Deep sequencing reveals as-yet-undiscovered small RNAs in Escherichia coli
© Shinhara et al; licensee BioMed Central Ltd. 2011
Received: 26 June 2011
Accepted: 24 August 2011
Published: 24 August 2011
In Escherichia coli, approximately 100 regulatory small RNAs (sRNAs) have been identified experimentally and many more have been predicted by various methods. To provide a comprehensive overview of sRNAs, we analysed the low-molecular-weight RNAs (< 200 nt) of E. coli with deep sequencing, because the regulatory RNAs in bacteria are usually 50-200 nt in length.
We discovered 229 novel candidate sRNAs (≥ 50 nt) with computational or experimental evidence of transcription initiation. Among them, the expression of seven intergenic sRNAs and three cis-antisense sRNAs was detected by northern blot analysis. Interestingly, five novel sRNAs are expressed from prophage regions and we note that these sRNAs have several specific characteristics. Furthermore, we conducted an evolutionary conservation analysis of the candidate sRNAs and summarised the data among closely related bacterial strains.
This comprehensive screen for E. coli sRNAs using a deep sequencing approach has shown that many as-yet-undiscovered sRNAs are potentially encoded in the E. coli genome. We constructed the Escherichia coli Small RNA Browser (ECSBrowser; http://rna.iab.keio.ac.jp/), which integrates the data for previously identified sRNAs and the novel sRNAs found in this study.
RNA molecules are known to be key genetic regulators in diverse organisms. In bacteria, these regulatory RNAs are generally referred to as small RNAs (sRNAs) because they usually range from 50 to 200 nt in length . Recent studies have suggested that most sRNAs modulate target gene expression at the post-transcriptional level by base pairing to mRNAs . It has been reported that the majority of bacterial sRNAs are synthesised under very specific growth conditions and that these RNAs are regulators of gene expression in response to environmental stresses [2, 3] such as low iron , oxidative stress  and elevated glucose-phosphate levels . Currently, approximately 400 sRNAs have been detected in 70 microbial genomes, including of the Escherichia, Shigella and Salmonella genera . In Escherichia coli (E. coli), the most exhaustive and diverse searches for sRNAs have been conducted with several methods: high-throughput computational searches [7–13], shotgun cloning [14, 15] and tiling array analyses [16, 17]. As a result, 80 sRNAs have been experimentally verified and registered in RegulonDB , and many more sRNAs are predicted to exist . In recent years, deep sequencing has emerged as a new and powerful experimental method for transcriptome analysis . In eukaryotic organisms, this approach has been commonly used for transcriptome analyses, including microRNA detection . In contrast, in bacteria, this approach has only recently been used for transcriptome analysis, although it has the potential to increase our insight into transcriptional and post-transcriptional events in microorganisms dramatically [20–22]. For instance, several sRNAs have been identified using a deep sequencing-based approach in Salmonella[23, 24], Vibrio cholerae, Helicobacter pylori, Burkholderia cenocepacia, Bacillus anthracis and quite recently, E. coli.
In this study, we report novel sRNAs identified in E. coli with a deep sequencing analysis, focusing on low-molecular-weight RNAs (< 200 nt). Applying this approach, we successfully detected most of the previously known sRNAs, and discovered thousands of novel (fragmentary) transcribed regions. We selected 229 novel transcribed regions as candidate sRNAs with computational or experimental evidence of transcription initiation. Furthermore, we detected the expression of seven intergenic sRNAs and three cis-antisense sRNAs by northern blot analysis. Interestingly, five newly identified sRNAs are expressed from prophage regions. We conducted an evolutionary conservation analysis of our candidate sRNAs and summarised the data among closely related bacterial strains. Finally, we generated a platform (Escherichia coli Small RNA Browser; ECSBrowser) for sRNA research in bacteria.
Deep sequencing-based identification of cis-antisense and intergenic sRNAs
To predict novel transcribed regions, the resulting reads were classified into one of five groups (Groups A-E), according to the numbers (single or multiple) and types (annotated or non-annotated) of the mapped regions (Additional File 1B, step 4). We then eliminated the non-annotated mapped regions in Group E, because the reads in Group E matched multiple regions and because it was difficult to determine which one(s) was actually expressed. Finally, all overlapping mapped regions were assembled, thus generating 6,079 novel transcribed regions and 16,895 known transcribed regions (Additional File 1B, step 5). We assumed that some of these regions were still fragmentary. It is noted that 67 of 80 known sRNAs registered in RegulonDB 6.3 were detected with multiple deep sequencing reads, indicating the reliability of our method. The remaining 13 known sRNAs have been reported to be expressed only under specific conditions (Additional File 2).
Then, we attempted to extract novel sRNAs from the 6,079 novel transcribed regions. For this purpose, we focused on novel transcription units. Generally, sRNAs can be generated either as primary transcripts or by processing from longer precursor transcripts . In this study, we ignored the latter case because we could not discriminate the processed sRNAs from RNA degradation products in the current experiment. We used both computational and experimental data to reveal the novel transcriptional units. Firstly, we searched for potential promoter and terminator sequences at the genome level, as described previously , and predicted 47,630 sigma 70 promoter sequences and 5,290 rho-independent terminator sequences. The observed-to-expected (O/E) ratio of the predicted sigma 70 promoters for the 6,079 novel transcribed regions was lower than that for the 741 known genes with annotated sigma 70 promoters (Additional File 3A), but it is striking that their undulating patterns were very similar. These results suggest that many of the novel transcribed regions contain predicted sigma 70 promoters, as do the known genes. However, in the case of rho-independent terminators, the O/E ratio of the 6,079 novel transcribed regions was not significant (Additional File 3B), suggesting that most of the novel transcribed regions may not have a defined transcription termination site or may lack the 3' sequence. Based on these results, we decided to use only the transcription initiation information to extract the candidate sRNAs. Next, we collected published data on experimentally determined transcription start sites (TSSs) [30–33] and RNA polymerase-binding regions (RBRs) . Finally, we selected 229 novel transcribed regions as candidate sRNAs. The criteria for this selection process were as follows (a flow chart is shown in Figure 1B): (i) computational or experimental evidence of transcription initiation; and (ii) a length of more than 50 nt. These candidate sRNAs were designated ECS001-229. Of these candidate sRNAs, 117 were located in intergenic regions (intergenic candidate sRNAs) and the remaining 112 were located on the opposite strand of annotated regions (cis-antisense candidate sRNAs). We also classified these 229 candidate sRNAs into three classes (A-C) according to the evidence for transcription initiation (Figure 1B). The details of these candidate sRNAs are summarised in Additional File 4.
It has been reported that small functional proteins of less than 50 amino acids [34, 35] and several regulatory sRNAs encode small functional peptides (so called "dual-function sRNAs") in bacteria [36, 37]. Therefore, we scanned the sequence of each candidate sRNA for an open reading frame (ORF) and corresponding RBS using a computational approach, to determine whether our novel candidate sRNAs could encode small peptides. We determined that 29 candidate sRNAs had both an ORF and a corresponding RBS, suggesting that these sRNAs may encode small peptides (Additional File 5A, Type 1 & Additional File 5B). Seventy-five candidate sRNAs had putative start codons and corresponding RBSs, although they did not have stop codons (Additional File 5A, Type 2). Fifty-five candidate sRNAs had putative ORFs, although they did not have corresponding RBSs (Additional File 5A, Type 3). Collectively, our data suggest that additional small proteins might yet be discovered among these candidate sRNAs. In contrast, the remaining 70 candidate sRNAs did not have ORFs of an appropriate length or RBSs, suggesting that these function as non-coding RNAs (ncRNAs). The details of these candidate sRNAs are summarised in Additional File 4.
Expression of novel sRNAs analysed by northern blot hybridisation
Deep sequencing analysis revealed a large number of novel transcribed regions, identifying 229 intergenic and cis-antisense candidate sRNAs. We used northern blot analysis to confirm the expression and determine the sizes of these sRNAs, with strand-specific probes for the 25 most abundant candidate sRNAs with predicted sigma 70 promoters (ECS001-025; Additional File 4). Sixteen of these were intergenic sRNAs and nine were cis-antisense sRNAs. We used both total RNA and low-molecular-weight RNA from E. coli cells grown to either exponential or stationary phase. We defined the expected size of each sRNA as the length of the continuous genomic region mapped with at least one read.
The 16 most abundant intergenic candidate sRNAs with predicted sigma 70 promoters
Mean no. of
Northern blot band size*b(nt)
The nine most abundant cis-antisense candidate sRNAs with predicted sigma 70 promoters
Mean no. of
Putative target genes*b
Northern blot band size*c (nt)
ymfL & ymfM
e14 prophage; predicted protein
pseudogene; predicted protein
small RNA (unknown function)
23S ribosomal RNA 5-methyluridine methyltransferase
predicted ribosomal protein
glutamate synthase & periplasmic protein
It is well known that cis-antisense sRNAs and their target mRNAs exhibit complete complementarity, and that their base-pairing causes either translational regulation of the target RNA by changing the ratio of its ribosomal-binding activity or degradation of the target RNA by certain ribonuclease(s) . Furthermore, it has been reported that a subset of sRNAs regulates transcription . For example, the ECS003 sRNA is encoded on the cis-antisense strand of the ymfM RBS, and might regulate ymfM translation by masking the RBS. The ECS017 sRNA is encoded on the cis-antisense strand of the sigma 70 promoter (ykgMp) of the ykgM-ykgO operon and the ECS012 sRNA is encoded on the cis-antisense strand of the ybhJ mRNA (protein-coding region), suggesting that these cis-antisense sRNAs might regulate their corresponding sense transcript. In contrast, the ECS011 sRNA is encoded on the cis-antisense strand of the IS128 sRNA. It has previously been reported that the IS128 sRNA overlaps the IS129 sRNA in the same chromosomal region as a cis-antisense sRNA pair . Both the IS128 and IS129 sRNA transcripts (90-95 nt) were detected by northern blot analysis with a double-stranded probe that could not discriminate between the strands. We found that only IS128 is registered in the RegulonDB version 6.3 database. However, in our study, no IS128 sRNA was detected by northern blotting with strand-specific probes (data not shown), although IS129 was detected as part of ECS011. Therefore, we suggest that the IS128 sRNA is an artefact or is only expressed under specific conditions. We also suggest a re-annotation of IS129 as ECS011, a member of the intergenic sRNAs.
Novel sRNAs expressed from prophage regions and their possible roles
The E. coli K12 genome is reported to include 10 prophages . We noted that 27 of 229 candidate sRNAs were expressed from prophage regions (Additional File 4). To date, the isrC, C0293 and dicF sRNAs are known to be expressed from prophage regions. The functions of the isrC and C0293 sRNAs are totally unknown, although the dicF sRNA is known to bind the RBS of the ftsZ mRNA, thus inhibiting ribosome entry. Because ftsZ encodes a protein (Z ring) involved in cell division, dicF sRNA binding causes a defect in E. coli cell division . To investigate our sRNAs expressed from prophage regions, we focused on five sRNAs detected by northern blot analysis (Figures 2 and 3).
Nucleotide conservation of novel candidate sRNAs in bacteria
Construction of the ECSBrowser
In this study, our first priority was to detect novel transcribed regions, so we filtered unreliable reads quite strictly and only selected 3,065,206 high-quality reads from the original 12,473,172 reads (Additional File 1B, steps 1-3). We then discovered 6,079 novel (fragmentary) transcribed regions and selected 229 candidate sRNAs (Figure 1B). It is likely that these candidate sRNAs are actually expressed under the conditions used in this study because the deep sequencing-based approach directly determines the cDNA sequence. Traditionally, the expression of previously known sRNAs was ultimately confirmed by northern blot analysis . Therefore, we also conducted northern blot analysis, detected the expression of 14 of the 25 examined candidate sRNAs (Tables 1 and 2, Figures 2 and 3) and identified 10 of them as novel sRNAs. We cannot explain why the remaining 11 candidate sRNAs could not be detected by northern analysis. It might be necessary to optimise the hybridisation conditions for each sRNA. In addition, we could not intrinsically distinguish the novel sRNAs from stable degradation products in our current study. Therefore, we established criteria for extracting sRNAs (Figure 1B). In this way, we selected 229 candidate sRNAs with computational or experimental evidence of transcription initiation and a length of more than 50 nt (Figure 1B). Because several sRNAs generated by processing from larger precursor RNAs have been reported [14, 15] and fairly small sRNAs (≤ 50 nt) are known in a few cases , the remaining novel transcribed regions might contain additional sRNAs. Collectively, at least 10 previously unreported sRNAs were identified, and the expression of many other candidate sRNAs was implied by this study, although the functions of the vast majority of these candidate sRNAs are not yet clear.
In the last decade, several groups have searched extensively for sRNAs in E. coli using computational and experimental approaches [15–23, 29]. Therefore, we examined the overlaps between the candidate sRNAs in this study and those identified in previous studies. Forty-seven of the 229 novel candidate sRNAs overlapped with previously predicted sRNAs, but the remaining 182 could not be predicted by traditional approaches. The details are summarised in Additional File 4 and the ECSBrowser. Moreover, we suggested that 29 of the 229 novel candidate sRNAs might encode small proteins because they have both an ORF (≥ 4 amino acids [aa]) and corresponding RBS. Interestingly, 10 out of 29 potential small proteins in this study overlapped with the dataset of small proteins that have been computationally predicted based on sequence conservation and RBS models . We show six examples of the candidate sRNAs that encode putative peptides in Additional File 4.
The primary limitation of most traditional methods of identifying sRNAs is a bias toward intergenic regions . Therefore, it is likely that the previous list of cis-antisense sRNAs is far from complete. Currently, 24 of the 80 previously known sRNAs registered in the RegulonDB database are located on the opposite strand from known genes in E. coli. Among them, 12 are reported to act as cis-antisense sRNAs [14, 15, 47–51]. These sRNAs interact with the 5'-UTR [14, 48, 49, 51], the 3'-UTR  or the whole area  of each target mRNA by a base-pairing mechanism. Two sRNAs are known to interact with each other by a cis-antisense mechanism . In this study, we found 112 novel cis-antisense candidate sRNAs (Figure 1B). We noted that 74 of these (such as ECS012; Figure 3) are encoded on the opposite strand from protein-coding regions, but are not in the 5'- or 3'-UTRs of the mRNAs (Additional File 4). The target sites of a few trans-antisense sRNAs are reported to occur in protein-coding regions and these sRNAs cause endonucleolytic mRNA destabilisation rather than the typical inhibition of translational initiation , although most trans-antisense sRNAs interact with the 5'-UTR of the target mRNA by a base-pairing mechanism, blocking translation initiation . These results suggest that novel cis-antisense sRNA species, as well as trans-antisense sRNAs, may regulate a coding gene on the opposite strand.
Three previously identified sRNA genes (isrC, C0293 and dicF) are expressed from prophage regions in E. coli. Our results show that at least five additional sRNAs are expressed from prophage regions, and that the transcript levels of some of these sRNAs are high (Figure 4). Recently, it was reported that prophage regions contain several sRNAs and that the expression levels of some sRNAs are also high in Bacillus subtilis. This may represent a common feature of prokaryotic sRNAs. Four of five prophage-derived sRNAs were modestly activated by temperature stress (Figure 4), although the expression levels of these sRNAs were unchanged under conditions of SOS induction, such as mitomycin C treatment and UV irradiation (data not shown). The SOS response is known to trigger significantly higher induction of prophages .
Previously, it was reported that the majority of intergenic sRNAs in E. coli regulate the expression of genes in response to environmental stresses, and that these integral elements of the stress response are usually very tightly regulated . In contrast, the novel sRNAs in the current study are constantly expressed under our conditions, although the transcript levels of several novel sRNAs were also moderately activated in response to temperature stress, nutrient limitation and growth phase (Figures 2, 3, 4). To obtain more functional information on the novel sRNAs, we generated single-deletion mutants for 13 novel intergenic sRNA regions and conducted growth analysis of these mutants on rich and minimal medium. However, the growth of these deletion mutants was not significantly different from that of wild-type E. coli. Typical data are shown in Additional File 7. We concluded that many of these sRNAs are not necessary for the growth of E. coli, although they are clearly expressed. We are now generating mutants from all the sRNAs for functional analysis (to be published separately). Recently, a deep sequencing analysis of novel regulatory RNAs in E. coli has been reported . The authors identified 10 new sRNAs and 9 new regulatory leader sequences in the intergenic regions of E. coli. We found that four of our sRNAs (ECS026, ECS181, ECS183 and ECS224) overlapped with their new sRNAs, and another four of our sRNAs (ECS028, ECS031, ECS080 and ECS210) overlapped with their new regulatory leader sequences (summarised in Additional File 4). These results also support the concept that there are many as-yet-undiscovered sRNAs or transcribed regions in E. coli.
Finally, we emphasise that the ECSBrowser is the only browser that specifically focuses on the transcribed regions of the E. coli genome (Figure 6). This browser includes information about transcript elements, such as promoters, terminators, RBRs and TSSs, which have been predicted and/or identified at the genome-wide level by various methods. Furthermore, this browser includes information about sRNAs that were predicted in eleven previous papers [7, 9–17, 29]. Using this browser, it is also possible to integrate a number of deep sequencing data that have been obtained under various culture conditions; we intend to use this to compare the dynamic changes in the transcriptome under each specific culture condition. We anticipate that the ECSBrowser will contribute to future transcriptome analysis, including the detection of novel sRNAs.
In previous research, many bacterial sRNAs have been characterised under certain stress conditions such as heat shock, cold shock or oxidative stress. In this paper, we comprehensively analysed sRNAs expressed in normal E. coli growth conditions and discovered 229 novel candidate sRNAs (≥ 50 nt) with computational or experimental evidence of transcription initiation. We also found several novel sRNAs that are highly expressed in prophage regions in the E. coli genome. We conducted an evolutionary conservation analysis of the candidate sRNAs and summarised the data among closely related bacterial strains. Furthermore, we made 13 sRNA deletion mutants and characterised their growth, although the growth of these deletion mutants was not significantly different from that of wild-type E. coli. Finally, we generated a platform (Escherichia coli Small RNA Browser; ECSBrowser) for sRNA research in bacteria. Previously, other studies have reported hundreds of candidate sRNAs encoded in the E. coli genome. However, there is no consensus among them and the wealth of information is confusing. In ECSBrowser, users can access the results of the 11 primary E. coli sRNA prediction papers plus our new data in a uniform database format. We believe this will be invaluable to researchers in this field.
The complete genome sequence of E. coli K12 strain MG1655 (NC_000913.2) was downloaded from the National Center for Biotechnology Information (NCBI) Reference Sequence database http://www.ncbi.nlm.nih.gov/RefSeq/. Information on the annotated regions (CDSs, tRNAs, rRNAs, sRNAs, pseudogenes, phantom genes, operons, promoters, terminators and RBSs) was obtained from RegulonDB version 6.3 http://regulondb.ccg.unam.mx/. In this study, we also predicted additional RBSs with RBSfinder using a window size of 50 bp , and defined them as an annotated region. Data on experimentally determined TSSs and RBRs were obtained from the literature [30–33].
Bacterial strains and growth conditions
E. coli K12 strain BW25113 (wild type)  and the hfq knockout strain from the Keio collection  were used in this study. E. coli was cultured overnight at 37°C under microaerobic conditions in LB medium or modified M63 medium (glucose minimal medium: (K2HPO4 53.6 g, KH2PO4 26.2 g, (NH4)2SO4 10.0 g, FeSO4.7H2O 2.5 mg, D-glucose 4 g, thiamine·HCl 5.0 g and MgSO4.7H2O 50.0 g per litre). The overnight culture was diluted 20-fold in fresh medium and grown at 37°C to the indicated optical densities at 600 nm (OD600). For the heat shock treatment, the cells were grown at 30°C to an OD600 of 0.6 and then at 42°C for 15 min. For the cold shock treatment, the cells were grown at 37°C to an OD600 of 0.6 and then at 15°C for 30 min. These cultured cells were harvested, immediately treated with RNAprotect Bacteria Reagent (Qiagen, Valencia, CA), according to the manufacturer's protocol, and stored at -80°C until RNA extraction.
Two types of RNA samples (total RNA and low-molecular-weight RNA) were prepared from the cultured cells. The total RNA was extracted with an RNeasy Midi Kit (Qiagen), with a slight modification. Instead of using an RNeasy Midi column, we used phenol-chloroform extraction to collect a complete set of RNAs, including low-molecular-weight RNAs. The low-molecular-weight RNA was extracted using the enrichment procedure for small RNAs included in the mirVana™ miRNA Isolation Kit (Ambion, Austin, TX), according to the manufacturer's protocol.
Deep sequencing of low-molecular-weight RNAs
We used the low-molecular-weight RNAs extracted from E. coli K12 strain BW25113 cells grown to an OD600 of 0.76 (late exponential phase) for deep sequencing analysis (Additional File 1A). A cDNA library was prepared and the deep sequencing analysis was performed at the Post Genome Institute (Hongo, Tokyo, Japan). Briefly, the RNA samples were treated with tobacco acid pyrophosphatase to convert the 5' terminus from a triphosphate to a monophosphate. The resulting RNAs were then ligated with 5'- and 3'-specific RNA adapters, reverse transcribed and amplified as described in the manufacturer's protocol. The nucleotide sequences were determined with the Illumina/Solexa 1G Genome Analyzer system (Illumina, San Diego, CA), which generates 35-base reads. The deep sequencing data have been submitted to the DDBJ Sequence Read Archive http://trace.ddbj.nig.ac.jp/dra/index_e.shtml under accession number DRA000221.
Analysis of the deep sequencing data and prediction of the transcribed regions
We used a series of filters to remove unreliable deep sequencing reads. The flow chart for these filtering approaches is summarised in Additional File 1B. In step 1, reads containing the character 'N' (indeterminable A, T, G or C) were discarded from the total reads. In step 2, reads with low-quality base calls were also discarded based on the GERALD module (Illumina). In step 3, we collected the reads that could be mapped to the E. coli genome: (i) the first 32 bases of each read were mapped to the E. coli genome with SOAP v1 , allowing up to two mismatches; and (ii) the remaining 33rd-35th bases were mapped to the E. coli genome by forcing perfect alignments, beginning at the 33rd base and retaining the longest region that could be aligned to the genome. In this study, we defined the "mapped region" as the genomic region that could be mapped by some read(s). To predict novel transcribed regions, we classified the filtered reads into one of five groups (Groups A-E) according to the numbers (single or multiple) and types (annotated or non-annotated) of the mapped regions (Step 4). We eliminated the non-annotated mapped regions in Group E because we could not determine whether these mapped regions were actually expressed. Finally, we defined the "transcribed region" as the minimum continuous mapped region, and all overlapping mapped regions were assembled into novel and known transcribed regions (Step 5). After assembling the mapped regions in step 5, we noticed that some non-annotated mapped regions in Groups B and D were contiguous with the known transcribed regions. Thus, we manually transferred these regions to the "known transcribed regions" category.
Prediction of novel candidate sRNAs
To characterise the transcribed regions and extract novel candidate sRNAs, we first predicted possible promoter and terminator sequences at the genome level using bioinformatics approaches: (i) the consensus sequences of sigma 70 promoters (-35 and -10 box) were predicted with pftools 2.3  using a profile of E. coli sigma 70 promoters ; and (ii) the consensus sequences of rho-independent terminators were predicted with RNAmotif  based on information from previous studies [13, 61].
We calculated the observed number of predicted sigma 70 promoters and rho-independent terminators for each transcribed region. We also calculated the expected number of predicted sigma 70 promoters and rho-independent terminators for random genomic position sets that were generated 100 times using the Blum-Blum-Shub algorithm obtained from CPAN http://search.cpan.org/. The O/E ratio was obtained by dividing the observed numbers by the average expected numbers. As a positive control, the same analysis was also performed for each known gene with an annotated sigma 70 promoter or rho-independent terminator, obtained from RegulonDB version 6.3.
Candidate sRNAs were extracted from all novel transcribed regions based on the following two criteria (a flow chart is shown in Figure 1B). First, the novel transcribed regions with computational or experimental evidence of transcription initiation were extracted (Step 1). We used predicted sigma 70 promoters as the computational evidence. The definition of a possible promoter was that the -10 box consensus sequence of the sigma 70 promoter occurred within 10 bases upstream of a novel transcribed region. We also employed published data on TSSs [30–33] and RBRs  as experimental evidence. We used the data only when the TSS was located within 10 bases upstream or downstream of a novel transcribed region and the RBR was found within 50 bases upstream or 10 bases downstream of a novel transcribed region. Second, the novel transcribed regions that were larger than 50 nt were extracted as novel candidate sRNAs (Step 2). Finally, novel candidate sRNAs were classified according to their evidence for transcription initiation (computational or experimental) and their coding positions in the E. coli genome (intergenic or cis-antisense). Furthermore, we scanned the sequence of each candidate sRNA for a start codon (AUG, UUG or GUG), stop codon (UAA, UGA or UAG), appropriate reading frame (≥ 4 aa) and RBS.
Northern blot analysis
Both total RNA and low-molecular-weight RNA were prepared from E. coli cells grown to an OD600 of either 0.6 (the middle of the exponential phase) or 1.2 (the early stationary phase) in our growth conditions. These RNAs (total RNA, 20 μg per lane; low-molecular-weight RNA, 2 μg per lane) were separated on denaturing 6% polyacrylamide gels containing 8 M urea and transferred onto Hybond-N+ membrane (GE Healthcare, Piscataway, NJ) by electroblotting. The 3' ends of specific oligonucleotide probes (summarised in Additional File 9) were labelled using a Biotin 3' End DNA Labeling Kit (Pierce Biotechnology, Rockford, IL). The membranes were hybridised with these probes in ULTRAhyb-Oligo hybridisation buffer (Ambion) at room temperature. The hybridisation was conducted at 42°C for ECS020 sRNA or 65°C for 5S rRNA. The washing temperature was the same as the respective hybridisation temperature. The non-isotopic blots were visualised using a BrightStar BioDetect Kit (Ambion) using the ECF™ Substrate (GE Healthcare). The images were captured using a Molecular Imager FX Pro (Bio-Rad Laboratories, Hercules, CA).
Nucleotide conservation analysis of both known and novel candidate sRNAs in bacteria
For the conservation analysis, BLAT v.34  was used to compare the nucleotide sequences of the candidate sRNAs in E. coli with those of 1,378 complete bacterial genomes downloaded from the NCBI ftp server (http://ftp.ncbi.nlm.nih.gov; 2011/02/11). An sRNA-encoding gene was considered "conserved" in another organism if the sequence met the following three criteria: (i) an E-value lower than 0.01; (ii) a coverage higher than 70%; and (iii) an identity higher than 70%. Then, the nucleotide sequence conservation score was calculated using the following formula: ((nucleotide match-length)*(nucleotide identity/100))/(nucleotide length of the candidate sRNA). These scores are summarised in Additional File 10. Next, we performed hierarchical clustering from the conservation score of candidate sRNAs using Cluster 3.0 software  with centroid linkage, and visualised the results using Java TreeView 1.1.5 . For phylogenetic tree construction, we first extracted each 16S rRNA sequence from their corresponding genomic sequences by BLAT using the E. coli rrsA gene (16S rRNA) as the query. Then, a multiple alignment of the 16S rRNA sequences was constructed using Clustal W version 1.83  and the phylogenetic tree was depicted using Interactive Tree Of Life .
Construction of the Escherichia coli Small RNA Browser (ECSBrowser)
The database for E. coli sRNA was constructed using the Generic Genome Browser version 2.0 of the Generic Model Organism Database project http://gmod.org/wiki/Main_Page, and named the Escherichia coli Small RNA Browser (ECSBrowser). This browser provides all the information from the current analysis.
References for Additional Files
Twenty-one references for the Additional Files are listed in Additional File 11.
List of abbreviations
Blast-Like Alignment Tool
- E. coli:
National Center for Biotechnology Information
open reading frame
RNA polymerase-binding region
ribosomal binding site
transcriptional start site
Acknowledgements and Funding
We thank Dr. Kosuke Fujishima and Ms. Yuki Takai (Institute for Advanced Biosciences, Keio University, Japan) for critical reading of the manuscript and for technical assistance, respectively. We also thank Drs. Teppei Morita and Hiroji Aiba (Suzuka University of Medical Sciences, Japan), and all the members of the RNA group (Institute for Advanced Biosciences, Keio University, Japan) for their insightful discussions. This work was supported by the Japan Society for the Promotion of Science; a grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan (The 21st Century Centre of Excellence Program, entitled Understanding and Control of Life's Function via Systems Biology); by research funds from the Yamagata Prefectural Government; and by research funds from Tsuruoka City, Japan.
- Huang HY, Chang HY, Chou CH, Tseng CP, Ho SY, Yang CD, Ju YW, Huang HD: sRNAMap: genomic maps for small non-coding RNAs, their regulators and their targets in microbial genomes. Nucleic Acids Res. 2009, 37 (Database): D150-154. 10.1093/nar/gkn852.PubMedView ArticleGoogle Scholar
- Waters LS, Storz G: Regulatory RNAs in bacteria. Cell. 2009, 136 (4): 615-628. 10.1016/j.cell.2009.01.043.PubMedPubMed CentralView ArticleGoogle Scholar
- Wassarman KM: Small RNAs in bacteria: diverse regulators of gene expression in response to environmental changes. Cell. 2002, 109 (2): 141-144. 10.1016/S0092-8674(02)00717-1.PubMedView ArticleGoogle Scholar
- Masse E, Gottesman S: A small RNA regulates the expression of genes involved in iron metabolism in Escherichia coli. Proc Natl Acad Sci USA. 2002, 99 (7): 4620-4625. 10.1073/pnas.032066599.PubMedPubMed CentralView ArticleGoogle Scholar
- Altuvia S, Weinstein-Fischer D, Zhang A, Postow L, Storz G: A small, stable RNA induced by oxidative stress: role as a pleiotropic regulator and antimutator. Cell. 1997, 90 (1): 43-53. 10.1016/S0092-8674(00)80312-8.PubMedView ArticleGoogle Scholar
- Vanderpool CK, Gottesman S: Involvement of a novel transcriptional activator and small RNA in post-transcriptional regulation of the glucose phosphoenolpyruvate phosphotransferase system. Mol Microbiol. 2004, 54 (4): 1076-1089. 10.1111/j.1365-2958.2004.04348.x.PubMedView ArticleGoogle Scholar
- Argaman L, Hershberg R, Vogel J, Bejerano G, Wagner EG, Margalit H, Altuvia S: Novel small RNA-encoding genes in the intergenic regions of Escherichia coli. Curr Biol. 2001, 11 (12): 941-950. 10.1016/S0960-9822(01)00270-6.PubMedView ArticleGoogle Scholar
- Carter RJ, Dubchak I, Holbrook SR: A computational approach to identify genes for functional RNAs in genomic sequences. Nucleic Acids Res. 2001, 29 (19): 3928-3938.PubMedPubMed CentralGoogle Scholar
- Chen S, Lesnik EA, Hall TA, Sampath R, Griffey RH, Ecker DJ, Blyn LB: A bioinformatics based approach to discover small RNA genes in the Escherichia coli genome. Biosystems. 2002, 65 (2-3): 157-177. 10.1016/S0303-2647(02)00013-8.PubMedView ArticleGoogle Scholar
- Rivas E, Klein RJ, Jones TA, Eddy SR: Computational identification of noncoding RNAs in E. coli by comparative genomics. Curr Biol. 2001, 11 (17): 1369-1373. 10.1016/S0960-9822(01)00401-8.PubMedView ArticleGoogle Scholar
- Saetrom P, Sneve R, Kristiansen KI, Snove O, Grunfeld T, Rognes T, Seeberg E: Predicting non-coding RNA genes in Escherichia coli with boosted genetic programming. Nucleic Acids Res. 2005, 33 (10): 3263-3270. 10.1093/nar/gki644.PubMedPubMed CentralView ArticleGoogle Scholar
- Wassarman KM, Repoila F, Rosenow C, Storz G, Gottesman S: Identification of novel small RNAs using comparative genomics and microarrays. Genes Dev. 2001, 15 (13): 1637-1651. 10.1101/gad.901001.PubMedPubMed CentralView ArticleGoogle Scholar
- Yachie N, Numata K, Saito R, Kanai A, Tomita M: Prediction of non-coding and antisense RNA genes in Escherichia coli with Gapped Markov Model. Gene. 2006, 372: 171-181.PubMedView ArticleGoogle Scholar
- Kawano M, Reynolds AA, Miranda-Rios J, Storz G: Detection of 5'- and 3'-UTR-derived small RNAs and cis-encoded antisense RNAs in Escherichia coli. Nucleic Acids Res. 2005, 33 (3): 1040-1050. 10.1093/nar/gki256.PubMedPubMed CentralView ArticleGoogle Scholar
- Vogel J, Bartels V, Tang TH, Churakov G, Slagter-Jager JG, Huttenhofer A, Wagner EG: RNomics in Escherichia coli detects new sRNA species and indicates parallel transcriptional output in bacteria. Nucleic Acids Res. 2003, 31 (22): 6435-6443. 10.1093/nar/gkg867.PubMedPubMed CentralView ArticleGoogle Scholar
- Tjaden B, Saxena RM, Stolyar S, Haynor DR, Kolker E, Rosenow C: Transcriptome analysis of Escherichia coli using high-density oligonucleotide probe arrays. Nucleic Acids Res. 2002, 30 (17): 3732-3738. 10.1093/nar/gkf505.PubMedPubMed CentralView ArticleGoogle Scholar
- Zhang A, Wassarman KM, Rosenow C, Tjaden BC, Storz G, Gottesman S: Global analysis of small RNA and mRNA targets of Hfq. Mol Microbiol. 2003, 50 (4): 1111-1124. 10.1046/j.1365-2958.2003.03734.x.PubMedView ArticleGoogle Scholar
- Gama-Castro S, Jimenez-Jacinto V, Peralta-Gil M, Santos-Zavaleta A, Penaloza-Spinola MI, Contreras-Moreira B, Segura-Salazar J, Muniz-Rascado L, Martinez-Flores I, Salgado H, et al: RegulonDB (version 6.0): gene regulation model of Escherichia coli K-12 beyond transcription, active (experimental) annotated promoters and Textpresso navigation. Nucleic Acids Res. 2008, 36 (Database): D120-124.PubMedView ArticleGoogle Scholar
- Hershberg R, Altuvia S, Margalit H: A survey of small RNA-encoding genes in Escherichia coli. Nucleic Acids Res. 2003, 31 (7): 1813-1820. 10.1093/nar/gkg297.PubMedPubMed CentralView ArticleGoogle Scholar
- Marguerat S, Bahler J: RNA-seq: from technology to biology. Cell Mol Life Sci. 2010, 67 (4): 569-579. 10.1007/s00018-009-0180-6.PubMedView ArticleGoogle Scholar
- van Vliet AH: Next generation sequencing of microbial transcriptomes: challenges and opportunities. FEMS Microbiol Lett. 2010, 302 (1): 1-7. 10.1111/j.1574-6968.2009.01767.x.PubMedView ArticleGoogle Scholar
- Croucher NJ, Thomson NR: Studying bacterial transcriptomes using RNA-seq. Curr Opin Microbiol. 2010, 13 (5): 619-624. 10.1016/j.mib.2010.09.009.PubMedPubMed CentralView ArticleGoogle Scholar
- Perkins TT, Kingsley RA, Fookes MC, Gardner PP, James KD, Yu L, Assefa SA, He M, Croucher NJ, Pickard DJ, et al: A strand-specific RNA-Seq analysis of the transcriptome of the typhoid bacillus Salmonella typhi. PLoS Genet. 2009, 5 (7): e1000569-10.1371/journal.pgen.1000569.PubMedPubMed CentralView ArticleGoogle Scholar
- Sittka A, Lucchini S, Papenfort K, Sharma CM, Rolle K, Binnewies TT, Hinton JC, Vogel J: Deep sequencing analysis of small noncoding RNA and mRNA targets of the global post-transcriptional regulator, Hfq. PLoS Genet. 2008, 4 (8): e1000163-10.1371/journal.pgen.1000163.PubMedPubMed CentralView ArticleGoogle Scholar
- Liu JM, Livny J, Lawrence MS, Kimball MD, Waldor MK, Camilli A: Experimental discovery of sRNAs in Vibrio cholerae by direct cloning, 5S/tRNA depletion and parallel sequencing. Nucleic Acids Res. 2009, 37 (6): e46-10.1093/nar/gkp080.PubMedPubMed CentralView ArticleGoogle Scholar
- Sharma CM, Hoffmann S, Darfeuille F, Reignier J, Findeiss S, Sittka A, Chabas S, Reiche K, Hackermuller J, Reinhardt R, et al: The primary transcriptome of the major human pathogen Helicobacter pylori. Nature. 2010, 464 (7286): 250-255. 10.1038/nature08756.PubMedView ArticleGoogle Scholar
- Yoder-Himes DR, Chain PS, Zhu Y, Wurtzel O, Rubin EM, Tiedje JM, Sorek R: Mapping the Burkholderia cenocepacia niche response via high-throughput sequencing. Proc Natl Acad Sci USA. 2009, 106 (10): 3976-3981. 10.1073/pnas.0813403106.PubMedPubMed CentralView ArticleGoogle Scholar
- Passalacqua KD, Varadarajan A, Ondov BD, Okou DT, Zwick ME, Bergman NH: Structure and complexity of a bacterial transcriptome. J Bacteriol. 2009, 191 (10): 3203-3211. 10.1128/JB.00122-09.PubMedPubMed CentralView ArticleGoogle Scholar
- Raghavan R, Groisman EA, Ochman H: Genome-wide detection of novel regulatory RNAs in E. coli. Genome Res. 2011,Google Scholar
- Cho BK, Zengler K, Qiu Y, Park YS, Knight EM, Barrett CL, Gao Y, Palsson BO: The transcription unit architecture of the Escherichia coli genome. Nat Biotechnol. 2009, 27 (11): 1043-1049. 10.1038/nbt.1582.PubMedView ArticleGoogle Scholar
- Dornenburg JE, Devita AM, Palumbo MJ, Wade JT: Widespread Antisense Transcription in Escherichia coli. MBio. 2010, 1 (1):
- Gama-Castro S, Salgado H, Peralta-Gil M, Santos-Zavaleta A, Muniz-Rascado L, Solano-Lira H, Jimenez-Jacinto V, Weiss V, Garcia-Sotelo JS, Lopez-Fuentes A, et al: RegulonDB version 7.0: transcriptional regulation of Escherichia coli K-12 integrated within genetic sensory response units (Gensor Units). Nucleic Acids Res. 2011, 39 (Database): D98-105. 10.1093/nar/gkq1110.PubMedView ArticleGoogle Scholar
- Mendoza-Vargas A, Olvera L, Olvera M, Grande R, Vega-Alvarado L, Taboada B, Jimenez-Jacinto V, Salgado H, Juarez K, Contreras-Moreira B, et al: Genome-wide identification of transcription start sites, promoters and transcription factor binding sites in E. coli. PLoS One. 2009, 4 (10): e7526-10.1371/journal.pone.0007526.PubMedPubMed CentralView ArticleGoogle Scholar
- Hemm MR, Paul BJ, Miranda-Rios J, Zhang A, Soltanzad N, Storz G: Small stress response proteins in Escherichia coli: proteins missed by classical proteomic studies. J Bacteriol. 2010, 192 (1): 46-58. 10.1128/JB.00872-09.PubMedView ArticleGoogle Scholar
- Hemm MR, Paul BJ, Schneider TD, Storz G, Rudd KE: Small membrane proteins found by comparative genomics and ribosome binding site models. Mol Microbiol. 2008, 70 (6): 1487-1501. 10.1111/j.1365-2958.2008.06495.x.PubMedPubMed CentralView ArticleGoogle Scholar
- Morita T, Aiba H: Small RNAs making a small protein. Proc Natl Acad Sci USA. 2007, 104 (51): 20149-20150. 10.1073/pnas.0710634105.PubMedPubMed CentralView ArticleGoogle Scholar
- Wadler CS, Vanderpool CK: A dual function for a bacterial small RNA: SgrS performs base pairing-dependent regulation and encodes a functional polypeptide. Proc Natl Acad Sci USA. 2007, 104 (51): 20454-20459. 10.1073/pnas.0708102104.PubMedPubMed CentralView ArticleGoogle Scholar
- Storz G, Opdyke JA, Wassarman KM: Regulating bacterial transcription with small RNAs. Cold Spring Harb Symp Quant Biol. 2006, 71: 269-273. 10.1101/sqb.2006.71.033.PubMedView ArticleGoogle Scholar
- Casjens S: Prophages and bacterial genomics: what have we learned so far?. Mol Microbiol. 2003, 49 (2): 277-300. 10.1046/j.1365-2958.2003.03580.x.PubMedView ArticleGoogle Scholar
- Bouche F, Bouche JP: Genetic evidence that DicF, a second division inhibitor encoded by the Escherichia coli dicB operon, is probably RNA. Mol Microbiol. 1989, 3 (7): 991-994. 10.1111/j.1365-2958.1989.tb00249.x.PubMedView ArticleGoogle Scholar
- Tetart F, Bouche JP: Regulation of the expression of the cell-cycle gene ftsZ by DicF antisense RNA. Division does not require a fixed number of FtsZ molecules. Mol Microbiol. 1992, 6 (5): 615-620. 10.1111/j.1365-2958.1992.tb01508.x.PubMedView ArticleGoogle Scholar
- Liu JM, Camilli A: A broadening world of bacterial small RNAs. Curr Opin Microbiol. 2010, 13 (1): 18-23. 10.1016/j.mib.2009.11.004.PubMedView ArticleGoogle Scholar
- Kent WJ: BLAT--the BLAST-like alignment tool. Genome Res. 2002, 12 (4): 656-664.PubMedPubMed CentralView ArticleGoogle Scholar
- Ackermann HW, Audurier A, Berthiaume L, Jones LA, Mayo JA, Vidaver AK: Guidelines for bacteriophage characterization. Adv Virus Res. 1978, 23: 1-24.PubMedView ArticleGoogle Scholar
- Gottesman S: The small RNA regulators of Escherichia coli: roles and mechanisms. Annu Rev Microbiol. 2004, 58: 303-328. 10.1146/annurev.micro.58.030603.123841.PubMedView ArticleGoogle Scholar
- Irnov I, Sharma CM, Vogel J, Winkler WC: Identification of regulatory RNAs in Bacillus subtilis. Nucleic Acids Res. 2010Google Scholar
- Fozo EM, Kawano M, Fontaine F, Kaya Y, Mendieta KS, Jones KL, Ocampo A, Rudd KE, Storz G: Repression of small toxic protein synthesis by the Sib and OhsC small RNAs. Mol Microbiol. 2008, 70 (5): 1076-1093. 10.1111/j.1365-2958.2008.06394.x.PubMedPubMed CentralView ArticleGoogle Scholar
- Kawano M, Aravind L, Storz G: An antisense RNA controls synthesis of an SOS-induced toxin evolved from an antitoxin. Mol Microbiol. 2007, 64 (3): 738-754. 10.1111/j.1365-2958.2007.05688.x.PubMedPubMed CentralView ArticleGoogle Scholar
- Kawano M, Oshima T, Kasai H, Mori H: Molecular characterization of long direct repeat (LDR) sequences expressing a stable mRNA encoding for a 35-amino-acid cell-killing peptide and a cis-encoded small antisense RNA in Escherichia coli. Mol Microbiol. 2002, 45 (2): 333-349. 10.1046/j.1365-2958.2002.03042.x.PubMedView ArticleGoogle Scholar
- Opdyke JA, Kang JG, Storz G: GadY, a small-RNA regulator of acid response genes in Escherichia coli. J Bacteriol. 2004, 186 (20): 6698-6705. 10.1128/JB.186.20.6698-6705.2004.PubMedPubMed CentralView ArticleGoogle Scholar
- Pedersen K, Gerdes K: Multiple hok genes on the chromosome of Escherichia coli. Mol Microbiol. 1999, 32 (5): 1090-1102. 10.1046/j.1365-2958.1999.01431.x.PubMedView ArticleGoogle Scholar
- Pfeiffer V, Papenfort K, Lucchini S, Hinton JC, Vogel J: Coding sequence targeting by MicC RNA reveals bacterial mRNA silencing downstream of translational initiation. Nat Struct Mol Biol. 2009, 16 (8): 840-846. 10.1038/nsmb.1631.PubMedView ArticleGoogle Scholar
- Waldor MK, Friedman DI: Phage regulatory circuits and virulence gene expression. Curr Opin Microbiol. 2005, 8 (4): 459-465. 10.1016/j.mib.2005.06.001.PubMedView ArticleGoogle Scholar
- Suzek BE, Ermolaeva MD, Schreiber M, Salzberg SL: A probabilistic method for identifying start codons in bacterial genomes. Bioinformatics. 2001, 17 (12): 1123-1130. 10.1093/bioinformatics/17.12.1123.PubMedView ArticleGoogle Scholar
- Datsenko KA, Wanner BL: One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci USA. 2000, 97 (12): 6640-6645. 10.1073/pnas.120163297.PubMedPubMed CentralView ArticleGoogle Scholar
- Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL, Mori H: Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2006, 2: 2006 0008-PubMedPubMed CentralView ArticleGoogle Scholar
- Li R, Li Y, Kristiansen K, Wang J: SOAP: short oligonucleotide alignment program. Bioinformatics. 2008, 24 (5): 713-714. 10.1093/bioinformatics/btn025.PubMedView ArticleGoogle Scholar
- Bucher P, Karplus K, Moeri N, Hofmann K: A flexible motif search technique based on generalized profiles. Comput Chem. 1996, 20 (1): 3-23. 10.1016/S0097-8485(96)80003-9.PubMedView ArticleGoogle Scholar
- Mulligan ME, Hawley DK, Entriken R, McClure WR: Escherichia coli promoter sequences predict in vitro RNA polymerase selectivity. Nucleic Acids Res. 1984, 12 (1 Pt 2): 789-800.PubMedPubMed CentralView ArticleGoogle Scholar
- Macke TJ, Ecker DJ, Gutell RR, Gautheret D, Case DA, Sampath R: RNAMotif, an RNA secondary structure definition and search algorithm. Nucleic Acids Res. 2001, 29 (22): 4724-4735. 10.1093/nar/29.22.4724.PubMedPubMed CentralView ArticleGoogle Scholar
- Lesnik EA, Sampath R, Levene HB, Henderson TJ, McNeil JA, Ecker DJ: Prediction of rho-independent transcriptional terminators in Escherichia coli. Nucleic Acids Res. 2001, 29 (17): 3583-3594. 10.1093/nar/29.17.3583.PubMedPubMed CentralView ArticleGoogle Scholar
- de Hoon MJ, Imoto S, Nolan J, Miyano S: Open source clustering software. Bioinformatics. 2004, 20 (9): 1453-1454. 10.1093/bioinformatics/bth078.PubMedView ArticleGoogle Scholar
- Saldanha AJ: Java Treeview--extensible visualization of microarray data. Bioinformatics. 2004, 20 (17): 3246-3248. 10.1093/bioinformatics/bth349.PubMedView ArticleGoogle Scholar
- Thompson JD, Higgins DG, Gibson TJ: CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994, 22 (22): 4673-4680. 10.1093/nar/22.22.4673.PubMedPubMed CentralView ArticleGoogle Scholar
- Letunic I, Bork P: Interactive Tree Of Life (iTOL): an online tool for phylogenetic tree display and annotation. Bioinformatics. 2007, 23 (1): 127-128. 10.1093/bioinformatics/btl529.PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.