Spatial Regulation of Histidine Kinases Governing Biofilm Formation in Bacillus subtilis

Abstract

Bacillus subtilis is able to form architecturally complex biofilms on solid medium due to the production of an extracellular matrix. A master regulator that controls the expression of the genes involved in matrix synthesis is Spo0A, which is activated by phosphorylation via a phosphorelay involving multiple histidine kinases. Here we report that four kinases, KinA, KinB, KinC, and KinD, help govern biofilm formation but that their contributions are partially masked by redundancy. We show that the kinases fall into two categories and that the members of each pair (one pair comprising KinA and KinB and the other comprising KinC and KinD) are partially redundant with each other. We also show that the kinases are spatially regulated: KinA and KinB are active principally in the older, inner regions of the colony, and KinC and KinD function chiefly in the younger, outer regions. These conclusions are based on the morphology of kinase mutants, real-time measurements of gene expression using luciferase as a reporter, and confocal microscopy using a fluorescent protein as a reporter. Our findings suggest that multiple signals from the older and younger regions of the colony are integrated by the kinases to determine the overall architecture of the biofilm community.

Undomesticated strains of Bacillus subtilis form architecturally complex communities on solid medium (colonies) and at the surfaces of standing cultures (pellicles) (3). The cells in these biofilms are in long parallel chains that are bound together by an extracellular matrix consisting of exopolysaccharide and an amyloid-like fiber composed of the TasA protein (3, 6, 12, 18). The exopolysaccharide is produced by the gene products of the epsA-epsO operon, and the TasA protein is encoded within the yqxM-sipW-tasA operon (referred to below as yqxM). The expression of the matrix operons is controlled by phosphorylation of the response regulator Spo0A, but this regulation is indirect (10). The matrix operons are under the direct negative control of two repressors, AbrB and SinR (10, 12). Spo0A∼P shuts off the synthesis of AbrB and directs the synthesis of an antirepressor, SinI, that binds to, and inactivates, SinR, thereby derepressing the epsA-epsO and yqxM operons (2, 4, 12, 21). Thus, the phosphorylation of Spo0A and its control are critical for biofilm formation (8, 14, 23). This phosphorylation is governed by the activity of five histidine sensor kinases via a multicomponent phosphorelay consisting of the phosphorelay proteins Spo0F and Spo0B (11, 13). To better understand the contributions of these kinases to biofilm formation, we undertook a study of the spatial organization of their activities.

The histidine kinases are KinA, KinB, KinC, KinD, and KinE. KinA and KinE are cytoplasmic proteins, whereas KinB, KinC, and KinD are membrane proteins. KinB functions in association with a lipoprotein known as KapB (7). Each of the kinases responds to environmental cues, but the nature of the signals that the kinases directly recognize is largely mysterious. A clue in the case of KinC comes from the observation that diverse but unrelated natural products that cause potassium leakage, including the B. subtilis cyclic lipopeptide surfactin, specifically activate KinC via its cytoplasmic PAS-PAC sensor domain (16, 22).

Here we report that four of the kinases, KinA, KinB, KinC, and KinD, contribute to the formation of architecturally complex communities on solid medium but that these contributions are partially masked by redundancy in a pairwise fashion, with KinA and KinB representing one pair and KinC and KinD representing the other. We also show that the kinases' activities show spatial organization in the colonies, with KinA and KinB directing matrix gene expression in the inner zones of colonies and KinC and KinD directing expression in the peripheral zones.

MATERIALS AND METHODS

Bacterial strains, culture media, and culture conditions.

The Bacillus subtilis strains used in this study are listed in Table S1 in the supplemental material. The ΔkinA, ΔkinB, ΔkinC, and ΔkinD mutations were constructed in PY79 (9, 16) and were transduced into NCIB3610 and kinase-mutant strains by using the Spp1 bacteriophage according to previously described methods (25).

The strains were routinely manipulated in LB medium (Sigma), TY medium (LB medium with the addition of 10 mM MgSO₄ and 100 μM MnSO₄), or MSgg medium (5 mM potassium phosphate, 100 mM morpholinepropanesulfonic acid, pH 7 [MOPS], 2 mM MgCl₂, 50 μM MnCl₂, 50 μM FeCl₃, 700 μM CaCl₂, 1 μM ZnCl₂, 2 μM thiamine, 0.5% glycerol, 0.5% glutamate, and 50 μg/ml [each] threonine, tryptophan, and phenylalanine). Solid medium contained 1.5% Bacto agar. Colony biofilms were grown on solid MSgg medium for 3 days at 30°C, and pellicle biofilms were grown for 3 days in 7 ml of liquid MSgg medium in 6-well plates at 22°C. We used different temperatures, because we observed the greatest range of colony phenotypes on plates incubated at 30°C and the greatest range of pellicle phenotypes in cultures grown at 22°C. Colony and pellicle images were optimized for contrast and brightness using Adobe Photoshop. Antibiotics were added, when necessary, at the following concentrations: 100 μg/ml spectinomycin, 10 μg/ml tetracycline, 5 μg/ml chloramphenicol, 0.5 μg/ml erythromycin, 2.5 μg/ml lincomycin, and 5 μg/ml kanamycin.

Construction of luciferase and green fluorescent protein (GFP) reporter strains.

The promoter for sinI expression was PCR amplified from genomic DNA; for the primer sequences, see Table S2 in the supplemental material. The PCR product was cloned into the EcoRI and SalI restriction sites of pAH321 (20). The resulting reporter was allowed to integrate into the neutral sacA locus of laboratory strain PY79 by natural competence and double recombination (19). The resulting integrated reporter was then transduced into NCIB3610 and kinase-mutant strains by using the Spp1 bacteriophage according to previously described methods (25).

The promoters for kinA, kinB, kinC, and kinD were PCR amplified from genomic DNA (see Table S2 in the supplemental material) and were ligated to plasmid pDG1728 (4) using the EcoRI and BamHI restriction sites. The resulting lacZ fusions were integrated into the amyE locus of laboratory strain PY79, and the integrated reporter was then transduced into NCIB3610 as described above.

Amplified kinB DNA was cloned into the BglII and EcoRI sites of PYC121 (5). The resulting kinB-gfp fusion was integrated into the amyE locus of laboratory strain PY79, and the integrated reporter was then transduced into NCIB3610 as described above.

A plasmid containing a kinD-gfp fusion (a gift from A. Banse) was introduced into the chromosome by single-reciprocal recombination. The fusion was created by introducing a fragment containing kinD into pkL147 (lab stock), which had been cut with EcoRI and XhoI.

The reporter amyE::P_yqxM-cfp (Spec) (24) was introduced into strains IKG120 (ΔkinA::mls ΔkinB::kan), IKG125 (ΔkinC::mls ΔkinD::tet), and IKG127 (ΔkinB::kan ΔkinC::mls) by using SPP1 phage-mediated transduction.

Real-time luciferase measurements of gene expression.

Starter LB cultures of desired strains were diluted 1:50 in double-distilled water (ddH₂O), and then 112.5 μl was plated onto 1.5 ml solid MSgg medium in each well (white with a clear bottom) of a 24-well polystyrene Visiplate (Wallac). The cultures were then grown for 48 h at 30°C in a BioTek Synergy 2 plate reader, and the optical density at 600 nm (OD₆₀₀) and luminescence (sensitivity setting, 200) were measured every 10 min. The data shown in the figures are averages for two replicate wells from a single representative experiment, where luminescence values were normalized to the highest luminescence recorded.

Confocal microscopy to determine the localization of matrix gene expression.

B. subtilis strains carrying amyE::P_yqxM-cfp (Spec) were grown on cover slides between two pieces of MSgg agar. Cells were grown to the logarithmic phase in LB medium, and then 2 μl of the culture was placed on a slide covered with 40 μl MSgg agar. After the droplet dried, it was overlaid with 350 to 400 μl MSgg agar (see Fig. S3 in the supplemental material). The cells were grown for 60 h at room temperature in petri dishes under humidity-controlled conditions. The setup allowed sufficient aeration for growth. The slides were examined by confocal microscopy, and a whole colony was imaged using a Leica SP5 confocal microscope. Due to the restrictive growth conditions, the average diameter of a colony was about 1 mm, and the average thickness was about 60 μm. To allow comparison between the behaviors of different strains, all measurements were taken using the same acquisition conditions (laser intensity, gain pinhole radius, exposure time). Twenty to 40 Z-stacks covering the entire height of the biofilm were collected at 2-μm intervals and were then tiled together to give a millimeter-sized 3-dimensional (3D) image. The image was then averaged along the z axis to obtain the 2D x-y spatial distribution of P_yqxM expression.

RESULTS AND DISCUSSION

Mutants of individual kinases exhibit subtle alterations in colony morphology.

B. subtilis 3610 produces highly structured colonies on a solid, biofilm-inducing medium. These biofilms exhibit wrinkles and surface projections whose formation is dependent on the phosphorylation of Spo0A via the phosphorelay. The removal of individual phosphorelay kinases through mutation caused subtle but detectable alterations in colony morphology. In what follows, we distinguish between the peripheral zone of the colony—the narrow band near the colony edge—and the inner zone, representing the remainder of the colony. For single mutants of kinA or kinB, little effect was seen in the peripheral zone, but close inspection revealed a somewhat reduced pattern of wrinkles and ridges within the inner zone (Fig. 1 A). In contrast, the inner zones of kinC or kinD mutant colonies were slightly more wrinkled than that of the wild type, while the peripheral zones exhibited a flatter morphology (Fig. 1A; for a close-up of the inner zone, see Fig. S1 in the supplemental material). Under our conditions, the colonies of a kinE mutant were indistinguishable from those of the wild type; hence, kinE will not be considered further here.

FIG. 1.

Pairs of kinases make distinct contributions to biofilm formation. (A) Top-down images of colonies of the wild type (WT) and mutant strains grown on solid MSgg medium. The strains are NCIB3610 (wild type), FC63 (kinA), FC64 (kinB), ALM104 (kinC), FC25 (kinD), and IKG85 (spo0F). (B) Colonies of the wild type and the indicated double mutants. The strains are NCIB3610, IKG121 (kinA kinB), ALM105 (kinC kinD), IKG126 (kinA kinC), FC112 (kinA kinD), ALM136 (kinB kinD), and IKG127 (kinB kinC). The rightmost image in the top row is a composite of two split images, one of the kinC kinD double mutant (left) and one of the kinA kinB double mutant (right).

It should be noted that kinB and kinC are followed in the chromosome by genes in the same orientation (kapB and ykqA, respectively), raising the possibility that the phenotypes observed were due to polar effects on the expression of the downstream genes. However, the kinC mutation could be complemented with a wild-type copy of the gene that had been integrated into the chromosome at the amyE locus, a finding that indicates that the biofilm phenotype of the kinC mutant was due solely to the loss of KinC function (data not shown). In the case of kinB, the function of the kinase is known to be dependent on the product of the downstream gene kapB (7), and indeed, a kapB mutation exhibited the same effect on biofilm formation as that observed for the kinB mutation.

We conclude that KinA, KinB, KinC, and KinD each contribute to the structural complexity of the colony but that the contributions of individual kinases are subtle. This is in contrast to the findings for a mutant of the phosphorelay protein Spo0F, which is downstream of KinA, KinB, KinC, and KinD (17); this mutant forms smooth, unstructured colonies.

The kinase contributions to colony morphology are masked by redundancy.

We hypothesized that the contributions of KinA, KinB, KinC, and KinD were masked by redundancy and that elimination of multiple kinases would result in more-extreme changes in colony morphology. Because both KinA and KinB seemed to influence the inner zones of the colonies and KinC and KinD seemed to influence the peripheral zones, we first examined the colony phenotypes of kinA kinB and kinC kinD double mutants. Indeed, the peripheral zone of the kinA kinB double mutant was relatively unaltered, but the inner zone was much less structured than the corresponding region of the wild type (Fig. 1B). In fact, the kinA kinB double mutant completely lacked the typical wrinkles and ridges in this inner zone. An even more striking two-zone phenotype was seen for the kinC kinD double mutant, whose peripheral zone was almost completely flat but whose inner zone was even more wrinkled than the corresponding region of the wild type. The contrast between the two double mutants is seen especially clearly in the split images of the rightmost panel in the top row of Fig. 1B. We also examined the phenotypes of kinA kinC, kinA kinD, kinB kinC, and kinB kinD double mutants. Here effects on both the inner and outer zones could be seen, but in none of the cases were these effects as apparent as for the other two double mutants.

These results reinforce the view that all four kinases contribute to colony structural complexity but that the contributions are partially masked by redundancy. In addition, the contributions of the kinases seem to fall into two classes, with KinA and KinB contributing mainly to the structural complexity of the inner zones of the colonies and KinC and KinD contributing mainly to the peripheral zones. It has been shown previously that the expression of the kinases does not vary dramatically during growth; thus, the differences observed are likely due to the activities of the kinases (11). We interpret these results as indicating that the microenvironments in the peripheral and inner zones differ due to growth and the consumption of nutrients and that KinA and KinB respond to (unknown) environmental signals found in the inner zone while KinC and KinD respond to signals principally confined to the peripheral zone.

We also examined the effects of double mutations on pellicle formation. Pellicles are floating biofilms that develop from a uniformly inoculated liquid culture and are exposed to a common medium where the diffusion of nutrients and signaling molecules occurs rapidly. This means that all locations on the surfaces of pellicles are likely to experience similar microenvironments, minimizing the spatial differences observed across biofilms formed on solid medium. Indeed, we obtained little evidence for zonal effects of single and double mutants. Nonetheless, the examination of 3-day-old pellicles confirmed the expectation that the contributions of the different kinases to biofilm formation were masked by redundancy. That is, all of the double mutants were more impaired in pellicle formation than the single mutants, with the kinA kinB double mutant exhibiting the most conspicuous defect in pellicle formation (Fig. 2; see also Fig. S1 in the supplemental material).

FIG. 2.

The contributions of kinases to pellicle formation are redundant. Shown are top-down images of floating biofilms formed at the air-liquid interfaces of standing cultures on MSgg medium incubated at 22°C for 3 days.

Monitoring of gene expression in real time reveals that all four kinases contribute to sinI expression.

Matrix production is regulated by the antirepressor SinI (4), which is synthesized under the control of Spo0A∼P (1). We were thus interested in determining the effect of the kinase gene disruptions on sinI expression. To monitor sinI gene expression in real time in growing colonies and pellicles, we used a P_sinI-luciferase fusion and measured light production in a plate reader with a luminometer. When grown on solid MSgg medium, wild-type cells expressed sinI robustly, beginning at about 12 h and peaking about 24 h after plating of the cells (Fig. 3, dark blue curve). The expression of the P_sinI-luciferase fusion was reduced by single mutations in kinA, kinB, kinC, and kinD, indicating that all four kinases are necessary for maximal sinI expression. In addition, different mutations caused different effects on the temporal pattern of sinI expression. Mutations in kinA or kinB caused an overall reduced level of sinI expression, whereas mutations in kinC and kinD reduced maximal levels of expression at early times but also caused the appearance of a secondary peak of expression (Fig. 3A).

FIG. 3.

Measurement of the transcription of sinI in real time using luciferase as a reporter reveals redundancy in the contributions of the kinases. Shown are the time courses of light emission from colonies of the wild type and the indicated single mutants (A) or double mutants (B) harboring sacA::P_sinI-luciferase (see Materials and Methods). The strains used are derivatives of the corresponding strains used in Fig. 1 harboring the luciferase construct.

The use of double mutants reinforced the view that the four kinases contribute to sinI expression and that they can be grouped into two functional pairs. A kinA kinB double mutant was severely impaired in sinI expression, and a kinC kinD double mutant exhibited an even more pronounced late-appearing peak of expression than either single mutant (Fig. 3B). In contrast, combining mutations from each of the two phenotypic classes (kinB and kinC or kinA and kinD) led to less-dramatic impairment of sinI expression. The patterns of sinI expression in pellicles were somewhat different; we did not see sharp peaks of expression in the wild type, nor did we see delayed peaks of expression in double kinase mutants. However, we did still see redundancy between the kinases in that double kinase mutants showed marked decreases in sinI expression; the kinA kinB mutant had the most severe phenotype (see Fig. S2 in the supplemental material).

Use of confocal microscopy reveals that kinase activity is subject to spatial regulation.

Based on the phenotypes of the kinA kinB double mutant and the kinC kinD double mutant, we hypothesized that KinA and KinB are active in the inner zone of the colony while KinC and KinD are active in the outer zone. To investigate this possibility, we used fluorescence reporters under the control of the yqxM matrix operon promoter in conjunction with confocal microscopy to visualize the localization of cells expressing matrix genes across the colony from its midpoint (Fig. 4, left side of the images) to its outer edge (right side). In wild-type colonies, matrix gene expression was distributed in a relatively uniform pattern across the colony. In comparison, a kinA kinB double mutant colony displayed the highest matrix gene expression in the peripheral zone of the colony. Conversely, in a colony from a kinC kinD double mutant, matrix gene expression occurred principally in the inner zone. Finally, and in contrast, a colony of a kinB kinC double mutant did not exhibit a well-defined spatially restricted pattern of matrix gene expression: expression was seen across the entire colony but at a level lower than that of the wild type. These results are consistent with the idea that KinA and KinB are active in directing matrix gene expression in the inner zone while KinC and KinD direct matrix gene expression in the outer zone, providing a simple explanation for the spatial effects on colony morphology seen for mutants of the two gene pairs.

FIG. 4.

The use of confocal microscopy and a fluorescent reporter reveals the spatial control of kinase activity. Shown are fluorescent images from 3-day-old colonies of the indicated wild-type (WT) and double mutant strains harboring a fusion of the gene for cyan fluorescent protein (CFP) to the promoter of the yqxM operon (P_yqxM-cfp). The cartoon indicates the region of the colonies from which the images were taken. Bar, 100 μm.

Finally, we considered the possibility that the apparent spatial control of the kinases was exerted at the level of the expression of the kinase genes rather than the activities of the kinases themselves. To address this, we used in-frame fusions of gfp to kinB and kinD in order to visualize the expression of the kinase genes spatially across colonies. Detecting expression was technically challenging, since the kinase genes are expressed at relatively low levels. Nonetheless, we were able to obtain clear evidence that the fusions were expressed throughout the colonies, with some evidence for enrichment of kinB-gfp expression in the edges and in the inner zone (see Fig. S5 in the supplemental material). Thus, at least for kinB and kinD, the observed spatial patterns of matrix gene expression cannot be attributed simply to a corresponding spatial restriction in kinase gene expression.

Spatial regulation.

How are we to explain the apparent spatial regulation of KinA, KinB, KinC, and KinD in colonies of biofilm-forming cells? The simplest interpretation of our results is that each of the kinases senses distinct signals and that the signals sensed by KinA and KinB are largely confined to the inner zone of the colony while those sensed by KinC and KinD are largely confined to the outer zone. We note that the central region of the colony represents the oldest portion of the biofilm. Hence, consumption of nutrients is likely to be greater in the central zone than closer to the periphery, where cell growth is still taking place. Perhaps, then, KinA and KinB respond to nutrient depletion, which is most acute in the colony center, or to metabolic by-products that are present at elevated levels in the older portions of the biofilm community (Fig. 5). Because KinC and KinD are most active at the outer edges of the colony, we hypothesize that these kinases respond to threshold levels of signals that are generated mainly near the edges of the colony (Fig. 5). For example, surfactin, which is known to stimulate the activity of KinC (16), might be present at levels optimal for maximal KinC activation in the peripheral zone of the community rather than in the inner regions. Just such a peripheral pattern of surfactin accumulation was recently observed by use of imaging mass spectrometry for a colony grown on a sporulation-inducing medium (15).

FIG. 5.

Model for the spatial control of kinase activity. The cartoon depicts a structured colony with wrinkles, and the blow-ups represent pairs of kinases that are active in different zones due to the differential distribution of unknown signaling molecules (represented by the small symbols in colors corresponding to those of their respective kinase).

Clearly, the activation of the regulatory circuitry governing biofilm formation is a complex and nuanced process. Important challenges for the future are both to identify the signals that each of the kinases recognizes and to determine how these signals are distributed across the biofilm.