INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Various organisms have developed sophisticated signaling systems for eliciting a variety of adaptive responses to their environment. Adaptive response systems of a class present in prokaryotes, lower eukaryotes, and plants each consist of at least two signal transduction proteins and are therefore referred to as two-component regulatory systems (12, 19, 20). A typical two-component regulatory system consists of two types of a sensor kinase and its cognate response regulator, which usually functions as a transcriptional factor. The sensor kinase receives some environmental signal, which induces autophosphorylation of a histidine residue. The phosphoryl group on the histidine residue is then transferred to a conserved aspartate residue on the cognate response regulator, resulting in modulation of the expression of its target genes in response to the environmental signal.

Genomic sequencing of various microorganisms has revealed the presence of many two-component regulatory systems in every species. In Bacillus subtilis, 36 sensor kinases and 35 response regulators have been found, among which each of 30 kinase-regulator pairs resides in an operon on the genome (4, 11). Of the B. subtilis two-component regulatory systems, approximately 10 have been characterized so far as to their roles (4, 14). Very recently, the DesK-DesR (formerly YocF-YocG) system was reported to be involved in thermosensing and signal transduction at low temperatures (1). However, the functions of the other systems remain unknown (4, 14). In a directed gene knockout study of the response regulators of unknown two-component systems, only the YycG-YycF system was found to be essential for growth (5, 8). The other response regulator null mutations did not noticeably affect colony morphology, growth, or sporulation on laboratory media (5; see also the websites of the Japan functional analysis network for B. subtilis [JAFAN {http://bacillus.genome.ad.jp/}] and the Microbial Advanced Database Organization [Micado {http://locus.jouy.inra.fr/micado}], which are provided by the Japanese and European consortia for Functional Analysis of the B. subtilis Genome, respectively). It is probably safe to conclude that most two-component regulation is used to enhance the versatility of the response of an organism to environmental stimuli through regulation of normally unexpressed genes.

Recently, DNA chip technology involving high-density arrays of open reading frame (ORF)-specific DNA fragments has been rapidly developed for transcriptome analysis of various microorganisms whose genome sequences have been determined. With respect to array analysis of B. subtilis two-component systems, the Spo0A and ResD regulons have been investigated (6, 25), and the respective global changes in gene expression in strains with disruptions in the spo0A and resD genes have been revealed.

The B. subtilis DegU, ComA, and PhoP regulons were recently analyzed using a DNA microarray (15). For the analysis, a strategy was used by which the response regulator genes were cloned downstream of the spac promoter (Pspac) in plasmid pDG148 (21) in Escherichia coli, and the constructed plasmids were then introduced into B. subtilis strains with disruptions in the cognate sensor kinase genes; this plasmid contains Pspac as well as lacI, E. coli, and B. subtilis replication origins derived from plasmids pS11 (24), pBR322, and pUB110, respectively (21). As the response regulator genes were placed under the control of Pspac and negatively regulated by LacI, the response regulators could be overproduced upon the addition of isopropyl--D-galactopyranoside (IPTG) to the medium. It was expected that this overproduction of the response regulators might result in altered expression of the target genes in the absence of the environmental signals responsible for their phosphorylation (15). Even if unknown signals that cause autophosphorylation of uncharacterized sensor kinases are present under the growth conditions employed, one can eliminate phosphorylation of the cognate response regulators by using the sensor kinase gene disruptants (15). Thus, one can examine alteration of the expression of the target genes of two-component regulatory systems only upon overproduction of the nonphosphorylated response regulators. Determination of the gene expression ratios of the fluorescence intensities obtained by DNA microarray analysis with RNA from cells grown with IPTG compared to those of cells grown without IPTG allowed for the detection of target gene candidates for the three response regulators, which included most of their known target genes as well as various unknown ones (15). Although some known target genes escaped detection (probably because they were also under the control of other regulatory mechanisms under the experimental conditions used), this strategy was certainly applicable to the detection of the target genes of many uncharacterized two-component systems of B. subtilis for which environmental signals for activation are unknown.

In this work, we comprehensively analyzed 24 B. subtilis two-component regulatory systems by means of the strategy described above. The target gene candidates for the 24 two-component regulatory systems detected in this work, which may be used to identify their target genes, are heuristic but provide valuable information for investigation of the functions of unknown two-component systems as well as further studies of the known ones.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids and their construction. B. subtilis strains deficient in each of the sensor kinases are listed in Table 1. These mutants were constructed by means of integrational disruption of one of the pMUTIN plasmids (22) through a single-crossover event (26) at the 5' site of each of the sensor kinase genes of B. subtilis strain 168 trpC2 (3) by the Japanese and European consortia for Functional Analysis of the B. subtilis Genome (see the websites of JAFAN and Micado for details of their construction). Derivatives of plasmid pDG148 (23) carrying the respective cognate response regulator genes (Table 1) were constructed as follows. DNA regions encompassing the structural genes for the response regulators and their Shine-Dalgarno (SD) sequences were amplified by PCR using the primer pairs listed in Table 1. The amplified DNA fragments were digested with HindIII (or SalI) and SalI (or SphI) and then cloned into plasmid pDG148, which had been digested with the same restriction enzymes, resulting in derivatives of plasmid pDG148 (Table 1). For this DNA manipulation, E. coli strain DH5 cells were grown in liquid or agar Luria-Bertani (LB) medium (16), and the transformants exhibiting ampicillin resistance (50 µg/ml) were selected. After the nucleotide sequences of the cloned DNA regions had been confirmed by sequence determination, B. subtilis strains with disruptions in the sensor kinase genes were transformed to have kanamycin resistance (10 µg/ml) by plasmids carrying the respective cognate response regulator genes.

Growth conditions and RNA isolation. Cells of the strains deficient in sensor kinases, which carried the respective plasmids containing the cognate response regulator genes, were grown overnight in LB medium. After inoculation of the cells into 100 ml of Schaeffer medium (17) or LB medium in two 500-ml Erlenmeyer flasks, the cells were grown at 37°C to the early logarithmic phase. IPTG was added to one of the flasks at a concentration of 1 mM, and the cells were harvested after 2 h. The detailed growth conditions are available at the KEGG Expression Database website (http://www.genome.ad.jp/kegg/expression). Total RNA was isolated from the cells essentially as described previously (27).

Preparation of fluorescently labeled cDNA. The fluorescently labeled cDNA probes used for hybridization to DNA microarrays were prepared by a two-step procedure, as described previously (15).

Hybridization and microarray analysis. DNA microarrays were prepared as described previously (27). The microarrays that we used in this study contained 4,055 protein genes, 45 not being spotted due to a problem with DNA amplification by PCR, as well as 39 calf thymus DNA spots used as negative controls. The hybridization and microarray analysis were performed as described previously (15, 27). The mean Cy3 and Cy5 fluorescence intensities for each spot were calculated, the background being taken as the average of the intensities of the 39 calf thymus DNA spots. After subtracting the background from the intensities of each of the B. subtilis gene spots and normalizing their intensities using the total Cy3 and Cy5 intensities, we calculated the expression ratios (Table 2). To get reliable ratios, we ignored the spots with intensities used as numerators that were less than the background, and replaced the intensities used as denominators with the standard deviation of the average intensity of the negative controls if they were less than the standard deviation.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

B. subtilis two-component regulatory systems subjected to DNA microarray analysis. Out of the 35 response regulators found in the B. subtilis genome, three well-characterized regulons (DegU, ComA, and PhoP) were analyzed previously by means of the strategy described above (15). DNA microarray analysis of the DegU regulon revealed 116 target gene candidates whose expression was affected by DegU overproduction, including known target genes such as aprE, nprE, and ispA (15). Out of the target gene candidates, several (bpr, yukL, ycdA, and murD) were newly found to be under DegU regulation (15). Microarray analysis of the ComA and PhoP regulons revealed 33 and 23 target candidates, respectively. Target candidates for ComA included the known target genes srfAA, srfAB, srfAC, srfAD, and rapA, while those for PhoP included phoA, phoB, ydhF, phoD, tuaB, tuaC, tuaD, tuaE, tuaF, tuaG, tuaH, pstS, pstA, pstC, pstB1, pstB2, glpQ, and PhoR (15). Out of these target candidates, rapF was newly found to be under ComA regulation and yycP and yjdB were newly found to be under PhoR regulation. These results underscored the validity of our strategy of selecting target genes of the two-component systems even if their functions are not known (15).

43

Detection of target gene candidates by DNA microarray analysis. To find target gene candidates for the 24 two-component regulatory systems, cells deficient in each of the sensor kinases, which carried a plasmid pDG148 derivative with its cognate response regulator gene, were grown with and without IPTG, and then their RNAs were subjected to DNA microarray analysis. The B. subtilis gene expression intensities obtained with RNA from cells grown with IPTG were divided by those obtained with RNA from cells grown without IPTG, which yielded their expression ratios. When RNAs were prepared from cells of strain 168 trpC2 bearing plasmid pDG148 that had been grown with and without IPTG, and gene expression profiles were obtained by means of microarray experiments involving them, none of the genes exhibited ratios greater than 1.8-fold upon the addition of IPTG. The DNA microarray data obtained on analysis of the response regulators are available at the KEGG Expression Database website.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Our strategy for detecting target gene candidates for known and unknown two-component regulatory systems by DNA microarray analysis was based on the assumption that overproduction of a nonphosphorylated response regulator in cells deficient in the cognate sensor kinase might affect the expression of its target genes (15). This strategy was previously applied to the three well-characterized systems (DegS-DegU, ComP-ComA, and PhoR-PhoP), and it appeared to work so well for the detection of their target genes (15) that further attempts were not immediately made to examine whether alternate strategies, such as one involving the use of the disruptant of the response regulator gene, also work for target gene detection. As all of the kinase-regulator gene pairs of the two-component systems examined in this study reside in the same operons on the B. subtilis genome (4, 11), our strategy is applicable to the analysis of all of them. We comprehensively analyzed 24 two-component systems by means of this strategy (Table 2). The expression levels of the target gene candidates for each of the two-component regulatory systems listed in Table 2 were affected by at least fourfold upon overexpression of the regulatory protein, providing valuable information for further investigation of the functions of the unknown two-component regulatory systems analyzed.

However, the following limitations of our present comprehensive analysis of the two-component systems should be kept in mind for any further investigation. (i) We were able to detect only the target genes that are solely regulated by each of the examined two-component systems. In other words, we could not detect the fraction of the target genes unexpressed under the present experimental conditions due to additional negative and/or positive regulation mediated by other regulatory proteins. (ii) We detected the target gene candidates for each of the two-component regulatory systems, the expression of which was affected upon overproduction of the nonphosphorylated response regulator. As discussed previously (15), it is conceivable that in some instances the up- and downregulation of gene expression might not reflect the in vivo regulation occurring when an actual signal that causes phosphorylation of the response regulator is transduced. Nevertheless, most known target genes for the three well-characterized systems were regulated in a manner similar to that which has been observed in physiological signal transduction (15). (iii) It is obvious that the target gene candidates listed in Table 2 include not only direct targets to which the response regulators bind, but also secondary target genes, the expression of which is a consequence of alteration of the expression of the direct target genes. To find only direct target genes for the two-component systems, further molecular genetic analyses of the interactions between the cis elements of the target gene candidates and the respective response regulators, involving such methods as gel retardation and DNA footprinting, are necessary. (iv) For the present analysis of two-component regulatory systems, we used the integrants of plasmid pMUTIN (22) into the sensor kinase genes, which carried plasmid pDG148 derivatives containing the cognate response regulator genes. Thus, IPTG addition induced the transcription not only of genes upregulated upon overproduction of the respective response regulators but also of the truncated sensor kinase genes themselves and their downstream genes, which reside in the same operons as those encoding the sensor-regulator pairs. (v) We have attempted to improve the DNA microarray analysis technique in order to reduce cross hybridization between paralogues, which affects the gene expression ratios (15, 27). However, these procedures are unlikely to eliminate this cross hybridization completely. In addition, our DNA microarrays are deficient in 45 genes which could not be amplified by PCR. Therefore, the potential problems of our microarray analysis itself must be taken into consideration upon interpretation of the target gene candidates listed in Table 2.

Some interesting features can be seen in the list of the target gene candidates for various unknown two-component regulatory systems (Table 2). There is a striking overlap between the respective target gene candidates of the YxjM-YxjL and YvqE-YvqC systems; out of 19 target gene candidates of YxjM-YxjL upregulated, 17 are included in those of YvqE-YvqC (Table 2). The target gene candidates for the DesK-DesR system also overlap those of YvfT-YvfU. The four response regulators (YxjL, YvqC, DesR, and YvfU) belong to the NarL family of response regulators (4). YxjL and DesR exhibited the highest similarity (e⁴³ and e⁵⁸) to YvqC and YvfU, respectively, among the 35 B. subtilis response regulators on a BLASTP search (2). Furthermore, the target gene candidates for the YvqE-YvqC system include yxjL as well as yhcZ, the response regulator of the YhcY-YhcZ system, while those of YxjM-YxjL include yvqC and yhcZ (Table 2). As shown in Fig. 1A, we can deduce an interaction between these two-component systems, which might explain this extensive overlapping of the target gene candidates for the YvqE-YvqC and YxiM-YxiL systems. Moreover, as shown in Fig. 1B, the target gene candidates of the DesK-DesR system include yvfU (Table 2), which likely explains the overlapping of the target gene candidates for the DesK-DesR and YvfT-YvfU systems. It would be quite interesting to determine how and when these two-component systems that interact with each other evolved from a common ancestor.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   Probable interactions between some two-component regulatory systems. (A) Interaction between the YvqE-YvqC, YxjM-YxjL, and YhcY-YhcZ systems. (B) Regulation of the YvfT-YvfU system by the DesK-DesR system. DesK is the sensor of a thermosignal (1).

The functions of the target genes for three unknown two-component systems (YdbG-YdbF, YufL-YufM, and YvrG-YvrH) were examined by consulting JAFAN and SubtiList, the B. subtilis genome analysis websites. The target gene candidates for the YdbG-YdbF system included the mcpA, mcpB, flgK, and flgM genes (Table 2), suggesting that this system might be related to chemotaxis. The YufL-YufM system is likely to be involved in competence, because its target gene candidates include most members of the ComK regulon (9), such as comK, nucA, comGG, comGF, comGE, comGD, comGC, comGB, comGA, comEC, comEA, comER, comC, comFC, comFB, and comFA (Table 2). Furthermore, the YvrG-YvrH system appears to be related to cell membrane and cell wall function, because various genes encoding membrane proteins (YfhI, CsbB, YfhO, YkcB, YkcC, and YvrG), transporters (YcbN, YkfD, SunT, and YxdL), and wall-associated proteins (WprA, YocH, DltA, and WapA) are included among its target gene candidates. The target gene candidates for other unknown two-component systems also include more than a few genes whose functions are known or suggested (Table 2), providing valuable clues for unveiling the functions of these unknown two-component systems.