All strains (Supplementary Table S1) were, if not indicated otherwise, routinely cultivated at 37 °C with agitation (180 rpm) in lysogeny broth (LB) media (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl). LB agar plates were prepared by the addition of 1.5% (w/v) agar-agar. Selection of B. subtilis strains harboring a resistance marker were carried out by using 5 μg/mL chloramphenicol, 10 μg/mL kanamycin, 1 μg/mL erythromycin along with 25 μg/mL lincomycin for MLS, or a combination of these antibiotics, respectively. For cultivation of E. coli strains carrying a plasmid 100 μg/mL ampicillin or 50 μg/mL kanamycin were used. Long time storage at −70 °C was achieved by supplementing the overnight culture with 50% (v/v) glycerol at a ratio of 1:1.

All restriction endonucleases, T4 DNA ligase, Q5 DNA polymerase, and OneTaq® DNA polymerase originated from New England Biolabs (Ipswich, MA, United States) and cloning procedures followed the respective protocols. DNA fragments were purified using Hi Yield® PCR Clean-up/ Gel Extraction kit (Süd-Laborbedarf GmbH, Gauting, Germany) or NucleoSpin® Gel and PCR Clean-up kit (Macherey-Nagel, Düren, Germany) and plasmids were isolated applying Hi Yield® Plasmid mini kit (Süd-Laborbedarf GmbH, Gauting, Germany) according to the corresponding manufacturer’s recommendations. Genomic DNA of B. subtilis was isolated either with the NucleoSpin® Microbial DNA kit (Macherey-Nagel, Düren, Germany) used as PCR template or by SC lysis for transformation (Harwood, 1990). Transformation of E. coli and B. subtilis was carried out as described previously (Harwood, 1990; Sambrook and Russell, 2001)^. Successful genetic or genomic modification was verified by colony PCR using OneTaq® DNA polymerase and sequencing. Plasmids and oligonucleotides used were listed in Table S2 and Table S3, respectively.

Targeted introduction of single nucleotide polymorphisms (SNPs), which resulted in stop codons within comS, was carried out using the CRISPR-Cas9 vector pJOE8999 (Altenbuchner, 2016), which contains the cas gene under control of a mannose-inducible promoter (P _manP ), a customizable guide RNA (gRNA) module driven by the semi-synthetic P _vanP promoter, a kanamycin resistance marker, and dual origins of replication (oriR): one from plasmid pUC19 for replication in E. coli and pET194^ts, a temperature sensitive oriR for replication in B. subtilis (Supplementary Figure S1a).

To generate the gRNA expression construct (pJOE8999_gRNA), 5′-phosphorylated oligonucleotides TM7980 and TM7981 purchased from Sigma-Aldrich (St. Louis, Missouri (MO), USA) were annealed by heating to 98 °C for 3 min and cooling to 25 °C at 0.2 °C/s in 1X r2.1 buffer (NEB, Ipswich, MA, United States). The resulting double-stranded DNA was ligated into BsaI-digested pJOE8999 using golden gate assembly (Engler et al., 2008; Supplementary Figure S1b). The reaction (10 μL) contained 50 ng vector, 1 μL BsaI, 1 μL T4 DNA ligase, 1 μL ligase buffer, 0.2 μL bovine serum albumin (BSA), and 5 μL of annealed oligonucleotides. Thirty cycles of digestion (37 °C, 5 min) and ligation (16 °C, 10 min) were followed by final digestion (37 °C, 1 h) and enzyme inactivation (80 °C, 10 min). The entire reaction was applied for transformation of E. coli DH10β and correct gRNA insertion was confirmed. The verified pJOE8999_gRNA plasmid was digested with SalI and Xbal, and the vector backbone was gel-purified from 1% (w/v) agarose 1X TAE gel (1 mM EDTA disodium salt, 40 mM TRIS, 20 mM acetic acid). Following, homology-directed repair (HDR) templates were assembled from PCR-amplified upstream and downstream regions flanking the comS target site, as well as a central fragment containing the desired SNP and a modified protospacer adjacent motif (PAM) sequence to prevent re-cleavage by Cas9 (Supplementary Figure S1c).

All fragments were amplified with oligonucleotides containing BsaI recognition sites. The use of BsaI as type IIS restriction enzyme that cleaves outside of its recognition site, enabled control over the resulting DNA overhangs and allowed markerless assembly of the HDR fragments while simultaneously introducing the desired point mutations including stop-gain SNP and PAM sequence alteration (Table 1; Supplementary Figure S1d).

Subsequently, golden gate assembly was performed to insert the HDR fragments into the vector using a 1:10 molar ratio of vector to each fragment (upstream, PAM exchange, and downstream part). Following, the constructed plasmids (pJOE8999_gRNA_comS C11A, C74A, C74G, T83A, and G128A) were used to transform B. subtilis. Transformants were selected on LB agar plates supplemented with kanamycin and 0.2% (w/v) mannose to induce Cas9 expression. Plates were incubated at 30 °C to maintain replication by the temperature-sensitive pE194^ts oriR. Colonies harboring the plasmid were re-streaked on LB agar plates without antibiotics and incubated at 50 °C overnight to promote plasmid backbone loss. Single colonies were patched onto LB agar plates with and without kanamycin, and screened for kanamycin sensitivity. The presence of the desired point mutations was verified by sequencing.

Luciferase-based reporter and implementation strains were constructed following the protocols described by Radeck et al. (2013) and Popp et al. (2017). Transcriptional luxABCDE (lux) fusions were generated using the vector pBS3Clux, into which the desired promoter fragments were inserted. The P _comGA , P _comK , and P _srfAA DNA fragments were amplified by using the oligonucleotides TM8067 & TM8068, TM7811 & TM7812, and TM7512 & TM7514, respectively.

The minimal medium MSggN (5 mM sodium phosphate buffer pH 7.0; 100 mM MOPS; 2 mM MgCl₂; 700 μM CaCl₂; 50 μM MnCl₂; 700 μM FeCl₃; 1 μM ZnCl₂; 2 μM thiamine hydrochloride; 0.5% (v/v) glycerol; 0.5% (w/v) K-glutamate; 100 μM KCl) (Fall et al., 2006), a modified MSgg medium (Branda et al., 2001), was used to analyze sliding motility. For solidification, 0.35% (w/v) agarose were added to the MSggN broth. The original MSgg medium (5 mM potassium phosphate buffer pH 7.0; 100 mM MOPS; 2 mM MgCl₂; 700 μM CaCl₂; 50 μM MnCl₂; 700 μM FeCl₃; 1 μM ZnCl₂; 2 μM thiamine hydrochloride; 0.5% (v/v) glycerol; 0.5% (w/v) K-glutamate; 50 μg/mL tryptophan, 50 μg/mL phenylalanine) was selected to form biofilms and pellicles. For pellicle generation 6 mL MSgg broth were transferred into each well of a 6-well plate (Greiner Bio-One, Frickenhausen, Germany). Biofilms were grown on MSgg plates solidified with 1.5% (w/v) agar-agar. Swarming behavior was investigated using LB soft agar plates with a final concentration of 0.7% (w/v) agar-agar. All agar plates were allowed to dry for at least 45 min before usage. For inoculation of the different media B. subtilis strains were standardized cultivated overnight without the addition of antibiotics. Pellicle formation was carried out by inoculation of MSgg broth to final optical density at 600 nm wavelength (OD_6oo) of 0.1 and incubation at 25 °C for 5 days.

All solid differentiation agar plates were inoculated at the center of the plate with 5 μL of cell suspension, which was previously diluted to OD₆₀₀ 0.1 with LB broth. While LB soft-agar plates (swarming assay), and MSggN agar plates (sliding assay) were incubated at 30 °C for 24 h, the MSgg agar plates (biofilm assay) were grown at 25 °C for 7 days. Digital documentation of the resulting morphologies was performed with either the PCAM camera (TU Dresden) or the binocular microscope WILD M3Z (Leica Microsystems GmbH, Wetzlar, Germany) with a ProgRes SpeedXT ^core®5 camera (Jenoptik, Jena, Germany).

The luciferase activity of the strains carrying promoter luxABCDE fusions (P _comK -lux; P _srfAA -lux; P _comGA -lux) was analyzed using the Synergy Neo 2 multi-mode microplate reader connected to the BioSpa 8 automated incubator from Agilent Technologies (Santa Clara, CA, United States), and controlled by the Gen5™ software. The assays were undertaken in 96-well plates (black walls, clear bottom, Greiner Bio-One, Frickenhausen, Germany) and cell growth was monitored at OD₆₀₀. Relative luminescence units (RLU) were measured and normalized to the optical density, whereby each RLU data point was divided by the corresponding OD₆₀₀ (RLU/OD₆₀₀). Strains were routinely cultivated overnight and used for inoculation of 5 mL minimal medium MNGE (2% (w/v) glucose; 0.2% (w/v) glutamate; 11 μg/mL Fe(III)-ammonium citrate; 3 mM MgSO₄; 92% (w/v) 1X MN) to an OD₆₀₀ of 0.1. Subsequently, the culture was grown at 37 °C with agitation (180 rpm) till OD₆₀₀ 0.2–0.4 and diluted to OD₆₀₀ 0.05. Following, 100 μL of the cell suspension were transferred into the 96-well plate and measurements were performed at 37 °C, whereby the plated was linear shaken for 10 s before each measurement. The OD₆₀₀ and luminescence were monitored every 9 min for at least 18 h. Data analysis and preparation of graphs was carried out by using GraphPad Prism 5.0 software (GraphPad, San Diego, CA, United States).

Biofilms were grown on mixed cellulose ester membranes with 0.22 μm average pore size and a thickness of 150 μm following the procedure described earlier (Friebel et al., 2025). The membranes were used to ensure easy detachment of the biofilms from the MSgg agar and to simplify subsequent inactivation as described previously (Friebel et al., 2025; Brockmann et al., 2019). For fixation and safe inactivation, bacterial biofilms were cut out and wetted with a solution of 10% (v/v) formaldehyde (Carl Roth GmbH & Co. KG, Karlsruhe, Germany) in water. After 30 min, fixation was complete and the biofilms were washed two times with Milli-Q water (Merck KGaA, Darmstadt, Germany) and then glued together with the membrane on a microscope slide using super glue (UHU GmbH & Co KG, Bühl, Germany). A VS200 microscope slide scanner with a 4X objective (Evident GmbH, Hamburg, Germany) and ORCA-Fusion C14440 20UP camera (Hamamatsu photonics, Hamamatsu City, Japan) was used for reflected light microscopy of biofilms. For microscope image visualization, OlyVia (version 3.4.1, Evident, Hamburg, Germany) was used. A solution of 7 mg/mL 2,5-dihydroxyacetophenone (Merck KgaA, Darmstadt, Germany) dissolved in 75% (v/v) acetonitrile (Carl Roth GmbH & Co. KG, Karlsruhe, Germany), 10% (v/v) methanol (Carl Roth GmbH & Co. KG, Karlsruhe, Germany), 10% (v/v) trifluoroacetic acid (Carl Roth GmbH & Co. KG, Karlsruhe, Germany) and 5% (v/v) Milli-Q water was sprayed on the biofilm as MALDI matrix using a SunCollect pneumatic spray roboter (SunChrom, Friedrichsdorf, Germany). The matrix was applied within 22 spraying cycles at a nitrogen back pressure of 2.5 bar in a meandering pattern. The flow rate followed a gradient and started at 15 μL/min in the first cycle and increased to 20 μL/min and 30 μL/min for the second and third spraying cycle, respectively. From the fourth spraying cycle onwards, the flow rate was 50 μL/min until the end. The spray nozzle moved with a velocity of 700 mm/min and a line distance of 2 mm. The distance between the spray nozzle and sample surface was set to 44 mm.

All MALDI-MSI measurements were conducted with a timsTOF fleX MALDI-2 QTOF mass spectrometer (Bruker Daltonics GmbH & Co. KG, Bremen, Germany), equipped with a Smartbeam3D™ laser emitting at 355 nm providing a quadratic focal spot size of 5 × 5 μm diameter. For all imaging experiments, the resulting field size was 50 μm using beam scan function and the “M5 small” setting with 200 laser shots per pixel with a frequency of 1 kHz and a nitrogen partial pressure in the MALDI source of 2.5 mbar. With the analytical focus on the detection of surfactin along with the peptide toxins EPE, SKF, and SDP, all herein reported MSI data were acquired in the “high sensitivity detection” mode within a m/z range of 1,000–4,500 in positive ion mode and with the MALDI-2-postionisation disabled. MSI data was processed using SCiLS Lab MVS (vs. 2026a Pro, Bruker Daltonics GmbH & Co. KG, Bremen, Germany) and ion images visualized as false color images using the same software. MSI experiments were performed in biological and technical duplicates.

In B. subtilis, competence development and surfactin production are genetically coupled though co-transcription of comS and srfAB. However, their gene products exhibit distinct physiological roles. While ComS stabilizes ComK by preventing its proteolytic degradation, initiating the competence pathway (Figure 1), srfAB encodes components for the non-ribosomal peptide synthetase complex responsible for surfactin biosynthesis (Figure 2). To investigate the role of surfactin independent of comS expression, we introduced targeted single nucleotide polymorphisms (SNPs) that lead to premature stop codons within the comS coding sequence (Figure 3).

These point mutations were strategically selected to truncate comS translation while preserving the amino acid sequence of the overlapping srfAB. Five distinct mutations within comS - 11C > A (Ser4*), 74C > A (Ser25*), 74C > G (Ser25*), 83 T > A (Leu28*), and 128G > A (Trp43*) - hereafter referred to as comS C11A, C74A, C74G, T83A, and G128A, were introduced into the comS coding sequence by CRISPR-Cas9-based genome editing, employing the vector pJOE8999 (Altenbuchner, 2016; Figure 3). This system allowed precise incorporation of the desired point mutations into the homology repair templates, which mediated repair of the Cas9-induced double-strand break at the target site within srfAB (Supplementary Figure S1). All mutations occur within codons that encode either isoleucine or leucine in the overlapping srfAB gene, allowing synonymous substitutions that preserve SrfAB protein function. Codon usage at the mutated sites in srfAB was evaluated based on Nakamura et al. (2000) to ensure translational efficiency was not compromised (Table S4). Although codon usage frequencies were either slightly reduced (C11A, C74G, T83A, and G128A) or modestly increased (C74A) compared to the wild type, all resulting codons retained a usage frequency of at least approximately 5 per thousand, suggesting they remain within the range of commonly used codons in B. subtilis (Nakamura et al., 2000). A summary of codon usage before and after mutagenesis is provided in Table S4. The mutations resulted in truncated ComS peptides of varying lengths: C11A resulted in the shortest version with 2 amino acids, whereas G128A led to almost complete translation of ComS lacking only the final three residues (Figure 3).

The phenotypes of the premature stop codon mutations in comS were subsequently compared to different srf mutants to investigate their respective roles in multicellular differentiation, including biofilm formation, pellicle development at the air-liquid interface, sliding motility, and swarming behavior (Figure 4).

On solid biofilm-inducing medium (MSgg; Branda et al., 2001), wild type strain DK1042 formed structured, wrinkled colonies, indicative of robust biofilm development. These morphology characteristics are associated with matrix production, endospore formation, and programmed cell death, features that collectively contribute to the spatial organization and wrinkling of mature biofilms (Gingichashvili et al., 2017; Arnaouteli et al., 2021). All comS stop codon mutants (C11A, C74A, C74G, T83A, and G128A) retained complex colony architecture comparable to the wild type, suggesting that truncation of ComS does not impair biofilm formation at the colony level (Figure 4a). In contrast, mutants harboring disruptions in the srf operon (srfAA:kan, srfAB:kan, srfAC:kan, srfAD:kan) showed reduced vein-like wrinkling and limited colony expansion. Moreover, the strains exhibited diminished spatial organization, with poorly defined inner, middle and outer zones. Notably, mutants carrying a comS deletion, accompanied with srfAB disruption within the overlapping sequence (comS:mls) or expressing comS in the absence of srfAABCD (srf:comS-mls) displayed similar altered phenotypes. The colonies were flatter and smaller than the wild type, but retained some degree of structural organization (Figure 4a).

Interestingly, the srfAA deletion mutant entirely lacked inner colony structuring and exhibited disorganized wrinkling across the colony surface, deviating from the typical vein-like pattern observed in wild type biofilms. The srfAC and srfAD mutants maintained central ring-like structure, but showed minimal pigmentation, reduced colony size, and nearly absent wrinkling. Overall, while all mutations in the srf operon exhibited severely impaired biofilm morphology and differentiation, the introduction of targeted stop codons in comS had no noticeable effect on colony architecture. This confirmed that surfactin production, rather that ComS-mediated competence regulation, is critical for maintaining structured multicellular biofilm formation on solid surfaces (Figure 4a).

At the air-liquid interface, the wild type and all comS stop codon mutants formed robust, wrinkled pellicle with reddish-brown color (Figure 4b). This pigmentation is presumably caused by the complex formation of pulcherriminic acid and iron to the reddish-brown pulcherrimin (Yuan et al., 2020). In contrast, strains lacking genes of the srf operon or carrying a full comS deletion failed to form stable pellicles. Instead, they produced thin, fragile surface films or remaining entirely dispersed in the medium (Figure 4b). The most severe defects were observed for the srfAC, srfAD and srf operon deletion strains, which were completely unable to establish pellicles and exhibited a pale white color in place of the typical red-brown pigmentation seen in the wild type (Figure 4b). These observations confirm that surfactin, but not ComS, is essential for pellicle integrity, and that comS stop mutants maintain sufficient surfactin expression to support pellicle formation (Figure 4b).

MSggN (Fall et al., 2006) soft agar plates were employed to study sliding motility, which is a form of passive, flagella-independent surface expansion (Kearns, 2010). Under these conditions, all comS point mutation strains displayed colony spreading comparable to the dendritic patterns of the wild type (Figure 4c). In contrast, surfactin-deficient mutants exhibited severely impaired expansion and extreme reduction of dendritic structures, demonstrating that sliding motility depends on surfactin, but is independent of ComS. Notably, srfAC and srfAD mutants showed the strongest defect: Colonies remained compact and lacked any visible spreading (Figure 4c).

Swarming, a flagellum-dependent form of multicellular movement on semi-solid surfaces, remained unaffected in comS point mutation strains, which exhibited swarming patterns comparable to the wild type (Figure 4d). This behavior is characterized by circular wave-like expansion, reflecting coordinated motility. In contrast, strains carrying deletions in srfAB, comS, srfAC, or srfAD exhibited impaired swarming motility, characterized by disorganized colony morphology with irregular offshoots and uneven radial expansion. Interestingly, srfAA deletion resulted in increased colony expansion, yet the pattern remained disordered, indicating a similar disruption in coordinated movement. These phenotypes implicate an essential role of surfactin in facilitating collective swarming behavior. Notably, deletion of the entire srf operon combined with reintroduction of comS (srf:comS-mls) led to a complete loss of swarming ability, further reinforcing the requirement of surfactin for motility (Figure 4d).

All of the above phenotypic results indicate that the introduced stop codons in comS did not affect surfactin production. To validate this assumption, biofilms were subsequently cultivated on mixed cellulose ester membranes (Supplementary Figure S2) and analyzed by matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MS imaging, MALDI-MSI) (Friebel et al., 2025). Colonies grown under these conditions displayed comparable morphologies to biofilms grown on MSgg medium only (Figure 4a), suggesting that the use of mixed cellulose ester membranes does not alter biofilm architecture (Supplementary Figure S2). The MALDI-MSI technique allowed spatial visualization of surfactin distribution within the colony at a resolution of about 50 μm (with the used settings). All identified surfactin variants (C13-C15) were detected predominantly in the positive ion mode as potassium adduct ions ([M + K]⁺). For example, the three surfactins are detected at mass-to-charge ratios (m/z) of 1046.6, 1060.63, and 1074.6, respectively. In all cases this is within 12 ppm of the calculated m/z values. In Figure 5, the distribution of surfactin C14 is shown as a representative signal, however the other isoforms showed essentially identical spatial distributions in line with previous MALDI-MS imaging studies on wild type B. subtilis strains (Friebel et al., 2025; Brockmann et al., 2019; Si et al., 2016).

While MALDI-MSI is not a quantitative method without use of internal standards or additional quantitation of extracts, the relative signal intensities can still provide spatial insights into the localization and relative abundance of metabolites, in particular if the ion type is investigated in comparably prepared samples (Figure 5). Interestingly, surfactin appeared to accumulate more strongly in the center of the biofilm and extended into the middle zones, although the signal was detectable across the entire colony. This pattern suggests a central production or accumulation zone with broader distribution throughout the biofilm structure (Figure 5).

Consistent with previous phenotypic observations, surfactin was detected in all comS point mutation strains (comS C11A, C74A, C74G, T83A, and G128A) carrying premature stop codons, confirming that the introduced mutations did not impair srf operon function. In contrast, strains with deletions of comS, srfAA, srfAB, or srfAC, as well as the srf:comS-mls mutant, did not produce surfactin (Figure 5; Supplementary Figures S3, S4). However, the absence of surfactin, resulting in a black image, does not indicate a lack of growth, as colonies were present for all these strains (Supplementary Figure S2). Notably, surfactin was still detected, although at low levels, in the srfAD mutant (Figure 5). This gene encodes a type II thioesterase that acts as a proofreading enzyme to recycle misacylated PCP domains rather than directly participating in surfactin assembly. Accordingly, surfactin is still produced in the absence of SrfAD. However, the observed impairment in multicellular behavior indicates that the surfactin biosynthetic machinery may not be fully functional in the absence of srfAD, potentially due to reduced efficiency or accumulation of defective intermediates (Figure 5; Supplementary Figures S3, S4).

Taken together, swarming and sliding assays, biofilm and pellicle formation, as well as MALDI-MS imaging confirmed that the stop codon mutations introduced into comS did not disrupt srf operon functionality and demonstrate that it should be possible to decouple competence regulation from surfactin-mediated multicellular behavior through precise genomic modification. Next, we therefore analyzed the impact of comS mutations on competence development.

A comparative analysis of competence-related promoter activities was conducted, focusing on expression patterns and intensities in the laboratory strain W168 and the biofilm-forming strain DK1042. Previous studies have shown that W168 exhibits approximately 1,000-fold higher transformation efficiency than DK1042 (Konkol et al., 2013; Miras and Dubnau, 2016), indicating that W168 should be more suitable for detailed investigation of competence regulation. Interestingly, the two B. subtilis strains displayed notable differences in growth behavior under competence-inducing conditions. When cultured in MNGE medium, strain DK1042 entered stationary phase after approximately 8 h of growth, whereas W168 grew more slowly and did not reach a stable stationary phase within the 18 h measurement period (Figure 6a).

Competence is controlled by quorum sensing-mediated activation of ComA, which initiates expression of the srf operon and comS by binding to the P _srfAA promoter region (Roggiani and Dubnau, 1993). Since expression of this locus is essential for competence development, P _srfAA activity was first examined in both strains (Figure 6a), which uncovered striking differences: In W168, promoter activity peaked around 5 h and then maintained an almost constitutive expression pattern (Figure 6a, II). In contrast, DK1042 showed an early peak in P _srfAA activity at about 2 h, with promoter activity declining after reaching its peak, indicating a brief and tightly regulated window of surfactin and comS expression (Figure 6a, I). Moreover, P _srfAA activity was approximately 30-fold lower in DK1042, compared to that of W168 (Supplementary Figure S5a), which is in agreement with the overall lower competence in this strain.

Since ComS has a higher binding affinity for the proteolytic complex MecA-ClpCP, it competitively inhibits ComK degradation, allowing ComK to accumulate and activate the competence regulon (D'Souza et al., 1994). P _comK was therefore chosen as an indicator of competence initiation and P _comGA as representative of the ComK-regulated, so-called late competence genes. Similar to the differences observed for P _srfAA activity between the two strains, both P _comK and P _comGA activities differed significantly (Supplementary Figure S5a). In W168, P _comK exhibited a relatively high basal activity compared to P _empty -lux control and showed an approximately ten-fold increase, reaching its maximum after around 12 h of growth (Figure 6b). In contrast, DK1042 showed the maximum P _comK activity already after around 4 h, followed by a decline to baseline levels by the onset of stationary phase (Figure 6b). However, the maximum was around 12-fold lower in comparison to W168. The P _comGA activity in W168 mirrored the expression pattern of P _comK , though it was approximately four-fold lower (Supplementary Figure S5b).

Surprisingly, no P _comGA activity above the background reporter signal was detected in DJ1042, which was approx. 90-fold lower when compared to W168 (Supplementary Figure S5c). While this does not conclusively prove the absence of P _comGA expression, it suggests that the activity is extremely low and below the sensitivity of the assay applied (Figure 6b).

Taken together, these results demonstrate that W168 exhibits markedly stronger and more sustained activity of competence-associated promoters (P _srfAA , P _comK , and P _comGA ) compared to DK1042, where promoter activities were transient and substantially weaker (Supplementary Figure S5, S6; Figure 6). Thus, W168 represents a more suitable strain for the analysis of competence-related gene expression and hence competence development.

The target-specific stop mutations in comS were subsequently introduced into B. subtilis W168 P _comK -lux and P _comGA -lux reporter strains to investigate the functional relevance on competence development. These constructs were compared with the wild type background as well as with comS and comK mutants. A promoterless P_empty-lux strain served as a background control. In the wild type strain, P _comK activity reached its maximum after approximately 12 h of growth (Figure 7a).

In contrast, comS mutant strains exhibited drastically reduced P _comK activity, up to around 14-fold decrease relative to the wild type at the peak time point (Figure 7b). All comS point mutations resulted in P _comK activity comparable to the comS deletion strain, indicating that the introduction of premature stop codons within comS is sufficient to disrupt competence initiation (Figure 7b). Deletion of comK led to even lower P _comK activity, likely due to the absence of ComK-mediated positive autoregulation of its own promoter (Figures 7a,b) (Berka et al., 2002; Van Sinderen and Venema, 1994). Interestingly, comS and comK mutations also affected growth dynamics, with the mutant strains entering a stationary phase after approximately 13 h and 9 h, whereas the wild type did not reach the stationary phase during the 18 h measurement period. This growth behavior resembles that of the biofilm-forming strain DK1042 (Figure 6a). Furthermore, P _comGA activity, which reflects activation of the ComK regulon, was significantly reduced in all comS point mutation strains, again reaching levels comparable to the comS deletion mutant (Figure 7c; Supplementary Figure S5d). This strongly suggests that these mutations completely disrupt the ComS-dependent activation of downstream competence genes.

Overall, these results indicate that the stop mutations in comS severely compromise ComS function in activating the ComK-dependent competence pathway. Even though the mutations do not alter the amino acid sequence of the overlapping srfAB, these mutations mirror a full comS deletion in terms of P _comK and P _comGA activity, highlighting the critical importance of comS integrity for competence development.