Complete genome sequence of the alkaliphilic bacterium Bacillus halodurans and genomic sequence comparison with Bacillus subtilis

Molecular cloning and sequencing of the gene encoding an exopolygalacturonase of a Bacillus isolate and properties of its recombinant enzyme

An exopolygalacturonase (exo-PGase; EC 3.2.1.82) was found in the culture broth of a Bacillus isolate. The gene encoding the exo-PGase, pehK, was cloned by polymerase chain reaction using mixed primers designed from N-terminal and internal amino acid (aa) sequences of the enzyme (PehK). The determined nucleotide (nt) sequence of pehK revealed a 2940 bp open reading frame (980 aa) that encoded a putative signal sequence (27 aa) and a mature protein (953 aa; 103810 Da). The recombinant enzyme was purified to homogeneity from a culture broth of Bacillus subtilis harboring a pehK-containing plasmid. It had a molecular mass of 105 kDa and a pI value of 5.0. The maximum activity was observed at pH 8 and 55 degrees C in Tris-HCl buffer. The degradation products from polygalacturonic or oligogalacturonic acids were digalacturonic acid, like the exo-PGases, PehX of Erwinia chrysanthemi and PehB of Ralstonia solanacearum. The deduced aa sequence of PehK exhibited moderate homology to those of PehX and PehB with approx. 30% identity for both. High homology was observed in a suitably aligned internal region of the three enzymes (65% identity), and some of the conserved aa residues appeared to form the catalytic core of the enzymes.

Biodegradation of aromatic compounds by Escherichia coli

Although Escherichia coli has long been recognized as the best-understood living organism, little was known about its abilities to use aromatic compounds as sole carbon and energy sources. This review gives an extensive overview of the current knowledge of the catabolism of aromatic compounds by E. coli. After giving a general overview of the aromatic compounds that E. coli strains encounter and mineralize in the different habitats that they colonize, we provide an up-to-date status report on the genes and proteins involved in the catabolism of such compounds, namely, several aromatic acids (phenylacetic acid, 3- and 4-hydroxyphenylacetic acid, phenylpropionic acid, 3-hydroxyphenylpropionic acid, and 3-hydroxycinnamic acid) and amines (phenylethylamine, tyramine, and dopamine). Other enzymatic activities acting on aromatic compounds in E. coli are also reviewed and evaluated. The review also reflects the present impact of genomic research and how the analysis of the whole E. coli genome reveals novel aromatic catabolic functions. Moreover, evolutionary considerations derived from sequence comparisons between the aromatic catabolic clusters of E. coli and homologous clusters from an increasing number of bacteria are also discussed. The recent progress in the understanding of the fundamentals that govern the degradation of aromatic compounds in E. coli makes this bacterium a very useful model system to decipher biochemical, genetic, evolutionary, and ecological aspects of the catabolism of such compounds. In the last part of the review, we discuss strategies and concepts to metabolically engineer E. coli to suit specific needs for biodegradation and biotransformation of aromatics and we provide several examples based on selected studies. Finally, conclusions derived from this review may serve as a lead for future research and applications.

Bacillus subtilis glutamine synthetase controls gene expression through a protein-protein interaction with transcription factor TnrA

Bacillus subtilis TnrA, a global regulator of transcription, responds to nitrogen availability, but the specific signal to which it responds has been elusive. Genetic studies indicate that glutamine synthetase is required for the regulation of TnrA activity in vivo. We report here that the feedback-inhibited form of glutamine synthetase directly interacts with TnrA and blocks the DNA binding activity of TnrA. Mutations in the tnrA gene (tnrA(C)) that allow constitutive high level expression of tnrA-activated genes were isolated and characterized. Feedback-inhibited glutamine synthetase had a significantly reduced ability to block the in vitro DNA binding by three of the TnrA(C) proteins. Thus, glutamine synthetase, an enzyme of central metabolism, directly interacts with and regulates the DNA binding activity of TnrA.

Comparative genomics and bioenergetics

Bacterial and archaeal complete genome sequences have been obtained from a wide range of evolutionary lines, which allows some general conclusions about the phylogenetic distribution and evolution of bioenergetic pathways to be drawn. In particular, I searched in the complete genomes for key enzymes involved in aerobic and anaerobic respiratory pathways and in photosynthesis, and mapped them into an rRNA tree of sequenced species. The phylogenetic distribution of these enzymes is very irregular, and clearly shows the diverse strategies of energy conservation used by prokaryotes. In addition, a thorough phylogenetic analysis of other bioenergetic protein families of wide distribution reveals a complex evolutionary history for the respective genes. A parsimonious explanation for these complex phylogenetic patterns and for the irregular distribution of metabolic pathways is that the last common ancestor of Bacteria and Archaea contained several members of every gene family as a consequence of previous gene or genome duplications, while different patterns of gene loss occurred during the evolution of every gene family. This would imply that the last universal ancestor was a bioenergetically sophisticated organism. Finally, important steps that occurred during the evolution of energetic machineries, such as the early evolution of aerobic respiration and the acquisition of eukaryotic mitochondria from a proteobacterium ancestor, are supported by the analysis of the complete genome sequences.

Analysis of codon usage diversity of bacterial genes with a self-organizing map (SOM): characterization of horizontally transferred genes with emphasis on the E. coli O157 genome

With increases in the amounts of available DNA sequence data, it has become increasingly important to develop tools for comprehensive systematic analysis and comparison of species-specific characteristics of protein-coding sequences for a wide variety of genomes. In the present study, we used a novel neural-network algorithm, a self-organizing map (SOM), to efficiently and comprehensively analyze codon usage in approximately 60,000 genes from 29 bacterial species simultaneously. This SOM makes it possible to cluster and visualize genes of individual species separately at a much higher resolution than can be obtained with principal component analysis. The organization of the SOM can be explained by the genome G+C% and tRNA compositions of the individual species. We used SOM to examine codon usage heterogeneity in the E. coli O157 genome, which contains 'O157-unique segments' (O-islands), and showed that SOM is a powerful tool for characterization of horizontally transferred genes.

Loss-of-function mutations in yjbD result in ClpX- and ClpP-independent competence development of Bacillus subtilis

Mutations in clpP and clpX have pleiotropic effects on growth and developmentally regulated gene expression in Bacillus subtilis. ClpP and ClpX are needed for expression of comK, encoding the competence transcription factor required for the expression of genes within the competence regulon. ClpP, in combination with the ATPase ClpC, degrades the inhibitor of ComK, MecA. Proteolysis of MecA is stimulated by a small protein, ComS, which interacts with MecA. Suppressor mutations (cxs) were isolated that bypass the requirement for clpX for comK expression. These were found also to overcome the defect in comK expression conferred by a clpP mutation. These mutations were identified as missense mutations (cxs-5, -7 and -12) and a nonsense (UAG) codon substitution (cxs-10) in the yjbD coding sequence in a locus linked to mecA. That a yjbD disruption confers the cxs phenotype, together with its complementation by an ectopically expressed copy of yjbD, indicated that the suppressor alleles bear recessive, loss-of-function mutations of yjbD. ClpP- and ClpX-independent comK expression rendered by inactivation of yjbD was still medium-dependent and required ComS. MecA levels in a clpP-yjbD mutant were lower that those of clpP mutant cells and ComK protein concentration in the clpP mutant was restored to wild-type levels by the yjbD mutation. Consequently, the yjbD mutation bypasses the defect in competence development conferred by clpP and clpX. YjbD protein is barely detectable in wild-type cells, but is present in large amounts in the clpP mutant cells. The results suggest that the role of ClpP in competence development is to degrade YjbD protein so that ComS can productively interact with the MecA-ClpC-ComK complex. Alternatively, the result could suggest that YjbD has a negative effect on regulated proteolysis and that MecA is degraded independently of ClpP when YjbD is absent.

Pentapeptide regulation of aspartyl-phosphate phosphatases

Aspartyl-phosphate phosphatases are integral components of the phosphorelay signal transduction system for sporulation initiation in Bacillus subtilis. The Rap and Spo0E families of protein phosphatases specifically dephosphorylate the sporulation response regulators Spo0F and Spo0A, respectively. The phosphatases interpret regulatory signals antithetical to sporulation and the Rap phosphatases are subject to inactivation by specific pentapeptides generated from an inactive peptide precursor. Additional regulatory signals are brought about by the complex activation circuit that generates the Phr pentapeptide inhibitors of Rap phosphatases. Phr peptide's recognition of the Rap phosphatase targets is remarkably specific. Specificity is dictated by the amino acid sequence of the pentapeptide. The identification of tetratricopeptide repeats in the Rap proteins may explain the mechanism by which Phr peptides bind to and inhibit the activity of Rap phosphatases.

Ongoing evolution of strand composition in bacterial genomes

We tried to identify the substitutions involved in the establishment of replication strand bias, which has been recognized as an important evolutionary factor in the evolution of bacterial genomes. First, we analyzed the composition asymmetry of 28 complete bacterial genomes and used it to test the possibility that asymmetric deamination of cytosine might be at the origin of the bias. The model showed significant correlation to the data but left unexplained a significant portion of the variance and indicated a systematic underestimation of GC skews in comparison with TA skews. Second, we analyzed the substitutions acting on the genes from five fully sequenced Chlamydia genomes that had not suffered strand switch since speciation. This analysis showed that substitutions were not at equilibrium in Chlamydia trachomatis or in C. muridarum and that strand bias is still an on-going process in these genes. Third, we identified substitutions involved in the adaptation of genes that had switched strands after speciation. These genes adapted quickly to the skewed composition of the new strand, mostly due to C-->T, A-->G, and C-->G asymmetric substitutions. This observation was reinforced by the analysis of genes that switched strands after divergence between Bacillus subtilis and B. halodurans. Finally, we propose a more extended model based on the analysis of the substitution asymmetries of CHLAMYDIA: This model fits well with the data provided by bacterial genomes presenting strong strand bias.

A cofactor-dependent phosphoglycerate mutase homolog from Bacillus stearothermophilus is actually a broad specificity phosphatase

The distribution of phosphoglycerate mutase (PGM) activity in bacteria is complex, with some organisms possessing both a cofactor-dependent and a cofactor-independent PGM and others having only one of these enzymes. Although Bacillus species contain only a cofactor-independent PGM, genes homologous to those encoding cofactor-dependent PGMs have been detected in this group of bacteria, but in at least one case the encoded protein lacks significant PGM activity. Here we apply sequence analysis, molecular modeling, and enzymatic assays to the cofactor-dependent PGM homologs from B. stearothermophilus and B. subtilis, and show that these enzymes are phosphatases with broad substrate specificity. Homologs from other gram-positive bacteria are also likely to possess phosphatase activity. These studies clearly show that the exploration of genomic sequences through three-dimensional modeling is capable of producing useful predictions regarding function. However, significant methodological improvements will be needed before such analysis can be carried out automatically.

The tripartite ATP-independent periplasmic (TRAP) transporters of bacteria and archaea

Until recently, extracytoplasmic solute receptor (ESR)-dependent uptake systems were invariably found to possess a conserved ATP-binding protein (the ATP-binding cassette protein or ABC protein), which couples ATP hydrolysis to the translocation of the solute across the cytoplasmic membrane. While it is clear that this class of ABC transporter is ubiquitous in prokaryotes, it is now firmly established that other, unrelated types of membrane transport systems exist which also have ESR components. These systems have been designated tripartite ATP-independent periplasmic (TRAP) transporters, and they form a distinct class of ESR-dependent secondary transporters where the driving force for solute accumulation is an electrochemical ion gradient and not ATP hydrolysis. Currently, the most well characterised TRAP transporter at the functional and molecular level is the high-affinity C4-dicarboxylate transport (Dct) system from Rhodobacter capsulatus. This consists of three proteins; an ESR (DctP) and small (DctQ) and large (DctM) integral membrane proteins. The characteristics of this system are discussed in detail. Homologues of the R. capsulatus DctPQM proteins are present in a diverse range of prokaryotes, both bacteria and archaea, but not in eukaryotes. The deduced structures and possible functions of these homologous systems are described. In addition to the DctP family, other types of ESRs can be associated with TRAP transporters. A conserved family of immunogenic extracytoplasmic proteins is shown to be invariably associated with TRAP systems that contain a large DctQM fusion protein. All of the currently known archaeal systems are of this type. It is concluded that TRAP transporters are a widespread and ancient type of solute uptake system that transport a potentially diverse range of solutes and most likely evolved by the addition of auxiliary proteins to a single secondary transporter.

The acid-stress response in Lactobacillus sanfranciscensis CB1

Lactobacillus sanfranciscensis CB1, an important sourdough lactic acid bacterium, can withstand low pH after initial exposure to sublethal acidic conditions. The sensitivity to low pH varied according to the type of acid used. Treatment of LB: sanfranciscensis CB1 with chloramphenicol during acid adaptation almost completely eliminated the protective effect, suggesting that induction of protein synthesis was required for the acid-tolerance response. Two constitutively acid-tolerant mutants, CB1-5R and CB1-7R, were isolated using natural selection techniques after sequential exposure to lactic acid (pH 3.2). Two-dimensional gel electrophoresis analysis of protein expression by non-adapted, acid-adapted and acid-tolerant mutant cells of LB: sanfranciscensis showed changes in the levels of 63 proteins. While some of the modifications were common to the acid-adapted and acid-tolerant mutant cells, several differences, especially regarding the induced proteins, were determined. The two mutants showed a very similar level of protein expression. Antibodies were used to identify heat-shock proteins DnaJ, DnaK, GroES and GrpE. Only GrpE showed an increased level of expression in the acid-adapted and acid-tolerant mutants as compared with non-adapted cells. The N-terminal sequence was determined for two proteins, one induced in both the acid-adapted and mutant cells and the other showing the highest induction factor of those proteins specifically induced in the acid-adapted cells. This second protein has 60% identity with the N-terminal portion of YhaH, a transmembrane protein of Bacillus subtilis, which has 54 and 47% homology with stress proteins identified in Listeria monocytogenes and Bacillus halodurans. The constitutively acid-tolerant mutants showed other different phenotypic features compared to the parental strain: (i) the aminopeptidase activity of CB1-5R decreased and that of CB1-7R markedly increased, especially in acid conditions; (ii) the growth in culture medium at 10 degrees C and in the presence of 5% NaCl was greater (the same was found for acid-adapted cells); and (iii) the acidification rate during sourdough fermentation in acid conditions was faster and greater.

Multiple genes for the last step of proline biosynthesis in Bacillus subtilis

The complete Bacillus subtilis genome contains four genes (proG, proH, proI, and comER) with the potential to encode Delta(1)-pyrroline-5-carboxylate reductase, a proline biosynthetic enzyme. Simultaneous defects in three of these genes (proG, proH, and proI) were required to confer proline auxotrophy, indicating that the products of these genes are mostly interchangeable with respect to the last step in proline biosynthesis.

Alkaline shock induces the Bacillus subtilis sigma(W) regulon

When confronted with a stress factor, bacteria react with a specific stress response, a genetically encoded programme resulting in the transiently enhanced expression of a subset of genes. One of these stress factors is a sudden increase in the external pH. As a first step to understand the response of Bacillus subtilis cells towards an alkali shock at the transcriptional level, we attempted to identify alkali-inducible genes using the DNA macroarray technique. To define the appropriate challenging conditions, we used the ydjF gene, the orthologue of the Escherichia coli pspA, as a model gene for an alkali-inducible gene. Hybridization of 33P-labelled cDNA to a DNA macroarray revealed induction of more than 80 genes by a sudden increase in the external pH value from 6.3 to 8.9. It was discovered that a large subset of these genes belong to the recently described sigmaW regulon, which was confirmed by the analysis of a sigW knockout. A comparison of B. subtilis wild type with the congenic sigW knockout also led to the discovery of new members of the sigmaW regulon. In addition, we found several genes clearly not belonging to that regulon. This analysis represents the first report of an extracellular stimulus inducing the sigmaW regulon.

Identification and distribution of new insertion sequences in the genome of alkaliphilic Bacillus halodurans C-125

Fifteen kinds of new insertion sequences (ISs), IS641 to IS643, IS650 to IS658, IS660, IS662, and IS663, and a group II intron (Bh.Int) were identified in the 4,202,352-bp genome of alkaliphilic Bacillus halodurans C-125. Out of 120 ISs identified in the C-125 genome, 29 were truncated, indicating the occurrence of internal rearrangements of the genome. The ISs other than IS650, IS653, IS660, and IS663 generated a 2- to 9-bp duplication of the target site sequence, and the ISs other than IS650, IS653, and IS657 carry 14- to 64-bp inverted repeats. Sequence analysis revealed that six kinds of ISs (IS642, IS643, IS654, IS655, IS657, and IS658) belong to a separate IS family (IS630, IS21, IS256, IS3, IS200/IS605, and IS30, respectively) as a new member. Also, IS651 and IS652 were characterized as new members of the ISL3 family. Significant similarity was found between the transposase (Tpase) sequences between IS650 and IS653 (78.2%), IS651 and IS652 (56.3%), IS656 and IS662 (71.0%), and IS660 and IS663 (44.5%), but the others showed no similarity to one another. Tpases in 28 members of IS651 in the C-125 genome were found to have become diversified. Most of the IS elements widely distributed throughout the genome were inserted in noncoding regions, although some genes, such as those coding for an ATP-binding cassette transporter/permease, a response regulator, and L-indole 2-dehydrogenase, have been mutated through the insertion of IS elements. It is evident, however, that not all IS elements have transposed and caused rearrangements of the genome in the past 17 years during which strain C-125 was subcultured under neutral and alkaline conditions.

Multiple mobile promoter regions for the rare carbapenem resistance gene of Bacteroides fragilis

Two novel insertion sequences (IS), IS1187 and IS1188, are described upstream from the carbapenem resistance gene cfiA in strains of Bacteroides fragilis. Mapping, with the RACE procedure, of transcription start sites of cfiA in these and two other previously reported IS showed that transcription of this rarely encountered gene is initiated close to a variety of B. fragilis consensus promoter sequences, as recently defined (D. P. Bayley, E. R. Rocha, and C. J. Smith, FEMS Microbiol. Lett. 193:149-154, 2000). In the cases of IS1186 and IS1188, these sequences overlap with putative Esigma(70) promoter sequences, while in IS942 and IS1187 such sequences can be observed either upstream or downstream of the B. fragilis promoters.

Microbial genomes: dealing with diversity

We have now complete genome sequences of several pairs of closely related prokaryotes (conspecific strains or congeneric species). Surprisingly, even strains of the same species can differ by as much as 20% in gene content. Conceptual and methodological approaches for dealing with such diversity are now being developed, and should transform microbial genomics.

Organization of the genes involved in dimethylglycine and sarcosine degradation in Arthrobacter spp.: implications for glycine betaine catabolism

The nucleotide sequences of two cloned DNA fragments containing the structural genes of heterotetrameric sarcosine oxidase (soxBDAG) and dimethylglycine dehydrogenase (dmg) from Arthrobater spp. 1-IN and Arthrobacter globiformis, respectively, have been determined. Open reading frames were identified in the soxBDAG operon corresponding to the four subunits of heterotetrameric sarcosine oxidase by comparison with the N-terminal amino-acid sequences and the subunit relative molecular masses of the purified enzyme. Alignment of the deduced sarcosine oxidase amino-acid sequence with amino-acid sequences of functionally related proteins indicated that the arthrobacterial enzyme is highly homologous to sarcosine oxidase from Corynebacterium P-1. Deletion and expression analysis, and alignment of the deduced amino-acid sequence of the dmg gene, showed that dmg encodes a novel dimethylglycine oxidase, which is related to eukaryotic dimethylglycine dehydrogenase, and contains nucleotide-binding, flavinylation and folate-binding motifs. The recombinant dimethylglycine oxidase was purified to homogeneity and characterized. The DNA located upstream and downstream of both the soxBDAG and dmg genes is predicted to encode enzymes involved in the tetrahydrofolate-dependent assimilation of methyl groups. Based on the sequence analysis reported herein, pathways are proposed for glycine betaine catabolism in Arthrobacter species, which involve the identified folate-dependent enzymes.

Identification of the cobalamin-dependent methionine synthase gene, metH, in Vibrio fischeri ATCC 7744 by sequencing using genomic DNA as a template

To confirm the presence of cobalamin-dependent methionine synthase (CDMS) in luminous bacteria, which is a prerequisite for the substantiation of our proposals on the physiological function of the lux operon, we identified the CDMS gene (metH) in Vibrio fischeri ATCC 7744. Two partial metH sequences, one located near the 5'-terminus of the gene and the other near the 3'-terminus, were sequenced by a PCR based method. To design a new set of PCR primers located on the two flanking regions of the gene, the genomic DNA was sequenced by SUGDAT method (sequencing using genomic DNA as a template) upstream or downstream from the respective partial gene sequences. Subsequently a 4.2 kb DNA fragment containing the whole metH was amplified by PCR and sequenced. The number of amino acid residues comprising the protein (1226 amino acids) was comparable to those of known CDMSs. The deduced amino acid sequence showed 85, 74, 55, 31, 30, 52, or 52% identity with that of Vibrio cholerae, Escherichia coli, Deinococcus radiodurans, Synechocystis PCC6803, Mycobacterium tuberculosis, Caenorhabditis elegans or Homo sapiens, respectively. All the predicted amino acid residues for the binding of cobalamin and S-adenosylmethionine were conserved. In the regulatory region of the V. fischeri metH, the binding site of the met repressor, MetJ, was present, although the site is atypically not present in E. coli metH or Salmonella typhimurium metH. It was shown that nucleotide sequences, even long ones, can be determined without a cloning step, if only parts of the DNA fragment to be sequenced are amplified by PCR.

Nucleotide and deduced amino acid sequences of an alkaline pullulanase from the alkaliphilic bacterium Bacillus sp. KSM-1876

The nucleotide sequence of an alkaline pullulanase-encoding gene from alkaliphilic Bacillus sp. strain KSM-1876 was determined. The open reading frame of the gene encoded 1142 amino acids with a calculated molecular mass of 128739 Da. The alkaline pullulanase showed very limited homology (<32% identity) to previously reported debranching enzymes from prokaryotes and eukaryotes. It contained unique tandem repeats in both the N-terminal and the C-terminal regions.

Prediction of co-regulated genes in Bacillus subtilis on the basis of upstream elements conserved across three closely related species

BACKGROUND

Identification of co-regulated genes is essential for elucidating transcriptional regulatory networks and the function of uncharacterized genes. Although co-regulated genes should have at least one common sequence element, it is generally difficult to identify these genes from the presence of this element because it is very easily obscured by noise. To overcome this problem, we used conserved information from three closely related species: Bacillus subtilis, B. halodurans and B. stearothermophilus.

RESULTS

Even though such species have a limited number of clearly orthologous genes, we obtained 1,884 phylogenetically conserved elements from the upstream intergenic regions of 1,568 B. subtilis genes. Similarity between these elements was used to cluster these genes. No other a priori knowledge on genes and elements was used. We could identify some genes known or suggested to be regulated by a common transcription factor as well as genes regulated by a common attenuation effector.

CONCLUSIONS

We confirmed that our method generates relatively few false positives in clusters with higher scores and that general elements such as -35/-10 boxes and Shine-Dalgarno sequence are not major obstacles. Moreover, we identified some plausible additional members of groups of known co-regulated genes. Thus, our approach is promising for exploring potentially co-regulated genes.

Resistance of spores of Bacillus species to ultraviolet light

Dormant spores of the various Bacillus species, including B. subtilis, are 5 to 50 times more resistant to UV radiation than are the corresponding growing cells. This elevated spore UV resistance is due to: a) the photochemistry of DNA within spores, as UV generates few if any cyclobutane dimers, but rather a photoproduct (Fig. 1) called spore photoproduct (SP; 5-thyminyl-5,6-dihydrothymine); and b) DNA repair, in particular SP-specific repair, during spore germination. The novel UV photochemistry of spore DNA is largely due to its saturation with a group of small, acid-soluble proteins (SASP), which are unique to spores and whose binding alters the DNA conformation and thus its photochemistry. SP-specific repair is also unique to spores and is carried out by a light-independent SP-lyase, an iron-sulfur protein that utilizes S-adenosylmethionine to catalyze SP monomerization without DNA backbone cleavage.

DBTBS: a database of Bacillus subtilis promoters and transcription factors

With the completion of the determination of its entire genome sequence, one of the next major targets of Bacillus subtilis genomics is to clarify the whole gene regulatory network. To this end, the results of systematic experiments should be compared with the rich source of individual experimental results accumulated so far. Thus, we constructed a database of the upstream regulatory information of B.subtilis (DBTBS). The current version was constructed by surveying 291 references and contains information on 90 binding factors and 403 promoters. For each promoter, all of its known cis-elements are listed according to their positions, while these cis-elements are aligned to illustrate their consensus sequence for each transcription factor. All probable transcription factors coded in the genome were classified with the Pfam motifs. Using this database, we compared the character of B.subtilis promoters with that of Escherichia coli promoters. Our database is accessible at http://elmo.ims.u-tokyo.ac.jp/dbtbs/.