Transcriptional Downregulation of Rice rpL32 Gene under Abiotic Stress Is Associated with Removal of Transcription Factors within the Promoter Region

Background The regulation of ribosomal proteins in plants under stress conditions has not been well studied. Although a few reports have shown stress-specific post-transcriptional and translational mechanisms involved in downregulation of ribosomal proteins yet stress-responsive transcriptional regulation of ribosomal proteins is largely unknown in plants. Methodology/Principal Findings In the present work, transcriptional regulation of genes encoding rice 60S ribosomal protein L32 (rpL32) in response to salt stress has been studied. Northern and RT-PCR analyses showed a significant downregulation of rpL32 transcripts under abiotic stress conditions in rice. Of the four rpL32 genes in rice genome, the gene on chromosome 8 (rpL32_8.1) showed a higher degree of stress-responsive downregulation in salt sensitive rice variety than in tolerant one and its expression reverted to its original level upon withdrawal of stress. The nuclear run-on and promoter:reporter assays revealed that the downregulation of this gene is transcriptional and originates within the promoter region. Using in vivo footprinting and electrophoretic mobility shift assay (EMSA), cis-elements in the promoter of rpL32_8.1 showing reduced binding to proteins in shoots of salt stressed rice seedlings were identified. Conclusions The present work is one of the few reports on study of stress downregulated genes. The data revealed that rpL32 gene is transcriptionally downregulated under abiotic stress in rice and that this transcriptional downregulation is associated with the removal of transcription factors from specific promoter elements.


Introduction
Plants encounter a number of abiotic stresses which limit agricultural production. Classical breeding programs, as also the initial attempts with genetic engineering to raise stress tolerant plants have met with limited success [1]. Through recent advances in genomics and proteomics, some knowledge about stress perception, stress signaling and transcription regulators has been gained from model organisms such as Arabidopsis, Synechocystis and Clamydomonas [2][3][4]. Yet a more detailed understanding of stress response and adaptation is desirable, especially for monocot crop plants like rice [5]. An in depth knowledge about the regulation of stress-responsive genes will help in developing strategies to increase the stress tolerance level of socially and economically important crops.
It is well known that growth is severely affected in plants under stress conditions. Ribosomal proteins are essential for proper growth and development of any organism. Not much was known about regulation of expression of genes encoding ribosomal proteins in plants until recently. Many ribosomal proteins have been reported to be up-or downregulated under various stress conditions [6][7][8][9][10]. Recently, chloroplast encoded rpL33 was shown to be necessary for plant survival under cold stress in Arabidopsis [11]. Pathways of translational downregulation of ribosomal proteins as observed in mammals and transcriptional regulation of ribosomal proteins, somewhat similar to yeast, have been described in plants earlier [12][13][14][15][16][17]. Post-transcriptional regulation for ribosomal proteins has also been suggested [7,[18][19][20][21]. Although altered expression of ribosomal proteins under stress has been shown to involve the post-transcriptional and translational regulatory mechanisms, stress-specific transcriptional regulation of ribosomal proteins is still to be reported in plants.
In this work, we have studied the transcriptional regulation of genes encoding 60S ribosomal protein rpL32 in rice under salt stress. Homologues of rpL32 have been shown to bind RNA in archaea and maize, possess transcriptional transactivation activity, mediate telomere silencing in yeast, and promote proper growth and development in Drosophila [22][23][24][25]. From the analysis of differential cDNA libraries of rice (previous work in our lab; unpublished) and barley [26], the downregulation of this gene under salt stress was postulated. The present work has shown that rpL32 gene is downregulated at the transcriptional level under abiotic stress and that this downregulation is linked to the removal of proteins from cis-elements in the promoter region. Our work presents a study of transcriptional downregulation of genes which has been far less studied than upregulation of genes during abiotic stress in plants.

Rice genome contain four genes encoding ribosomal protein rpL32
A BLASTn homology search using partial coding sequence of rpL32 in the rice genome database of NCBI (http://www.ncbi. nlm.nih.gov/) and MSU Rice Genome Annotation Project v6.1 (RGAP 6.1; http://rice.plantbiology.msu.edu/) indicated the presence of four rpL32 genes (Fig. S1). These loci have been annotated as putative 60S rpL32 gene in the NCBI database. One of these genes was present on chromosome 8 and was named rpL32_8.1. The other three were named rpL32_9.1, rpL32_9.2, and rpL32_9.3 and located contiguously on chromosome 9. The CDS and the predicted protein of all the four rpL32 members share a high percentage of identity (Fig. S2A,B). However, the 39 UTR sequence of these genes were dissimilar (Fig. S3). BLASTn homology search using 39UTR specific probes for each of the predicted rpL32 genes in NCBI database, revealed the presence of homologous expressed sequence tags (ESTs) and mRNA sequences. This confirmed that all these four genes are expressed in rice. The gene structure is also similar for all the rpL32 genes of rice, each having three introns (Fig. S2C). The first intron is the smallest one and resides within the 59 UTR of the genes. The location of the remaining introns, on the other hand differs in each gene.
Rice rpL32 genes are downregulated in shoots under salt stress Northern hybridization using radiolabeled 39UTR probe of rpL32_8.1 revealed that this gene is highly expressed under control (unstressed) conditions in shoots of salt tolerant indica rice Pokkali (Fig. 1A). After 1 to 12 h of salt stress, a 40-60% decrease in steady-state level of the transcript was observed relative to the unstressed condition (t0.0).The level of rpL32_8.1 decreased further to 20% of control after 24 h of salt treatment. In the salt sensitive PB1 variety of indica rice also, the level of rpL32_8.1 decreased in response to salt stress except that the response was more rapid and of greater magnitude than that observed in the salt tolerant pokkali variety (Fig. 1B). The root tissues of Pokkali and PB1 rice seedlings did not exhibit any steady decrease in their transcript level (Fig. 1C,D).
rpL32_9.3 and rpL32_9.2 also exhibited downregulation of their expression in response to salt stress in shoots of Pokkali rice ( Fig. 2A,B). However, the decrease in transcript level of rpL32_9.1 was relatively less under salt stress than the other three rpL32 genes (Fig. 2C). The transcript of this gene was also faintly detected in northern blots as compared to other rpL32 genes.
A semiquantitative RT-PCR, using elongation factor-1a (EF-1a) as internal control gene, reconfirmed the differential downregulation of rpL32 genes in shoots of PB1 and Pokkali rice plants under salt stress ( Fig. S4A-C). Though small differences in the degree of downregulation of these genes by RT-PCR and northern analyses were observed, yet the faster rate of shoot-specific downregulation of rpL32_8.1 under salt stress in PB1 rice variety was validated. Using RT-PCR, downregulation of rpL32_9.1 under salt stress was also verified.

Downregulation of rpL32 genes under stress is transcriptional in origin
To determine if the downregulation of rpL32 genes happens at transcriptional level, nuclear run-on assay was performed using nuclei isolated from unstressed and stressed shoots (Fig. 3). Nuclear run-on assay of nuclei from unstressed shoots revealed that all four rpL32 genes in rice were expressed, with the strongest transcription being of rpL32_8.1 (nearly 8-10 folds higher than the other rpL32 genes; Fig. 3B,D). Although such differences might result from different hybridization efficiencies of the 39UTR specific DNA fragments used in the dot blot for the detection of rpL32 genes, it appeared least likely as these 39UTR regions are similar in length and GC content. Moreover, during RT-PCR analysis also, rpL32_8.1 was detected with lesser number of PCR cycles (than that used for other rpL32 genes). Transcription was greatly reduced in stressed shoots (treated with 200 mM NaCl for 24 h), with the signal for all the rpL32 genes being barely visible (Fig. 3C,D). This indicated that these genes are transcriptionally downregulated. As the results of this assay indicated that the majority of the rpL32 transcript in rice is contributed by rpL32_8.1, further studies were focused on the regulation of this gene.

rpL32_8.1 gene expression is responsive to various treatments
To investigate if rpL32_8.1 gene is responsive to other factors, its expression was studied under a variety of conditions (Fig. 4A). As compared to untreated samples, about 40% and 60-85% reduction in the transcript level was observed in response to sucrose (4.5%) and cold (4uC) treatments (for 3 and 24 h), respectively. With more than 90% reduction in the transcript level by 3 h, the most drastic effect on rpL32_8.1 expression was observed in response to drought stress (air-drying). Although some reduction in the transcript level was evident after 3 h of abscisic acid (ABA; 100 mM) treatment, such downregulation was not observed after 24 h.
The transcript level of rpL32_8.1 gene was also measured in the shoots of seedlings during stress recovery (Fig. 4B). The seedlings were given 200 mM NaCl stress for 24 h, after which they were placed back onto the medium lacking NaCl. After 24 h of salt stress, expression of rpL32_8.1 was reduced as observed earlier ( Fig. 4B; Lane2). However, within 1 h of stress recovery, the transcript level increased and reached the levels found in unstressed shoots ( Fig. 4B; . This indicates that the regulation of this gene is dynamic and reversible, and the gene is downregulated only till such time as the plants are exposed to stress. Semiquantitative RT-PCR analysis reconfirmed the decline of this transcript under cold and drought stress, and its increase during recovery from NaCl stress. (Fig. S4D-F). However, ABA treatment did not show any significant downregulation of rpL32_8.1 expression. RT-PCR showed a smaller decrease (20-30%) in transcript levels under sucrose treatment than that observed by northern analysis. Similar to northern studies, this method also indicated that the extent of downregulation of rpL32_8.1 expression was greater in drought stress than under salt and cold stress.
The transcriptional regulation of rpL32_8.1 is governed by its promoter To further confirm the transcriptional downregulation of rpL32_8.1, promoter:reporter assays were performed. A ,1.23 kb (approx), DNA fragment corresponding to the promoter region of rpL32_8.1 gene was PCR amplified using genomic DNA from Pokkali indica rice variety and cloned in TA cloning vector. The sequence of this fragment did not show any significant difference from the sequence of the same region in japonica rice, which is available at RGAP 6.1 database (Fig. 5A). An in silico search using  To determine the transcription start site (TSS) in this cloned fragment, a primer extension assay was carried out using total RNA from shoots. The result indicated the presence of three major TSSs which were named as TSS1, TSS2 and TSS3 (Fig. 5B,C). The TSS1 was considered as the +1 TSS. No conventional TATA-box was found in the vicinity of the TSSs, but a TATA-like sequence TATAAGA was found 27 bp upstream of the +1 TSS ( Fig. 5B,C).
To identify the minimum functional promoter region responsible for the expression and stress responsiveness of rpL32_8.1, promoter:GUS constructs were prepared in pCAMBIA 1391z vector ( Fig. 6A-C) using a 1170 bp DNA fragment (DF0; considered as the full length promoter fragment) and its four  serial deletions (DF1, DF2, DF3 and DF4). These promoter:GUS constructs were then used to raise transgenics of tobacco. Seeds collected from the GUS-positive tobacco plants were germinated on medium containing hygromycin and GUS histochemical staining was done after 2 h of transferring one-week-old seedlings to medium containing 200 mM NaCl or lacking it. The assay was performed for 3-5 different transgenic lines of each of the promoter fragments and the results are summarized with a representative picture for each transgenic ( Fig. 6D-I). Plants bearing DF0 and DF1 fragments hardly showed any GUS staining in the presence or absence of salt stress (Fig. 6D,E). However, DF2 transgenic plants showed a strong GUS expression in both shoots and roots which decreased significantly in shoots upon salt treatment (Fig. 6F). DF3 transgenics also showed a strong expression of GUS but only in shoots which also decreased upon stress treatment (Fig. 6H). Plants bearing DF4 fragment showed a faint GUS expression in shoots which further decreased when subjected to salt stress (Fig. 6I). Thus, the promoter:GUS assays indicate that the transcriptional regulation of rpL32_8.1 gene originates within the promoter region of the gene. A significantly larger difference of GUS expression in unstressed shoots of DF3 and DF4 plants indicated that cis-elements required for the higher level of transcription of this gene are concentrated within a 262 bp region in the promoter (region which is present in DF3 fragment but not in DF4). Though the DF4 fragment expressed GUS faintly, it downregulated the expression of the reporter when subjected to stress. It is possible that either a repressor element is present within DF4 fragment, or the downregulation of the associated gene is caused by removal of transcription factors from the cis-elements in DF3 and DF4 fragments. For further analysis, this 262 bp DNA region was studied to determine if it was indeed the binding region of transcription factors.
Downregulation of rpL32_8.1 gene is associated with removal of transcription factors DMS-LMPCR in vivo footprinting was used to investigate if any cis-element in the aforementioned 262 bp region was showing differential trans-factor binding in the presence and absence of salt. A marked difference in protein binding was observed between stressed (200 mM NaCl, 24 h.) and unstressed (control) samples, at many 'G' residues in the top and bottom strand of the scanned region of the promoter fragment (Fig. 7A, B). In all these sites, a factor binding under the control condition (Lane 2; Fig. 7A,B) was present in reduced amounts under salt stress (Lane 3; Fig. 7A,B) which was evident from the appearance of bands in the latter case (pointed by arrowheads). LMPCR products of piperidine cleaved in vitro DMS-treated genomic DNA was used as experimental control in this assay (Lane 1; Fig. 7A,B). By superimposing 'G' residues detected by in vivo footprinting of top and bottom strands, over the sequence of the promoter region, three different core sequences showing differential binding of transcription factors between control and stressed samples were identified (Fig. 7C). These were 59GGCCC39 (at 2282, 2254, 2173 and 2127), 59GCCC39 (at 2240, 2209 and 2149), and 59GGGCC39 (at 2297 and 2191). Differential binding of transcription factors on 59GGGCC39 element at 2297 bp and 59GGCCC39 element at 2282 bp was detected only in top strand footprinting.
It was observed that all the functional GGCCC sequences are followed by an 'A' residue ( Fig. 7C). Similarly, all GGGCC sequences are preceded by a 'T' residue and GCCC sequences are flanked by an 'A' residue on each side. Hence, the functional ciselements were considered as GGCCCA, TGGGCC, and AGCCCA. The sequences GGCCCA and TGGGCC are reverse complementary to each other and hence were considered as the same cis-element. Upon comparing the sequences of the remaining two cis-elements, we could assign them with a consensus sequence of (A/G)GCCCA. In PLACE database, this consensus sequence is denoted as SITE II element.
The DF4 fragment, which expressed GUS very faintly and contributed to further downregulation of GUS under salt stress, was found to have a 40 bp DNA region containing a telo-box, a GCCCR type SITE II element and a SORLIP1 site. On performing an EMSA using this 40 bp DNA fragment and nuclear proteins from shoots of unstressed and stressed (200 mM NaCl, 24 h) rice seedlings, a reduction in protein binding was observed under stressed condition (Fig. 8A). Similar results were also obtained when two other oligos containing different SITE II elements, on which reduced protein binding was observed under salt stress by in vivo footprinting assays, were used in EMSA experiments (Fig. 8B,C). This substantiates the previous observations and indicates that the reduction in expression of rpL32_8.1 is associated with removal of transcription factors from its promoter.

Discussion
The mechanism of downregulation of genes encoding ribosomal proteins under stress conditions has not been well studied. The present work indicates that the genes encoding rpL32 are transcriptionally downregulated under abiotic stress in rice and this might be mediated via removal of transcription factors from the promoter region.
In silico analysis indicated that rpL32 is encoded by four genes in rice and belongs to a small multigene family as also reported earlier in plants and yeast [27][28][29][30][31]. The northern blots and RT-PCR analyses indicated that all rpL32 genes are downregulated under stress conditions in shoots of rice seedlings. Changes in expression pattern of various ribosomal proteins have been reported from many plant systems under different stresses [6][7][8]10]. In differential cDNA libraries of barley also, this gene was found to downregulate under abiotic stress [26]. It appears that downregulation of genes encoding rpL32 under stress is a common phenomenon in monocot plants.
Ribosomal proteins in plants have been reported to be translationally downregulated under osmotic stress via pathways involving TOR kinase, RAPTOR, PDK1 and S6K [12][13][14][15]. Posttranscriptional regulation of ribosomal proteins S4, S6, S28, L3, C24 (L13), L16 and P2, has also been suggested in plants [7,[18][19][20][21]. Despite identifying cis-elements and transcription factors responsible for high expression of ribosomal protein genes in growing and dividing plant cells [16,17], none of the reports describes transcriptional regulation of these genes under stress. In the current study, using nuclear run-on assay transcriptional downregulation of all four genes of rice encoding rpL32 in response to salt stress was inferred. Relatively weak detection of rpL32_9.1 in unstressed shoots during expression and nuclear runon analyses might be due to restrictive transcription of this copy in a few specific cell types and/or at a different developmental stage, if not due to lower hybridization efficiency of the probe or the primers. In plants, variation in expression pattern among genes encoding the same ribosomal protein were also observed in previous studies [28,31]. Being the major transcript observed in our experimental conditions, the regulatory studies under salt treatment were performed for rpL32_8.1 gene.
Promoter-dependent stress-responsive transcriptional downregulation of rpL32_8.1 was confirmed by promoter:GUS assays in transgenic tobacco plants. The failure of DF0 and DF1 fragments to express GUS might have occurred due to the absence of an activator element in the isolated 1.2 kb promoter fragment which might have a role in suppressing the activity of a repressor element within the DF1 region. In fact, a Matrix Attachment Region was predicted 5 kb upstream of the ATG start codon of rpL32_8.1 gene.
The results of promoter:GUS assays suggested that majority of the cis-elements responsible for stronger expression of rpL32_8.1 in unstressed shoots were located within the DF3 fragment; more specifically, between 284 bp and 546 bp region upstream of the ATG codon. An in vivo footprinting assay of this region indicated the presence of many (G/A)GCCCA elements where reduced binding of transcription factors was observed under salt stress. This observation was also confirmed by EMSA. (G/A)GCCCA (reverse complementary of TGGGC(C/T) sequences are variants of SITE II elements and are the binding sites of Class-I TCP-domain proteins. (G/A)GCCCA and GCCCR SITE II elements are enriched in promoters of ribosomal protein genes and have been reported to be necessary for high expression of ribosomal protein and other genes in actively growing and dividing cells of Arabidopsis and rice [16,17,[32][33][34][35]. EMSA also detected reduced binding of proteins from shoots of stressed plants over a 40 bp portion of DF4 fragment containing a telo-box (TAGGGTTT), GCCCR type SITE II element (AGCCCG) and SORLIP1 site (GCCAC). This might be responsible for the decrease in GUS expression in DF4 transgenics under salt stress. Our work, hence, strongly suggests that the transcriptional downregulation of rpL32_8.1 might be linked to the removal of transcription factors from the TCPdomain protein binding sites and other related elements. In silico analysis had also predicted the presence of GGCCC and GCCAC elements in promoters of stress downregulated genes [36]. It is being hypothesized that a similar mechanism might be mediating the decrease in transcription of other genes encoding rpL32 in rice under stress as similar cis-elements were also detected in their promoters by in silico methods. (Fig. S6).
Since many ribosomal proteins are coordinately regulated, a comparison was performed between the the abundance of (A/ G)GCCCA elements within 1 kb region upstream of the ATG codon of different cytoplasmic ribosomal protein genes of rice and the level of expression of these genes under drought and salt stress. The expression data was obtained from Microarray Experiment ID GSE6901 (www.ricearray.org). Of 140 genes encoding ribosomal proteins, having a unique Affymetrix probe ID and at least two (A/G)GCCCA elements, 64.25% and 72.85% genes were found to downregulate (more than 1.2 times) under drought and salt stress, respectively. This indicates that (A/G)GCCCA elements are probably enriched in stress-downregulated genes (Fig.  S7). However, the influence of other associated cis-elements on the promoter cannot be ignored as nearly 25-30% genes do not show significant change in expression under stress conditions (more than 1.2-fold up-or downregulation) and 4% of these genes were found to be upregulated significantly, even though they were highly enriched in (A/G)GCCCA elements (5 or more).
Expression analysis indicated relatively moderate decrease in rpL32_8.1 transcripts in response to external sucrose treatment. No differential protein binding by in vivo footprinting was observed at the sugar-responsive elements (SRE) within the analyzed promoter fragment. In response to ABA no consistent decrease in the transcript level of rpL32_8.1 was observed. These results indicated that sugar and ABA might not be directly involved in the stable downregulation of rpL32_8.1 in rice under stress conditions. Based on the available literature and the present study, it appears that ribosomal proteins are downregulated both at transcriptional and translational level under abiotic stress conditions. The rate of downregulation of rpL32_8.1 transcripts seems to correlate with the severity of stress and degree of tolerance of the plant as the rate of downregulation was highest for drought stress and was higher in salt sensitive variety PB1. Gene(s) encoding rpL32 is present in all organisms ranging from archaea to higher eukaryotes and the sequence also shows high conservation among various groups of organisms (Fig. S8,S9), thereby indicating that this protein is needed for proper growth and development of all organisms. In fact in Drosophila, the deletion of this gene was found to result in various growth and developmental defects [22]. This could be the reason for its downregulation only transiently for the time for which the stress was applied and rapid return to its normal level as soon as the stress was withdrawn. rpL32_8.1 was found to get downregulated in shoots but not in roots upon exposure to salinity stress. Though the exact mechanism is not known, differential expression of genes in different tissues under various abiotic stresses is well known. The stress perception and the signaling mechanisms do vary in different tissues of plants. The prevailing environmental factors, under which plants are subjected to salinity stress, might also generate differential signals like calcium in shoots and roots [37]. Whether the rate of downregulation of such ribosomal proteins determines the stress tolerance level of a plant variety, needs further experimentation.

Plant material, growth conditions and sample collection
Rice seedlings were grown under control conditions in growth chamber at 2862uC and 16 h photoperiod in a hydroponics culture system supplied with modified Yoshida medium [38]. At three leaf stage, different treatments (dissolved in Yoshida medium) were given and tissues were collected at various time points. All experiments were done with indica rice variety Pokkali unless stated otherwise.

Amplification and cloning of different genomic fragments
The 39 UTR region of different rpL32 genes and the promoter fragment of rpL32_8.1 gene were amplified from the genomic DNA by two rounds of PCR (30 cycles each), the second one being done with a set of nested primers (see Table S1 and S2 for primer sequences) and cloned in TA-cloning vector (TOPO-TA, Invitrogen). The 59 serial deletions of the promoter fragment were generated by PCR using a fixed reverse primer R1 and different forward primers (primers sequences are given in Table S3). These PCR products were cloned in pCAMBIA 1391z plant transformation vector between the PstI and BamHI sites to make the promoter:GUS constructs.
Northern blotting 20 mg of total RNA (isolated by using Trizol reagent from Invitrogen) was blotted onto nylon membranes according to Sambrook et al [39]. The membrane was pre-hybridized for 6 h at 55-60uC in buffer containing 5X SSC, 5X Denhardt reagent, 0.1% SDS, 100 mg/ml denatured salmon sperm DNA and 10% dextran sulphate. Thereafter, radio-labeled pre-denatured probe (generated by PCR) was added to the pre-hybridization solution. After 16-18 h of incubation at 60uC in the hybridization solution, membranes were washed once with 2X SSC at room temperature, once with 2X SSC, 0.5% SDS for 15 min and twice with 0.5X SSC, 0.5% SDS for 15 min at the hybridization temperature. The blots were imaged using a phosphorimager (Typhoon Scanner; GE Lifescience). All densitometric analyses were done using ImageJ and Microsoft excel software.
Semiquantitative RT-PCR cDNA was prepared from 4 mg of total RNA using oligo dT cDNA synthesis kit (Fermentas). PCR conditions and required number of PCR cycles were standardized for each gene using 5fold, 10-fold, 50-fold and 200-fold diluted cDNA. rpL32_8.1 was amplified using 50-fold diluted cDNA for 26 cycles. rpL32_9.2 and rpL32_9.3 were amplified using 10-fold diluted cDNA for 27 cycles. rpL32_9.1 was amplified using 5-folds diluted cDNA for 30 cycles. EF-1a was used as a control gene and was amplified using 10-fold diluted cDNA for 25 cycles. All PCRs were performed with 1 ml of the respective cDNA in 20 ml of reaction mixture. 10 ml of the amplified products were electrophoresed and imaged using gel doc. All further densitometric analyses were done using ImageJ and Microsoft excel software. The details of the primers used in RT-PCR analysis are mentioned in Table S4.

Isolation of nuclei from rice shoots
Nuclei were isolated by minor modification of the protocols published earlier [40,41]. Briefly, 10 g of frozen rice shoots were crushed to fine powder in liquid N 2 and was homogenized with 100 ml of nuclei isolation buffer (10 mM PIPES, pH 7.0; 10 mM MgCl 2; 1 M hexylene glycol; 10 mM b-mercaptoethanol; 2.5% Ficoll; 0.5% triton X-100; 2 mM spermine). After filtration through 3 layers of cheese cloth and 2 layers of miracloth (Calbiochem), and centrifugation (3300 g for 8 min), the pellet was washed thrice with washing buffer (10 mM PIPES, pH 7.0; 10 mM MgCl 2; 1 M hexylene glycol; 10 mM b-mercaptoethanol), twice in presence and once in absence of triton X-100. Finally the nuclear pellet was resuspended in 500 ml of resuspension buffer (10 mM PIPES, pH 7.0; 10 mM MgCl 2; 1 M hexylene glycol; 10 mM b-mercaptoethanol; 25% glycerol), snap frozen in liquid N 2 and stored at 280uC until used. DNA from the isolated nuclei were measured according to Wanner and Gruissem [42].

Nuclear run-on assay
The nuclear run-on assay was done according to Hirose and Yamaya [43] and Kanazawa et al [44] using Rnasin-treated (with 100 units; Fermentas) 50 mg DNA equivalent nuclei preparation in 200 ml of 1X nuclear transcription buffer that includes 10 ml each of 10 mM ATP, GTP and CTP (Ambion) and 10 ml of a 32 P-UTP (3000Ci/m mol). The synthesized radiolabeled RNA was isolated in presence of 25 mg yeast t-RNA (Ambion) and equal counts of the heat denatured RNA were added to dot blot that were preincubated for 6 h at 55uC with 2 ml of hybridization buffer (250 mM sodium phosphate pH 7.2 and 100 mg/ml ssDNA). For preparing these dot blots, equal amount (500 ng) of PCR generated 39 UTR portions of different rpL32 genes and 18S rRNA, and pBSK plasmid was used. The blots were washed at 55uC in a similar way as was done for Northern hybridization and then imaged using a phosphorimager. All densitometric analyses were done using ImageJ and Microsoft excel software.

Transcription start site (TSS) mapping
For mapping TSS, primer extension was done with cDNA prepared from 5-10 mg of total RNA using MMLV-Reverse Transcriptase H-(Fermentas) as per manufacturer's instruction and 50 ng of c 32 P end labeled primer (59-TTCGGATCGGA-GAGGGAAGC-39), designed from the 59UTR just upstream of the first intron of rpL32_8.1 gene. The primer annealing step was done at 55uC. The primer extension products were run in parallel to a marker generated by manual sequencing (Sanger's dideoxy method manual sequencing kit; USB) of the cloned promoter fragment of rpL32_8.1 using the same primer (but unlabeled), on a 40 cm long 6% denaturing urea-PAGE gel at 60W power (,2000 V) and were compared with each other to find the TSS.

Tobacco transformation via Agrobacterium
Leaf discs (,1 cm 2 ) obtained from four-week-old tobacco plants (grown axenically) were infected with secondary culture of Agrobacterium tumefaciens (A 600 0.6-1.0) bearing recombinant constructs. Shoot regeneration was done in 1X MS-Agar (including vitamins) containing 0.1 mg/L NAA and 1 mg/L BAP 250 mg/L carbenicillin, 25 mg/L hygromycin. Rooting was obtained in 1X MS-Agar (without vitamin and hormones). GUS histochemical staining and PCR with uidA specific primers was done from leaves to confirm transgenics. The seeds, collected from these T 0 plants, were germinated on filter sheets soaked with 1X MS medium containing 50 mg/L hygromycin. One-week-old seedlings were treated with 200 mM NaCl solution for 2 h and GUS histochemical staining was carried out for salt treated and untreated seedlings.

GUS histochemical staining assay
Seedlings were first washed twice with 50 mM sodium phosphate pH 7.0. The GUS staining buffer (50 mM sodium phosphate pH 7.0, 0.5 mM potassium ferrocyanide, 0.5 mM potassium ferricyanide, 0.1% triton X-100 and 1 mg/ml X-Gluc; Bioworld) was then vacuum infiltrated (for 10-15 min) into the samples. After this the samples were incubated in the same buffer under dark conditions at 37uC till the blue color was developed. The chlorophyll from the samples was bleached away using 70% ethanol before being photographed.

In vivo Dimethyl sulfate-Ligation mediated PCR (DMS-LMPCR) footprinting
The DMS in vivo footprinting was done according to Busk and Pages [45]. Sequence of the linker primers (Linker1 and Linker2) and the primers for the blunt end generation, amplification and the final extension of top strand (R1, R2 and R3, respectively) and bottom strand (F1, F2 and F3) LMPCR are given in Table S5. For all steps of the LMPCR, Vent polymerase was used with 4 mM MgSO 4 in the LMPCR buffer (20 mM Tris-HCl (pH 8.9), 40 mM NaCl, 0.009% NaCl and 0.09% Triton X-100). The amplification step was done for 18 cycles. For the final extension step, c 32 P-dATP end labeled primer F3 and R3 were used. After electrophoresing, the gel was dried and imaged using a phosphorimager. Sanger's dideoxy sequencing reaction products of the cloned promoter fragment using the final extension primer was used as markers. The position of n th guanidine residue in the footprinting lane was determined as a measure of its distance from the +1 TSS by deducting 25 nucleotides from the position of coinciding base in the marker lane because the LMPCR products only contained the 25 bp linker DNA.

Electrophoretic mobility shift assay (EMSA)
10 mg of nuclear proteins, prepared according to He et al [46] were incubated with c-32 P end labeled double stranded oligos (50000 cpm) for 30 min in presence of 10 mM HEPES (pH 7.5), 1 mM MgCl 2 , 50 mM KCl, 10% glycerol, 0.5 mM DTT, 0.5 mM PMSF and 1 mg poly dI-dC in a 20 ml binding reaction at room temperature. The products were then run on a non-denaturing 4.5% polyacrylamide gel and imaged using a phosphorimager. Figure S1 Rice genome contains four genes encoding rpL32. One of the genes is located on chromosomes 8 (rpL32_8.1) and three others on chromosome 9 (rpL32_9.1, rpL32_9.  Figure S7 Stress induced expression change and the abundance of (A/G)GCCCA elements in ribosomal protein genes. Analysis presented here is of 140 genes encoding cytoplasmic ribosomal protein having unique Affymetrix probe ID and at least two (A/G)GCCCA elements. The top and middle graph shows the fold-change in expression of genes under drought and salt stress, respectively. The lower graph shows the abundance of (A/G)GCCCA elements in the promoter of these genes. The genes have been categorized on the basis of the number of (A/G)GCCCA elements in the promoter region as represented by vertical dashed lines and the corresponding abundance of the cis-element is also written below the lower graph. The red dashed lined box in the top two graphs represents insignificant change in gene expression (,1.2 fold up-or downregulation). A lower level of fold-change in fold expression was considered as significant as this microarray was done at a early time point and also includes roots where down regulation of rpL32_8.1 was not observed.