A master regulator for biofilm formation by Bacillus subtilis


E-mail losick@mcb.harvard.edu; Tel. (+1) 617 495 4905; Fax (+1) 617 496 4642.


Wild strains of Bacillus subtilis are capable of forming architecturally complex communities of cells known as biofilms. Critical to biofilm formation is the eps operon, which is believed to be responsible for the biosynthesis of an exopolysaccharide that binds chains of cells together in bundles. We report that transcription of eps is under the negative regulation of SinR, a repressor that was found to bind to multiple sites in the regulatory region of the operon. Mutations in sinR bypassed the requirement in biofilm formation of two genes of unknown function, ylbF and ymcA, and sinI, which is known to encode an antagonist of SinR. We propose that these genes are members of a pathway that is responsible for counteracting SinR-mediated  repression.  We  further  propose  that  SinR is a master regulator that governs the transition between a planktonic state in which the bacteria swim as single cells in liquid or swarm in small groups over surfaces, and a sessile state in which the bacteria adhere to each other to form bundled chains and assemble into multicellular communities.


Biofilms are architecturally complex communities of microorganisms in which the cells are held together by an extracellular matrix, typically containing exopolysaccharides (EPSs), proteins and even nucleic acids (Hall-Stoodley et al., 2004). An attractive organism in which to identify, and investigate the function of, genes involved in biofilm formation is Bacillus subtilis. This Gram-positive, spore-forming, soil bacterium is highly accessible to manipulation by the techniques of classical and molecular genetics, and wild (undomesticated) strains of B. subtilis form robust biofilms both at liquid/air interfaces and on solid surfaces (Branda et al., 2001). In standing liquid medium, cells of B. subtilis switch from a submerged, highly motile planktonic state in which the bacteria swim as single cells, to a non-motile state in which the cells grow as bundled chains that rise to the surface and form a robust pellicle. On the surface of agar plates, the cells form colonies with elaborate architecture, including aerial structures that resemble fruiting bodies and that preferentially produce spores at their tips (Branda et al., 2001). Pellicles and colonies share the common feature that growth in filamentous chains of cells is associated with, if not essential for, architectural complexity. Indeed, this capacity to grow in bundled chains was recognized by Ferdinand Cohn in his report on the discovery of B. subtilis (Cohn, 1872).

Previous work has revealed a large number of genes that govern biofilm formation in B. subtilis (Branda et al., 2001; 2004; Hamon and Lazazzera, 2001; Hamon et al., 2004). These include regulatory genes involved in the early stages of sporulation (spo0A, spo0H and abrB), genes that are putatively involved in EPS production (yhxB and the 15-gene-long yveK-yvfF operon, which is herein renamed epsA-O1), a gene encoding a putative phosphatase (yqeK), a gene involved in the production of the surfactant, surfactin (sfp), a gene encoding a signal peptidase (sipW), a gene encoding an ABC transporter subunit (ecsB), and two genes of largely unknown function whose inferred products exhibit substantial amino acid-sequence similarity to each other (ylbF and ymcA). Key challenges for the future are to comprehensively identify genes involved in biofilm formation, to determine which genes fall into common pathways, and to elucidate how these pathways operate.

Here we describe one such pathway that governs the switch from growth as motile cells to growth as biofilm-forming chains of cells. In addressing the issue of the role of cell chains in biofilm formation, our attention was drawn to the regulatory genes sinR and sinI. In commonly used laboratory strains, a sinR mutation causes the formation of rugose colonies in which cells grow constitutively as chains of non-motile cells (Fein, 1979; Gaur et al., 1986; Sekiguchi et al., 1988; 1990). Cells of a sinI mutant, in contrast, are always motile and do not form chains (Bai et al., 1993). The sinR gene is known to encode a DNA-binding protein, and sinI is known to encode an antagonist of the sinR gene product (SinR) with which it forms a complex (Gaur et al., 1991; Bai et al., 1993; Lewis et al., 1998). To investigate the effects of mutations in these genes on biofilm formation, we created and introduced null mutations of sinI and sinR into the wild strain 3610 and examined both pellicle formation in standing liquid medium and the formation of fruiting bodies on solid medium. We report that the sinI mutation blocked biofilm formation and that, oppositely to sinI, the sinR mutation caused the formation of robust, rugose, multicellular structures. Helping to explain the role of sinI and sinR, we show that SinR binds to the promoter region of the eps operon, thus repressing transcription of genes believed to be responsible for production of the EPS component of the extracellular matrix. Finally, we show that sinI as well as ylbF and ymcA are members of a pathway(s) that is responsible for counteracting SinR-mediated repression and thereby activating the eps operon and other genes involved biofilm and fruiting body formation.


SinI and SinR have opposing effects on the formation of multicellular communities

The results of Fig. 1 show that in liquid medium, the sinR mutant derivative of the wild strain 3610 grew as bundled chains and formed robust rugose pellicles. These structures were hardy and difficult to disrupt mechanically when probed with a toothpick. In contrast, the sinI mutant lacked bundled chains and was completely defective (at least initially; see below) in biofilm formation (Fig. 1). On solid medium, the sinR mutant formed extremely rough colonies in which it was difficult to discern individual fruiting bodies (Fig. 1). The sinI mutant, on the other hand, produced flat, featureless colonies devoid of fruiting bodies (Fig. 1). Consistent with the idea that SinI is an antagonist of SinR, the phenotype of a sinI sinR double mutant, both in liquid and on solid medium, was indistinguishable from that of a sinR mutant (Fig. 1).

Figure 1.

Effect of mutations in sinR, sinI and espH on chain bundling, pellicle formation, colony surface architecture and swarming motility. ‘Broth culture’ column depicts 1000× phase contrast images of cells grown to mid-exponential phase in LB broth and reveals effects of mutations in the indicated genes on the bundling of chains of cells. Scale bar is 10 µm. ‘Pellicle’ column depicts top-down images of microtitre wells (6-well plate) in which cells have been grown in MSgg medium for 3 days at 22°C and reveals effects of mutations on pellicle formation. Scale bar is 1 cm. ‘Colony’ column depicts 10× images of individual colonies grown on MSgg medium for 3 days at 22°C and reveals effects on colony surface architecture. Scale bar is 1 mm. ‘Swarming motility’ column depicts top-down images of Petri plates containing LB and 0.7% agar, centrally inoculated and incubated at 37°C overnight. Swarm plates were filmed against a black background such that zones of bacterial colonization appear white and uncolonized agar appears black. Scale bar is 1 cm. The indicated wild-type and mutant strains were as follows: WT (3610), sinI (DS91), sinR (DS92), sinI sinR (DS93), epsH (DS76) and epsH sinR (DS207).

Despite the severe effect of the sinI mutation in blocking biofilm formation, upon prolonged incubation in liquid medium the sinI mutant eventually produced a thick pellicle (Fig. 2A). The delay in pellicle formation led us to suppose that the sinI mutant eventually acquired suppressor mutations that restored its capacity to form a biofilm. In support of this hypothesis, cells that were isolated from the mutant biofilm, clonally purified and tested on solid medium, were found to exhibit an extremely rough phenotype that strongly resembled that of a sinR mutant, rather than the smooth, featureless phenotype of the original sinI mutant (Fig. 2B). Given that mutations in sinR suppress the effects of a sinI mutation (Fig. 1), one possibility was that the spontaneous suppressors were second-site mutations within sinR. To investigate this possibility, we amplified and sequenced the sinR gene from three independently arising sinI suppressor mutants. In each case, the mutants harboured (different) frameshift mutations within the sinR open reading frame (ORF) (Fig. 3). Interestingly, each of the spontaneous sinR frameshift mutations lie within, or adjacent to, a homopolymeric track of repetitive nucleotides. Tracks such as these are subject to high frequency mutation resulting from slipped-strand mispairing during DNA replication and have recently been determined to be responsible for phase regulation of the B. subtilis swarming motility gene, swrA (Levinson and Gutman, 1987; Kearns et al., 2004). The nature of the spontaneous sinR mutations raises the possibility that, like swarming motility, biofilm formation in B. subtilis is also regulated by phase variation.

Figure 2.

Prolonged incubation leads to the appearance of suppressors that restore biofilm formation to a sinI mutant.
A. WT (3610) and sinI (DS91) mutant cells were grown in MSgg medium at 22°C for the indicated times. Notice that by 102 h the sinI mutant had formed a pellicle. Scale bar is 1 cm.
B. Surface features of colonies grown on MSgg medium. Shown are colonies of a sinI mutant (DS91) and three independently isolated suppressor strains derived from pellicles of the sinI mutant at 126 h: sinI sup1 (FC139), sinI sup2 (FC140) and sinI sup3 (FC141). Scale bar is 1 mm.

Figure 3.

Suppressor mutations in sinR that restore biofilm formation to sinI, ylbF and ymcA mutants. DNA sequence of the sinR ORF is given in lower case. The SinR amino acid sequence is in upper case above the DNA sequence. Mutations indicated below the DNA sequence are suppressors of a sinI mutation: deletion of an A:T base pair within the underlined stretch of A:T base pairs (FC139), deletion of a T:A base pair with the underlined stretch of T:A base pairs (FC140) and a duplication of the sequence indicated by the downward pointing arrow (FC141). Mutations indicated above the DNA sequence by the upward arrows are missense suppressors of ylbF (P42S, SSB581) and ymcA (V50A, SSB584) mutations.

Surface motility and the formation of multicellular communities are alternative physiological states

In addition to the formation of multicellular communities, wild strains, but not standard laboratory strains, exhibit a second multicellular behaviour called swarming motility (Kearns and Losick, 2003). Swarming is a form of motility that takes place on solid medium in which small groups of cells (rafts) migrate along the surface of a substratum. As the swarm expands, a second internal population arises that is enriched in chains of non-motile cells. As mutations in sinR and sinI had profound effects on biofilm and fruiting body formation, we wondered whether the mutations might also influence swarming motility. The sinR mutant was blocked in swarming motility (Fig. 1), a finding consistent with previous observations that sinR mutants are non-motile (Fein, 1979; Pooley and Karamata, 1984; Sekiguchi et al., 1990). In contrast, the sinI mutant swarmed as readily as the wild type but lacked the characteristic population of chains of non-motile cells observed in the centre of swarms of the wild-type parent (Fig. 1; Fig.S1). Once again, the effects of the sinR mutation were epistatic to those of the sinI mutation (Fig. 1).

In summary, our results are consistent with the idea that B. subtilis exists in two mutually exclusive, physiological states in which cells either grow as bundled chains or as single motile cells. Each state is associated with an alternative multicellular behaviour in that growth in bundled chains promotes biofilm and fruiting body formation, whereas growth as single cells promotes swarming. We propose that SinR and SinI play a central role in regulating the switch between these two states and in this capacity determine which multicellular behaviour is adopted by the cells.

SinR is a negative regulator of genes involved in EPS synthesis

In earlier work we found that cells that contained mutations within the eps operon were defective in bundling of cell chains and produced fragile pellicles that would frequently shatter and sink to the bottom of the culture vessel. To investigate the relationship of eps genes to SinR, a mutation of epsH (formerly yveR) was introduced into a SinR mutant, which, as described above, is locked in the state of forming bundled chains. Consistent with the idea that the eps operon produces a substance involved in bundling, the epsH mutation was epistatic to the sinR mutation in that the double mutant resembled the epsH single mutant with respect to its inability to form bundled chains, pellicles, or architecturally complex colonies (Fig. 1). These data indicate that the eps genes (or at least epsH) act downstream of SinR and raise the possibility that SinR is a negative regulator of the expression of the eps operon. Interestingly, the epsH mutation did not restore swarming to the SinR mutant, a finding that indicates that the motility defect caused by the absence of SinR is not simply a consequence of tethering the cells to each other by extracellular polysaccharides (Fig. 1).

To investigate the idea that SinR is a negative regulator of the eps operon, we constructed a transcriptional fusion of lacZ to the promoter region (PepsA) upstream of the first gene in the eps operon (epsA), and inserted the construct into the chromosome at the amyE locus (amyE::PepsA-lacZ). As the absence of SinR causes constitutive bundling, cells of a sinR mutant tend to clump, interfering with efforts to measure cell number and gene expression. To alleviate this problem, a mutation in the epsH gene was introduced into the reporter strain to ensure dispersed growth even in the presence of the sinR mutation. The epsH mutation had no effect on expression of the PepsA-lacZ reporter and hence did not interfere with our efforts to investigate the influence of a sinR mutation on expression of the reporter (Table 1). We found that in medium (MSgg) that promoted biofilm formation, the sinR mutation increased expression of the PepsA-lacZ construct by greater than 20-fold (Table 1). A sinI mutation, in contrast, had the opposite effect, reducing PepsA-lacZ expression by at least 20-fold. We conclude that SinR is a potent negative regulator of the expression of the eps operon.

Table 1.  SinR represses expression of the eps operon.
GenotypeaActivity (MU)b
  • a

    . The following PepsA-lacZ-containing strains were used: FC5, FC13, FC14, FC15, DS698, DS699.

  • b

    .β-Galactosidase activity, presented as Miller units (MU), was measured for mid-exponential phase cells and is the average of three replicas.

Wild type 30 ± 2
ΔepsH 27 ± 3
ΔsinR ΔepsH674 ± 92
ΔsinI ΔepsH  1 ± 0.4
ΔylbF  4 ± 0.7
ΔymcA  4 ± 0.3

SinR binds to the regulatory region for the eps operon

To determine whether SinR represses the eps operon by directly binding to its regulatory region, we carried out electrophoretic mobility shift assays (EMSAs) with purified SinR and several promoter-containing DNA fragments (see Experimental procedures). The results presented in Fig. 4 show that SinR retarded the electrophoretic mobility of PepsA-containing DNA, and did so with an apparent binding affinity similar to that for DNA containing the promoter region for aprE, a known direct target of SinR (Fig. 4A, Gaur et al., 1991). Interestingly, SinR generated several species of PepsA-containing DNA with reduced electrophoretic mobility, suggesting that the repressor binds to multiple sites within the epsA promoter region. As a negative control, little or no retardation in electrophoretic mobility was observed for DNA containing the promoter for an arbitrarily chosen gene (yvbA) not known nor believed to be under SinR control (Fig. 4A). As a further indication of a specific interaction with the radioactive PepsA-containing DNA, SinR binding was inhibited by the addition of an excess of unlabelled PepsA-containing competitor DNA, but not by an excess of PyvbA-containing DNA (Fig. 4B). SinR has also been reported to bind to the promoter regions for the sporulation genes spo0A and spoIIA (Mandic-Mulec et al., 1992; 1995). However, we detected little or no effect on the mobility of Pspo0A-containing DNA and only a weak effect on PspoIIA-containing DNA (Fig. 4A). If these sporulation genes are indeed direct targets of SinR, then they evidently have much weaker binding sites for the repressor than do epsA and aprE.

Figure 4.

SinR binds to DNA containing the promoter region for the eps operon.
A. Electrophoretic mobility shift assays (EMSAs) in which radiolabelled DNAs (indicated in the upper left corner of each panel) were mixed with purified SinR at the following concentrations: 0 nM (lane 1), 4 nM (lane 2), 10 nM (lane 3), 40 nM (lane 4), 100 nM (lane 5), 400 nM (lane 6) and 1 µM (lane 7).
B. EMSA experiment in which 40 nM of SinR was mixed with either unlabelled PepsA-containing DNA (lanes 2–6) or unlabelled PyvbA-containing DNA (lanes 7–11) added at 1× (lanes 2 and 7), 4× (lanes 3 and 8), 10× (lanes 4 and 9), 40× (lanes 5 and 10), and 100× (lanes 6 and 11) concentration relative to the concentration of radiolabelled PepsA (100 nM).
C. EMSA experiment in which 40 nM of SinR was mixed with its antagonist SinI at the following concentrations: 0 nM (lane 2), 10 nM (lane 3), 40 nM (lane 4), 100 nM (lane 5) and 400 nM (lane 6) prior to the addition of radiolabelled PepsA-containing DNA (100 nM). Lane 1 had PepsA-containing DNA without added proteins and lane 7 had a mixture of PepsA-containing DNA and 400 nM SinI but no SinR. All lanes for panel C were taken from the same gel.

SinI is an antagonist of SinR that inhibits the capacity of the repressor to bind to its targets, such as the promoter region of aprE (Bai et al., 1993). To determine if SinI would similarly interfere with the binding of SinR to PepsA-containing DNA, SinR was mixed with various concentrations of purified SinI and then tested for DNA-binding activity by EMSA. As the amount of SinI added to the reaction was increased, the ability of SinR to retard the electrophoretic mobility of DNA containing PepsA was diminished (Fig. 4C). This was particularly evident from the accumulation of the unbound (unretarded) labelled probe, which was almost entirely depleted in the absence of the antagonist. Nevertheless, even at a 10-fold excess of SinI to SinR, some binding to PepsA-containing DNA could still be detected by the persistence of a species with retarded mobility (Fig. 4C, lane 6). SinI alone had no effect on the electrophoretic mobility of PepsA-containing DNA (Fig. 4C, lane 7).

Before attempting to localize the precise binding site(s) for SinR within the eps regulatory region, we carried out primer extension analysis to map the transcriptional start site for the operon. The results show that the start site, which is designated as position +1, was located 56 bp upstream of the first gene in the operon (Fig. 5B). The canonical recognition sequences for promoters used by RNA polymerase containing the housekeeping sigma factor σA are TATAAT for the ‘−10 sequence’ and TTGACA for the ‘−35 sequence’ (Moran et al., 1982). B. subtilis promoters frequently have an AT-rich sequence or UP element  located  just  upstream  of  the  −35  sequence (Ross et al., 1993; Fredrich et al., 1995). Centred at position −10 relative to the transcriptional start site is a perfect match (TATAAT) to the canonical −10 sequence and centred at position −32 is a sequence (TTTTAA) that somewhat (three out of six positions) conforms to a canonical −35 sequence. We also note the presence of an AT-rich sequence just upstream of the putative −35 sequence that could correspond to a UP element.

Figure 5.

Identification of multiple binding sites for SinR in the eps operon promoter region.
A. A footprinting experiment in which radiolabelled DNA containing PepsA was mixed with SinR at the following concentrations: 0 nM (lanes 1 and 8), 4 nM (lane 2), 10 nM (lane 3), 20 nM (lane 4), 40 nM (lane 5), 100 nM (lane 6) and 400 nM (lane 7) and then treated with DNase I as described in the Experimental procedures. Vertical bars indicate regions that were protected from enzyme digestion.
B. The results of a primer extension experiment carried out with RNA purified from mid-exponential phase cells from strain 3610 growing on MSgg. The RNA was used as a template for extension of radiolabelled primer ‘+94 eps R’ as described in Experimental procedures. The extension product is labelled with an asterisk. Also presented are dideoxy sequencing ladders generated with the same primer and terminated with ddTTP (T), ddGTP (G), and ddCTP (C). The corresponding DNA sequence is given between the two panels and was determined from a DNA sequencing ladder. Note: The sequencing ladder from panel B is complementary to the sequence presented here. The open arrow indicates the eps operon transcriptional start site. The numbers indicate the position of regions of interest relative to the transcriptional start site. The solid box indicates a canonical −10 element for the transcription factor σA while the dashed box indicates a weak −35 element. Solid arrows flanking the DNA sequence indicate predicted recognition sequences for SinR from within the regions that were protected from DNase I.

To localize the binding site(s) for SinR in the eps regulatory region, we carried out footprinting experiments in which radiolabelled PepsA-containing DNA was mixed with various concentrations of SinR. Next, the DNA–SinR complexes were treated with DNase I, and the digestion products were subjected to electrophoresis. The results show that SinR protected two regions of PepsA from the action of DNase I (See Fig. 5A). The region proximal to epsA contained an inverted repeat of the sequence GTTCTCT centred at positions −58 and −67 relative to the transcriptional start site. The extended region of protection distal to epsA contained three nearly identical direct repeats of the same sequence centred at positions −129, −150 and −159 relative to the transcriptional start site. We propose that these are operator sites for SinR, and that the motif GTTCTCT is the recognition sequence for SinR, which differs from that previously published (Shafikhani et al., 2002). We note the presence of a similar sequence (GTTCTCA) in the binding site for SinR within the regulatory region for aprE (Gaur et al., 1991). Both regions of SinR binding are too far upstream to compete with the binding of RNA polymerase to the −10 and −35 elements of the eps promoter. Conceivably, SinR acts by occluding the UP element or by preventing the binding of an unknown activator protein.

Mutations in sinR bypass genes required for biofilm formation

We recently identified two genes of unknown function, ylbF and ymcA, that are required for biofilm formation (Tortosa et al., 2000; Branda et al., 2004). Consistent with the flat featureless colonies produced by ylbF and ymcA mutants, mutation of either gene resulted in a marked reduction in eps gene expression in both liquid and solid medium (Table 1, Fig. 6A). As indicated above, a mutation in sinR reverses the block in biofilm formation caused by a mutation in sinI. We therefore wondered whether a sinR mutation would similarly bypass the requirement for ylbF or ymcA in biofilm formation. Whereas single mutants of each of the aforementioned genes produced flat featureless colonies, the introduction of a sinR mutation restored the formation of colonies with complex surface features (Fig. 6) and rescued the ability of the mutants to form robust pellicles in standing liquid minimal medium (data not shown).

Figure 6.

A sinR mutation restores complex colony architecture to mutants blocked in biofilm formation.
A. Colonies of PepsA-lacZ-containing strains FC5 (wild type), DS698 (ylbF) and DS699 (ymcA) grown on MSgg medium supplemented with Xgal to detect β-galactosidase activity.
B. Colonies of strains 3610 (wild type), DS92 (sinR), SSB136 (ylbF), SSB561 (ylbF sinR), SSB581 (ylbF sup1), SSB132 (ymcA), SSB563 (ymcA sinR) and SSB584 (ymcA sup1). SSB581 and SSB563 are suppressor mutant isolates obtained from pellicles arising after prolonged incubation of the ylbF and ymcA mutants. Scale bar is 1 mm.

Reinforcing the view that ylbF and ymcA lie upstream of sinR are the following observations concerning the appearance of spontaneous suppressors of mutations in ylbF and ymcA. Whereas ylbF and ymcA mutants in standing liquid medium were initially defective in pellicle formation, prolonged incubation resulted in the appearance of suppressor mutants that had regained the ability to form a pellicle. Moreover, after clonal purification, these suppressor mutants were found to produce colonies with complex architectural features (Fig. 6B). Suspecting that the suppressors had acquired second-site mutations in sinR (as we had observed for similar suppressors of sinI), we amplified by polymerase chain reaction (PCR) and sequenced sinR from each of the suppressor mutants. In both cases, the suppressors were found to contain a missense mutation in the ORF for sinR (Fig. 3). The simplest conclusion from these findings is that like sinI, the genes ylbF and ymcA lie upstream of sinR in a pathway that is responsible for reversing SinR-mediated repression of the eps operon.


The results of this investigation lead us to conclude that SinR is a central regulator in the assembly of B. subtilis cells into multicellular communities. SinR plays this role by repressing the transcription of the eps operon, which directs EPS biosynthesis, as well as other, yet-to-be-identified target genes. EPS is, in turn, responsible for the adhesion of chains of cells into bundles. These bundles appear essential to give rise to complex architectural features of biofilms, such as fruiting bodies.

We further propose that activation of the eps operon (and other SinR targets) requires the action of SinI, YlbF and YmcA, which are collectively responsible for reversing, or otherwise counteracting, SinR-mediated repression. SinI antagonizes the binding of SinR to DNA in vitro (Bai et al., 1993; Lewis et al., 1998), as confirmed by the results presented herein. However, our genetic analyses indicate that, while necessary, SinI was insufficient to fully overcome SinR-mediated repression in vivo. High-level expression of the eps operon additionally required the action of YlbF and YmcA. These proteins are similar to each other in their amino acid sequences, and it is appealing to imagine that they interact to form a heteromeric complex. Conceivably, YlbF and YmcA help to counteract SinR-mediated repression and do so by interacting (separately or as a heteromeric complex) with SinI or SinR (or both) or by stimulating the synthesis or stability of SinI. Alternatively, YlbF and YmcA could influence the expression of the eps operon in some other indirect manner. Indeed, YlbF and YmcA mutants display some pleiotropy in that they grow slightly more slowly than the wild type (S. S. Branda, D. B. Kearns, R. Losick and R. Kolter, unpubl. results).

Interestingly, sinI is under the positive control of the sporulation regulatory proteins Spo0A and σH (Gaur et al., 1988; Shafikhani et al., 2002). Previous work has shown that mutations in the genes (spo0A and spo0H, respectively) for Spo0A and σH block biofilm and fruiting body formation (Branda et al., 2001; Hamon and Lazazzera, 2001). In light of the role assigned to SinI in the present investigation, we suppose that Spo0A and σH contribute to multicellularity in part by stimulating the synthesis of the SinR antagonist. Hamon and Lazazzera (2001) have reported that the requirement for Spo0A in biofilm formation can be bypassed by a mutation in abrB, a repressor gene that is under the negative control of Spo0A. We have found that a mutation in sinI is epistatic to an abrB mutation (that is, an abrB mutation did not relieve the requirement for sinI in biofilm formation in an abrB sinI double mutant; D. B. Kearns, S. S. Branda, R. Kolter and R. Losick, unpubl. results). We therefore conclude that, like YlbF, YmcA and SinI, AbrB somehow acts to counteract SinR-mediated repression.

Finally, we note that SinR has traditionally been studied in the context of domesticated, laboratory strains, which are incapable of swarming motility and form neither robust pellicles in standing liquid medium nor architecturally complex colonies on solid medium. Such studies have implicated SinR in diverse processes, such as genetic competence (Hahn et al., 1994; Liu, L. et al., 1996), production of autolysins (Sekiguchi et al., 1988; Kuroda and Sekiguchi, 1993; Rashid and Sekiguchi, 1996), production of secreted proteases (Gaur et al., 1991; Bai et al., 1993; Olmos et al., 1997), swimming motility (Fein, 1979; Pooley and Karamata, 1984; Sekiguchi et al., 1990; Barilla et al., 1994; Márquez-Magaña et al., 1994; Fredrick and Helmann, 1996) and sporulation (Gaur et al., 1986; Mandic-Mulec et al., 1992; Louie et al., 1992; Cervin et al., 1998). The direct targets of SinR action in most of these cases are unknown, and the overall physiological significance of SinR has been difficult to elucidate. In the context of wild B. subtilis strains, however, SinR can be seen as a master regulator that governs entry into two alternative, and mutually exclusive, physiological states: a motile state in which the bacteria are capable of swimming as single cells in liquid or swarming in groups of cells on surfaces, and a non-motile state in which the bacteria form sessile, multicellular communities.

Experimental procedures

Strains and growth conditions

Bacillus subtilis PY79, 168 and 3610 were grown in Luria–Bertani (LB) 10 g tryptone, 5 g yeast extract, 5 g NaCl per L broth  or  LB  plates  supplemented  with  1.5% agar  at  37°C. For pellicle formation experiments, 12 µl mid-log phase culture was inoculated into 12 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) and incubated at 22°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 at 22°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 (ml).

‘Swarm agar’ plates (25 ml) containing LB fortified with 0.7% agar were prepared fresh and the following day were dried for 30 min in a laminar flow hood. Each plate was toothpick inoculated from an overnight colony and scored for swarming motility after 24 h incubation at 37°C (Kearns and Losick, 2003). Plates were visualized with a Bio-Rad geldoc system and digitally captured using Bio-Rad Quantity One software. A tripod mounted Sony CoolPix950 digital camera was used to capture swarm images illuminated by an oblique transmitted light source.

The handling of sinI, ylbF and ymcA mutants required special care as prolonged incubation in standing liquid minimal medium or as colonies on solid minimal medium readily led to outgrowth of suppressor mutants. As a precaution, we frequently returned to frozen stocks of our strains and conducted genetic backcrosses when necessary to ensure that strains were free of suppressor mutations.

Strain construction

All insertion deletion mutations were generated using long flanking homology PCR (using primers indicated in TableS2) and transformed into competent cells of strain PY79 (Wach, 1996). DNA containing a spectinomycin drug resistance gene (pDG1726) was used as a template for marker replacement (Guérout-Fleury et al., 1995). Mutations were transferred to the 3610 background using SPP1 mediated generalized transduction (Yasbin and Young, 1974). All strains used in this study are listed in TableS1.

SPP1 phage transduction

To 0.2 ml of dense culture grown in TY broth (LB broth supplemented after autoclaving with 10 mM 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 of 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 of TY broth at 37°C. A total of 0.9 ml of cells were mixed with 10 µl of SPP1 donor phage stock. A total of 9 ml 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. A total of 100 µl of cell suspension was then plated on TY fortified with 1.5% agar, the appropriate antibiotic, and 10 mM sodium citrate.

Sequencing sinR

A PCR product containing the sinR gene was amplified from B. subtilis chromosomal DNA (either from strain 3610 or the appropriate suppressor strain) using the primers −33 sinR F/+512 sinR R. The sinR PCR product was then sequenced using either primer individually.

Reporter and protein expression constructs

All primers used in the construction of plasmids are listed in Table S2. To generate the PepsA-lacZ reporter construct pFC1, a PCR product containing the PepsA promoter was amplified from B. subtilis 3610 chromosomal DNA using primers PepsAF and PepsAR. The PCR product was cloned into the EcoRI  and  BamHI  sites  of  plasmid  pDG268,  which  carries a chloramphenicol-resistance marker and a polylinker upstream of the lacZ gene between two arms of the amyE gene (Antoniewski et al., 1990).

To generate plasmids for the expression of N-terminal 6-histidine translation fusions to SinR and SinI (pDP90 and pDP91), PCR products containing the sinR and sinI genes were amplified from B. subtilis 3610 chromsomal DNA using primers H6sinRF/H6sinRR and H6sinIF/H6sinIR respectively. The PCR products were then cloned into the NheI/XhoI sites of pET28a(+) (Novagen).

β-Galactosidase assay

One millilitre of cells were harvested from a mid-log phase (OD600∼0.5) culture grown in MSgg broth shaken at 37°C, harvested and resuspended in an equal volume of Z buffer (40 mM NaH2PO4, 60 mM Na2HPO4, 1 mM MgSO4, 10 mM KCl and 38 mM β-mercaptoethanol). To each sample, lysozyme was added to a final concentration of 0.2 mg ml−1 and incubated at 30°C for 15 min. Each sample was diluted appropriately in 500 µl of Z buffer and the reaction was started with 100 µl of 4 mg ml−1 2-nitrophenyl β-D-galactopyranoside (in Z buffer) and stopped with 250 µl 1 M Na2CO3. The OD420 of the reaction mixtures was recorded and the β-galactosidase specific activity was calculated according to the equation: [OD420/time × OD600)] × dilution factor × 1000.

Protein purification

Plasmids pDP90 and pDP91 were transformed into the Invitrogen overexpression strain ‘CodonPlus’, a BL21 DE3 derivative carrying a plasmid bearing three rare codons (arginine, isoleucine and leucine) for optimal expression of heterologous proteins. A total of 500 ml of culture was grown in LB broth supplemented with 25 µg ml−1 kanamycin and 50 µg ml−1 chloramphenicol at 30°C until an OD600 of 0.5 was obtained, at which point IPTG was added to a final concentration of 1 mM. The cultures were incubated for 2 h at 30°C, washed once in 25 ml of PBS buffer, harvested and resuspended in 18 ml of lysis/binding buffer (50 mM Tris HCl, 500 mM NaCl and 10 mM imidazole, pH 8.5). Two millilitres of Novagen bug buster lysis solution and 10 µl of Novagen benzonase nuclease (25 U µl−1) was added to the suspension, and rotated end over end for 1 h at room temperature. The lysate was centrifuged at 5000 r.p.m. for 5 min to remove most of the cell debris and then the supernatant was ultracentrifuged at 35 000 r.p.m. for 30 min at 4°C.

One millilitre of Ni-NTA agarose beads (Qiagen) was added to the cleared lysate and rotated for 1 h at 4°C. The lysate/bead mixture was then loaded onto a column and washed five times, each with two bed volumes of wash buffer (50 mM Tris HCl, 500 mM NaCl and 20 mM imidazole, pH 8.5). The beads were recollected in 1 ml of elution buffer (10 mM Tris HCl, 10 mM MgCl2, 1 mM EDTA, 0.3 mM DTT, 5% glycerol, 1 mM PMSF, pH 8.5) (Gaur et al., 1991), to which 3 µl of 1.4 U µl−1 biotinylated thrombin (Novagen) was added and rotated at room temperature overnight. The eluate was recovered by passaging the bead slurry over a fresh column.

To remove the biotinylated thrombin, 130 µl of streptavidin agarose was added to the protein-containing supernatant, and rotated for 1 h at room temperature. The supernatant was recovered by passaging over a fresh column. Finally, the protein samples were dialysed against dialysis buffer (10 mM Tris HCl, 10 mM MgCl2, 1 mM EDTA, 0.3 mM DTT, 50% glycerol, 1 mM PMSF, pH 8.5), aliquoted and stored at −80°C.

Electrophoretic mobility shift assay (EMSA)

DNA probes were generated by PCR using chromosomal DNA from B. subtilis 3610 and the following primer combinations: PepsAF2/PepsAR2 (PepsA probe), PaprEF/PaprER (PaprE probe), ECH245/ECH246 (PyvbA probe), omf187/omf188 (PspoIIA probe) and Pspo0AF/Pspo0AR (Pspo0A probe). Each probe was gel purified and 5′ end labelled with 10 µCi of [γ−32P]-ATP (NEG002A, New England Nuclear) and polynucleotide kinase (New England Biolabs). DNA binding reactions were conducted in 30 µl of binding buffer (10 mM Tris HCl, 50 mM NaCl, 1 mM EDTA, 5% glycerol, 1 mM DTT, 10 µg ml−1 BSA) containing 25 µg ml−1 polydeoxyinosinic-deoxycytidylic acid (poly dI-dC). Various concentrations of SinR were added to approximately 100 nM radiolabelled DNA probe and incubated for 20 min at room temperature. A total of 10 µl of each binding reaction was loaded on a 6% polyacrylamide 0.5× TBE gel and resolved for 1 h at 200 mV. When appropriate, SinI and SinR were combined and incubated for 15 min at room temperature prior to the addition of radiolabelled probe.

Primer extension assay

Total RNA was isolated from mid-log B. subtilis 3610 cells grown in MSgg liquid medium. RNA was isolated using the hot acid/phenol method (protocol available at http://mcb.harvard.edu/losick/fawcettpaper/RNAprep.htm).

The primer extension product was obtained using the primer + 94epsAR (TableS2) 5′ end labelled with 40 µCi [γ-32P]-ATP and polynucleotide kinase. The radiolabelled primer (0.2 pmol) was annealed to 20 µg of total RNA in a 10 µl reaction volume of 1× first strand buffer (Invitrogen SuperScript First Strand Synthesis System for RT PCR). The annealing reaction was heated to 95°C for 1 min, transferred to 70°C for 10 min, and then put on ice for 2 min. The extension reaction was carried out using 5 µl of the annealing reaction, 0.01 M DTT, 1 mM dNTP’s, 1× first strand buffer and 10 units of Invitrogen SuperScript II RNase H Reverse Transcriptase. The reaction was incubated at 45°C and stopped with 5 µl of formamide loading buffer (80% deionized formamide, 10 mM EDTA, 1 mg ml−1 xylene cyanol FF, 1 mg ml−1 bromophenol blue).

A sequencing ladder was generated to a DNA fragment generated by PCR using primers −565 epsAF/+94epsAR (Table S2). A total of 2 µl of PCR product was denatured in 0.2 N NaOH and 100 µM EDTA for 5 min at room temperature before adding ammonium acetate to a final concentration of 0.8 M. The DNA was ethanol precipitated and resuspended in 7 µl of water, to which sequenase (USB) reaction buffer and 0.4 pmol of radiolabelled primer were added. The mixture was heated to 65°C for 2 min and cooled to room temperature for 10 min. To the annealing mix, 2 µl 1.5 µM dNTP’s, 1 µl 0.1 M DTT, and 3.2 units of sequenase (USB) were added. The reactions were incubated at room temperature for 90 s. A total of 3.5 µl of the reaction was transferred to tubes containing 200 pmol dNTP's with 20 pmol of the appropriate ddNTP. After incubating the termination reactions at 37°C for 5 min, the reactions were stopped with 4 µl of formamide loading buffer.

Prior to loading the reactions on an 8% sequencing gel, the primer extension and sequencing reactions were heated to 95°C for 5 min. A total of 7 µl of the primer extension and 4 µl of the sequencing reaction were resolved on the gel and visualized by phosphoimaging.

DNase I footprinting assay

Primer PepsAF2 (100 pmol) was 5′ end labelled with 70 µCi of [γ−32P]-ATP and polynucleotide kinase. A PCR product containing the PepsA promoter region was generated using 20 pmol radiolabelled PepsAF2 primer and 20 pmol unlabelled PepsAR2 primer using B. subtilis 3610 chromosomal DNA as a template. The PCR product was gel purified and subjected to scintillation counting such that, for each reaction, 30 000 c.p.m. of radiolabelled PCR product was mixed with SinR protein in 100 µl of footprinting buffer (20 mM Tris pH 8.0, 5 mM MgCl2, 5 mM CaCl2, 0.1 mM DTT, 0.1 mM EDTA,  50 µg ml−1 bovine  serum  albumin)  containing  5 µg ml−1 poly dI-dC and incubated for 15 min at room temperature. To each mixture, 2 µl of DNase I (1:50 dilution of  1 U µl−1 stock, Invitrogen) was added and incubated for 30 s at room temperature before digestion was inhibited by the addition of 25 µl of stop solution (1.5 M sodium acetate pH 5.3, 20 mM EDTA and 400 µg ml−1 glycogen). Each reaction was ethanol precipitated and resuspended to a final volume of 8 µl in formamide running buffer (80% deionized formamide, 10 mM EDTA, 1 mg ml−1 xylene cyanol and 1 mg ml−1 bromophenol blue). A total of 4 µl of each sample was loaded on an 8% sequencing gel (SequaGel Sequencing System, National Diagnostics) and resolved for 2 h at 35 mW.


We thank D. Dubnau, members of the Losick and Kolter laboratories, and, in particular, M. Fujita, E. Hobbs, A. Handler and B. Gorbatyuk for helpful advice and discussions. This work was supported by an NIH National Research Service Award GM66612 to D.K., National Institutes of Health grants GM18568 to R.L. and GM58213 to R.K. and American Cancer Society postdoctoral fellowship PF0033201MBC and The Medical Foundation/Charles A. King Trust (Fleet National Bank) to S.S.B.

Supplementary material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/mmi/mmi4440/mmi4440sm.htm

Fig.S1. Effect of a sinI mutation on the formation of a centrally located population of non-motile cells during swarming.

TableS1  Strains and plasmids.

TableS2  Primers.


  • 1

    We noticed that the predicted products of two adjacent genes (originally named yvfA and yvfB; Kunst et al., 1997) showed sequence similarity to the N and C terminal regions respectively, of an orthologous group of proteins (Tatusov et al., 1997) (COG2244, which includes E. coli WzxE and other proteins thought to mediate export of O-antigen polysaccharides; Liu, D. et al., 1996; Rick et al., 2003), suggesting the possibility that a single ORF had been annotated as two resulting from a sequencing error. Indeed, we found that B. subtilis 168 and 3610 contain a G-C insertion in codon 101 of the reported yvfA sequence, resulting in the merger of yvfA and yvfB into a single, 505-codon-long ORF, which we have named epsK.