A WYL domain transcription factor regulates Lactiplantibacillus plantarum intestinal colonization via perceiving c-di-GMP

Abstract

Cyclic diguanosine monophosphate (c-di-GMP) functions as a crucial bacterial second messenger to control diverse biological functions. Although numerous studies have reported the health effects of Lactiplantibacillus plantarum, the regulatory role of c-di-GMP in L. plantarum remains elusive. Here we show that c-di-GMP functions as an important signal molecule for intestinal colonization of L. plantarum. The intracellular c-di-GMP pool in this probiotic is governed principally by the diguanylate cyclases DgcB, DgcC, and DgcD and the phosphodiesterases PdeA and PdeD. Moreover, we reveal that the WYL domain transcription factor MbpR is a c-di-GMP effector in L. plantarum WCFS1. MbpR reduces the transcription level of mucin-binding proteins (MucBPs) via binding to a special motif within the coding sequences. Perception of c-di-GMP by the WYL domain reversed the inhibitory effect of MbpR on the expression of MucBPs, resulting in increased adherence to intestinal epithelial cells by L. plantarum. Overall, our study provides evidence that a WYL domain transcription factor participates in probiotic colonization by sensing c-di-GMP.

Introduction

Lactiplantibacillus plantarum is a lactic acid bacterium orally administrated to improve host health as numerous studies have reported the probiotic effects of L. plantarum^1,2,3,4. For example, L. plantarum L168 was reported to ameliorate colorectal tumorigenesis by producing indole-3-lactic acid¹. A large double-blind trial supported the use of L. plantarum P9 to manage chronic diarrhea in young adults². Our recent study showed that the candidate probiotic L. plantarum HNU082 can relieve ulcerative colitis by affecting the gut microbiota and alleviate the inflammatory response induced by pathogen invasion^3,4. The surface of the intestinal tract is covered with a thick layer of mucus, which is not only the first line of protection of the intestinal tract but also provides a desirable habitat for bacteria^5,6. To appropriately interact with the host and impart health benefits, it is crucial for probiotics to successfully colonize in the intestinal tract⁷. However, this step is greatly challenged by harsh conditions in the gastrointestinal tract, including the presence of gastric acids in the stomach (pH 1.5–4.5), bile salts in the intestine (0.3%–0.5% w/v), and various digestive enzymes^8,9,10. Additionally, survival probiotics must also compete with the indigenous bacteria inhabiting the mucosa for space and nutrition, especially in the colon, which harbors a dense bacterial population ranging from 10¹¹ to 10¹² CFU/mL¹¹.

The bacterial second messenger bis-(3’-5’)-cyclic guanosine monophosphate (c-di-GMP) is enzymatically synthesized by diguanylate cyclases (DGCs) harboring GGDEF domains and hydrolyzed by specific phosphodiesterases (PDEs) containing HD-GYP or EAL domains¹². This intracellular signal modulates various bacterial biofunctions, including biofilm formation, motility, and virulence¹³. A recent study showed that elevated intracellular c-di-GMP promoted the fitness and persistence of Pseudomonas lurida within its host Caenorhabditis elegans, revealing the important role of c-di-GMP in host-microbe symbiosis¹⁴. However, whether and how L. plantarum uses c-di-GMP to regulate colonization remains elusive.

c-di-GMP mediates bacterial biological processes via interaction with classes of specific receptors¹⁵. Proteins containing a PilZ domain, an inactive GGDEF domain with an inhibitory site (I site), or a degenerated EAL domain are the widely described c-di-GMP receptors^16,17,18. This signal molecule can also be recognized by other types of effectors, such as proteins containing a AAA domain, MshEN domain, or REC domain^19,20,21. In addition, two classes of c-di-GMP-specific riboswitches have been identified recently²². These results suggested that c-di-GMP receptors have substantial structural and sequence diversity. Additionally, c-di-GMP receptors usually orchestrate downstream processes via diverse regulatory mechanisms. In Acinetobacter baumannii, for example, c-di-GMP mediates the translation of proteins containing consecutive prolines via its interaction with elongation factor P²³. In Caulobacter crescentus, c-di-GMP, when presents at elevated levels, inhibits glucose consumption by completing with (p)ppGpp to binding SmbA²⁴. In P. aeruginosa, c-di-GMP promotes adherence and virulence by stimulating pilus assembly through the activation of FimW upon encountering surfaces¹⁶.

Recently, a class of bacterial transcription factors carrying a conserved Trp–Tyr–Leu (WYL) sequence motif has been identified across various microbes^25,26,27. Studies on the structure and function of the WYL domain transcription factors indicate that the WYL domain might serve as a module that controls the activity of transcription factors participating in bacterial immunity and DNA damage response^28,29,30. Here, we report that LP_RS00300, a WYL domain transcription factor, controls the intestinal colonization behavior of L. plantarum WCFS1. We named this protein mucin binding protein regulator (MbpR). Interestingly, MbpR was found to bind a special motif within the coding DNA sequence (CDS), reducing the mRNA of mbpA (LP_RS05345) and mbpB (LP_RS05225), which encode mucin binding proteins (MucBPs) and enhance the adherence of L. plantarum WCFS1 to the intestinal mucosa. Interaction with c-di-GMP triggered the dissociation of MbpR from the CDSs of mbpA and mbpB, consequently increasing their expression. In conclusion, our results reveal a c-di-GMP receptor that enables the intracellular c-di-GMP signal to mediate probiotic colonization.

Results

c-di-GMP Regulates the Intestinal Colonization Behavior of L. plantarum

As an important nucleotide second messenger, c-di-GMP plays a regulatory role in diverse bacterial biological processes³¹. To detect the c-di-GMP in L. plantarum WCFS1, we analyzed the intracellular extract by using liquid chromatography-mass spectrometry (LC-MS). A peak with an m/z value of 691.1023 at a retention time of 1.5 min was observed in positive-ion mode (Fig. 1a), which was consistent with the peak for the standard c-di-GMP. The peak was further verified by triple-quadrupole MS in a multiple reaction monitoring (MRM) model. The three characteristic mass transitions (m/z 691 → 540, 691 → 248, and 691 → 152) and retention time of the substance, confirmed that it was c-di-GMP (Fig. 1b). The intracellular c-di-GMP concentration of wild-type L. plantarum WCFS1 was detected at 32 ± 4 pmol/mg protein (three biological replicates).

Fig. 1: c-di-GMP promoted intestinal colonization of L. plantarum.

a, b High-resolution mass spectra (a) and triple-quadrupole mass spectra in a multiple reaction monitoring model (b) of c-di-GMP extracted from L. plantarum WCFS1. Left panel: c-di-GMP standard; right panel: extracted from L. plantarum WCFS1. The experiments were performed three times with similar results. c–g Overexpression of active c-di-GMP synthetase WspR or the active c-di-GMP phosphodiesterase RocR influenced the acid- (c) and bile salt (d) tolerance, intracellular c-di-GMP level (e), adhesive ability to mucus-secreting HT-29 cells (f), and colonization ability within mice colon (g) of L. plantarum. The genes rocR and wspR were cloned into the pLH01 plasmid and then electroporated into the wild-type strain of L. plantarum WCFS1. WT(pLH01) is the wild-type strain containing pLH01 plasmid as a control vector. Data are presented as mean values ± standard deviations (SD). Error bars indicate the SD. Dots represent individual data. Statistical significance in (c–g) was determined using one-way ANOVA (e–g) with Tukey’s multiple-comparison test. In (c, d, f, g): n = 5 biological replicates. In (e): n = 3 biological replicates. Source data are provided as a Source Data file.

To obtain insight into the regulatory role of c-di-GMP in L. plantarum, we overexpressed the active c-di-GMP synthetase WspR or the active c-di-GMP degradation protein RocR from P. aeruginosa in the wild-type strain of L. plantarum WCFS1. Overexpression of WspR resulted in improved cell viability of L. plantarum WCFS1 in pH 3.0 [degree of freedom (df) = 19, p = 7.81E-5, effect size (η²) = 0.76, 95% Confidence Intervals (CI) = −20.64 to −7.50] or 0.2% bile salt (df = 19, p = 4.89E-7, η² = 0.95, 95% CI = −32.52 to −17.08) environments and increased the intracellular c-di-GMP concentration by 245% (df = 11, p = 1.36E-8, η² = 0.99, 95%CI = −81.55 to −65.15)(Fig. 1c–e). In contrast, overexpression of RocR led to reduced acid- (df = 19, p = 4.51E-3, η² = 0.62, 95% CI = 2.75 to 15.89) and bile salt- (df = 19, p = 6.80E-4, η² = 0.78, 95% CI = 5.78 to 21.23) resistance abilities and decreased the cellular c-di-GMP level by 87% (df = 11, p = 3.64E-5, η² = 0.98, 95% CI = 17.69 to 34.09)(Fig. 1c–e). The gastrointestinal tract is covered by a mucus layer that provides a desirable habitat for bacteria⁵. The mucin-secreting HT-29 cell line, a human colon cancer cell line, was used as an in vitro model to test the adhesive ability of L. plantarum WCFS1³². In trans expression of WspR facilitated the adherence of L. plantarum WCFS1 to HT-29 cells (df = 19, p = 7.68E-7, η² = 0.88, 95% CI = −5.79 to −2.97), whereas overexpression of RocR inhibited its ability to adhere to HT-29 cells (df = 19, p = 1.58E-3, η² = 0.65, 95% CI = 0.85 to 3.67) (Fig. 1f). Moreover, through in vivo colonization experiments, we observed that the number of L. plantarum cells overexpressing WspR or RocR that colonized the colon was 340% (df = 14, p = 7.93E-8, η² = 0.94, 95% CI = −1813 to −1179) and 15% (df = 14, p = 0.012, η² = 0.82, 95% CI = 92.89 to 727.1), respectively, of that of the wild-type strain (Fig. 1g). These results suggest that c-di-GMP promoted intestinal colonization by improving survival and adhesive ability of L. plantarum.

Diguanylate Cyclases and c-di-GMP-Specific Phosphodiesterases in L. plantarum WCFS1

By using the Basic Local Alignment Search Tool (BLAST) algorithm (https://blast.ncbi.nlm.nih.gov/Blast.cgi), we found dgcA (LP_RS00050), dgcB (LP_RS12710), dgcC (LP_RS12345), and dgcD (LP_RS10755) encodes GGDEF domains; pdeA (LP_RS03290), pdeB (LP_RS03320), pdeC (LP_RS11455), pdeD (LP_RS12350), and pdeE (LP_RS03570) encode EAL domains; but did not find any HD-GYP domain-encoding genes in the genome of L. plantarum WCFS1 (Fig. 2a). InterPro analysis of the protein sequences revealed the presence of multiple transmembrane domains in all GGDEF-containing proteins. In addition to a GGDEF domain, DgcA additionally possessed a PAS domain, a DDH domain, and a DHHA1 domain, which might be involved in the degradation of c-di-AMP³³. None of the EAL domains exhibited crosslinking with other domains, thus, these proteins are predicted to be cytoplasmic.

Fig. 2: The major diguanylate cyclases and c-di-GMP phosphodiesterases in L. plantarum WCFS1.

a Domain structure of predicted c-di-GMP-metabolizing proteins in L. plantarum WCFS1. aa: amino acid. b Intracellular c-di-GMP level in L. plantarum WCFS1 with single deletion of predicted c-di-GMP-metabolizing genes. Data are presented as mean values ± standard deviations (SD). Error bars indicate the SD. Dots represent individual data. The results were analyzed using one-way Analysis of Variance (ANOVA) with Tukey’s multiple-comparison test (n = 3 biological replicates). Source data are provided as a Source Data file.

To reveal the enzymatic activity of these c-di-GMP metabolic enzymes, the GGDEF and EAL domains were purified to investigate the in vitro DGC or PDE activity. No substrate was converted by the DgcA GGDEF domain within two hours, while approximately 38% (df = 8, p = 9.94E-6, η² = 0.98, 95% CI = 30.45 to 45.22), 34% (df = 8, p = 2.09E-4, η² = 0.93, 95% CI = 22.29 to 46.09), and 60% (df = 8, p = 2.81E-7, η² = 1.00, 95% CI = 54.30 to 65.97) of GTP was converted into c-di-GMP after incubation with GGDEF domains of DgcB, DgcC, and DgcD, respectively (Supplementary Fig. 1a–d). Both the PdeA (df =8, p = 4.41E-8, η² = 1.00, 95% CI = 92.10 to 107.1) and PdeD (df = 8, p = 1.39E-13, η² = 1.00, 95% CI = 96.01 to 103.60) EAL domains degraded all of the c-di-GMP to pGpG within 30 min, whereas PDE activity was not observed for the PdeB, PdeC, and PdeE EAL domains (Supplementary Fig. 1e–i).

To investigate the effect of c-di-GMP metabolic enzymes on the intracellular c-di-GMP concentration in L. plantarum WCFS1, we constructed single-gene deletion mutants by using clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated protein 9 (Cas9)-based gene editing. Consistent with the in vitro DGC and PDE results, the intracellular c-di-GMP levels in the ∆dgcB, ∆dgcC, and ∆dgcD mutants decreased by 42% (df = 56, p = 5.94E-6, η² = 0.88, 95% CI = 18.81 to 64.86), 28% (df = 56, p = 6.97E-3, η² = 0.72, 95% CI = 4.52 to 50.57), and 43% (df = 56, p = 3.70E-6, η² = 0.91, 95% CI = 19.75 to 65.80), respectively, compared with those in the wild-type strain (Fig. 2b). In contrast, the intracellular c-di-GMP levels in the pdeA and pdeD deletion mutants increased by 74% (df = 56, p = 2.02E-12, η² = 0.95, 95% CI = −96.99 to −50.94) and 46% (df = 56, p = 9.02E-7, η² = 0.88, 95% CI = −68.62 to −22.57), respectively. In-frame deletion of dgcA, pdeB, pdeC, and pdeE did not induce a change in the intracellular c-di-GMP level in L. plantarum WCFS1. These results indicate that DgcB, DgcC, and DgcD are active DGCs, while PdeA and PdeD are active PDEs, in L. plantarum WCFS1.

MbpR is a c-di-GMP effector protein

To identify the underlying mechanism by which L. plantarum WCFS1 senses c-di-GMP to control biological functions, we conducted an affinity pull-down experiment by using biotinylated c-di-GMP to enrich potential c-di-GMP receptor from soluble protein extracts of L. plantarum WCFS1. Three highly enriched protein bands were identified (Fig. 3a), and these bands corresponded to PdeB, PdeC, and MbpR based on mass spectrometry analysis (Supplementary Data 1). PdeB and PdeC are EAL domain-containing proteins with conserved c-di-GMP binding sites (Fig. 2a), while MbpR harbors an HTH domain and a WYL domain (Fig. 3b). As no information on the relationship between the WYL domain and c-di-GMP has been reported, we focused on the role of MbpR in the c-di-GMP signaling system in L. plantarum.

Fig. 3: MbpR is a c-di-GMP receptor in L. plantarum.

a SDS-PAGE analysis of proteins enriched by biotinylated c-di-GMP (+) or control (-) from L. plantarum WCFS1 lysates via affinity pull-down assay. M: marker. b Domain structure analysis of MbpR. HTH: helix-turn-helix (HTH) domain; WYL: WYL domain. c Bacterial two-hybrid assay (BTHA) analysis of the interaction of MbpR and itself. d Native polyacrylamide gel electrophoresis of MbpR. M: Marker. e BTHA analysis of the interaction of MbpR and RpoD. f–h Isothermal titration calorimetry (ITC) analysis of the dissociation constants (K_D) between c-di-GMP and MbpR (f), or polypeptides containing only the WYL domain (MbpR_WYL) (g) or HTH domain (MbpR_HTH) (h) of MbpR. N indicates binding stoichiometry. ND: not detected. i Surface representation of the structural model of MbpR in complex with c-di-GMP. c-di-GMP is shown as red sticks. The figure is prepared using PyMOL. j The interactions between MbpR and c-di-GMP were analyzed by autodocking vina software and protein-ligand interaction profiler. The binding was visualized by Pymol. k Binding of c-di-GMP to MbpR mutants. The binding affinity was measured by microscale thermophoresis (MST). The K_D values are presented as mean ± standard deviations (SD) of 3 biological replicates. l–o Effects of mbpR on the adhesive ability to HT-29 cells (l), colonization ability (m), acid (n), and bile salt tolerance (o) of L. plantarum WCFS1.The data in (a, c–h) were performed three times with similar results. Data in (l–o) are presented as mean values ± SD. Error bars indicate the SD. Dots represent individual data. Statistical significance in l-o was analyzed using one-way Analysis of Variance (ANOVA) with Tukey’s multiple-comparison test (n = 5 biological replicates). Source data are provided as a Source Data file.

The WYL domain transcription factor DriD in Caulobacter crescentus forms a homodimer and PafBC in Mycobacterium smegmatis forms a heterodimer^28,34. We did not identify other WYL domain transcription factors in L. plantarum WCFS1. Thus, to investigate whether MbpR forms a homodimer, we performed a bacterial two-hybrid assay (BTHA). As shown in Fig. 3c, no interaction was observed between the MbpR and itself. Moreover, the native polyacrylamide gel electrophoresis of MbpR showed a molecular weight of 40 kDa with or without c-di-GMP (Fig. 3d), indicating that MbpR did not form a homodimer or oligomer in physiological conditions. PafC mediates protein-protein interactions with the RNA polymerase sigma factor SigA via its HTH domain³⁵. However, we found that MbpR does not bind to the RpoD, the homologue of SigA in L. plantarum WCFS1 (Fig. 3e). Together, those results indicated that MbpR acts as a monomer.

Isothermal titration calorimetry (ITC) revealed that c-di-GMP binds MbpR at a 1:1 stoichiometry with a calculated dissociation constant (K_D) of 0.22 μM (Fig. 3f), while related nucleotides such as c-di-AMP, 3’,5’-cAMP, 3’,5’-cGMP, AMP, and GMP failed to bind, as determined by microscale thermophoresis (MST) (Supplementary Fig. 2a–f). In addition, we also purified two polypeptides containing only the WYL domain or HTH domain of MbpR and found that it is the WYL domain binds to c-di-GMP (Fig. 3g, h). To reveal how MbpR interacts with c-di-GMP, we modeled the structure of MbpR by using Alphafold2. Potential ligand-binding sites were predicted using AutoDock Vina 1.1.2 software. The lowest binding energy between the two molecules was −9.1 kcal/mol, suggesting that c-di-GMP closely contacts with MbpR. Nine amino acid residues, namely, Arg192(R192), Thr195 (T195), Tyr191 (Y191), Phe177 (F177), Tyr271 (Y271), Glu268 (E268), Tyr220 (Y220), Asp267 (D267), and Ile190 (I190), might be vital for the interaction between MbpR and c-di-GMP (Fig. 3i, j). Using protein-ligand interaction profiler (PLIP), we found that a salt bridge formed by the side chains of E268 and a guanine base (G1) of the c-di-GMP, two π-stacking interaction between the Y271 and another guanine base (G2) of the c-di-GMP, and several hydrogen bonds with the R192, Y220, and I190 residues (The model is available in Source Data). Nine single-point mutants were generated and were used to measure their ability to interact with c-di-GMP by using MST. Results showed that mutation of R192 (df = 29, p = 1.03E-9, η² = 0.93, 95% CI = −95.79 to −55.16) or E268 (df = 29, p = 1.21E-4, η² = 0.89, 95% CI = −68.49 to −4.26) resulted in a marked reduction in the c-di-GMP-binding affinity of MbpR (Fig. 3k), suggesting that these two residues directly participated in c-di-GMP binding. Sequences alignment shows that the E268 of DriD, PafB, and PafC has deviated from MbpR, however, both the R192 and E268 are conserved in the WYL domain transcription factors from Lactiplantibacillus, Lactobacillus, Lactococcus, Staphylococcus, and Enterococcus (Supplementary Fig. 3). We thus reasonably suspected those WYL domain transcription factors might also be involved in c-di-GMP recognition.

To test the impacts of mbpR on the phenotypes of L. plantarum, we then generated an mpbR-null mutant by deleting the region encoding MbpR from the L. plantarum WCFS1 genome. In-frame deletion of mbpR had no impact on the intracellular c-di-GMP concentration (Supplementary Fig. 4a) but exhibited a marked increase in the ability to adhere to HT-29 cells and colonize the colon (Fig. 3l, m). Interestingly, the differences in acid- and bile salt-resistance ability between the ΔmbpR mutant and wild-type strains were negligible (Fig. 3n, o). In trans expression of the wild-type version of MbpR rescued the mutant phenotypes of ΔmbpR. At the same protein level, MbpR^R192A which almost lost activity to bind c-di-GMP, even exhibited an increased effect on the complement of phenotypes (Fig.3l, m, and Supplementary Fig. 4b, c). Overexpression of wspR did not impact the adhesive ability of the mbpR^R192A complemented strain to HT-29 cells, indicating that c-di-GMP promotes the adhesive ability via MbpR (Fig. 3l). However, overexpression of wspR enhanced the colonization ability of the mbpR^R192A complemented strain (df = 34, p = 0.040, η² = 0.82, 95% CI = −815.0 to −13.02), suggesting that there is at least one other c-di-GMP effector contributing to the colonization phenotype different from MbpR. To gain insight into the regulatory role of mbpR, the transcriptomes of the wild-type, ΔmbpR, ΔmbpR(mbpR), and ΔmbpR(mbpR^R192A) strain were analyzed and compared by using RNA sequencing (RNA-seq). The expressions of 94 genes were elevated while those of 3 genes were decreased (absolute value of log₂ Fold Change was >1.5, p < 0.05) by deletion of mbpR (Supplementary Fig. 5a and Supplementary Data 2). Most of these differentially expressed genes were complemented by the wild-type MbpR and MbpR^R192A. Clusters of orthologous groups (COG) analysis revealed that plenty of those genes encode proteins with unknown functions or are associated with metabolism (Supplementary Fig. 5b).

MbpR binds to specific DNA sequences

Given that MbpR belongs to a member of DNA binding proteins with a HTH domain, we performed Chromatin immunoprecipitation sequencing (ChIP-Seq) to identify the binding sites for MbpR at a genome-wide scale in the L. plantarum WCFS1 genome using ΔmbpR harboring a plasmid encoding MbpR with a 3 × FLAG-tag at C-terminally. The expression of 3 × FLAG-MbpR is controlled by the P23 promoter, a strong promoter in L. plantarum. Western blots showed that MbpR is highly expressed under the control of P23 promoter (df = 11, p = 5.55E-8, η² = 0.98, 95% CI = −136.1 to −102.3) (Supplementary Fig. 4b, c), and overexpression of MbpR markedly inhibited the ability of L. plantarum to adhere HT-29 cells (df = 34, p < E-15, η² = 0.99, 95% CI = 7.00 to 9.55) and colonize gastrointestinal tract (df = 34, p = 6.16E-11, η² = 0.93, 95% CI = 1070 to 1872) (Fig. 3l, m), suggesting that 3 × FLAG-MbpR is functional. The ChIP-Seq peaks for 3 × FLAG-MbpR showed widespread occupancy across the L. plantarum WCFS1 chromosome (Fig. 4a). The DNA loci and enrichment fold changes for these peaks are shown in Supplementary Data 3. To experimentally verify the ChIP-Seq results, we randomly selected 3 peaks (fold enrichment >2) and a fragment with no peak as a control to perform electrophoretic mobility shift assays (EMSAs). The amount of DNA probes of peak 367, peak 336, and peaks 114 bound by MbpR increased with increasing amount of MbpR, while no DNA-protein complex was formed between the control probe and MbpR (Supplementary Fig. 6a–d), suggesting that MbpR binds the nucleoid sequence-specifically. In addition, we observed that it was the HTH domain rather than the WYL domain bound to the target DNA fragments (Supplementary Fig. 6e, f).

Fig. 4: MbpR binds DNA sequence specifically.

a 3 × FLAG-MbpR bound regions on the L. plantarum WCFS1 chromosome. b Potential MbpR-binding motif identified by MEME-ChIP (https://meme-suite.org/meme/doc/meme.html) based on the ChIP-Seq peaks. c Electrophoretic mobility shift assays (EMSA) analysis of MbpR binding to DNA probes of wild-type and mutated peak 114. d Chromatin immunoprecipitation sequencing (ChIP-Seq) signal density at the transcription start site. TSS: transcription start site. e The number of MbpR-binding sequences located at coding DNA Sequence (CDS), 3’ ends of two genes, and promoter region. The data in (a, c) was performed three times with similar results. f Real-time quantitative reverse transcriptase-polymerase chain reaction (RT-qPCR) analysis of MbpR bound gene expression in WT and ΔmbpR. Transcription levels are shown as mean ± standard deviations (SD) of 3 biological replicates. Error bars indicate the SD. Dots represent individual data. Statistical significance was analyzed using one-way Analysis of Variance (ANOVA) with Tukey’s multiple-comparison test. Source data are provided as a Source Data file.

Multiple expectation maximization for motif elicitation (MEME) analysis of the ChIP data revealed a potential MbpR-binding motif, which was identified as 5’-YAATGGTGCCA-3’ (Fig. 4b). To test whether MbpR binds the sequence specifically, we performed EMSAs with mutated DNA fragments from peak 336 and peak 114 in which the potential MbpR-binding sites 5’-CAATGGTGACA-3’ and 5’-GAAAGGTGTCA-3’ were deleted. As expected, no DNA-protein complex was formed between MbpR and the mutated DNA probes (Supplementary Fig. 6g). Several additional peaks were also chosen for the EMSAs. The amount of DNA probe from peak 194 and peak 142 bound by MbpR increased with increasing amounts of MbpR, while deletion of the MbpR binding sequence 5’-TAATGGTGCCA-3’ or 5’- CAATGATGCCA-3’ from peak 194 and peak 142 resulted in MbpR being unable to bind the DNA probes (Supplementary Fig. 7a–d). We also added the MbpR-binding sequence of peak 144 to the DNA probe of no peak, and it was found that the addition of the MbpR-binding sequence is sufficient for binding by MbpR (Supplementary Fig. 7e). To investigate the most critical base in the MbpR-binding motif, we mutated individual nucleotide of peak 144 and found that mutation of the adenine in the second position, thymine in the seventh position, the guanosine in eighth position, or the cytosine in the tenth position led to an obvious decrease in the complex formed by MbpR and DNA probe (Fig. 4c). Using Find Individual Motif Occurrences (FIMO), a software tool for scanning DNA sequences with motifs described as position-specific scoring matrices, we identified the MbpR-binding sequences in the majority of MEME-unpredicted peaks (Supplementary Data 3). Some peaks did not harbor the MbpR-binding sequence which might be due to that MbpR binds with some sequence non-specifically at high concentrations.

Analysis of the ChIP-Seq data revealed that the FLAG-tagged MbpR signal was widely distributed at transcription start sites (TSSs) with a sharp single peak after TSSs (Fig. 4d). And 94.0% (255) and 2.2% (4) of the MbpR binding boxes were located at the CDS, or the 3’ ends of the two adjacent genes; only 3.8% (14) were enriched in the promoter region (Fig. 4e). RT-qPCR analysis was performed to determine whether MbpR represses the transcription of the bound genes. Among the ten genes tested, which all had an MbpR-binding sequence in the CDS, LP_RS04695 (df = 8, p = 5.66E-5, η² = 0.96, 95% CI = -2.44 to -1.43), LP_RS04100 (df = 8, p = 1.88E-3, η² = 0.90, 95% CI = −3.20 to −1.09), and LP_RS06770 (df = 8, p = 6.21E-3, η² = 0.82, 95% CI = −2.98 to −0.70) showed a significant increase in intracellular mRNA levels upon MbpR depletion, while the other seven genes were unaffected by mbpR depletion (Fig. 4f). RNA-Seq analysis revealed that only 22 out of 273 genes identified by ChIP-Seq were remarkedly changed by mbpR deletion (Supplementary Data 3). All these differentially expressed genes were upregulated and contain an MbpR-binding sequence within their CDS region, except LP_RS03970 which harbors an MbpR-binding sequence within its promoter. This suggests that MbpR might not function as a traditional transcriptional regulator, which is consistent with the finding that the majority of ChIP-Seq peaks are found within CDSs.

To exclude the impact of c-di-GMP and further enhance MbpR binding across the genome, we performed ChIP-Seq again with MbpR^R192A which cannot bind c-di-GMP. 528 peaks were identified by the c-di-GMP-blind version (Supplementary Data 4). The number of peaks identified by MbpR was substantially fewer than that of MbpR^R192A, and the majority of them coincide with the peaks found by MpbR^R192A (Supplementary Fig. 7f), suggesting that c-di-GMP might prevent MbpR binding genome.

MbpR binding reduced the transcription level of MucBPs

Since MbpR mediated the HT-29 adherence and intestinal colonization behavior of L. plantarum WCFS1, we searched all the peaks for genes that might be involved in such functions and found that LP_RS12930 (peak 316), mbpA (peak 181), and mbpB (peak 186) encode MucBPs (Fig. 5a). LP_RS12930 is a BspA-like surface protein while its expression was not controlled by mbpR (Fig. 5b and Supplementary Data 2). However, the intracellular mRNA levels of mbpA (df = 8, p = 3.69E-4, η² = 0.93, 95% CI = −3.83 to −1.79) and mbpB (df = 8, p = 1.27E-4, η² = 0.96, 95% CI = −4.53 to −2.43) were significantly increased by mbpR depletion. Thus, we generated the ΔmbpA, ΔmbpB, and ΔmbpAB mutant by in-frame deletion of mbpA, mbpB, and both mbpA and mbpB, respectively. The acid- and bile salt-resistant capability was unaffected by the deletion of mbpA or mbpB (Supplementary Fig. 8a, b), while the number of ΔmbpA, ΔmbpB, and ΔmbpAB mutant cells that adhered to HT-29 cells were 48% (df = 19, p = 3.16E-8, η² = 0.92, 95% CI = 1.03 to 1.74), 29% (df = 19, p = 7.19E-6, η² = 0.80, 95% CI = 0.57 to 1.28), and 15% (df = 19, p = 2.96E-9, η² = 0.95, 95% CI = 1.27 to 1.98) of that of the wild-type strain, respectively (Fig. 5c). When the mucus layer on HT-29 cells was removed by pretreatment with N-acetylcysteine, the number of ΔmbpA, ΔmbpB, and ΔmbpAB cells that adhered to pretreated HT-29 cells was similar to that of the wild-type strain. To further confirm the mucin-binding abilities of MbpA and MbpB, we purified their mucin-binding domain and performed MST assays with mucin. Results showed that the mucin binding domain of MbpA and MbpB binds mucin with K_D values of 0.53 ± 0.17 μM and 0.69 ± 0.21 μM, respectively (Fig. 5d, e).

Fig. 5: MbpR mediates expression of mucus-binding protein via impeding transcription.

a Domain structure analysis of LP_RS12930, MbpA, and MbpB. MucBP: mucin-binding-protein domain; LRR: Leucine-rich repeat; TM: Transmembrane domain. b Relative transcription of mbpA and mbpB in L. plantarum wild-type and ΔmbpR mutant measured by real-time quantitative reverse transcriptase-polymerase chain reaction (RT-qPCR). c Effects of mbpA and mbpB on the ability of L. plantarum WCFS1 to adhere HT-29 cells. d, e MST was used to analyze the binding affinity between mucin and the mucin-binding domains of MbpA (d) and MbpB (e). Data are presented as mean values ± standard deviations (SD) of three replicates. Error bars indicate the SD. f Location of MbpR binding motif on the mbpA and mbpB coding DNA sequence (CDS) regions. g, h EMSA analysis of MbpR binding to mbpA (g) and mbpB (h) CDS regions. i, j EGFP expression from the mbpA (i) and mbpB (j) promoter. k, l Effect of MbpR-binding motif in the CDS region on the expression EGFP under the control of mbpA (k) and mbpB (l) promoter. The strains used in (i–l) contain reporting plasmid. m In vitro transcription of wild-type mbpA CDS and mutated mbpA CDS in which the MbpR-binding sequence is deleted. The mRNA yield was measured after 2 h of reaction. Data in (b, c, i–m) are presented as mean values ± SD. Error bars indicate the SD. Dots represent individual data. Statistical significance in (b, c, m, i–l) was analyzed using one-way ANOVA Tukey’s multiple-comparison test. In (b, m): n = 3 biological replicates. In (c, i–l): n = 5 biological replicates. The data in (g, h) were performed three times with similar results. Source data are provided as a Source Data file.

We then tested whether MbpR binds to the promoters of mbpB and mbpA. However, no DNA‒protein complex formation was observed between MbpR and the promoters of mbpB and mbpA (Supplementary Fig. 8c, d). The MbpR binding sequences were located at the CDSs of mbpA and mbpB (Fig. 5f) and EMSAs revealed that MbpR bound the mbpB and mbpA CDS DNA probes (Fig. 5g, h). Deletion of MbpR binding sequences abolished the binding of MbpR to DNA probes (Supplementary Fig. 8e, f). Thus, we hypothesize that MbpR negatively regulates the transcription of mbpB and mbpA by binding the CDS. To confirm this, a reporter plasmid containing the mbpB or mbpA promoter and egfp was transformed into the wild-type strain and the ΔmbpR mutant. The fluorescence intensity of EGFP from the mbpA or mbpB promoter was unaffected by mbpR depletion (Fig. 5i, j). Insertion of MbpR-binding sequences of mbpA (df = 19, p = 2.91E-10, η² = 0.95, 95% CI = −4.81 to −3.30) or mbpB (df = 19, p = 2.87E-8, η² = 0.92, 95% CI = −2.56 to −1.52) before the egfp resulted in a significantly weaker EGFP fluorescence intensity in the wild-type strain compared to the ΔmbpR mutant (Fig. 5k, l). The EGFP level in ΔmbpR was complemented by the wild-type of MbpR, while MbpR^R192A which almost lost its activity to bind c-di-GMP exhibited an enhanced effect on the complement of EGFP. We first suspected that MbpR might act on transcription elongation. Thus, we added a MbpR-binding sequence of mbpA between the open reading frames of EGFP and mCherry, and the fusion protein was expressed under the control of the P23 promoter. If our supposition is true, the mCherry fluorescence intensity should be enhanced by deleting mbpR, while the in-frame deletion of mbpR should have an ignorable change in the EGFP fluorescence intensity. However, we observed that mbpR deletion led to a similar improvement in the intensity of EGFP and mCherry (Supplementary Fig. 8g), indicating that MbpR binding reduces the transcription of the bound gene but did not cause transcription termination. As expected, the deletion of mbpR did not cause any changes in the intensity of EGFP and mCherry which were expressed from the egfp-mCherry fusion gene without MbpR-binding sequence (Supplementary Fig. 8h). To further verify that MbpR acts as a transcription inhibitor and exclude the impact of the native promoter, we performed an in vitro transcription assay using wild-type mbpA CDS regions or mutated mbpA CDS without MbpR-binding sequence as templates which are expressed under the control of the T7 promoter. MbpR reduced the mRNA synthesis of wild-type mbpA CDS (df = 14, p = 1.78E-4, η² = 0.95, 95% CI = 38.69 to 101.00) but did not impact that of mutated mbpA CDS. As expected, c-di-GMP eliminated the inhibitory effect of MbpR (df = 35.65, p = 1.04E-3, η² = 0.90, 95% CI = −87.38 to −25.11), while it did not have the same impact on MbpR^R192A (Fig. 5m).

c-di-GMP mediated intestinal colonization of L. plantarum via Inhibiting MbpR Binding

As noted, elevated c-di-GMP levels promoted intestinal colonization by L. plantarum, and both mbpB and mbpA were involved in this effect. We then examined the impact of c-di-GMP on the MbpR binding to the mbpA and mbpB CDS DNA fragments by EMSA. The amounts of DNA probes bound by MbpR decreased with increasing amounts of c-di-GMP (Fig. 6a, b). Moreover, mutation of the c-di-GMP-binding site (R192A) eliminated the effect of c-di-GMP on the ability of MbpR to bind the mbpA and mbpB CDS (Fig. 6c, d). Similarly, c-di-GMP abolished the binding of MbpR to the other DNA fragments tested in the EMSAs (Supplementary Fig. 9a–d). To further examine the impact of c-di-GMP binding on the transcriptional levels of mbpB and mbpA, we generated a DGC-null mutant of L. plantarum WCFS1 by deleting the active dgcB, dgcC, and dgcD. The intracellular c-di-GMP concentration in the Δ3DGC mutant was below the detectable threshold (Fig. 6e). An in-frame deletion mutant strain was also constructed by deleting pdeA and pdeD, and the generated Δ2PDE mutant showed a 2.51-fold increase (df = 8, p = 2.33E-4, η² = 0.93, 95% CI = −201.8 to −100.6) in the intracellular c-di-GMP concentration. RT-qPCR analysis revealed that the transcription of mbpA and mbpB in the Δ3DGC mutant were downregulated, while that in the Δ2PDE strain were significantly elevated (Fig. 6f). EGFP fluorescence intensity from the mbpB and mbpA promoters in the wild-type, Δ2PDE, and Δ3DGC mutant strains were not significantly different (Fig. 6g, h). Thus, these results suggested that c-di-GMP promotes the expression of mbpB and mbpA by inhibiting MbpR binding.

Fig. 6: c-di-GMP reversed the inhibitory effects of MbpR.

a–d EMSA detection of MbpR (a, b) and MbpR^R192A (c, d) binding to the mbpB or mbpA coding DNA sequence (CDS) probe in the presence of different amounts of c-di-GMP. The probe was incubated with protein at room temperature for 30 min. The experiments were performed three times with similar results. e Intracellular c-di-GMP level of the wild type, Δ3DGC, and Δ2PDE mutants. f Relative mbpA and mbpB transcription level in Δ3DGC and Δ2PDE strains analyzed by RT-qPCR. g, h Fluorescence intensity of EGFP, which was expressed from the reporting plasmid under the control of mbpA (g) and mbpB (h) promoter, in the wild-type, Δ3DGC, and Δ2PDE strains. The Δ3DGC strain was constructed by deleting dgcD, dgcC, and dgcB. The Δ2PDE strain mutant was generated by deleting pdeD and pdeA. i A pathway for c-di-GMP mediating L. plantarum colonization (Created in BioRender. Zhang, J. (2025) https://BioRender.com/g28k816). At low c-di-GMP level, MbpR binds to the mucin-binding protein (MucBP) CDS and inhibits its transcription. After sensing unknown signals, the activated diguanylate cyclases or inhibited c-di-GMP phosphodiesterases maintain the intracellular c-di-GMP level at a high level. The elevated c-di-GMP level results in MbpR releasing from the CDS of MucBP and promotes MucBP expression, promoting the colonization of L. plantarum. RNAP: RNA polymerase. Data in (e–h) are presented as mean values ± standard deviations (SD). Error bars indicate the SD. Dots represent individual data. Statistical significance in (e, g, h) was analyzed using one-way Analysis of Variance (ANOVA) with Tukey’s multiple-comparison test. In (e, f): n = 3 biological repeats, In (g, h): n = 5 biological repeats. Source data are provided as a Source Data file.

Discussion

Based on the homology of the entire domain and the conservation of the extended signature motif, GGDEF domains can be categorized into three major classes: enzymatically functional domains (class I); enzymatically functional domains in combination with an EAL domain (class II); and enzymatically nonfunctional domains (class III)³⁶. DgcB, DgcC, and DgcD could be differentiated into class I GGDEF domain, while DgcA could be differentiated into class III GGDEF domain since the GGDEF motif of DgcA is degenerated to GSRTF (Fig. 2a). As expected, the DGC activity of DgcB, DgcC, and DgcD were observed, while DgcA was not (Supplementary Fig. 1). Some degenerated GGDEF domains act as c-di-GMP receptors¹³, we thus expressed the GGDEF of DgcA but did not observe its ability to bind c-di-GMP (Supplementary Fig. 10). Based on conservation of catalytic residues and loop 6, the EAL domains can be differentiated into three classes³⁶. Class 1 EAL domain with conserved catalytic residues and loop6 are enzymatically active; class 2 EAL domain with conserved catalytic residues and degenerated loop6 are potentially active, whereas class 3 EAL domain with degenerated catalytic residues and loop6 are catalytically inactive. Due to the degenerated loop 6, PdeA, PdeB, PdeD, and PdeE belong to the class 2 EAL domain. PdeA and PdeD exhibited catalytic ability, while PdeB and PdeE did not (Supplementary Fig. S1). We speculate that PdeB and PdeE were enzymatically inactive or their capabilities are too weak to detect. The catalytic base of the PdeC EAL domain (the first glutamate residue in the EGVE motif) is mutated to methionine. Therefore, PdeC belongs to the class 3 EAL domain and is catalytically inactive.

Conserved features facilitate the identification of c-di-GMP synthetase or hydrolase. However, since diverse classes of c-di-GMP effectors do not share sequence or structural similarity, only a few types of c-di-GMP receptors could be discerned solely through conserved motifs, such as the ExLxR motif in the EAL domain, the RxxD motif in the GGDEF domain, and the RxxxR and (D/N)x(S/A)xxG motifs in the PilZ domain. This is particularly true for L. plantarum, in which no c-di-GMP effectors have been reported yet. In this study, we identified MbpR, a WYL domain transcription factor, as a c-di-GMP receptor protein via multiple approaches (Fig. 3). WYL domains have an Sm-like SH3 β-barrel fold, which is commonly encountered in RNA-binding proteins and is predicted to bind ligands²⁶. Previous research showed the ability of the WYL domain to interact with ssDNA^27,34, and our results revealed that the WYL domain of MbpR binds to c-di-GMP with high affinity (Fig. 3f, g). These results indicated that the WYL domain might respond to different nucleic acids.

WYL domain transcription factors are frequently mentioned in connection with bacterial immunity due to their colocalization with antiphage systems, including CRISPR–Cas adaptive immunity systems or the BREX system^29,37,38. For example, DriD in Caulobacter crescentus was reported to regulate DNA damage responses independent of the canonical SOS pathway²⁵. CapW in E. coli strongly represses cyclic oligonucleotide-based anti-phage signaling system gene expression in uninfected cells²⁹. BrxR in E. fergusonii was reported as a regulator of pEFER phage defence island³⁰. These WYL regulators form a homodimer and mediate gene transcription by binding to specific sequences in promoter regions. Our results revealed that MbpR acts as a monomer and binds to the CDS rather than the promoter regions of mbpB and mbpA (Figs. 3, 5), thus inhibiting transcription of bound genes (Fig. 5i–l). An increasing number of transcription factors have been found to bind the CDS region in prokaryotes. For example, ChIP-Seq data revealed that the transcription factors from bacteria, including P. aeruginosa, P. syringae, and E. coli K-12, have 92.6% (19,707) of their binding peaks located in CDS³⁹. However, the functions of CDS binding by transcription factors are little illustrated. A recent study revealed that transcription factor AlgR and VqsM interact with the CDS region and are involved in producing antisense transcripts and regulating the expression of bound genes³⁹. Our research indicated that MbpR reduced the mRNA production but did not cause transcription termination of bound genes (Supplementary Fig. 8g, h). Moreover, the inhibitory effect of MbpR binding on mRNA production was observed not only in L. plantarum in vivo but also in the in vitro transcription assay (Fig. 5i–m). We thus speculated that MbpR might lower the transcription rate of bound genes given that the transcription elongation rate varies between regions within a gene^40,41. To verify our hypothesis, we detected the mRNA synthesis rate of the wild-type mbpA CDS and mutated mbpA CDS without the MbpR-binding sequence under the control of the T7 promoter. It was observed that MbpR indeed weakened the transcription rate of wild-type mbpA CDS (Supplementary Fig. 11). Moreover, c-di-GMP eliminated the inhibitory effect of MbpR on the transcription rate of the wild-type mbpA CDS. Those results revealed that MbpR binding inhibited the transcription rate of the bound gene and c-di-GMP reversed the inhibitory effect of MbpR. Though our study is limited by the unresolved mechanism by which MbpR inhibits transcription rate, our study provided a glance into the function of CDS binding by transcription factors.

To establish attachment on the biotic or abiotic surfaces, many bacteria produce adhesins to allow themselves glue to the solid substrate. A critical signal controlling the processes is the second messenger c-di-GMP, which mediates the production of various adhesins in different bacteria. The most reported adhesin includes exopolysaccharides and fimbria^42,43. In addition, some extracellular proteins such as LapA and CdrA were also secreted to form biofilm. LapA is required for the initial adhesion of P. putida to a surface and expression of lapA is under the control of intracellular c-di-GMP via the GacS/GacA two-component regulatory system⁴⁴. The transcription of CdrA, another extracellular protein known to mediate cell-cell aggregation and biofilm maturation, is also in a c-di-GMP-dependent manner⁴⁵. Our results revealed that c-di-GMP also mediated the expression of MucBP to adhere the mucin on intestinal epithelial cells. Those data suggested that c-di-GMP controls bacterial attachment via diverse pathways. It’s worth noting that high c-di-GMP levels on the adhesive phenotype fully go through MbpR as mbpR^R192A is fully epistatic to wspR overexpression (Fig. 3l). In contrast, the high c-di-GMP affects the colonization phenotype only partially through MbpR, as wspR overexpression increases colonization in a mbpR^R192A background (Fig. 3m). This means that most likely there is at least one other c-di-GMP effector contributing to the colonization phenotype different from MbpR. Together, this also implies that increased adherence to mucus is not the only factor contributing to better colonization in the high c-di-GMP regime.

We analyzed the genomes of 277 L. plantarum strains and found though 36% of strains lack the homologous gene of mbpA, 99% of the strains concurrently harbor the homologous genes of mbpR, dgcC, dgcD, dgcB, and mbpB (Supplementary Data 5). Based on these findings, we propose a regulatory pathway in which c-di-GMP controls intestinal colonization of L. plantarum (Fig. 6i). MbpR binds to the MucBPs CDS and reduces their transcriptional at low level of c-di-GMP. After sensing unknown signals, the intracellular c-di-GMP level elevated and that was perceived by the WYL domain of MbpR. This interaction causes MbpR to detach from the CDSs of MucBPs, thereby promoting the expression of MucBPs and facilitating the colonization of L. plantarum. While sugar, bile salts, and bicarbonate were shown to affect intracellular c-di-GMP level^46,47,48, certain details in this model remain to be elucidated, particularly the exact signals detected by GGDEF/EAL domain-containing proteins in L. plantarum. However, the model offers a foundation for understanding the intestinal colonization of L. plantarum. In conclusion, our results may inspire further studies to reveal the regulatory roles of c-di-GMP in the probiotic-microbe interaction.

Methods

Bacterial growth conditions

The bacterial strains and plasmids used in this work are listed in Supplementary Data 6. All primers were synthesized by GENEWIZ Inc. and listed in Supplementary Data 7. L. plantarum WCFS and its derivatives were cultured at 37 °C in de Man, Rogosa, Sharpe (MRS) medium. When required, the following antibiotics were added: chloramphenicol, 10 μg/mL; erythromycin, 10 μg/mL. E. coli strains were grown in lysogeny broth (LB) culture medium. For selection, 50 μg/mL of kanamycin or 20 μg/mL of chloramphenicol were supplemented when necessary.

Preparation of electrocompetent cells and transformation

Briefly, 2 mL of the overnight culture of L. plantarum was inoculated into 100 mL MRS supplemented with 0.2 M sucrose and 3 % glycine at 37 °C. The cells were grown to an optical density of 600 nm (OD₆₀₀) = 0.4–0.6. Cells were sequentially washed twice with 1 mM MgCl₂ and once with electroporation buffer (952 mM sucrose, 3.5 mM MgCl₂)⁴⁹, and resuspended in 1 mL of electroporation buffer. 200 ng of pLH01 plasmid and 60 μL of electrocompetent cells were mixed and transferred into a 2 mm cuvette (Bio-Rad, USA). The electroporation was performed with GenePulser Xcell (Bio-Rad, USA) with the following parameters: 2.1 kV, 200 Ω, 25 μF. One milliliter of recovery buffer (MRS with 0.5 M sucrose and 0.1 M MgCl₂) was added into the cuvette and transferred into 2 mL of centrifuge tube. After incubating at 37 °C for 3 h, the cells were spread on MRS agar plates containing 10 µg/mL chloramphenicol. The clone containing pLH01 plasmid was confirmed by PCR using the primer listed in Supplementary Data 7.

Genome editing and complementation

To generate genome-editing plasmids, the guide RNA framework and two 1000 bp fragments flanking target gene were amplified from pHSP02 plasmid⁴⁹ and L. plantarum WCFS1, respectively, by PCR using primers listed in Supplementary Data 7. The abovementioned gRNA and homologous arms were cloned into pHSP02 digested with ApaI and XbaI by using In-Fusion cloning. Then, L. plantarum WCFS1 containing pLH01 plasmid was cultured to an OD₆₀₀ of 0.3. The inducing peptide (amino acid sequence: MAGNSSNFIHKIKQIFTHR, synthesized by GenScript Biotech Corp., China) was added (final concentration of 100 ng/mL) to induce the expression of RecE/T until OD₆₀₀ reached 0.6. Then, the cells were generated into electrocompetent cells and transformed with 1.0 µg pHSP plasmids (listed in Supplementary Data 6) followed by 2-3 h recovery at 30 °C. The recovery mixture was plated on the MRS agar plate supplemented with 10 µg/mL erythromycin and 10 µg/mL chloramphenicol and incubated at 30°C for colony formation. The mutants were screened by PCR using primers listed in Supplementary Data 7. The mutant was sub-cultured in MRS without antibiotics for two generations to cure the editing plasmids. Then, a fraction of the culture was streaked onto an MRS agar plate and incubated at 37 °C for 48 h. PCR was performed to ensure the successful curing of editing plasmids. For complementation analysis, the coding region of target gene with its native promoter was amplified by PCR using the primers listed in Supplementary Data 7 and cloned into the pLH01 digested with EcoRI. The generated plasmid was electroporated into the mutant as described above.

Chromatin immunoprecipitation sequencing (ChIP-Seq)

To construct the vectors for ChIP-Seq, the pP23-FLAG plasmid was generated by inserting a DNA fragment, which is synthesized by GENEWIZ Inc., containing the P23 promoter, a multiple cloning site, and a 3 × FLAG tag was cloned into the EcoRI site of pLH01 plasmid by In-Fusion cloning. The mbpR and mbpR^R192A were amplified from L. plantarum WCFS1 genomic DNA. The purified PCR products were cloned into the pP23-FLAG vector digested with EcoRI and NotI to generate pP23-mbpR-FLAG and pP23-mbpR^R192A-FLAG. The constructs were transformed into the ΔmbpR mutant by electroporation. ChIP assays were performed by Wuhan IGENEBOOK Biotechnology Co., Ltd⁵⁰. Briefly, ΔmbpR containing pP23-mbpR-FLAG or pP23-mbpR^R192A-FLAG was grown in MRS to the logarithmic phase. Formaldehyde was added to the medium at a final concentration of 1% (vol/vol), and incubation for 30 min. Glycine (at a final concentration of 125 mM) was then added to stop the cross-linking. The samples were left at room temperature for 10 min and washed twice with phosphate-buffered saline buffer (pH 7.4). The pellets were resuspended and incubated at 37 °C for 25 min in 0.5 mL of lysis buffer (100 mM Tris-Cl, pH 8.0, 300 mM NaCl) which contains 15 mg/mL of lysozyme and 1 × protease inhibitor (Thermo Fisher Scientific, USA). Subsequently, 0.5 mL immunoprecipitation (IP) buffer (50 mM Tris-Cl, pH 8.0, 250 mM NaCl, 1% Triton X-100, 1 × protease inhibitor) was added. The samples were chilled on ice and sonicated to shear chromosomal DNA. The sheared chromatin was immunoprecipitated by 10 μg of Anti-DDDDK tag (Binds to FLAG® tag sequence) antibody [EPR20018-251] (Abcam, ab205606) at 4 °C for 12 h. Then, 30 μL of protein beads were added for further incubation for 3 h. The beads were next washed once with 20 mM Tris/HCL (pH 8.1), 50 mM NaCl, 2 mM EDTA, 1% Triton X-100, 0.1% SDS; twice with 10 mM Tris/HCL (pH 8.1), 250 mM LiCl, 1 mM EDTA, 1% NP-40, 1% deoxycholic acid; and twice with 1 × TE buffer (10 mM Tris-Cl at pH 7.5. 1 mM EDTA). The bound material was then eluted from the beads in 300 μL of elution buffer (100 mM NaHCO₃, 1% SDS), treated with RNase A (final concentration 8 μg/mL) for 6 h at 65 °C and then treated with proteinase K (final concentration 345 μg/mL) overnight at 45 °C. Immunoprecipitated DNA was used to construct sequencing libraries following the protocol provided by the NEXTFLEX ChIP-Seq Library Prep Kit for Illumina Sequencing (NOVA-5143-02, Bioo Scientific) and sequenced on an Illumina NovaSeq 6000 with the PE 150 method. Trimmomatic (version 0.38) was used to filter out low-quality reads⁵¹. The clean reads were mapped to the L. plantarum WCFS1 genome by Bwa (version 0.7.15)⁵². SAMtools (version 1.3.1) was used to remove potential PCR duplicates⁵³. MACS2 software (version 2.1.1.20160309) was used to call peaks by default parameters (bandwidth, 300 bp; model fold, 5, 50; q value, 0.05). If the summit of a peak was located closest to the transcription start site (TSS) of one gene, the peak was assigned to that gene⁵⁴. MEME (version 3) was used to predict motif occurrence within peaks with default settings for a maximum motif length of 12 base pairs⁵⁵.

Overexpression of RocR and WspR

To overexpress rocR and wspR in L. plantarum WCFS1, the open reading frames (ORF) of the two genes were amplified from P. aeruginosa PAO1 genomic DNA and cloned into the EcoRI and the NotI site of pP23-FLAG vector by In-Fusion cloning. The generated pP23-rocR-FLAG and pP23-wspR-FLAG were transformed into the wild-type strain of L. plantarum WCFS1 by electroporation.

Electrophoretic mobility shift assay (EMSA)

Briefly, the DNA fragments or promoters were amplified by using primers listed in Supplementary Data 7 and purified by EZNA Cycle-Pure Kit (Omega Bio-tek, USA). PCR products were labeled by biotin using Biotin 3’ End DNA Labeling Kit (Thermo Fisher Scientific, USA). The binding reactions were performed in EMSA buffer (100 mM Tris pH 8.0, 10% glycerol, 300 mM NaCl, 50 mM MgCl₂) in the presence of indicated concentrations of protein and 50 ng of DNA probes. When required, the indicated amount of c-di-GMP was added. The mixture was incubated for 30 min at 37 °C and then sampled to a 5% polyacrylamide gel. After electrophoresis in 0.5 × TBE buffer at 4 °C, the DNA probe was transferred to the nylon membrane. The protein-DNA complex was crosslinked by UV. The probes were detected using the biotin luminescence detection kit (Thermo Fisher Scientific, USA) and imaged by ChemiDoc Imaging Systems (Bio-Rad, USA).

Extraction and quantification of intracellular c-di-GMP

Bacteria cells were grown to mid-log phase (OD₆₀₀ = 1.0) and harvested by centrifugation. The pellet was washed with cooled water followed by treatment with 0.6 M of perchloric acid (Sigma)⁵⁶. The pH of the soluble fraction was adjusted with KHCO₃. After centrifugation at 20,000 g for 5 min, the resulting supernatant was filtered, and c-di-GMP levels were measured using LC-MS/MS. The samples were separated on a UPLC Waters HSS T3 column (100 mm × 2.1 mm, 1.8 μm) using a binary pump system with solvent A (water containing 0.1% (v/v) formic acid) and eluent B (methanol containing 0.1% (v/v) formic acid). The gradient started at 12% eluent B and was held for 10 min at a flow rate of 0.3 mL/min. The analyte detection was carried out on a hybrid LCMS-IT-TOF liquid chromatograph mass spectrometer from Shimadzu (Kyoto, Japan) or an AB SCIEX Triple Quad™ 6500 system in positive ionization mode. The mass spectrometry parameters of LCMS-IT-TOF were set as follows: spray voltage, 4.50 kV; detector voltage, 1.65 kV; ion accumulation time, 20 ms; N₂ flow, 1.5 L/min; drying gas, 100.0 kPa; heat block temperature, 200 °C; CDL temperature, 200 °C; TOF pressure, 1.8 × 10⁻⁴ Pa; IT pressure, 1.8 × 10⁻² Pa; RP vacuum, 70.0 Pa; mass range was set from 350 to 1000 Da. The mass spectrometry parameters of Triple Quad 6500+ were set as follows: ion source gas 1, 50 psi; ion source gas 2, 50 psi; curtain gas, 25 psi; collision gas, high; positive ion spray voltage, 5500 V; source temperature, 500 °C. The following three mass transitions were detected in multiple reaction monitoring (MRM) model: 691 → 152 (qualifier), 691 → 540, and 691 → 248.

Protein purification

Briefly, target genes were amplified by using primers listed in Supplementary Data 7 and L. plantarum WCFS genomic DNA as the template. The PCR products were purified and cloned into the pET-28a vector. The fusion gene constructs were then transformed into E. coli strain BL21(DE3) for expression. After incubating the cultures at 37 °C to an OD₆₀₀ of 0.5, isopropyl-β-d-thiogalactoside (IPTG) at a final concentration of 1 mM was added. The cells were grown overnight at 16 °C and collected by centrifugation. The pellet was resuspended in Lysis buffer (100 mM Tris-HCl, pH 8.0, 300 mM NaCl) and was lysed by sonication. The lysate was then centrifuged at 4 °C, and the 6 × His fusion proteins were purified from the supernatant by using Ni-NTA resin (Thermo Fisher Scientific, USA). Finally, the His-tag was removed using thrombin enzyme.

Affinity pull-down assay

Screening of c-di-GMP effector protein in L. plantarum WCFS1 was performed using affinity pull-down method⁵⁷. Briefly, L. plantarum WCFS1 was inoculated in 100 mL of MRS supplemented with 0.2 M sucrose and 3 % glycine. The cells were grown at 37 °C until OD₆₀₀ = 0.6, and then were harvested by centrifugation. The pellet was re-suspended with 1 mL binding buffer [100 mM Tris-HCl buffer (pH 7.5), 300 mM NaCl, 10 mM MgCl₂, 0.5 mM DTT, 0.5% (vol/vol) Triton X-100, 1 × protease inhibitor, 1% (vol/vol) n-dodecyl-β-D-Maltopyranoside], which was subjected to disruption by ultrasonication. The supernatant was collected after centrifugation. Then, biotinylated c-di-GMP (B098, BioLog, USA) at a final concentration of 10 μM was added to 100 μg of soluble protein. The mixture was incubated at 4 °C overnight. Streptavidin dynabeads (Thermo Fisher Scientific, USA) were added to capture the biotin-c-di-GMP for 30 min and were repeatedly washed thrice with binding buffer. After each washing, the supernatant was removed. Then, the streptavidin dynabeads were resuspended in protein loading buffer and boiled for 5 min. The supernatant was used for SDS-PAGE analysis. The experiments were performed three times and all PAGE gels were collected and cut into 1–2 mm gel particles. After washing with ethanol-acetic acid-water (4:1:5, v/v), the gel particles were subjected to dehydration by 500 μL of acetonitrile twice. 10 mM DTT (1 M DTT: 25 mM NH₄HCO₃ = 1:100) was added and incubation at 56 °C for 1 h. The liquid was removed and 55 mM iodoacetamide (IAM) (0.55 M IAM:25 mM NH₄HCO₃ = 1:10) was added. After incubation for 45 min at room temperature in dark environment, the particles were dehydrated by acetonitrile again and digested by 0.01 μg/μL with 25 mM NH₄HCO₃ with Trypsin overnight at 37 °C. The resulting peptides were separated by Easy-nLC 1200 (Thermo Fisher Scientific, USA) with self-packed C18 column (75 μm internal diameter, 1.9 μm column size, 25 cm column length) at a flow rate of 200 nL/min by the following effective gradient: 0–3 min, 5% mobile phase B (80% acetonitrile, 0.1% formic acid); 3–45 min, mobile phase B linearly increased from 8% to 44%; 45–50 min, mobile phase B rose from 44% to 60%; At 50-53 min, mobile phase B rose from 60% to 100%; 53–60 min, 80% mobile phase B. The nanoliter liquid phase separation end was directly connected to the mass spectrometer. The separated peptides were ionized by a nanoESI source and then passed to a tandem mass spectrometer Oritrap Exploris 480 (Thermo Fisher Scientific, USA) for data dependent acquisition mode detection. The main parameters were set: ion source voltage was set to 2.1 kV, MS1 mass spectrometer scanning range was 350 ~ 1600 m/z; resolution was set to 60,000; MS2 starting m/z was fixed at 100. The ion screening conditions for MS2 fragmentation: charge 2⁺ to 7⁺, and the first second parent ions with the peak intensity exceeding 50,000. The ion fragmentation mode was HCD. The dynamic exclusion time was set to 30 seconds. The AGC was set to: MS1 1E6, MS2 1E5. Results from search engine were pre-processed and re-scored using Percolator (V2.04) to improve the matching accuracy. The output was then filtered by FDR 1% at spectral level (PSM-level FDR ≤ 0.01) to obtain a significant identified spectrum and peptide list. The Mascot 2.3.02 search engine was used to identify proteins with more than 3 unique peptides based on UniProt protein database. LC-MS analysis was performed once.

Measurement of fluorescence

To generate the EGFP reporting plasmids, the egfp and the promoter of mbpA or mbpB were amplified from pAb-egfp²³ and L. plantarum WCFS1 genomic DNA, respectively. The fragments were cloned into the EcoRI site of pLH01. To detect the EGFP expression under the control of mbpA or mbpB promoter in the mbpR complementary mutant, a fragment containing mbpA or mbpB promoter, egfp, and mbpR with its native promoter was amplified by overlap extension PCR and cloned into the EcoRI site of pLH01. To determine the expression of EGFP and mCherry under the control of P23 promoter, a fragment containing the MbpR-binding sequence of mbpA between egfp and mCherry was amplified by using overlap extension PCR and cloned into the pP23-FLAG vector digested with EcoRI and NotI. The ORF of mCherry was amplified from pT7-6 × His-MCS-mCherry vector (Miaolingbio, China). To determine the expression of EGFP and mCherry fusing protein in the mbpR complementary strain, the DNA fragment of mbpR with its native promoter and the fragment containing egfp, MbpR-binding sequence of mbpA, and mCherry were cloned into the pP23-FLAG vector digested with EcoRI and NotI. The generated reporting plasmids were transformed into L. plantarum by electroporation. Bacterial colonies containing reporter plasmids were picked and grown to OD₆₀₀ = 1.0 in MRS supplemented with 10 μg/mL of chloramphenicol at 37 °C. Cells were collected by centrifugation and washed with 3 × PBS. The signal of fluorescent protein was measured by Multimode Microplate Reader Synergy H1 (BioTek, USA) and normalized to the OD₆₀₀ at the time of sampling.

Western blot

The DNA fragments containing FLAG-tagged mbpR and rpoD with their respective promoters were amplified and cloned into the pLH01 plasmid digested with EcoRI. To determine the expression of MbpR from the pP23-mbpR-FLAG plasmid used for ChIP-Seq, another plasmid (pP23-mbpR-FLAG-rpoD) was constructed by the addition of a FLAG-tagged rpoD fragment in the NheI cutting site of pP23-mbpR-FLAG vector. The reconstructive plasmids were transformed into ΔmbpR by electroporation and the levels of FLAG-tagged proteins were detected by using western blot. Briefly, cells (strains used for WB were listed in Supplementary Data 6) were grown to logarithmic growth phase (OD₆₀₀ = 1.0) and were collected by centrifugation. After rinsing three times with PBS, the pellets were resuspended in 100 μL of Laemmli buffer, and boiled for 5 min. 20 μL of samples were run on 10% SDS-PAGE gels and transferred onto polyvinylidene fluoride membranes. The membranes were blocked in Tris-buffered saline containing 0.5% Tween 20 (TBST) and 5% non-fat milk, followed by overnight incubation with the Anti-DDDDK tag (Binds to FLAG® tag sequence) antibody [EPR20018-251] (Abcam, ab205606, diluted 1:1000) in blocking buffer at 4 °C. After three washes with TBST, the membranes were incubated with HRP-labeled Goat Anti-Rabbit IgG (Beyotime, A0208, diluted 1:1000) in TBST for 2 h. The bands were visualized by using SignalFire ECL Reagent (Cell Signaling Technology, USA).

In vitro transcription assay

The mbpA CDS region with T7 promoter was amplified using primers listed in Supplementary Data 7. The products were purified and 1 μg of the product was used as transcription template for in vitro transcription assay. In vitro transcription assay was performed by using the T7 Quick High Yield RNA Transcription Kit (Beyotime, China) as described by the manufacturer in a final volume of 20 μL. The reaction was initiated by the addition of templates and incubated for indicated times at 37 °C. Synthesized RNAs were purified and quantified using nanodrop (Thermo Fisher Scientific, USA).

Bacterial two-hybrid assay (BTHA)

BTHA is based on the reconstitution of adenylate cyclase activity in E. coli BTH101⁵⁸. The mbpR was amplified from L. plantarum WCFS1 genomic DNA and cloned into the pUT18 digested with HindIII and EcoRI. The mbpR and rpoD were amplified from L. plantarum WCFS1 genomic DNA and cloned into the pKNT25 digested with HindIII and EcoRI. To test protein-protein interactions, the pair of plasmids were co-electroporated into BTH101. The cells were grown on LB agar plates containing 100 μg/mL ampicillin, 50 μg/mL kanamycin, 0.5 mM IPTG, and 40 μg/L 5-bromo-4-chloro-3indolyl-β-D-galactopyranoside (X-gal) for 48 h at 30 °C. The interactions were then analyzed by the formation of blue clones.

RNA-Seq analysis

High-throughput RNA-Seq was performed by using double-stranded cDNA⁵⁹. Double biological replicates were sequenced for each strain. The trimmed sequence reads were aligned to the L. plantarum WCFS1 genome sequence using Bowtie2-2.2.3⁶⁰, and the normalized read counts were compared by HTSeq v0.6.1⁶¹. Then, the FPKM of each gene was calculated based on the 26 lengths of this gene, and the read count was mapped to the gene⁶¹. Differentially expressed genes were identified as those with a high or low expression level (absolute value of log₂ Fold Change was ≥1.5, p < 0.05) across all replicates with a false discovery rate (FDR) of <0.01.

Site-directed Mutagenesis

The sequences encoding protein mutants were generated by an overlapping PCR mutagenesis approach. Then, the purified PCR product was cloned into vector pET28a using In-Fusion cloning. After verification by DNA sequencing, all mutations plasmids were transformed into E. coli BL21 (DE3) strain, and the mutant proteins were expressed and purified using the same protocol described above.

Molecular docking analysis

To construct the 3D model, the protein sequence of MbpR (accession no: WP_011100868.1) was downloaded from the GenBank database in FASTA format and was submitted to the AlphaFold server (https://golgi.sandbox.google.com/). The output of the predicted model generated as a PDB file was downloaded for further analysis. 3D structure of c-di-GMP was extracted from the complex of syn_CdgR-(c-di-GMP) (PDB: 8HJA)⁶² by pymol. Molecular docking was performed by using AUTODOCK tools 1.5.6 (http://www.scripps.edu/mb/olson/doc/Autodock). Before docking, polar-H atoms were added to the target protein and c-di-GMP, and Gasteiger charge calculations were performed. The molecule files were then saved in PDBQT format. Ligand-centered maps were generated by the AutoGrid program with a spacing of 0.368 Å and grid dimensions of 126 × 126 × 126 Å3. To screen the potential binding sites, the grid box was prepared to wrap the entire WYL domain. The receptor grid was centered at center_x = 8.641, center_y = 1.573, center_z = -13.284, size_x = 20.83, size_y = 26.25, size_z = 26.25. The exhaustiveness was set at 50. The conformation with the lowest binding energy was chosen and the potential ligand-protein interaction was calculated by the protein-ligand interaction profiler (https://plip-tool.biotec.tu-dresden.de/plip-web/plip/index). Finally, the surface representation of the structural model of MbpR in complex with c-di-GMP and the potential amino acids of MbpR involved in c-di-GMP binding was visualized by Pymol.

RT-qPCR

Strains including WT(pLH01), ΔmbpR(pLH01), ΔmbpR(mbpR) were cultured to OD₆₀₀ = 1.0 and then harvested. RNA was isolated using a TRIZOL Reagent (Thermo Fisher Scientific, USA). cDNA synthesis and RT-qPCR analysis were performed with ChamQ^TM Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) according to the manufacturer’s instructions in a qTOWER 3 G (Analytic Jena, Jena, Germany). The expression levels of target genes were normalized to the level of the 16S RNA transcript for each experiment. The relative expression levels of the target genes were calculated using the comparative CT (2^-ΔΔCT) method⁶³.

Isothermal titration calorimetry (ITC) assay

The experiments were performed on MicroCal PEAQ-ITC microcalorimeter (Malvern Panalytical GmbH, Germany). Both the c-di-GMP (Sigma-Aldrich) and protein were dissolved in Tris buffer (100 mM Tris-HCl, 300 mM NaCl, pH 8.0). Titrations began with one injection of 0.4 μL of c-di-GMP (500 μM) solution into MbpR solution (50 μM) in the calorimeter cell at 25 °C. The volume of c-di-GMP injection was changed to 3 µL in the subsequent twelve injections. The data were calibrated with a buffer control. The instrument software (MicroCal PEAQ-ITC Analysis Software) was used for baseline adjustment, peak integration, and normalization of the reaction heats with respect to the molar amount of injected ligand as well as for data analysis and binding parameter evaluation.

Enzyme activity assays

DGC and PDE activity assays were performed with 10 μM purified protein (final concentration) in a final volume of 40 μL of reaction buffer [100 mM Tris-HCl (pH 8.0), 300 mM NaCl, 10 mM MgCl₂] at 37 °C. The reactions were induced by adding GTP or c-di-GMP at a final concentration of 0.1 mM followed by heating for 5 min at 100 °C. The products were detected by a high-performance liquid chromatography (HPLC) instrument equipped with a UV/Vis detector set to 253 nm. Separation was carried out using a reverse-phase C18 Waters column (4.6 × 250 mm; 5 μm) and a flow rate of 1 mL/min. Solvents containing methanol and 10 mM ammonium acetate at a ratio of 1/9 (v/v) were used.

Microscale thermophoresis (MST) assay

The protein binding affinities to c-di-GMP and other ligands were assessed by a Nano Temper 16 Monolith NT.115 instrument (NanoTemper technologies, Germany). In brief, purified proteins were labeled with the L001 Monolith NT.115 Protein Labeling Kit (Nano Temper). A mixture of labeled protein and varying concentrations of unlabeled c-di-GMP was prepared and loaded into standard treated silicon 19 capillaries (K002 Monolith NT.115. The fluorescence readings were acquired by a Nano Temper 16 Monolith NT.115 instrument, with measurements conducted at 20% LED power and 40% MST power.

Acid and bile salt tolerance assay

The cells of cultures growing in MRS were harvested by centrifugation at 4500 g for 10 min and then washed twice in PBS. Cell pellets were resuspended with MRS broth with different pH values or containing various concentrations of taurocholic acid sodium salt. The mixture was incubated at 37 °C for 2 h. The viable count was carried out on MRS agar, and the survival rate (%) was calculated.

Cell adhesion experiments

HT-29 cell (CBP60011) obtained from Nanjing Cobioer Biosciences Co. Ltd. was used to test the adhesive ability of L. plantarum. Cells (5 × 10⁵ cells/mL) were seeded in a 96-well culture plate and incubated at 37 °C under 5% CO₂ until the cells reached full confluence. L. plantarum was cultured until reaching mid-log phase. The bacterial cells were collected and washed twice with PBS, and then suspended in DMEM to obtain a final concentration of 10⁸ CFU/mL. After washing the HT-29 cells twice with PBS, the bacterial suspension was added to each well and incubated for 2 h. Cells containing attached bacteria were rinsed thrice with PBS and lysed by incubating with 1% (v/v) Triton X-100 at 4 °C for a duration of 20 min. The quantity of adhered L. plantarum was determined by plating serial dilutions of the lysed cells on MRS plates, followed by colony counting after incubation at 37 °C for 24 h.

Animal experiments

This study was approved by the Animal Ethics Committee of Hainan University (approval number: HNUAUCC-2024-00011), and all animal operations were carried out in accordance with the “Guidelines for The Care and Use of Experimental Animals” of Hainan University. As substantial researches have shown that male and female mice exhibit notable differences in metabolism, hormone levels, and immune responses^64,65,66, we thus only used male mice for experiments. Six-week-old male C57 BL/6 mice were purchased from Guangdong Vital River Laboratory Animal Technology. Randomly allocated mice of each group (n = 5 per group) were maintained at an ambient temperature of 25 ± 2 °C, relative humidity of 55% ± 5%, and a 12-hour light-dark cycle. The mice were allowed to drink and eat freely. The diet used is provided by Jiangsu Xietong Pharmaceutical Bio-Engineering Co., Ltd., catalog number XTI01WC-009. The main composition of the diet includes moisture ≤100 g/kg, crude protein ≥180 g/kg, crude fat ≥40 g/kg, crude fiber ≤50 g/kg, crude ash ≤80 g/kg, total phosphorus ranging from 6 to 12 g/kg, and calcium content between 10 and 18 g/kg. Additionally, the diet contains significant levels of vitamins such as vitamin A (13692 IU/kg), vitamin D (1339 IU/kg), and vitamin E (115.4 mg/kg). The total energy content is 3790 Kcal/kg, with a physiological energy of 3408 Kcal/kg. The protein, fat, and carbohydrate supply ratios are 21.5%, 11.1%, and 67.4%, respectively. Mice were fasted for 12 h before the experiment and then were gavaged with 5 × 10⁸ CFU of bacterial solution or an equal volume of MRS. The mice were euthanized by decapitation after 6 h of gavage, and the colon was collected at a length of 2 cm. After lavage trice with sterile PBS, the colon was stirred with a homogenizer. The samples were diluted and plated on MRS agar containing tetracycline (10 μg/mL), streptomycin (10 μg/mL), kanamycin (50 μg/mL), chloramphenicol (10 μg/mL) and vancomycin (30 μg/mL). Plates were placed in an incubator at 37 °C for 48 h to form single colonies.

Statistics & reproducibility

The data are presented as the means ± standard deviations (SDs) from at least three independent experiments and analyzed using Prism 8 software (GraphPad). The error bars indicate the SD. Dots represent individual data. Biological replicates and numbers of independent experiments were stated in the legends. Statistical significance was performed using one-way ANOVA with Tukey’s multiple-comparison test.