RemA is a DNA-binding protein that activates biofilm matrix gene expression in Bacillus subtilis

Summary

Biofilm formation in Bacillus subtilis requires expression of the eps and tapA-sipW-tasA operons to synthesize the extracellular matrix components, extracellular polysaccharide and TasA amyloid proteins, respectively. Expression of both operons is inhibited by the DNA-binding protein master regulator of biofilm formation SinR and activated by the protein RemA. Here we show that RemA is a DNA-binding protein that binds to multiple sites upstream of the promoters of both operons and is both necessary and sufficient for transcriptional activation in vivo and in vitro. We further show that SinR negatively regulates eps operon expression by occluding RemA binding and thus for the P_eps promoter SinR functions as an anti-activator. Finally, transcriptional profiling indicated that RemA was primarily a regulator of the extracellular matrix genes, but it also activated genes involved in osmoprotection, leading to the identification of another direct target, the opuA operon.

Introduction

Many types of bacteria are capable of forming multicellular communities known as biofilms (O’Toole et al., 2000; Branda et al., 2005; Kolter and Greenberg, 2006). Cells within a biofilm cohere to one another and adhere to surfaces by producing an extracellular matrix often made of a species-specific composite of polysaccharides, proteins and nucleic acids (Flemming and Wingender, 2010). In addition, bacteria often inhibit flagellar motility concomitant with biofilm formation, perhaps to permit the formation of nucleation centres for aggregates or to stabilize aggregates once they have formed (Kolter and Greenberg, 2006). Biofilms are important factors in biofouling and pathogenesis. Understanding the complex regulation of extracellular matrix components and the inhibition of motility is important to combat biofilm formation in industrial and clinical settings.

One organism capable of forming biofilms is the Gram-positive model organism Bacillus subtilis. Under laboratory conditions, B. subtilis biofilms form structured pellicles at the liquid–air interface or colonies with complex architecture at the solid–air interface (Branda et al., 2001). The biofilms consist of non-motile cells encased in an extracellular matrix composed of an extracellular polysaccharide (EPS) tethered by the amyloid protein TasA and made hydrophobic by the amphiphilic protein BslA (Branda et al., 2006; Guttenplan et al., 2010; Romero et al., 2010; Ostrowski et al., 2011; Kobayashi and Iwano, 2012). The TasA protein is encoded within the tapA-sipW-tasA operon that also encodes the TapA protein and SipW signal peptidase, which are required for TasA anchoring and processing respectively (Branda et al., 2004; Chu et al., 2006; Romero et al., 2011; Terra et al., 2012). The EPS is synthesized by gene products resembling sugar-related enzymes encoded within the 15 gene eps operon (Branda et al., 2001; Kearns et al., 2005; Guttenplan et al., 2010). Among the products encoded within the eps operon is the bifunctional glycosyltransferase EpsE that not only participates enzymatically in EPS synthesis but also inhibits motility by disengaging the flagellar rotor from the proton channels that power rotation (Blair et al., 2008; Guttenplan et al., 2010). Thus, regulation of the eps operon is sufficient to promote the motility-to-biofilm transition as eps expression results in biofilm matrix production and motility inhibition.

The eps and tapA-sipW-tasA operons are expressed by σ^A RNAP directed by the P_eps and P_tapA promoters respectively. Both the P_eps and P_tapA promoters are co-ordinately inhibited by SinR, a DNA-binding protein that serves as the master regulator for biofilm formation (Gaur et al., 1991; Kearns et al., 2005; Chu et al., 2006; Colledge et al., 2011). During biofilm initiation, SinR inhibition is relieved in a subpopulation of cells when the concentration of SinI exceeds a theoretical threshold, binds to SinR and prevents SinR binding to DNA (Bai et al., 1993; Kearns et al., 2005; Chai et al., 2008). SinI antagonism of SinR also turns on slrR, a gene upstream and oppositely oriented from the eps operon, that encodes SlrR, a secondary SinR antagonist that further drives the cells into a biofilm forming state (Chu et al., 2008; Chai et al., 2010). Cells mutated for sinR are non-motile because of the constitutive expression of EPS synthesis and the motility inhibitor flagellar clutch protein EpsE (Blair et al., 2008; Guttenplan et al., 2010).A genetic screen to identify mutations that restored motility to a sinR epsH mutant identified loss of function mutations that disrupted EpsE and two other proteins, RemA and RemB, that regulated the extracellular matrix by activating the eps, tapA and slrR genes (Blair et al., 2008; Winkelman et al., 2009).

Here we elucidate the mechanism by which RemA regulates the production of the extracellular matrix. Despite the fact that RemA does not appear to encode a known DNA-binding motif, it nonetheless bound specifically to and activated transcription from the P_eps and P_tapA promoters in vitro. RemA bound to multiple sites upstream of both promoters and may cause bending of the DNA. Transcriptome analysis suggested that few genes besides those involved in the synthesis of the biofilm matrix were strongly regulated by RemA but we show that the opuA operon, required for glycine betaine uptake and osmoprotection, is also a direct RemA target. Finally, we show that SinR inhibits the eps operon by competing with RemA for binding and thereby mechanistically functions as an anti-activator. RemA and SinR appear to form two parallel but converging pathways suggesting that EPS and TasA activation in B. subtilis may be governed by at least two independent environmental inputs.

Results

RemA directly activates transcription of the eps operon

The P_eps promoter of the eps operon, necessary to synthesize the biofilm EPS, is inhibited by SinR and activated by RemA (Kearns et al., 2005; Chu et al., 2006; Winkelman et al., 2009). Expression from a reporter in which the P_eps promoter was fused to the lacZ gene encoding β-galactosidase (P_eps–lacZ) increased 50-fold in the absence of the inhibitor SinR and was abolished in the absence of the SinR antagonist SinI or the activator protein RemA (Fig. 1, grey bars) (Winkelman et al., 2009). SinR, however, binds to the P_eps promoter from positions −78 to −52 relative to the transcriptional start site, which is too far upstream to occlude RNAP access, and the mechanism of RemA activation is unknown (Kearns et al., 2005). To explore P_eps regulation further, an ‘improved’ promoter was generated in which the third position in the P_eps −35 hexamer was mutated towards consensus for σ^A RNAP (TTTTAA to TTGTAA) and fused to lacZ. Expression from the improved reporter (P_eps^improved–lacZ) was increased in all genetic backgrounds tested and expression was restored to near wild-type levels even in the absence of SinI and RemA (Fig. 1, black bars). We conclude that improving the P_eps promoter closer to consensus relieves both the inhibition by SinR and the requirement for RemA.

Fig. 1.

RemA activates P_eps expression in vivo. β-Galactosidase activity of P_eps–lacZ (grey bars) and P_eps^improved–lacZ (black bars) are presented for the indicated genotypes. Error bars are the standard deviations of data from three replicates. Raw data are presented in Table S3 in Supporting information. The following strains were used to generate this figure: wild type (DS1882, DS9468), sinR (DS2609, DS9467), remA (DS2913, DS9469), sinR remA (DS2911, DS9466) and sinI (DS444, DK445).

Mutation of RemA was epistatic to mutation of SinR with respect to expression from the P_eps–lacZ reporter (Fig. 1, Winkelman et al., 2009). One way to explain the epistasis is if RemA was a transcriptional activator and SinR served as an anti-activator. To determine whether RemA activates transcription directly, we conducted in vitro transcription reactions with σ^A-containing RNAP holoenzyme, in the presence and absence of a purified RemA-maltose-binding protein fusion (RemA–MBP). Transcription from a plasmid template containing the P_eps or P_eps^improved promoter was strongly enhanced in the presence of RemA–MBP (Fig. 2A). Activation was specific for RemA–MBP, as a control protein fusion of β-galactosidase to maltose-binding protein (LacZ–MBP) failed to increase expression from the P_eps promoter (Fig. 2A).Addition of SinR abolished RemA-dependent activation but did not abolish basal expression from the P_eps^improved promoter (Fig. 2A). We conclude that RemA acts as a transcriptional activator of P_eps both in vivo and in vitro. We further conclude that SinR does not formally inhibit the P_eps core promoter but rather acts as a RemA anti-activator.

Fig. 2.

RemA activates transcription from specific promoters in vitro. In vitro transcription reactions were carried out in 100 mM KCl with 20 nM σ^A containing RNAP holoenzyme and 50 ng of supercoiled plasmid DNA with the indicated promoters cloned upstream of tandem transcription terminators.

A. RemA–MBP (700 nM), LacZ–MBP (700 nM) or SinR (400 nM) were added to in vitro transcription reactions containing the followingpromoters cloned upstream of tandem transcription terminators: P_eps (−452 to +35), P_eps^improved (−452 to +35 with the −35 element changed from TTTAAA to TTGAAA), lacUV5 (−60 to +40), P_tapA (−168 to +65) or P_opuA (−173 to +26).

B. In vitro transcription reactions were carried out with supercoiled plasmid containing P_eps promoter DNA with the indicated number of nucleotides upstream of the transcription start site cloned upstream of tandem transcription terminators. The downstream promoter end-point was +35 in each case. RemA–MBP (700 nM) or RemA–MBP storage buffer were added to each reaction. Relative transcription was determined by dividing the intensity in each lane by the intensity in lane 2. The indicated values for relative transcription are the averages from three different experiments.

C. In vitro transcription reactions were carried out with supercoiled plasmid containing 700 nM RemA–MBP and either WT P_eps (−452 to +35)or P_eps promoter fragments containing the indicated Δremsite34 or Δremsite56 deletions.

To determine whether RemA-dependent transcriptional activation was promoter-specific, we tested whether RemA–MBP could activate expression from a plasmid containing the heterologous Escherichia coli promoter, lacUV5. Expression from lacUV5 was unaffected by RemA–MBP, indicating that activation by RemA was specific to sequences found in the P_eps promoter (Fig. 2A). The promoter sequence determinants for activation were identified by creating nested 5' deletions upstream of P_eps in the context of the plasmid template used for the in vitro transcription reactions. There was no difference in the level of activation by RemA–MBP when there were 452 bp, 176 bp or 159 bp of P_eps sequence upstream of the eps transcriptional start site. There was a decrease in the ability of RemA–MBP to activate transcription, however, when only 121 bp of upstream sequence was present and a complete loss of RemA-dependent activation when only 107 bp of upstream sequence was present (Fig. 2B). We conclude that RemA requires, at least in part, a specific DNA sequence between 107 and 159 bp upstream of the P_eps transcription start site for full activation by RemA.

Most transcription factors specifically regulate expression of their target genes by binding to specific DNA sequences in or near the target promoters (Browning and Busby, 2004). To directly test if RemA specifically bound to DNA we carried out electrophoretic mobility shift assays (EMSAs) with purified RemA–MBP and radiolabelled DNA containing P_eps promoter fragments. The migration of the P_eps promoter fragment was shifted by RemA–MBP concentrations of 0.5 µM or greater (Fig. 3A). In contrast, RemA–MBP did not shift a control fragment containing the rrnB P1 promoter at any concentration tested (≤ 0.9 µM; Fig. 3B). We conclude that RemA is a DNA-binding protein that binds to sequences upstream of specific promoters. The change from no binding to binding occurred in a very narrow range of RemA concentrations, suggesting that RemA binding might be cooperative (Fig. 3A).

Fig. 3.

RemA binds to specific promoters. 5’-Radiolabelled DNA containing the indicated promoters were incubated with the indicated concentrations of RemA–MBP and electrophoresed under non-denaturing conditions. The size of each DNA fragment and the promoter end-points, respectively, were: (A) 252 bp containing P_eps from −159 to +35; (B) 182 bp containing rrnB P1 from −73 to +50; (C) 291 bp containing P_tapA from−168 to +65; and (D) 258 bp containing P_opuA DNA from −172 to +27 (P_opuA P2 from −135 to +64).

To determine where RemA bound relative to the P_eps core promoter, DNase I footprint analysis was performed. Increasing concentrations of RemA–MBP were added to radiolabelled P_eps-containing DNA, and the resulting complexes were subjected to nucleolytic cleavage by DNase I. Addition of RemA–MBP at 0.5 µM resulted in protection from approximately −110 to −50 on P_eps, but the pattern of protection was periodic: 7 bp regions of protection were separated by 3 bp of unaltered, or even enhanced, digestion (Figs 4A and 5). Based on these data, RemA appeared to bind to at least six sites (BS1–BS6) upstream of the P_eps core promoter within the region required for RemA-dependent stimulation of in vitro transcription (Figs 2B and 5). We also note that at high concentrations, RemA-dependent protection extended to positions closer to the transcriptional start site than −49 (Fig. 4A). Although we could not map these lower affinity sites to high resolution, they may be near the predicted position of RNAP binding. Finally, the presence of 0.4 µM SinR abolished all RemA-dependent protections and enhancements, and instead led to protection that corresponded to the known SinR binding sites that overlapped with RemA BS1 and BS2 (Figs 4A and 5; Kearns et al., 2005). We conclude that RemA binds in a specific pattern upstream of the Peps promoter and that SinR served as an anti-activator by occluding RemA binding.

Fig. 4.

RemA binds to multiple sites in target promoters. 5’-Radiolabelled DNA containing the indicated promoters was incubated with RemA–MBP, RemA–MBP storage buffer and/or SinR at the indicated concentrations before treating the samples with DNase I. Positions of cutting relative to the transcription start site are indicated on the left of each gel and were determined using an A+G sequence ladder specific for each DNA fragment (not shown). Regions protected from DNase I digestion by RemA–MBP are indicated as vertical black bars. Regions protected from DNase I digestion by SinR are indicated as vertical grey bars.

Fig. 5.

Summary of RemA binding sites and other relevant P_eps landmarks. The sequence of the P_eps promoter and intergenic region between the slrR gene and the eps operon is arranged from 5’ at the upper left to 3’ at the lower right. The slrR and epsA genes are indicated as grey arrows. SinR binding sites (SinR BS), the −35 box and −10 box of the P_eps promoter, and the +1 transcriptional start site of the eps operon are indicated by appropriately labelled grey boxes. Nucleotide positions relative to the transcriptional start site are indicated above the sequence at 20 bp intervals. Open boxes represent RemA binding sites (RemA BS1–6) as predicted from regions of DNase I protection by RemA–MBP presented in Fig. 4A. Black carets indicate regions of DNase I hyperdigestion caused by RemA–MBP presented in Fig. 4A. Dotted lines represent the regions of deleted bases corresponding to the Δremsite34 and Δremsite56 mutations as indicated. Bolded letters below the main line of sequence indicate the base substitutions identified by a forward genetic screen that abolished expression from a P_eps–lacZ reporter with the allele number in parentheses adjacent to the substitution. The delta symbol (Δ) represents the single base deletion found in the eps32 allele.

To determine the contributions of the RemA binding sites to gene activation, we first took an unbiased forward genetic approach. The P_eps promoter was randomly mutagenized by low fidelity PCR, cloned in front of a promoterless lacZ gene encoding β-galactosidase and inserted in the ectopic chromosomal amyE locus of a strain mutated for SinR and EpsH. SinR was mutated to relieve inhibition of the P_eps promoter and EpsH was mutated to circumvent the severe cell aggregation associated with P_eps expression that can result in loss of colony-forming units. In this genetic background, colonies with wild-type P_eps promoter expression were blue on media containing the chromogenic β-galactosidase substrate X-gal due to high level expression of the reporter. Out of 1600 colonies screened, 16 white colonies were clonally isolated and the P_eps promoters fused to the lacZ reporter were sequenced. After discarding non-unique mutations, six independent mutant alleles were isolated. Four of the alleles that disrupted expression (eps14, eps15, eps40, eps41) were mutations found within the −35 and −10 promoter elements of P_eps (Fig. 5). The fifth allele (eps32) was a deletion of a single T residue immediately upstream of the −35 sequence and the sixth allele (eps33) was a mutation in a region protected by RemA (BS6) (Fig. 5). Overall, few mutations were identified within the putative RemA binding sites perhaps because cooperative binding at some sites buffers against subtle changes at others.

Next, we took a reverse genetic approach to determine the contributions of the RemA binding sites upstream of the P_eps promoter. Two deletion mutations were generated by allelic replacement: P_epsΔ^remsite34 deleted the regions of RemA protection BS3 and BS4, and P_epsΔ^remsite56 deleted regions of RemA protection BS5 and BS6 (Fig. 5). Both Δremsite34 and Δremsite56 deletions abolished expression from corresponding P_eps promoter fragments in in vitro transcription assays (Fig. 2C). Finally, the Δremsite34 and Δremsite56 mutations had severe biofilm defects that phenocopied a remA mutation in an otherwise wild-type background (Fig. 6). We conclude that the BS3, BS4, BS5 and BS6 regions of P_eps bound by RemA are critical for transcriptional activation of the eps operon in vitro and biofilm formation in vivo.

Fig. 6.

Deletion of RemA binding sites upstream of the P_eps promoter abolishes biofilm formation. ‘Pellicle’ column depicts microtitre wells (six-well plate) in which cells have been grown in MSgg medium for 3 days at 25°C. Scale bar is 1 cm. ‘Colony’ column depicts 10× images of individual colonies grown on MSgg medium for 3 days at 25°C. Scale bar is 1 mm. The following strains were used to generate this figure: wild type (3610), remA (DS2679), Δremsite34 (DS8402), Δremsite56 (DS8403), epsH (DS76), tasA (DS3323) and opuA (DK220).

RemA activates and binds to promoters expressing the tasA and opuA operons

RemA is also required for the activation of the P_tapA promoter that drives expression of the tapA-sipW-tasA operon responsible for synthesis of the amyloid protein component of the B. subtilis biofilm matrix (Branda et al., 2006; Chu et al., 2006). As with the P_eps promoter, RemA– MBP activated in vitro transcription from a vector containing the P_tapA promoter region and caused an electrophoretic mobility shift of a P_tapA-containing DNA fragment within a narrow range of protein concentration (Figs 2A and 3C). DNase I footprinting assays showed that RemA– MBP protected multiple regions upstream of the P_tapA promoter with a periodicity reminiscent of that observed with the P_eps promoter (Fig. 4B). Unlike P_eps however, RemA binding was not occluded upstream of the SinR binding site and both SinR and RemA bound simultaneously (Fig. 4B). SinR was previously shown to bind both upstream and downstream of the P_tapA promoter and probably represses P_tapA expression by occluding promoter access (Chu et al., 2006). We conclude that RemA directly and co-ordinately activates both operons that encode structural components of the B. subtilis biofilm matrix but that SinR acts as an anti-activator for P_eps and a repressor for P_tapA.

To identify additional genes under RemA control we conducted transcriptional profiling comparing RNA from cells mutated for remA, sinR and epsH and cells mutated for sinR and epsH alone. The sinR mutation was introduced to relieve repression of potential RemA-activated genes that could mask the full effect of RemA on expression, and the epsH mutation was introduced to prevent cell clumping in the sinR background that can obscure cell density measurements. Consistent with previous reports, genes in the eps and tapA operons were activated 100-fold by RemA and these operons experienced the strongest level of transcriptional control in the regulon (Table 1). In addition, a number of other genes were activated including three operons involved in the uptake of the osmoprotectants glycine betaine and choline, opuA, opuB and opuC, which were expressed substantially higher in the presence of RemA (Kempf and Bremer, 1995; Kappes et al., 1999). Relatively few genes were repressed by RemA (Table 2). We infer from the strongest genes regulated by RemA that the primary function of RemA is to regulate extracellular matrix (rem) gene expression consistent with its name.

Table 1.

Genes activated by RemA.^a

Gene	Fold	Annotation	Gene	Fold	Annotation
		(Biofilm genes)b			(Miscellaneous cont.)
epsA	136.5	EPS synthesis, length regulator?	cgeA	10.5	Spore coat protein
epsB	87.3	EPS synthesis, tyrosine kinase?	cgeB	4.3	Maturation of the spore coat
epsC	123.3	EPS synthesis, epimerase	cotG	9.8	Spore coat protein
epsD	120.1	EPS synthesis, glycosyltransferase	cotJC	4.2	Spore coat protein
epsE	136.3	EPS synthesis and flagellar clutch	glgA	3.6	Glycogen synthase
epsF	136.4	EPS synthesis, glycosyltransferase	glgB	3.5	1,4-a-glucan branching enzyme
epsG	142.3	EPS synthesis, unknown	glgP	3.6	Glycogen phosphorylase
epsN	4.8	EPS synthesis, aminotransferase	sacT	3.6	Transcriptional antiterminator
epsO	6.9	EPS synthesis, pyruvyltransferase	sigL	3.9	Sigma factor L
tapA	116.3	TasA anchoring protein	sodF	4.6	Superoxide dismutase
sipW	145.7	TapA/TasA signal peptidase	spoVR	3.6	Involved in spore cortex synthesis
tasA	112.5	Extracellular amyloid fibre	spoVID	4.0	Assembly of spore coat
slrR	35.4	Regulator motility/autolysin genes	sqhC	3.3	Squalene-hopene synthase
bslA	4.2	Extracellular matrix protein	ycdA	7.5	Unknown
		(Osmoprotection)	ycsF	3.6	Similar to lactam utilization protein
opuAA	54.1	Glycine betaine ABC transporter	ycsG	3.4	Similar to amino acid transporter
opuAB	44.4	Glycine betaine ABC transporter	yfjB	7.2	Unknown
opuAC	43.7	Glycine betaine ABC transporter	yfjC	6.3	Unknown
opuBA	16.0	Choline ABC transporter	yfjD	6.5	Unknown
opuBB	15.4	Choline ABC transporter	yfjE	4.2	Unknown
opuBC	11.0	Choline ABC transporter	ykoL	7.5	Unknown
opuBD	12.8	Choline ABC transporter	ykzB	7.9	Unknown
opuCA	11.3	Glycine betaine/choline transporter	yobB	5.3	Unknown
opuCB	9.9	Glycine betaine/choline transporter	yozY	5.0	Unknown
opuCC	4.3	Glycine betaine/choline transporter	yqfZ	4.7	Unknown
opuCD	5.6	Glycine betaine/choline transporter	yrbD	9.9	Putative alanine transporter
		(Nitrogen metabolism)	ysnE	3.4	Similar to acetyltransferase
ansZ	12.0	Asparaginase	yueB	4.6	Surface receptor for phage SPP1
bpr	7.6	Bacillopeptidase F	yueC	4.4	Unknown
frlM	5.2	Amino sugar permease	yuiA	16.3	Unknown
frlN	5.1	Amino sugar permease	yuiB	10.9	Unknown
ispA	7.4	Intracellular protease	yukB	5.2	Unknown
nasB	9.7	Nitrate reductase (electron transfer)	yukC	4.6	Unknown
nasC	10.2	Nitrate reductase (catalytic subunit)	yukD	5.3	Putative bacteriocin
nasD	7.5	Assimilatory nitrate reductase	yukE	3.7	Unknown
nasE	6.3	Assimilatory nitrate reductase	yvbF	4.6	Unknown
nasA	14.0	Nitrate transporter	yvcB	3.7	Unknown
nrgA	5.6	Ammonium transporter	ywhK	4.0	Unknown
nrgB	6.6	Nitrogen regulated PII protein	ywqH	4.6	Unknown
pnbA	7.8	Para-nitrobenzyl esterase	ywqI	4.3	Unknown
pucA	7.8	Xanthine dehydrogenase	ywqJ	4.4	Putative toxin
pucB	11.6	Xanthine dehydrogenase	ywqK	4.1	Unknown
pucC	11.3	Xanthine dehydrogenase	ywqL	3.5	Endonuclease V
pucD	7.5	Xanthine dehydrogenase	ywrD	8.6	Unknown
pucE	8.3	Xanthine dehydrogenase	ywrK	5.4	Similar to arsenic exporter
pucI	6.3	Allantoin permease	ywzE	9.5	Unknown
pucJ	4.3	Uric acid permease	ywzF	8.6	Unknown
pucK	5.2	Uric acid permease	yxiB	5.4	Unknown
pucL	7.2	Uricase	yxiC	6.6	Unknown
ureA	9.9	Urease	yxiD	4.6	Toxin
ureB	9.7	Urease	yyzB	5.1	Unknown
ureC	7.0	Urease
vpr	3.1	Extracellular protease

^a.Microarray expression values of strain DS207 (sinR epsH) divided by microarray array expression values of strain DS9771 (remA sinR epsH).

^b.Genes epsH through epsM in the eps operon failed to show a threefold change in expression presumably due to the integration of thetetracycline resistance gene inserted into the epsH gene (epsH::tet).

Genes assigned to functional categories in bold. (σ^D) indicates genes previously identified as being expressed under the control of σ^D (Serizawa et al., 2004; Kearns and Losick, 2005).

Table 2.

Genes inhibited by RemA.^a

Gene	Fold	Annotation
		(σD-regulon)b
hag	4.9	Flagellin
hemAT	3.8	Methyl-accepting chemotaxis protein
lytD	3.5	Autolysin (glucosaminidase)
mcpA	3.9	Methyl-accepting chemotaxis protein
mcpC	3.1	Methyl accepting chemotaxis protein
pgdS	3.1	Gamma-DL-glutamyl hydrolase
tlpA	7.2	Methyl-accepting chemotaxis protein
ybdO	3.7	Unknown
yfmS	3.6	Methyl-accepting chemotaxis protein
ylqB	4.6	Unknown
yscB	3.6	Unknown
yxkC	14.4	Unknown
		(Miscellaneous)
braB	3.4	Amino acid transporter
comK	3.2	Transcriptional regulator of competence
sunA	3.3	Sublancin precursor
yjiC	4.6	Similar to macrolide glycosyltransferase
yomK	3.8	Unknown
yorD	4.2	Unknown
zinT	17.2	Unknown

^a.Microarray expression values of strain DS9771 (remA sinR epsH)divided by microarray array expression values of strain DS207 (sinR epsH).

^b.Genes previously identified as being expressed under the control of σ^D (Serizawa et al., 2004; Kearns and Losick, 2005).

The next strongest RemA regulatory effect after the biofilm matrix genes, was the 40-fold activation of the opuA operon. Regulation of opuA by RemA also appeared to be direct as purified RemA–MBP caused an electrophoretic mobility shift of the P_opuA promoter region in a concentration range similar to that observed with the P_eps and P_tapA promoters (Fig. 3D). Likewise, RemA–MBP protected DNA in a periodic repeating pattern reminiscent of the P_eps and P_tapA promoters (Fig. 4C). Unlike P_eps and P_tapA, however, RemA–MBP was not sufficient to activate expression from a P_opuA containing fragment in the in vitro transcription assay (Fig. 2A). Also different from P_eps and P_tapA, increasing concentrations of SinR had no effect on the pattern of RemA-dependent DNase I digestion on P_opuA and there were no SinR-dependent protections or enhancements of cutting (Fig. 4C). We conclude that RemA directly activates P_opuA expression in vivo but is not sufficient for activation in vitro perhaps suggesting that other regulatory factors may be required. We further conclude that not all genes activated by RemA are regulated by SinR.

Together, RemA bound to three different promoters and protected at least 17 regions of DNA that could be determined from the DNase I protection assays (Fig. 4; Fig. S1A). We attempted to determine a RemA consensus binding sequence from the regions of protection. Because the RemA binding sites did not appear to be clear discrete dyads, it was difficult to determine whether the binding sites represented a series of tandem inverted repeats or a series of direct repeats. The best alignment was obtained if the repeats were considered to be direct and a consensus prediction was conducted using Weblogo (Crooks et al., 2004). If the regions are in fact direct repeats, we predict that the consensus RemA binding site is AGNAAAA (Fig. S1B). We note that the consensus sequence is rather weak and often poorly conserved in each individual binding site, again perhaps owing to the apparently high level of cooperativity of RemA binding to DNA.

Discussion

The regulation of biofilm formation in B. subtilis is complex. SinR is a DNA-binding protein that co-ordinately binds to and represses expression of genes that encode the biofilm matrix (Kearns et al., 2005; Chu et al., 2006). Spo0A, a response regulator, both indirectly antagonizes SinR and inhibits the biofilm repressor AbrB, whereas DegU, another response regulator, activates expression of the matrix component BslA (Hamon et al., 2004; Kobayashi, 2007; Verhamme et al., 2007; Chai et al., 2011; Ostrowski et al., 2011). In addition, RemA and RemB are required to activate expression of matrix genes but their mechanism was unknown as they lacked helix–turn–helix motifs (Winkelman et al., 2009). Here we demonstrate that RemA is a DNA-binding protein that is both necessary and sufficient for the activation of the P_eps and P_tapA promoters that express the matrix biosynthesis eps and tapA-sipW-tasA operons respectively. Furthermore, transcriptional profiling indicated the matrix operons were the strongest RemA targets but also lead to the discovery that RemA regulates expression of operons involved in osmoprotection. Finally, we show that the master regulator SinR represses the eps operon by antagonizing the binding of RemA and therefore mechanistically functions as an anti-activator.

RemA bound to DNA with apparently high cooperativity and protected at least six regions upstream of both the P_eps and P_tapA promoters from DNase I digestion. We interpret each of the protected regions as a series of individual, direct repeat binding sites with a weak consensus of AGNAAAA that are perhaps compensated for by cooperative binding. Each direct repeat was separated by 3 bp of either unprotected or enhanced DNase I digestion. Enhancements of digestion are usually due to bending distortions that widen the major groove of DNA and increase the affinity of DNase I (Weston et al., 1992). We infer from the periodicity of DNase I protection and enhancement that RemA is binding to one face of the double helix and bending the DNA. DNA bending has been shown to be essential for some transcription factors to function (Gosink et al., 1993). How DNA bending promotes RemA-dependent activation is unclear but we note that the promoter-proximal binding site is precisely 16 bp upstream of the −35 elements for both P_eps and P_tapA (Fig. S1). Furthermore, deletion of a single base upstream of the P_eps −35 element was sufficient to abolish expression (eps32, Fig. 5). While the strict spacing between the two elements in the two promoters may be coincidental, it may suggest that the face of DNA to which RemA binds relative to the promoter is important for activation.

In addition to the matrix biosynthesis operons, we discovered that RemA also directly activates the opuA operon that encodes genes for the uptake of the osmoprotectant glycine betaine (Boch et al., 1994; Kempf and Bremer, 1995). High osmotic pressure may be experienced by cells in a biofilm due to the production of the extracellular matrix and perhaps osmoprotection operons are co-induced by RemA to physiologically compensate for matrix synthesis (Rubinstein et al., 2012). Whereas mutants defective in the RemA-matrix targets EpsH and TasA have severe biofilm defects, cells defective in OpuA formed biofilms like the wild type (Fig. 6). Biofilm defects in OpuA may only be detectable in environments with glycine betaine or the absence of OpuA may be compensated for by two other RemA-activated osmolyte uptake systems OpuB and OpuC (Kappes et al., 1999). Expression of the opuA operon is activated by high external salt concentrations and is inhibited by the osmoprotectants proline and glycine betaine but the regulators that mediate these transcriptional effects are unknown (Kempf and Bremer, 1995; Steil et al., 2003; Hoffmann et al., 2013). Whether RemA is responsible for the osmotic regulation of opuA is unknown but we note that DNA topology modulation by bending and twist, like that seemingly caused by RemA, has been implicated in the regulation of osmoprotection genes in other organisms (Hulton et al., 1990; Wang and Syvanen, 1992).

Systems-level transcriptomics suggests that expression of remA is nearly uniform save a decrease during late sporulation (Nicolas et al., 2012). Thus, we infer that the physiological input that controls RemA-dependent gene expression may occur at the level of RemA activity. Based on the regulation of the osmoprotection operons and the connection to the biofilm matrix, high osmolarity is one potential stimulus. Alternatively, RemA could respond to nitrogen levels or availability as a substantial number of genes related to nitrogen metabolism were modestly induced in the RemA transcriptional profiling experiment (Table 1). Phylogenetically, the RemA protein is both highly conserved and broadly distributed being encoded within the genomes of Firmicutes, Thermotogales, Cyanobacteria, δ-proteobacteria and Chloroflexi (Fig. S2A). The relationship of these diverse bacteria to one another is unclear but we explored the genetic neighbourhood surrounding RemA for further insight. In each genome we examined, RemA was encoded immediately upstream of the gmk gene encoding the Gmk guanylate kinase perhaps suggesting a connection between RemA and the cytoplasmic pool of GTP (Fig. S2B). Many bacteria form biofilms but the environmental and physiological signals that instruct biofilm formation is seldom known. In the case of B. subtilis, activation of the biofilm matrix genes appears to require the integration of at least two signals, one to antagonize SinR and the other to activate RemA.

Experimental procedures

Growth conditions

Bacillus subtilis strains were grown in Luria–Bertani (LB) (10 g tryptone, 5 g yeast extract, 5 g NaCl per litre) broth or on LB plates fortified with 1.5% Bacto agar at 37°C. When appropriate, antibiotics were included at the following concentrations: 10 µg ml⁻¹ tetracycline, 100 µg ml⁻¹ spectinomycin, 5 µg ml⁻¹ chloramphenicol, 5 µg ml⁻¹ kanamycin and 1 µg ml⁻¹ erythromycin plus 25 µg ml⁻¹ lincomycin (mls). Isopropyl (β-D-thiogalactopyranoside (IPTG, Sigma) was added to the medium at the indicated concentration when appropriate.

For pellicle formation experiments, 10 µl of culture grown overnight at room temperature in LB medium was inoculated into 10 ml minimal MSgg medium [5 mM potassium phosphate (pH 7), 100 mM MOPS (pH 7), 2 mM MgCl₂, 700 µM CaCl₂, 50 µM MnCl₂, 50 µM FeCl₃, 1 µM ZnCl₂, 2 µM thiamine, 0.5% glycerol, 0.5% glutamate, 50 µg ml⁻¹ tryptophan, 50 µg ml⁻¹ phenylalanine and 50 µg ml⁻¹ threonine] in six-well microtitre plates and incubated at 25°C (Branda et al., 2001). For colony architecture analysis, colonies were toothpick inoculated onto minimal MSgg medium fortified with 1.5% Bacto agar and incubated for 3 days at 25°C.

Strain construction

All constructs were first introduced into the domesticated strain PY79 by natural competence and then transferred to the 3610 background using SPP1-mediated generalized phage transduction (Yasbin and Young, 1974). All strains used in this study are listed in Table 3. All plasmids used in this study are listed in Supplemental Table S1. All primers used in this study are listed in Supplemental Table S2.

Table 3.

Strains.

Strain	Genotype	Reference
3610	Wild type
DK220	opuA::tet
DS444	sinI::spec amyE::Peps–lacZ cat
DK445	sinI::spec amyE::Pepsimproved–lacZ cat
DK549	sinR::spec epsH::tet amyE::Peps14–lacZ cat
DK550	sinR::spec epsH::tet amyE::Peps15–lacZ cat
DK551	sinR::spec epsH::tet amyE::Peps32–lacZ cat
DK552	sinR::spec epsH::tet amyE::Peps33–lacZ cat
DK553	sinR::spec epsH::tet amyE::Peps40–lacZ cat
DK554	sinR::spec epsH::tet amyE::Peps41–lacZ cat
DS76	epsH::tet	Kearns et al. (2005)
DS207	sinR::spec epsH::tet	Kearns et al. (2005)
DS1882	epsH::tet amyE::Peps–lacZ cat	Kearns et al. (2005)
DS2609	sinR::spec epsH::tet amyE::Peps–lacZ cat	Kearns et al. (2005)
DS2679	remA::TnYLB kan	Winkelman et al. (2009)
DS2911	sinR::spec remA::TnYLB kan epsH::tet amyE::Peps–lacZ cat	Winkelman et al. (2009)
DS2913	remA::TnYLB kan epsH::tet amyE::Peps–lacZ cat	Winkelman et al. (2009)
DS3323	tasA::Tn10 spec
DS8402	Δremsite34
DS8403	Δremsite56
DS9466	sinR::spec remA::TnYLB kan epsH::tet amyE::Pepsimproved–lacZ cat
DS9467	sinR::spec epsH::tet amyE::Pepsimproved–lacZ cat
DS9468	epsH::tet amyE::Pepsimproved–lacZ cat
DS9469	remA::TnYLB kan epsH::tet amyE::Pepsimproved–lacZ cat
DS9771	sinR::spec remA::TnYLB epsH::tet

opuA::tet

The opuA::tet insertion deletion allele was generated using a modified ‘Gibson’ isothermal assembly protocol (Gibson et al., 2009). Briefly, the region upstream of opuAA was PCR-amplified using the primer pair 3318/3319 and the region downstream of opuAC was PCR-amplified using the primer pair 3320/3321. DNA containing a tetracycline resistance gene (pDG1515; Guérout-Fleury et al., 1995) was amplified using universal primers 3250/3251. The three DNA fragments were combined at equimolar amounts to a total volume of 5 µl and added to a 15 µl aliquot of prepared master mix (see below). The reaction was incubated for 60 min at 50°C. The completed reaction was then PCR-amplified using primers 3318/3321 to amplify the assembled product. The amplified product was transformed into competent cells of PY79 and then transferred to the 3610 background using SPP1-mediated generalized transduction. Insertions were verified by PCR amplification using primers 3318/3321.

A 5× isothermal assembly reaction buffer [500 mM Tris-HCL pH 7.5, 50 mM MgCl₂, 50 mM DTT (Bio-Rad), 31.25 mM PEG-8000 (Fisher Scientific), 5.02 mM NAD (Sigma Aldrich) and 1 mM of each dNTP (New England BioLabs)] was aliquoted and stored at −80°C. An assembly master mixture was made by combining prepared 5× isothermal assembly reaction buffer (131 mM Tris-HCl, 13.1 mM MgCl₂, 13.1 mM DTT, 8.21 mM PEG-8000, 1.32 mM NAD and 0.26 mM each dNTP) with Phusion DNA polymerase (New England BioLabs) (0.033 units µl⁻¹), T5 exonuclease diluted 1:5 with 5× reaction buffer (New England BioLabs) (0.01 units µl⁻¹ ), Taq DNA ligase (New England BioLabs) (5328 units µl⁻¹) and additional dNTPs (267 µM). The master mix was aliquoted as 15 µl and stored at −80°C.

Markerless deletions

To generate the Δremsite34 marker-less deletion construct, the region upstream of remsite34 was PCR-amplified using the primer pair 2713/2714 and digested with EcoRI and XhoI, and the region downstream of remsite1 was PCR-amplified using the primer pair 2716/2718 and digested with XhoI and BamHI. The two fragments were then simultaneously ligated into the EcoRI and BamHI sites of pMiniMAD2 which carries a temperature-sensitive origin of replication and an erythromycin resistance cassette to generate pDP376 (Patrick and Kearns, 2008). The plasmid pDP376 was introduced to 3610 by rare single-cross-over integration by transformation at the restrictive temperature for plasmid replication (37°C) using mls resistance as a selection. To evict the plasmid, the strain was incubated in 3 ml LB broth at a permissive temperature for plasmid replication (22°C)for 14 h. Cells were then serially diluted and plated on LB agar at 37°C. Individual colonies were patched on LB plates and LB plates containing mls to identify mls-sensitive colonies that had evicted the plasmid. Chromosomal DNA from colonies that had excised the plasmid was purified, PCR-amplified with primers 2713/2718, and screened by digestion with XhoI to determine which isolate had retained the remsite34 allele.

To generate the Δremsite56 marker-less deletion construct, the region upstream of remsite1 was PCR-amplified using the primer pair 2713/2715 and digested with EcoRI and XhoI, and the region downstream of remsite56 was PCR-amplified using the primer pair 2717/2718 and digested with XhoI and BamHI. The two fragments were then simultaneously ligated into the EcoRI and BamHI sites of pMiniMAD2 which carries a temperature-sensitive origin of replication and an erythromycin resistance cassette to generate pDP377. The plasmid pDP377 was integrated, excised, and the Δremsite56 allele was verified as described above for Δremsite34.

LacZ reporter construct

To generate the amyE::P_eps^improved– lacZ β-galactosidase reporter construct pRLG12296, −452 to +35 of P_eps^improved was amplified from plasmid pRLG11810 using primers 4831/6097, digested with EcoRI and BamHI and cloned into the EcoRI and BamHI site of pDG268 which carries a chloramphenicol-resistance marker and a polylinker upstream of the lacZ gene between two arms of the amyE gene (Guérout-Fleury et al., 1996).

MBP–RemA expression construct

The remA gene was PCR-amplified using primer pair 1527/1528 and 3610 chromosomal DNA, digested with EcoRI and BamHI and cloned into the EcoRI/BamHI sites of plasmid pMAL-c2X (New England Biolabs) to generate plasmid pKB116.

In vitro transcription plasmids

All promoters studied using in vitro transcription were PCR-amplified from NCIB3610 chromosomal DNA, digested with EcoRI and HindIII, and ligated into the EcoRI and HindIII sites of pRLG770 (Ross et al., 1990). To generate pRLG11810, −452 to +35 of P_epsA was amplified using primers 4831 and 5619. To generate pRLG12760, −176 to +35 of P_epsA was amplified with primers 6157 and 5619. To generate pRLG12759, −159 to +35 of P_epsA was amplified using primers 6158 and 5619. To generate pRLG12764, −121 to +35 of P_epsA was amplified using primers 6159 and 5619. To generate, p12757, −107 to +35 of P_epsA was amplified using primers 5620 and 5619. To generate pRLG12786, −168 to +65 of P_tapA was amplified using primers 6463 and 6464. To generate pRLG12784, −172 to +27 of P_opuA P1 (−135 to +64 of P_opuA P2) was amplified from 3610 chromosomal DNA using primers 6465 and 6466. To generate the P_epsA^improved, pRLG11810 was mutagenized using primer 5892 to generate plasmid pRLG11825.

Plasmids for fragment labelling

To generate pRLG7340 E. coli rrnB P1 promoter DNA was amplified from pRLG1616 using primers 4633 and 1620, digested with EcoRI and HindIII and ligated into EcoRI and HindIII digested pSL6. To generate RLG12766, −159 to +35 of P_eps was amplified from NCIB3610 chromosomal DNA using primers 6158 and 5619, digested with EcoRI and HindIII and ligated into digested pSL6. To generate pRLG12783, −172 to +27 of P_opuA was amplified from NCIB3610 chromosomal DNA using primers 6465 and 6466, digested with EcoRI and HindIII and ligated into digested pSL6. To generate pRLG12785, −168 to +65 of P_tapA was amplified from NCIB3610 chromosomal DNA using primers 6463 and 6464, digested with EcoRI and HindIII, and ligated into digested pSL6.

SPP1 phage transduction

To 0.2 ml of dense culture grown in TY broth (LB broth supplemented after autoclaving with 10 µM MgSO₄ and 100 µM MnSO₄), serial dilutions of SPP1 phage stock were added and statically incubated for 15 min at 37°C. To each mixture, 3 ml TYSA (molten TY supplemented with 0.5% agar) was added, poured atop fresh TY plates, and incubated at 37°C overnight. Top agar from the plate containing near confluent plaques was harvested by scraping into a 50 ml conical tube, vortexed and centrifuged at 5000 g for 10 min. The supernatant was treated with 25 µg ml⁻¹ DNase final concentration before being passed through a 0.45 µm syringe filter and stored at 4°C.

Recipient cells were grown to stationary phase in 2 ml TY broth at 37°C. 0.9 ml cells were mixed with 5 µl of SPP1 donor phage stock. Nine millilitres of TY broth was added to the mixture and allowed to stand at 37°C for 30 min. The transduction mixture was then centrifuged at 5000 g for 10 min, the supernatant was discarded and the pellet was resuspended in the remaining volume. One hundred microlitres of cell suspension was then plated on TY fortified with 1.5% agar, the appropriate antibiotic and 10 mM sodium citrate.

Protein purification

SinR was purified as previously described (Kearns et al., 2005) His-tagged SinR was overexpressed from plasmid pDP91 in BL21(DE3) cells. One litre of culture was grown in LB supplemented with 30 µg ml⁻¹ kanamycin at 37°C until OD₆₀₀ of 0.5 was obtained. IPTG was added to 1 mM and the cells were shifted to 22°C and grown for 3 h before harvesting by centrifugation. The cell pellet was resuspended in 40 ml lysis buffer (50 mM Tris-Cl pH 8.5, 500 mM NaCl, 10 mM imidazole 0.1 mM PMSF) and the cells were lysed by sonication. Cell debris was cleared by centrifugation and the resulting supernatant was applied to a column with 1 ml of NiNTA slurry that had been equilibrated in lysis buffer. The column was washed with 10 ml of wash buffer (50 mM Tris-Cl pH 8.5, 500 mM NaCl, 20 mM imidazole) and eluted with 1 ml of elution buffer (10 mM Tris-Cl pH 8.5, 10 mM MgCl₂, 1 mM EDTA, 300 mM DTT, 5% glycerol). To cleave the His-tag from SinR 3 µl of 0.9 U µl⁻¹ biotinylated thrombin (Novagen) was added and gently mixed overnight at room temperature before applying to a column with 0.5 ml of NiNTA agarose bead slurry equilibrated in elution buffer. One hundred micro-litres of streptavidin agarose was added to the eluate and gently mixed for 1 h before applying to a column. The eluate was dialysed against 10 mM Tris-Cl pH 8.5, 10 mM MgCl₂, 1 mM EDTA, 0.3 mM DTT, 50% glycerol, 1 mM PMSF.

RemA–MBP and LacZ–MBP were overproduced from plasmids pKB116 and pMal-c2X, respectively, in BL21(DE3) cells. Cells were grown in 1 l of LB supplemented with 0.2% glucose and 100 µg ml⁻¹ ampicillin at 37°C to an OD₆₀₀ of 0.5 before adding IPTG to 0.3 mM. The cells were grown for 2 more hours at 37°C before harvesting by centrifugation. Cell pellets were resuspended in 25 ml column buffer (20 mM Tris-Cl pH 7.4, 200 mM NaCl, 10 mM β-mercaptoethanol, 1 mM EDTA) and lysed by sonication. After clearing cell debris by centrifugation, the supernatant was loaded onto 1 ml slurry of amylose resin (NEB) that had been equilibrated in column buffer. The beads were washed with 12 ml of column buffer before eluting in 0.4 ml steps of elution buffer (column buffer + 10 mM maltose). Appropriate fractions were dialysed into RemA–MBP storage buffer (50 mM Tris-Cl pH 8.0, 150 mM NaCl, 0.1 mM EDTA, 1 mM DTT).

Bacillus subtilis RNAP, histidine-tagged on the β' subunit (Qi and Hulett, 1998), was purified from strain MH5636 similar to how it was previously described (Anthony et al., 2000). MH5636 cells were grown in 4 l of LB to an OD₆₀₀ of 1.0 before harvesting by centrifugation. The pellet was resuspended in 55 ml buffer P (300 mM NaCl, 50 mM Na₂HPO₄, 3 mM 2-mercaptoethanol, 5% glycerol, 1 mM PMSF) and cells were lysed by sonication followed by centrifugation. The supernatant was mixed with 5 ml of NiNTA slurry (Qiagen) that had been equilibrated in buffer P and gently mixed for 30 min at 4°C before applying to a polyprep column. The beads were washed with 60 ml of buffer P followed by washing with 60 ml buffer P +30 mM imidazole. Core RNAP was eluted in 1 ml steps with Buffer P +400 mM imidazole. Appropriate fractions were dialysed into storage buffer [50 mM Tris-Cl (8.0), 0.1 M NaCl, 3 mM 2-mercaptoethanol, 50% glycerol]. The σ^A subunit was a generous gift of Libor Krasny and had been purified as described (Juang and Helmann, 1994; López de Saro et al., 1999). Core RNAP and σ^A were reconstituted in storage buffer (50 mM Tris-HCl pH 8.0, 0.1 M NaCl, 3 mM 2-mercaptoethanol, 50% glycerol) for 30 min at 30°C using 20-fold molar excess of σ^A. Titration experiments were carried out to ensure saturation of core RNAP.

In vitro transcription

Twenty-five microlitres of reactions containing 50 ng supercoiled plasmid DNA (plasmid templates used for in vitro transcription are listed in Supplemental Table S1) were pre-incubated 1 min at 22°C with the appropriate combinations of 700 nM RemA–MBP, 700 nM LacZ–MBP and/or 400 nM SinR in 10 mM Tris-Cl pH 8.0, 100 mM KCl, 10 mM MgCl₂, 1 mM DTT, 0.1 mg ml⁻¹ BSA (NEB) 500 µM ATP, 200 µM GTP, 200 µM CTP, 0.1 µM UTP and [α-³²P]-UTP (PerkinElmer). Transcription was initiated by addition of purified RNA polymerase holoenzyme to a final concentration of 20 nM. Reactions were terminated after 15 min of incubation at 22°C by the addition of 25 µl of 95% formamide, 0.05% bromophenol blue and 0.025% xylene cyanole. Fifteen-microlitre aliquots of these reactions were electrophoresed on 5.5% acrylamide, 7 M urea, 1× TBE gels for 1.5 h at 200 V. The relative yield of transcript was determined using ImageQuant.

DNase I footprinting

The DNase I footprinting protocol was adapted from Bartlett et al. (1998). Plasmids pRLG12766 (containing a P_epsA promoter fragment, end-points −159 to +35), pRLG12785 (containing PtapA promoter fragment, end-points −168 to +65) or pRLG12783 (containing P_opuAA promoter fragment end-points −170 to +27) were digested with NcoI (NEB), end-labelled by filling-in with [α-³²P]-dCTP (Perkin-Elmer) using Sequenase (USB), and digested with NheI (NEB). The DNA was purified after each step by ethanol precipitation. The promoter fragments were run on a 5% acrylamide, 0.5× TBE gel for 1.5 h before being excised and diffused into low-salt Elutip-D buffer (0.2 M NaCl, 20 mM, Tris-Cl pH 7.4, 1 mM EDTA) O/N at 4°C. The DNA was purified using Elutip-D columns (Whatman), ethanol precipitated, and resuspended in 100 µl 10 mM Tris-HCl, pH 8.0. 125–1000 nM RemA–MBP (or RemA–MBP storage buffer) was added to ~ 0.2 nM template DNA in transcription buffer (10 mM Tris-HCl, pH 8.0, 30 mM KCl, 10 mM MgCl₂, 1 mM DTT, 0.1 µg µl⁻¹ BSA) for 10 min at 22°C. One microlitre of DNase I (to a final concentration of 1 µg ml⁻¹) was added for 30 s before the reaction was stopped by addition of 10 mM EDTA and 0.3 M sodium acetate. The samples were phenol extracted and the DNA was precipitated with ethanol, washed with 100% ethanol, dried and suspended in 5 µl loading buffer (7 M urea, 0.5× TBE 0.05% bromophenol blue, 0.025% xylene cyanole). Reactions were electrophoresed on 8.5% acrylamide, 7 M urea, 0.5× TBE gels for 2.5 h at 2000 V.

Electrophoretic motility shift assays

Radiolabelled DNA fragments were purified as above. 100–900 nM RemA–MBP (or RemA–MBP storage buffer) was added to ~0.2nM template DNA in transcription buffer (10 mM Tris-HCl, pH 8.0, 30 mM KCl, 10 mM MgCl₂, 1 mM DTT, 0.1 µg µl⁻¹ BSA) for 10 min at 22°C. DNA load dye was added and the reactions were electrophoresed on 5% acrylamide, 0.5× TBE gels for 1 h at 200 V.

Transcriptional profiling

Custom Agilent DNA microarrays were designed using the Agilent eArray application. The arrays include 15 744 60-mer probes which consist of the annotated protein-coding genes of the B. subtilis str. 168 genome, as well as the known non-coding RNAs entered into NCBI and reported in Rasmussen et al., 2009 and Irnov et al., 2010. Each protein-coding gene includes three probes and the non-coding RNAs have either two or three probes.

Bacillus subtilis strains DS207 and DS9771 were grown in LB at 37°C to an optical density of 1.0–1.1 at OD₆₀₀ and samples were immediately mixed with an equal volume of methanol (−20°C) and centrifuged to pellet cells. Cell pellets were stored at −80°C. RNA was isolated using a hot acid-phenol isolation procedure as described above. The sample is then treated with Qiagen’s RNase-Free DNase in solution and the RNA clean-up protocol of Qiagen’s RNeasy® kit was used. The quality of the RNA was checked by visualizing the integrity of the 23S and 16S rRNA bands on an agarose gel.

Labelled cDNA was generated from RNA using Agilent’s Fairplay® III Microarry Labelling Kit. The Agilent Two-Colour Microarray-Based Prokaryotic Analysis (FairPlay III Labelling) Protocol was used with the following changes: between 6 and 10 µg total RNA was used, the reverse transcription reaction was performed at 42°C for 2 h, and the NHS-ester dye-coupling reaction to amino allyl dUTP (using GE Healthcare Cy™3 and Cy™5 monofunctional reactive dyes) was incubated for 90 min. Hybridizations were incubated at 65°C for 17 h. Arrays were scanned by Agilent Technologies DNA Microarray Scanner with Surescan High-Resolution Technology.

Data were subjected to standard lowess normalization using Bioconductor and run in R was performed on the processed signal given by the Agilent software and the geometric mean of the probes was used to give the final value for each gene and nc-RNA.