way than normally happens to other citrus species when used as rootstocks.Those related genes are potentially interesting subjects for further studies, including development of gene specific markers for selection within hybrid populations generated by Rangpur lime interspecies crossings and genetic engineering of other citrus rootstocks in order to increase their resistance to water stress.

908
Figure 1 -MIPS functional categories for the 39 putative Rangpur lime water stress-related EST contigs.

Figure 2 -
Figure 2 -Expression profiles of putative Rangpur lime water stressrelated EST contigs.Scale represents relative abundance of reads in:1-non-stressed root library and 2-roots under osmotic stress.Each profile is followed by the contig ID, the number of the MIPS functional category and the description of the functional category.

Table 1 -
described that they are considered as stress proteins involved in the formation of plant protective reactions against dehydration.The Up-regulated genes under water stress in Rangpur lime roots.