The sugarcane signal transduction (SUCAST) catalogue: prospecting signal transduction in sugarcane

EST sequencing has enabled the discovery of many new genes in a vast array of organisms, and the utility of this approach to the scientific community is greatly increased by the establishment of fully annotated databases. The present study aimed to identify sugarcane ESTs sequenced in the sugarcane expressed sequence tag (SUCEST) project (http://sucest.lad.ic.unicamp.br) that corresponded to signal transduction components. We also produced a sugarcane signal transduction (SUCAST) catalogue (http://sucest.lad.ic.unicamp.br/private/mining-reports/QG/QG-mining.htm) that covered the main categories and pathways. Expressed sequence tags (ESTs) encoding enzymes for hormone (gibberellins, ethylene, auxins, abscisic acid and jasmonic acid) biosynthetic pathways were found and tissue specificity was inferred from their relative frequency of occurrence in the different libraries. Whenever possible, transducers of hormones and plant peptide signaling were catalogued to the respective pathway. Over 100 receptors were found in sugarcane, which contains a large family of Ser/Thr kinase receptors and also photoreceptors, histidine kinase receptors and their response regulators. G-protein and small GTPases were analyzed and compared to known members of these families found in mammalian and plant systems. Major kinase and phosphatase pathways were mapped, with special attention being given to the MAP kinase and the inositol pathway, both of which are well known in plants.


INTRODUCTION
The analysis of the complete Arabidopsis thaliana genome sequence (The Arabidopsis Genome Initiative, 2000) has revealed the striking conservation of genetic mechanisms required for developmental and physiological processes while pointing to the unique properties of individual plant systems. Comparison of the Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophila melanogaster and Arabidopsis thaliana genomes indicates that many signal transduction modules are conserved and even though the signals and end results of the pathways may be different many components are shared by these very diverse organisms (McCarty and Chory, 2000;The Arabidopsis Genome Initiative, 2000). The differences in the structure of plant hormones in comparison with animal hormones has led to the assumption that plants probably evolved an array of regulatory molecules that were totally different from those of animals, although the identification of signal transduction components in many plant systems has contradicted this view. In the work presented in this paper we describe putative signal transduction components of sugarcane revealed by a systematic mining of the whole EST data set generated in the SUCEST project (http://sucest.lad.ic.unicamp.br). The main categories were catalogued as well as the number of their related ESTs clus-ters, paving the way for the functional analysis of signal transduction pathways in sugarcane.

METHODS
The SUCEST database stores over 250,000 quality controlled 5' and 3' ESTs reads from 37 cDNA libraries prepared from several sugarcane tissues (calli, root, stalk, leaves, flowers, developing seed, etc.) grown under different environmental conditions (Vettore et al., 2001). The ESTs were organized by sequence similarity into clusters and singletons and automatically annotated according to their similarities to sequences in the National Center for Biotechnological Information (NCBI) non redundant protein database . The basic local alignment search tool (BLAST) was used to perform a bi-directional search against the SUCEST clusters consensi generated using the CAP3 algorithm  using query sequences from known signal transduction components or domains and the TBLASTN algorithm (Altschul et al., 1997). Hits with E-values in most cases lower than 10 -10 , were manually inspected after previous automatic BLASTX annotation which had been performed by the SUCEST bioinformatics team. Positive matches were aligned using the multiple sequence alignment CLUSTAL method (Jeanmougin et al., 1998) and their amino acid translated sequences were compared with protein families and domains in the PROSITE (Hofmann et al., 1999) and Protein Family (PFAM) (Bateman et al., 2000) databases. If 100% of the composition of their ESTs reads were derived from the same set of cDNA libraries the clusters were considered tissue-specific, whereas if only 80% of the reads belonged to a specific library they were considered tissue-enriched. The sequences of clusters and their identities are available at the SUCEST web site at http://sucest.lad.ic.unicamp.brand and at the SUCAST web site at http://sucest.lad.ic.unicamp.br/private/mining-reports/QG/QG-mining.htm). The main protein categories so far catalogued and the number of related clusters found are shown in Table I, supplementary Information is available at http://sucest.lad.ic.unicamp.br/private/mining-reports/QG/ QG-mining.htm.

Plant hormones and signaling peptides
Plant hormones and peptides transduce signals such as temperature, light, water, nutrient and microbe-plant interactions which induce cellular responses locally and/or throughout the plant. We have investigated the existence of the major sugarcane routes for the synthesis of ethylene, abscisic acid, auxins, gibberellins and jasmonates. The enzymes for which clusters have been found are shown in Figure 1. Structurally, the simplest plant hormone is the gas ethylene, which has numerous roles including plant development, sex determination, fruit ripening, flower and leaf senescence and defense (Johnson and Ecker, 1998). Ethylene is synthesized from S-adenosyl methionine (AdoMet) to make 1-aminocyclopropane-1-carboxylic acid (ACC) which is converted to ethylene, CO 2 , and HCN by ACC oxidase. Four putative ACC synthases were found in sugarcane and 6 ESTs clusters were identified as being similar to ACC oxidases. This was expected since these enzymes are encoded by multigene families in several plant species. ACC synthase genes are regulated by developmental signals, hormones and environmental stimuli. Control of ethylene synthesis is largely attributed to ACC synthase but altered expression patterns of ACC oxidase (ACO) genes in senescence, fruit ripening and wounding suggests that the latter contribute to regulation of ethylene production as well (Johnson and Ecker, 1998). The 4 clusters corresponding to ACC synthase ESTs are made up of 14 reads of which 9 come from root or root zone transition libraries, indicating that there are higher levels of this enzyme in roots. Ethylene has been implicated in the production of root hairs but to our knowledge increased levels of ACC synthase in roots has not yet been detected.
Indole-3-acetic acid (IAA) is the major naturally occurring auxin and has been implicated in the regulation of growth and development of many plant species. Two major routes have been described for IAA biosynthesis, the first being a tryptophan-dependent pathway where Trp is converted to indole-3-acetaldoxime (IAOx) and then via indole-3-acetonitrile (IAN) to IAA, and a second Trp-independent pathway which has not, as yet, been very well characterized (Hull et al., 1999). In the tryptophan-dependent pathway the first step is the conversion of Trp to IAOx catalyzed by cytochrome P450 (CYP79B2 and CYP79B3) leading to the conversion of IAN to IAA by nitrilases (NIT1, NIT2, NIT3 and NIT4). The nitrilase related clusters (a total of seven) were also present at higher levels in root tissue, and one of them was specific to libraries from sugarcane infected with Glucoacetobacter diazotrophicus. A well established effect of auxin stimulation is the induction of ACC synthase (an early-auxin-response gene) leading to ethylene biosynthesis. The presence of many reads of ACC synthase in roots together with the prevalence of ESTs related to auxin synthesis in this organ leads us to think that the pathway leading from auxin to ethylene production may occur predominantly in sugarcane roots. The auxin receptor is unknown but auxin binding to the plasma membrane elicits the activation of selective protein degradation by the ubiquitin-proteosome pathway. Ubiquitin conjugation requires the sequential activity of three protein complexes, ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2) and ubiquitin-protein ligase (E3) (del Pozo and Estelle, 2000). Putative candidates for transducers of the auxin response have been found, including E1, E2 and E3 complexes and their candidate targets, the IAA domain proteins. IAA homology domains are found in a large family of auxin induced proteins, with possibly over 30 members in sugarcane.
Gibberellins (GA) play key roles in plant growth and development, mediating light-stimulated seed germination through phytochromes (Kamiya and Garcia-Martinez, 1999), mobilization of reserves by aleurone cells (Lovegrove and Hooley, 2000), leaf expansion, stem elongation, flower initiation, and flower and fruit development (Sun, 2000). In the major pathway of gibberellin synthesis trans-geranyl diphosphate (GGGP) is converted to ent-copalyl diphosphate (CPP) by ent-copalyl diphosphate synthase (CPS) and then to ent-kaurene by ent-kaurene synthase (KS) (Hedden and Phillips, 2000), these two enzymes have few clusters (CPS has one cluster and KS two) which occur most often in flowers. Sequential oxidations catalyzed by the enzymes ent-kaurene 19-oxidase, ent-kaurenoic acid 7-B-hydroxilase and GA12-aldehyde synthase produce GA12-aldehyde. Reads with similarity to the ent-kaurene 19-oxidase GA3 from Arabidopsis were found in several sugarcane tissues (flowers, internode, apical meristem, roots and stem bark) indicating widespread distribution in sugarcane, much as was seen in Arabidopsis (Helliwell et al., 1998). Mono-oxygenases catalyze the conversion of GA12-aldehyde to GA12 and GA53, which are substrates for the biosynthesis of gibberellins. The bio-active gibberellins GA1, GA4, GA3 and GA7 are formed by the 2-oxoglutarate-dependent dioxygenases, GA 20-oxidase and GA 3-β-hydroxylase. The dioxygenase genes are major targets for light regulation of GA metabo- lism (Kamyia and Garcia-Martinez, 1999). One putative cluster with similarity to both GA 20-oxidase and GA 3-β-hydroxylase was found in sugarcane, with no apparent tissue specificity observed. The dioxygenase GA 2-oxidase, which deactivates gibberellins also was detected in sugarcane (2 clusters). Gibberellins are detected at the plasma membrane but no receptors have been cloned. There is evidence of increased cytosolic calcium and induction of calmodulin levels following stimulation of aleurone cells by gibberellins and that G proteins (see below) transduce gibberellin signals (Lovegrove and Hooley, 2000).
Plant peptides are involved in various signaling pathways, including cellular communication in meristems mediated by CLAVATA3 and phytosulfokines (PSKs), Rhizobium nodule formation (Enod40) and defense responses (systemin) (Bisseling, 1999). The CLAVATA3 peptide, together with its receptor CLAVATA1, coordinates the growth and differentiation of meristem cells (Fletcher et al., 1999). The CLAVATA1 gene encodes a putative receptor kinase that controls the size of shoots and floral meristems in Arabidopsis (Clark et al., 1997). Although we did not find any CLAVATA3-related sugarcane ESTs, we did find a large number of clusters for receptor serine/threonine kinases similar to CLAVATA1 (see below). Phytosulfokines are peptides, made up of 4-5 amino acids, which stimulate cell proliferation (Matsubayashi et al., 1999). A cDNA clone encoding a PSK precursor has been isolated from rice, but no PSK-related ESTs clusters were detected in the SUCEST database. Likewise no signif-icantly similar clusters were found for systemin, a defense polypeptide released by some plants upon damage. Enod40 is one of the first plant genes activated during Rhizobium induced nodule formation. It is induced in the root pericycle, and the only sugarcane ESTs cluster found that had any similarity to this gene was derived from the leaf/root-zone transition library. Receptors for PSK and Enod40 have not yet been identified in any system. Abscisic acid (ABA) is a cleavage product of carotenoids which is produced by plants in response to stress (Cutler and Krochko, 1999). Water stress induces accumulation of ABA, presumably from the cleavage of 9-cis-epoxycarotenoids (9-cis-violaxanthin and 9-cis-neoxanthin) to xanthoxin by the enzyme 9-cis-epoxycarotenoid dioxygenase. The 9-cis-epoxycarotenoids are produced from zeaxanthin by the enzyme zeaxanthin epoxidase. ABA seems to counteract the increase of calmodulin and calcium levels in aleurone signaling by gibberellins (Lovegrove and Hooley, 2000) but ABA receptors have not been defined yet. No significant enrichment of ABA biosynthesis clusters were found in the sugarcane cDNA libraries.
Jasmonate is a 12-carbon fatty acid-derivative (synthesized from linoleic acid via the octadecanoid pathway) which has a role in the defense against different types of microbial pathogens (Reymond and Farmer, 1998). Most of the enzymes involved in the synthesis of jasmonate appear to be expressed in sugarcane and some ESTs clusters for the 12-oxo-phytodienoate (PDA) reductase were found in libraries derived from plantlets infected with 28 Souza et al. Glucoacetobacter or Herbaspirillum. This agrees with the proposed role for jasmonate in the plant response to biotic and abiotic stresses as suggested by Reymond et al. (2000).

Receptors
Analysis of the Arabidopsis genome sequence has indicated that the largest and most diverse family of receptors in plants is the receptor serine/threonine kinase family, which has over 300 genes (The Arabidopsis Genome Initiative, 2000). We found 93 clusters similar to receptor kinases in sugarcane, of which 6 contain leucine-rich-repeat (LRR) domains, including one cluster for a receptor serine/threonine kinase containing a lectin domain and another cluster for a receptor serine/threonine kinase containing a tetratricopeptide repeat (TPR) domain.
It appears that plants have evolved different signaling pathways compared to mammals and other metazoans since no receptor tyrosine kinase or evidence of the Ras pathway has been found in plants ( McCarty and Chory, 2000). It is not surprising then that we did not find any EST clusters related to tyrosine kinases or Ras (see below), and only one cluster related to the G-protein coupled receptor family.
The two-component histidine kinase pathway transduces ethylene and cytokinin signaling (Urao et al., 2000). Our analysis revealed 6 histidine kinase clusters similar to the ethylene receptor and 5 similar to the cytokinin receptor. Moreover, we catalogued 3 clusters corresponding to phosphotransfer intermediate proteins and 13 clusters related to response regulators.
In Arabidopsis there are two cryptochromes and five phytochrome photoreceptors. They overlap in function and transduce the blue and far-red light which regulates gibberellin synthesis (Kamiya and Garcia-Martinez, 1999), inhibition of hypocotyl elongation, anthocyanin production and the sensitivity of flowering to the photoperiod (Cashmore et al., 1999). The phytochromes are an interesting family of proteins with light-dependent serine/threonine-specific kinase activity. It has been proposed that these photoreceptors have evolved from ancestral histidine kinases (McCarty and Chory, 2000). We have found 4 ESTs clusters similar to phyA-D and 6 clusters similar to the blue light receptors.

G-protein and small GTPases
The current sugarcane EST data set appears to contain clusters similar to the α, β and γ subunits of the G-protein, three clusters for the α-subunit, twelve clusters for the β-subunit and one cluster for the γ-subunit. The sequencing of the Arabidopsis genome has confirmed the existence of a single gene for each of the G-protein α and β-subunits and recent studies have identified the only γ-subunit (The Arabidopsis Genome Initiative, 2000). Our findings con-trast with previous observations that in plants G-protein subunits are not members of large gene families, indicating an increased number of these transducers in sugarcane.
In animal cells the Ras superfamily of small guanosine triphosphatases (GTPases) is categorized into the Ras, Rab, Arf, Ran and Rho families according to their guanosine triphosphate (GTP) binding domain, effector and insertion sequences. Plants do not appear to contain members of the Ras family (McCarty and Chory, 2000). Accordingly we did not find any cluster related to this group in sugarcane. The most predominant small GTPase family found in the SUCEST database was the Rab family (Table II) with 28 EST clusters mapping to this family. The most predominant member of the Rho family in plants are the Rac (or Rops) GTPases (Valster et al., 2000) but only four clusters, for RacA, RacB and RacC from Zea mays, and no bona fide Rho (i.e. one with a characteristic LKCD GTP-binding domain) were found in the SUCEST database. The second largest group of small GTPases which we found in sugarcane belonged to the Arf group with 13 EST clusters occurring in the SUCEST database. We also found 7 Ran family clusters. In general, the clusters were most similar to their Oriza and Zea counterparts, with the Rab family being the most diverse family with additional members similar to those found in Lycopersicum and Arabidopsis. The GTP-binding proteins are regulated by GTPase activating proteins (GAPs), GDP dissociation inhibitors (GDIs) and GTP-exchange factors (GEFs). We found several GAPs and GDIs in sugarcane but no GEFs of the Dbl-type, indicating that only regulation by inhibition can be inferred for these proteins so far.

Second messengers
Inositol signaling in plants has been shown to play a role in cell growth and elongation, mediating membrane trafficking and calcium levels (Stevenson et al., 2000). Production of inositol triphosphate (Ins(1,4,5)P3)) is a common response to salt and hyper-osmotic stress in plants as well as to the effects of gravity (gravi-stimulation). A search for the enzymes involved in inositol metabolism in sugarcane indicated the pathways shown in Figure 2. Several clusters related to these enzymes showed tissue specificity or were enriched in the root, root-zone transition, flower or infected plant libraries. No inositol triphosphate receptor was found in sugarcane, nor, to our knowledge, in any other plants, suggesting that the plant and animal receptors might not share much sequence similarity. Cyclic ADP ribose (cADPR) has also been shown to trigger the release of calcium from the intracellular compartments of plants, but we did not find any EST clusters similar to ADP-ribosyl cyclases or cyclic ADP-ribose hydrolases in the SUCEST database.
In our analyses, we found no EST clusters related to proteins containing a guanylate cyclase domain or to cyclic guanidine monophosphate (cGMP) dependent protein kinases. We did find one putative cluster with sequence similarity to the catalytic subunit of the cyclic adenosine monophosphate (cAMP) dependent protein kinase (PKA) but no clusters for the regulatory subunit (R subunit) or adenylyl cyclase were found. These cAMP and cGMP regulatory components seem to be absent in sugarcane but cyclases and R subunits typical of animal systems do not seem to be present in other plant species as well.
An early response to many physiological stimuli (light, cold, heat, movement, hypoxia, drought, phytohormones and pathogens) include an elevation in the level of free calcium in the cytosol. Plasma membrane and vacuolar Ca 2+ channels have been implicated in multiple signaling pathways involving calcium-binding proteins and calcium dependent protein kinases (CDPKs). We identified a few sugarcane clusters corresponding to calcium channels and calmodulin regulated channels, some of which were enriched in or specific to flower libraries. Clusters for the calcium binding proteins calmodulin, calreticulin and calnexin were also found. Sugarcane also express genes for the new family of calcium sensors called calcineurin B-like proteins (CLB) and we detected 6 EST clusters related to these proteins. It has been shown that Arabidopsis express distinct members of a CBL gene family and one of them (AtCBL-1) is strongly induced by stress signals (Kudla et al., 1999). In Arabidopsis, some members of the CBL family appear to interact with a subset of a family of protein serine/threonine kinases called CLB interacting protein kinases (CIPKs) while other CBLs are thought to target distinct proteins (Kim et al., 2000). Two CIPK-related clusters were found in the SUCEST database.

Kinases and phosphatases
A strong and common theme of signaling networks in eukaryotic organisms is regulation by reversible protein phosphorylation. The importance of protein kinases and protein phosphatases in the control of most of the pathways involved in growth and development as well as in responses to changes in the environment is reflected in the large number of genes which encode these enzymes in eukaryotes (Hunter, 1995). It has been reported that the Arabidopsis thaliana genome encodes nearly 1000 genes belonging to the protein kinase superfamily and almost 300 genes encoding protein phosphatases (The Arabidopsis Genome Initiative, 2000). As expected, a large number (so far 80 clusters have been annotated) of sugarcane EST clusters related to protein kinases were found in the SUCEST database, including receptor-like protein kinases and two-component histidine kinases (Table I). However, in accordance with previous observations for other plant species (McCarty and Chory, 2000), sugarcane appears to lack typical receptor tyrosine kinases.
The MAP kinase (MAPK) cascade is the most studied phosphorylation pathway in plants and appears to transduce a vast array of signals (Figure 3). The input receptors include the histidine kinases and receptor-like serine/ threonine kinases. The MAP kinases are targets of ethylene through the two-component system leading to changes in gene expression mediated by EIN3 (Johnson and Ecker, 1998), and a MAPKKK has been shown to be associated with a Ras-like small GTPase Rop in the signal transduction of the CLAVATA1 receptor-like kinase in meristem signaling (Valster et al., 2000). As shown in Figure 3, other signals transduced by the MAP kinase cascade include plant hormones and many environmental signals, however their complete pathways are still undefined.
Protein phosphatases that de-phosphorylate phosphoserine/threonine residues are encoded by the PPP and PPM gene families, which have distinct amino acid sequences and crystal structures. Members of the PPP family usually exist, in vivo, as multimeric holoenzymes where a limited number of catalytic subunits are largely controlled by the nature of the associated regulatory subunit (Barford, 1996;Cohen, 1997). We were able to assign 30 clusters to catalytic subunits of the PPP family, covering its major members with the exception of calcineurin (PP2B). To our knowledge the PP2B catalytic subunit has not yet been detected in other plants, despite the existence of the calcineurin regulatory subunit-like proteins known as CLBs (see above).
It may be possible that the number of genes encoding PPP catalytic subunits in sugarcane may actually be larger, and that their ESTs are clustered together due to the striking sequence conservation among members of this family. In addition, we found at least 19 EST clusters encoding regulatory subunits for members of the PPP family (including B-type PP2A regulatory proteins) which serves to remind us of the highly conserved regulation strategy of this family of enzymes.
PP2C are monomeric magnesium-dependent enzymes classified in the PPM family, several members of the PP2C group being related to the ABI1/ABI2 Arabidopsis enzymes implicated in the negative regulation of the abscisic acid pathway (Merlot et al., 2000). In the SUCEST database, we found 11 sugarcane EST clusters with sequences very similar to PP2C. One of them (from the apical meristem and flower libraries) was significantly similar to a kinase associated protein phosphatase (KAPP), an enzyme related to the PP2C family that has been found to be part of a signaling complex involving the CLAVATA1 receptor and Rho GTPaserelated protein (Trotochaud et al., 1999). It is tempting to speculate that a similar complex might exist in sugarcane.
No tyrosine specific kinase has been found in higher plants or yeast. However it has been demonstrated (Zhang and Klessig, 1997) that plant MAPK activation follows its phosphorylation by MAPK kinases, a major group of dualspecificity kinases that phosphorylate MAPKs at both threonine and tyrosine, similar to what is seen in mammalian and yeast cells. Moreover it has been demonstrated that plants express tyrosine phosphatases (PTP) and that Arabidopsis PTP1 is encoded by a stress-responsive gene (Fordham-Skelton et al., 1999;Xu et al., 1998). We found three sugarcane EST clusters that appear to encode PTP-like enzymes and nine clusters for a subfamily of PTPs known as dual specificity phosphatases (DSPP) which have also been implicated in the negative regulation of an Arabidopsis MAPK .

Signal transduction of plant-microbe interactions
Plants are constantly attacked by a wide variety of microorganisms and have developed an array of responses to either survive pathogen attacks or, in the case of endophytes, to profit from these interactions. An effective response depends on sensing and transducing a particular microorganism's presence, leading to a specialized gene expression response that, for example, confers disease resistance on the plant. A number of resistance genes are induced by salicylic acid, ethylene and jasmonic acid when plants are exposed to pathogens. Jasmonate, for instance, has been shown to be essential for the defense of tomato against hornworm larvae and Arabidopsis against flies and fungal attacks (Reymond and Farmer, 1998) inducing expression of defensins.
To begin to access the signaling mechanisms that may be involved in sugarcane-microbial interactions, we performed a search for signal transduction components specifically expressed when sugarcane plantlets were infected with Herbaspirillum rubrisubalbicans (the HR cDNA library) and Glucoacetobacter diazotroficans (the AD cDNA library). Both bacteria are diazotrophic endophytes that present a unique type of association with sugarcane, H. rubrisubalbicans appearing to cause mottled stripe disease in susceptible sugarcane cultivars (Reinhold-Hurek and Hurek, 1998). In our libraries, though, the H. rubrisubalbicans strain used for infection was non-pathogenic for the host sugarcane variety used (Vettore et al., 2001).
From the 650 signal transduction-related ESTs clusters we have so far catalogued, 23 were specifically found only in AD and HR libraries. The analysis of these clusters revealed that an enzyme involved in the jasmonate synthesis pathway, 12-oxo-phytodienoate reductase, was specific for these libraries, indicating that jasmonate synthesis is probably induced under these conditions. Ethylene signaling components were also represented in the AD and HR specific clusters. Two histidine kinase receptor clusters similar to the ethylene receptor and a response regulator were found specifically in the infected libraries. Six receptor serine/threonine kinases were specific to the AD or HR libraries, one cluster being similar to a wall-associated kinase from Arabidopsis induced when these plants are exposed to pathogens and postulated to protect Arabidopsis against the attacks by microbial pathogens (He et al., 1998). We also detected one cluster encoding an authentic type 1 serine/threonine phosphatase catalytic subunit (PP1c) which appeared to be AD and HR specific.
Inositol signaling is prominent in clusters specific to the AD and HR libraries, where we found a 1-phosphatidylinositol-4-phosphate kinase, a inositol(myo)-1(or 4)-monophosphatase and two phospholipases C, but calcium signaling is probably also involved, as indicated by the presence of one calnexin and one calreticulin.
It appears that G-protein coupled sensing is also involved in the transduction of plant-microbe interaction signaling because a G-protein β-subunit was specifically induced in the sugarcane tissues experimentally infected with H. rubrisubalbicans and A. diazotroficans.
This preliminary attempt to define the signal transduction components induced when diazotrophic endophytes associate with sugarcane suggests that the plant actively participates in this process instead of behaving as a silent host for the growth of these bacteria. Even so, the inferences we have drawn are based on comparing several different non-infected tissues with infected plantlets, and a wider gene expression analysis of sugarcane infected with these endophytes will be necessary to prove these assumptions.

Future perspectives
So far the current sugarcane EST collection (SUCEST) has enabled the identification of over 650 clusters for signal transduction components, the analysis of which indicates that most of the signaling modules typical of plants are conserved. The mining of the SUCEST database for signal transduction components is an on-going effort, and if the same number of components is found in sugarcane as has been found in Arabidopsis, we may expect at least 5 thousand genes to be included in this category. Many multi-gene families have also been detected, the receptor serine/threonine kinases being the most striking example with almost 100 members. The work reported in this paper summarizes the catalogued clusters and main signal transduction pathways that can be found at the SUCAST Web site (http://sucest.lad.ic.unicamp.br/private/miningreports/QG/QG-mining.htm). It is hoped that this resource will aid future functional analysis of the sugarcane genome.