ABA (abscisic acid), an important growth regulator during plant growth and development, has been demonstrated to be involved in many plant developmental processes, especially seed development including maturation and germination, and plant response to environmental factors. It has shown that ABA was accumulated rapidly during plant seed development or under the stress condition of water deficiency, salt and cold conditions and further induced the expressions of the related genes to improve the resistance of plants1, 2, 3, or through the regulation of PKABA, encoding a kind of protein kinase, and Gamyb to effect the ABA and GA antagonism during seed germination4. Studies by means of genetic and molecular technologies have resulted in the identification of ABA-regulated genes and corresponding transcription factors necessary for ABA-related signal transduction (for review, see5, 6). Up to now, 6 genes, shown to be necessary for the ABA-related signal transduction have been reported and their encoded proteins were classified as 2 transcription factors (VP1 of maize and ABI3 of Arabidopsis7, 8), 2 members of highly conserved protein phosphatase 2C family9, 10, 11, 1 transcription regulator harbouring APETALA 2 domain (ABI412) and 1 farnesyl transferase (ERA1 of Arabidopsis13). Studies have been focused on the physiological roles of ABA and its related signal transduction5. However, the complication of ABA signal network made it relatively difficult to isolate and analyze ABA-regulated genes with traditional methods14. Even until now, much was still unknown about the functional mechanism of ABA, especially the related signal transduction in plant cells.

The recently developed cDNA microarray technology enabled monitoring of cell-, tissue- and developmental stage-specific gene expression profiles and simultaneous quantitative analyses of expression levels of genes15, 16, 17, 18, 19, which could help people to further find out new genes possibly involved in the same process or signaling pathway. This will help people to have the general and full knowledges on certain developmental processes. cDNA microarray technology has also been used in plant research field for gene studies in recent years20. Schena et al. first experimentally made the array with 45 Arabidopsis genes and successfully detected the difference of low expressed genes21. Through this technology, Reymond et al. analyzed 150 wounded and insect-related genes of Arabidopsis22, and Grike et al. analyzed genes related to seed development of Arabidopsis and found 25% and 10% of the genes tested, in total 2600 arrayed genes, increased 2 and 10 times respectively during Arabidopsis seed development23. Seki et al. analyzed expression patterns of 1300 Arabidopsis full-length cDNAs under cold and drought conditions, and found 44 drought- and 19 cold-induced cDNAs, in which 30 and 12, respectively, were newly reported24. They also found 12 genes were induced by transcription factor DREB1A (dehydration-response element binding protein) and further analyses showed that drought response elements were present in these 12 genes. All these results suggest that cDNA microarray would be useful not only for isolation of new genes, which would certainly help people to study the general profiles of the tissue-specific or environment related cDNAs, but also for identification of target genes and potential cis elements of transcription factors.

We are interested in the ABA related signal transduction and its function mechanism in plants. Here we report, based on the comparison of the differences of gene expression patterns through cDNA macroarray technology, the isolation and identification of ABA regulated genes in rice. We hope this would possibly provide some hints for the mechanisms on ABA mediated signal transduction in plant cells.



ABA (abscisic acid) was obtained from Sigma (USA). 96-well PCR plates were from ABGENE and 96-well cell plates were from Nunc. Taq polymerase and standard T3 and T7 primers were from Sangon (Shanghai, China). dNTP were obtained from TaKaRa Biotechnology (Dalian, China). Radiochemical [α-33P]dCTP was obtained from ICN (Meckenheim, Germany). RNA reverse transcription labeling kit was obtained from GIBCO-BRL (USA). Primers for reverse transcription PCR (RT-PCR) were obtained from TIB Molbiol (Berlin, Germany).

Bacteria and plant materials

Escherichia coli JM109 cells were used for amplifying the cDNA library. Cells were cultivated at 37°C in LB medium supplemented with appropriate antibiotics using standard methods25. Oryza sativa cv. Zhonghua 11 were germinated on 1/2 MS medium and grown in water in phytotron with a 12 h light (26°C and 12-h dark (18 °C) period.

ABA treatment and RNA isolation

2 week old rice seedlings (grown in water) were treated with 100 μM ABA for 2, 4, 8, 12 hrs and then frozen in liquid nitrogen for further analysis. Total RNA was prepared according to the extraction procedures of acid guanidinium thiocyanate-phenol-chloroform25 with few modifications: 1 g seedlings (including same amounts of treated materials for different time scales) were powdered in liquid nitrogen, extracted with guanidinium thiocayanate-phenol-chloroform, then precipitated with ethanol and purified with LiCl and chloroform each time. Total RNA, suspended in RNase-free water finally, was stored at −70 after spectrophotometrical quantification at 260 nm.

cDNA library construction and macroarray preparation

Collected ABA treated materials were used for plasmid cDNA library construction using pBluescript vector by TaKaRa Biotechnology (Dalian, China). Titers of the cDNA library were calculated based on the white colonies obtained under the white/blue selection on the plate supplemented with isopropylthio-β-D-galactoside (IPTG) and X-Gal after transferred to E. coli cells via electroporation.

A total of 6144 white colonies from the library were randomly chosen from the plates (supplemented with appropriate concentration of ampicillin, X-Gal and IPTG) and cultured overnight in 96-well cell plates. cDNA insertions were then amplified with standard T3 and T7 primers in 96-well PCR plate with T-gradient PCR instrument (Biometra Company, GÖtingen, Germany) using E. coli culture as templates. PCR reactions were performed in a total volume of 60 μl (including 48 ml ddH2O, 6 ml 10×PCR buffer with 15 μM MgSO4, 2.5 μl 2.5 mM dNTP, 2 ml 20 pM primers, 0.5 μl Taq polymerase and 1 μl E. coli culture as template) as follows: 94°C for 3 min, 1 cycle; then with 40 cycles of 94°C for 1 min, 52°C for 1 min, 72°C for 3 min; and finally with an extension at 72°C for 10 min. PCR products were checked on 1% agarose gel by electrophoresis.

Nylon membranes (Hybond, Amersham) was precut and wetted with 2×SSC for DNA spotting. PCR products were arrayed on nylon membranes using Biomek 2000 HDRT system (Beckman, Fullerton, CA, USA) with the spotting procedures as follows: spotting needles stayed for 2 sec in DNA samples (in 96-well PCR plate) and 2 sec on nylon membranes, then 5 sec in ddH2O twice for washing, 5 sec in 75% ethanol and dried for 10 sec. DNA samples were spotted in 4×4 array and each sample was duplicated on crossway which resulted in the 1536 DNA spots (16×96) standing for 768 cDNA clones on each membrane. After spotting membranes were then denatured (0.5 M NaOH, 1.5 M NaCl) for 1-5 min, neutralized (1.5 M NaCl, 0.5 M Tris.Cl, pH 7.4) for 5 min and washed with 2×SSC for 2-5 min, and then incubated at 80°C for 2 h for stabilization of DNA on the membranes.

Hybridization and image analyses

5 mg rice total RNA, extracted from ABA-treated (for 2 and 8 h) and untreated 2-week-old rice seedlings, respectively, was labeled with [α-33P]dCTP through reverse transcription using RNA labeling kit (GIBCO-BRL, USA). Resulted products were quantified with liquid scintillation counting (Beckman LS6500, USA) and the labeled first strand cDNA with same amounts of radioactivities were used as hybridization probes. Hybridization was performed at 65°C for more than 40 h in 250 m M sodium phosphate buffer pH7.2 containing 7% SDS, 1% BSA, and 1 m M EDTA. Washes were performed at 65°C in 2×SSC, 0.1% SDS for 15 min, 1×SSC, 0.1% SDS for 15 min and 0.2×SSC, 0.1% SDS for 20 min26. After overnight exposure, the images of membrane were scanned using phosphoimager (FUJIFILM, Japan) and resulted profiles were analyzed with commercial image processing system AIS program (Imaging Research INC., USA). As each sample was spotted on the membrane with duplication, signals acquired were calculated in the average of each clone for further comparison. After twice hybridizations using independent treated samples, differential expressed cDNA clones with the ratio (treated/control) > 1.5 or < 0.5 were selected and sequenced for further analyses.

Reverse transcription PCR (RT-PCR) analysis

RT-PCR analyses were carried out to confirm the ABA regulation of selected clones revealed by cDNA macroarray. 5 mg total RNA, isolated form ABA-treated (for 0, 2, 4, 8 and 12 h) rice seedlings was reverse transcribted to first strand cDNAs using oligo (dT) primer in a total volume of 40 μl according to supplier s instruction (SuperScript Pre-amplification System, Promega). Resulted cDNAs were then used as templates for PCR amplification in a volume of 30 μl as follows: 94°C for 3 min; then 30-35 cycles of 94°C for 40 sec, 56°C for 40 sec, 72°C for 45 sec; and finally with an extension at 72°C for 10 min. Primers used for the selected clones in RT-PCR were listed in the Tab 1. The rice Rac1 actin coding gene was used as positive internal control with primers RAc1-1 and 2 (Tab 1). Amplified PCR products (10 μl) were electrophoresed on a 2.5% (w/v) agarose gel and monitored using the Gel Doc 2000 (Bio-Rad Company, USA).

Table 1 Primers used in reverse-transcription PCR


cDNA library construction and macroarray preparation

Plasmid cDNA library using vector pBluescript was constructed with 100 μM ABA-treated (for 2, 4, 8 and 12 h) rice seedlings and harboured a total of 3×105 white colonies under white/blue selection. cDNA insertions were amplified via PCR using standard T3 and T7 primers and results showed that the sizes of insertions were between 0.5-3 kb, in 1.2 kb average. A total of 6144 PCR products representing same numbers of independent clones randomly chosen from plates were arrayed on 8 nylon membranes, on which there were 1536 DNA samples (in a 4×4 array) spotted on each membrane for 768 cDNA clones with duplication on the crossway (Fig 1A) to ensure the results reproducible.

Figure 1
figure 1

cDNA macroarray and image analyses A. PCR products were arrayed on the nylon membranes in a 4 × 4 array. Each sample was duplicated on crossway; B. A typical image of cDNA macroarray after hybridization, red and blue arrows indicate up- and down-regulated clones respectively, left-control; right-ABA treated; C. Images were analyzed with “AIS” program. Image signals were transferred to digital characteristics for comparison between treated and control samples resulting in the regulation ratio. D. Most of the samples analyzed are within the ratio 0.5-2.0. Ratio was calculated through treated/control (in average) and ratio 0.5 (green), 1.0 (yellow) and 2.0 (red) were marked.

Hybridization and isolation of rice ABA-regulated genes

Rice total RNAs, extracted from ABA-treated (for 2, 4, 8 and 12 h) and untreated control materials (using liquid culture medium), were labeled with [α-33P]dCTP through reverse transcription. Resulted labeled first strand cDNA were quantified with liquid scintillation counting and same amounts of labelled cDNAs were used as probes for hybridization. Resulted images, as shown in Fig 1B, were then analyzed with AIS software through transferring the image signals to digital characteristics for detecting the differences between treated samples and control (Fig 1C). As each clone was spotted on the membrane with duplication, digital characteristics acquired were analyzed with control in the average with ratio (treated/control) to test the induction or suppression, i.e. <2.0 or <0.5 indicated the induction or suppression of the corresponding clones with one time level respectively. Hybridizations were repeated for two times using independent treated samples. As shown in Fig 1D most of the samples (around 93% of the total clones) tested were located in a range between 2.0 and 0.5 that indicated no response to ABA for most of the samples under 2-12 h ABA treatment. Thus, samples with the ratio > 1.5 (for detailed analysis with a extension from 2.0 to 1.5) or < 0.5 comparing to control were picked out and original cDNA insertions were sequenced. In a total of 6144 arrayed cDNA clones, after sequencing, 41 were found to be regulated by ABA, in which 31 and 10 were found to be up- and down-regulated respectively. Sequences of these 41 clones were then analyzed with BLAST program against database and 32 of them shared homologies with known proteins and 6 of them with unknown functions. The putative proteins encoded by identified clones were classified as different families involved in transcription (encoding bZIP, MADS, Myb transcription factors), metabolism and resistance (encoding transferase, synthase and kinase), photosynthesis, signal transduction (encoding kinase, calmodulin and WD-repeat proteins), and seed development. Results were shown in Tab 2.

Table 2 Characteristics of ABA-regulated genes in rice shoots through cDNA macroarray analysis (Id. Identities; Po. Positives). Rice shoots were treated with 100 μ M ABA for 2 and 8 h. (Clones isolated before were indicated as “reported”, and those not isolated before were as “new”. The nucleotide sequences have been submitted to GenBankTM/EMBL Data Bank with shown accession numbers).

Among the identified clones, some cDNAs have been reported from rice, including phytoene synthase and putative bZIP (basic/leucine zipper) transcription factors coding genes, have been demonstrated to be induced under stress situations 27 and cold stress28,29. Most of them were not reported before.

Confirmation of ABA regulation with selected clones

Some clones, which encoded putative proteins involved in signal transduction, seed development or metabolism, were chosen for further RT-PCR analyses (primers used were shown in Tab 1) to confirm the ABA regulation. To confirmed the materials used were in parallel and especially avoid of the affection of circadian regulation, control plants (at exactly identical time scales with treatment) and treated rice seedlings were used for RNA extraction and further RT-PCR analysis. In total tested 7 clones, which encoded putative ethylene-inducible protein (AJ437339), cold accumulation protein (Lip9) (AJ439990), sensory transduction histidine kinase (AJ437267), WD-repeat protein (AJ439992), glyoxalase I (AB017042), calmodulin (AF231026) and sphingosine kinase (AJ437338), as shown in Fig 2, were detected to be regulated by ABA. The regulation tendencies of these clones fit the ratio results obtained from macroarray.

Figure 2
figure 2

Confirmation of ABA regulation of selected clones through RT-PCR. Samples were collected after ABA treatment for 0, 2, 4, 8 and 12 h. Primers used were shown in table 1. Numbers showed the ratio of the selected clones comparing to control after hybridization and analyzed with “AIS” program. Bottom line showed the photodensity of the DNA products. Rice Rac1 gene was used as internal control.


cDNA macroarray and hybridization

cDNA microarray technology has been shown to be powerful to analyze expression profiles of numbers of genes in parallel so as to find out new genes in specific tissue and organs or during specific growth and developmental stages, or under different conditions such as cold, salt stress, drought and pathogen infection30, 31. One advantage of this technology was the possibility to find out the genes related to one process or developmental stage in one experiment in genome scale, this would certainly provide informations for the complicated network of signal transduction32. However, the cost for microarray experiment performance is very expensive. Although the density of spots in a certain space is less and quality of nylon will have more affection in the hybridization results than microarray, macroarray will be cheaper and convenient for general lab performance. Since some steps may affect the macroarray results, including the array spotting, hybridization and evaluation of the results. To overcome the possible problems, we first spotted the DNA samples into different membranes at same time to be sure the amounts of DNA with same samples spotted on different membranes are identical so that the comparison after hybridization was believable, and spotted the same sample with duplication to confirm the results; second, quantified the hybridization probes after labeling with liquid scintillation counter to be sure the radioactivities used during hybridization were identical to make the signals acquired comparative; and third, evaluated the image signals to digital characteristics to compare the differences more precisely.

In the experiments reported in this paper, 6144 cDNA clones, which possibly stand for the 10% or more cDNAs of rice were arrayed and tested, which provided the possibility to isolate genes related to ABA signaling pathways from rice in genome scale. Amplification of cDNA insertions with standard primers made it relatively easier to perform the amplification in large scales and the following spotting of PCR products. Hybridization of the arrayed nylon membranes using [33P]dCTP labeled first strand cDNAs (same amounts of radioactivities of treated samples and control) was performed parallelly after reverse transcription and the following analyses with AIS program would be helpful for finding out the differentially expressed clones. This was demonstrated, as shown in Fig 1d, that most of the clones tested within the range 0.5-2.0 in ratio.

ABA regulated genes

Genes selected from macroarray, which provide hints for differential expressed gene isolation, still need further confirmation through either Northern blot or RT-PCR analysis. Sequence analyses and homologous searches indicated that some clones, identified by cDNA macroarray, were indeed regulated by environmental factors, which suggested the possible relation to ABA and the involvement in ABA-related signal transduction pathways. Protein encoded by clone Xd423 shared high homology (95%) to maize phytoene synthase, which was demonstrated to be induced under stress situations (Steinbrenner and Linden, 2001). Protein encoded by clone Xd426 (X57325), a basic/leucine zipper transcription factor, was shown to be induced under cold treatment28,29. However, confirmation of the identified clones is still required. RT-PCR results, however, 30% of clones (3 of 10) tested did not give PCR products, which might be due to the reason by PCR itself.

Possible ABA related signaling pathways

Most of the clones identified by cDNA macroarray were unknown or unreported to be related or involved in ABA mediated signal transduction before. Based on RT-PCR confirmation for some clones related to different signal transduction pathways, we may suggest that some genes were involved in ABA mediated signaling pathways for stress response.

WD-repeat protein was demonstrated to be involved in many signal transduction pathways33, 34, 35 including SPA (inhibitor of PhyA36) and TRIP1 (subunit of eIF3, interacted with TGF-β receptor, and may involve in the brassinosteroid stimulated plant growth37). A putative ABA induced WD-repeat protein, with 64% identities (80% homologies) to an Arabidopsis one4,33, was indeed regulated by ABA.

Histidine kinase plays a role in many signal transduction pathways38, 39, 40. Histidine kinase domain was found in phytochrome and related to rhymes41. The recently reported CKI1, receptor of cytokinin in Arabidopsis, was a histidine kinase. Sphingosine kinase, together with phospholipase C, phospholipae D and MAPK, were involved in the cell response to environmental stimuli including growth factors, cytokinin, G-protein coupled stimulator and so on40, 42, 43, 44. Up-regulation of rice histidine kinase and sphingosine kinase by ABA treatment may suggest their roles in ABA function.

Cold accumulation protein: ABA increased rapidly under cold stress and further induced related gene expression for plant resistance. Cold-induced gene, Kin2, isolated from Arabidopsis, was induced by ABA, salt and drought45. A putative rice cold accumulation protein was isolated and confirmed to be induced by ABA treatment, which shared 72% and 52% homologies with rice and wheat cold accumulation protein Lip9 and WCOR410b.

Additionally, protein encoded by clone Xd481 (AJ439989) shared 68% homologies to rice ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco), which catalyzed the reaction of oxygen and ribulose-1,5-bisphosphate resulting the production of glycerate-3-phosphate and glycolic acid-2-phosphate. CO2 and NH3 were produced during the conversion of glycolic acid-2-phosphate to glycine. This reaction could prevent the overreduction of photosynthetic electron transport chain in light respiration under stress situation so as to prevent the light inhibition. Furthermore, glycine, the metabolite of photorespiration, is the precursor of the synthesis of glutathione, which also play a role in the responses of plant cells to environmental stress46. Induction of the expression of small chain precursor of Rubisco under ABA treatment may suggest that under stress condition, increase of the Rubisco would stimulate the plant photorespiration, reduce the photoinhibition and result in the plant self protection.

Interaction of ABA with auxin and BRs

It should be noticed that the cross talk between different signaling pathways were widely present in plant cells, especially on hormone regulated pathways. We hereby performed the analysis on the same samples under treatment with other hormones, i.e. auxin and brassinosteroid (Shen et al., submitted; Chu et al., submitted). Results of expression profiles showed that some clones, regulated by ABA, were regulated by IAA or BR as well (Tab 3). The possible multi-regulated clones encoded putative cytochrome P450, MADS box transcription factor, Proline-rich protein precursor, seed maturation protein and so on. Together with the results that ethylene-response protein was regulated by ABA (Fig 2, clone Xd439, AJ437339), although the detailed analysis on the regulation by both hormones is still required, these may indicate the complicated interactions of signal pathways with different hormones. For example, polyubiquitin expressed at variable amounts in many plant tissues with high levels in metabolically active and/or dividing cells, vascular tissues and was induced by wound, heat, or ethylene treatments47, 48, 49. Although there was no report about the connection of auxin and ubiquitin, but the promoter of GLP3b gene has a single putative auxin-response element, which was physically linked to the polyubiquitin 4 gene50. In our experiments, we gave evidences that polyubiquitin played a role in both auxin and ABA functions.

Table 3 Clones regulated by both ABA and auxin or brassinosteroids through cDNA macroarray. Auxin (IAA, 100 M) treatment was performed with 2-week rice shoots for 2 and 12 h and brassinosteroids (24-epi-brassinolide, 10 M) treatment was performed with 2-week rice shoots for 4 and 24 h.