The CarSR two-component system directly controls radD expression as a global regulator that senses bacterial coaggregation in Fusobacterium nucleatum

Abstract

Two-component systems (TCS) enable bacteria to sense and respond to environmental signals, facilitating rapid adaptation. Fusobacterium nucleatum, a key oral pathobiont, employs the CarSR TCS to modulate coaggregation with various gram-positive partners by regulating the expression of radD, encoding a surface adhesion protein, as revealed by RNA-Seq analysis. However, the direct regulation of the radD-containing operon (radABCD) by the response regulator CarR, the broader CarR regulon, and the signals sensed by this system remain unclear. In this study, chromatin immunoprecipitation followed by high-throughput DNA sequencing (ChIP-seq) identified approximately 161 CarR-enriched loci across the genome and a 17 bp consensus motif that likely serves as the CarR-binding site. Notably, one such binding motif was found in the promoter region of the radABCD operon. The interaction of CarR with this binding motif was further validated using electrophoretic mobility shift assays, mutagenesis, and DNase I footprinting analyses. Beyond regulating radABCD, CarR directly controls genes involved in fructose and amino acid (cysteine, glutamate, lysine) utilization, underscoring its role as a global regulator in F. nucleatum. Lastly, we discovered that RadD-mediated coaggregation enhances radD expression, and deletion of carS abolished this enhancement, suggesting that coaggregation itself serves as a signal sensed by this TCS. These findings provide new insights into the CarR regulon and the regulation of RadD, elucidating the ecological and pathogenic roles of F. nucleatum in dental plaque formation and disease processes.IMPORTANCEFusobacterium nucleatum is an essential member of oral biofilms acting as a bridging organism that connects early and late colonizers, thus driving dental plaque formation. Its remarkable ability to aggregate with diverse bacterial partners is central to its ecological success, yet the mechanisms it senses and responds to these interactions remain poorly understood. This study identifies the CarSR two-component system as a direct regulator of RadD, a major adhesin mediating coaggregation, and reveals its role in sensing coaggregation as a signal. These findings uncover a novel mechanism by which F. nucleatum dynamically adapts to polymicrobial environments, offering new perspectives on biofilm formation and bacterial communication in complex oral microbial ecosystems.

Keywords: CarSR; ChIP-seq; Fusobacterium nucleatum; RadD; coaggregation; two-component system.

Fig 1

Validation of FLAG-tagged CarR functionality in F. nucleatum. (A) Schematic of plasmid constructs showing the replacement of mCherry in pCWU6 with the gene of interest to generate expression plasmids for wild-type CarR, FLAG-tagged CarR (CarR-FLAG), and FLAG alone, each controlled by the native PcarR promoter. (B) FLAG-tagged CarR mirrors the function of wild-type CarR by restoring RadD expression levels in the ∆carR mutant to those observed in wild-type cells. Stationary-phase cultures of the indicated strains (1 mL each) were prepared for SDS-PAGE and western blotting using anti-RadD, anti-FLAG, and anti-FomA antibodies. FomA served as the loading control. (C) Coaggregation assays show that FLAG-tagged CarR supports coaggregation with A. oris MG-1 similar to untagged CarR. Representative images from three independent experiments demonstrate various coaggregation phenotypes across strains. (D) Diagram of the radABCD, carRS, and fruRBA operons, including their relative genomic positions and associated gene IDs (radD, carR, and fruB) as annotated on the National Center for Biotechnology Information RefSeq assembly: GCF_003019785.1. (E) FLAG-tag does not alter CarR’s regulatory effect on the radD and fruRBA operons at the mRNA level. Cultures of each strain were grown to an OD₆₀₀ of 1.0 in a TSPC medium, and then processed for RNA extraction. Transcript levels of radD and fruB were quantified by qRT-PCR, normalized to gyrB rRNA, and set relative to the wild type, with an arbitrary value of 1. Data represent the mean ± standard error of the mean from three biological replicates.

Fig 2

Identification of CarR binding sites in F. nucleatum using ChIP-seq. (A) A representative genome-wide ChIP-seq profile displaying CarR binding peaks across the F. nucleatum genome is visualized using Integrated Genome Viewer. Peak height corresponds to the number of reads from the immunoprecipitated samples (top) and input samples (bottom) mapped to the F. nucleatum ATCC 23726 genome, with enriched peaks indicating potential CarR regulatory regions. (B) Pie chart representing the distribution of 161 enriched CarR binding peaks, indicating the locations of binding sites within various genes. (C) Consensus CarR-binding motif generated from ChIP-seq data using MEME. The motif logo shows nucleotide frequency at each position, with letter height representing prevalence. The position frequency matrix illustrates nucleotide distribution across the consensus CarR binding sequence.

Fig 3

CarR regulation of the radABCD operon via direct binding to the radA promoter in F. nucleatum. (A) The ChIP-seq coverage track shows CarR binding enrichment at the radA promoter region visualized using the Integrated Genome Viewer. The immunoprecipitated samples (top) display a prominent peak at the radA promoter, indicating specific CarR binding. In contrast, the input samples (bottom) serve as controls to confirm the specificity of the enrichment. (B) Sequence analysis of the radA promoter region, with key regulatory elements annotated. The predicted CarR binding site highlighted in red lies between the −10 box and the transcription start site. Additional elements, including the transcriptional start site, ribosome binding site (rbs), start codon of radA, and potential −35 and −10 boxes, are also indicated. (C) Four synthesized oligonucleotide probes (P1, P2, P3, and P4) were designed for competitive electrophoretic mobility shift assay (EMSA) analysis. P1 represents the wild-type CarR binding site, while P2, P3, and P4 contain distinct mutations within the binding site. (D) EMSA demonstrating CarR binding to the intact radA promoter. Purified CarR protein was incubated with a Cy5-labeled radA promoter probe in the presence of a 10-fold excess of non-specific competitor DNA (poly[dI-dC]) and 25 ng of labeled probe. Competitive binding was tested by adding a 100-fold excess of unlabeled probes (P1, P2, P3, P4) as indicated, which were pre-incubated with the CarR mixture before the labeled probe. (E) DNase I footprinting analysis of CarR binding to the radA promoter. Electropherograms display the protected regions of the radA promoter following DNase I digestion in the presence of 7 µM CarR compared to without CarR, confirming the specific binding region of CarR on the promoter.

Fig 4

CarR regulation of the fruRBA operon via direct binding to the fruR promoter in F. nucleatum. (A) ChIP-seq analysis shows CarR binding at the fruR promoter. The ChIP samples (top) display a clear peak at the promoter, indicating CarR binding, while the input samples (bottom track) confirm this enrichment is specific. Data from the two experiments are consistent, as shown in the Integrated Genome Viewer. (B) Growth assay demonstrating that deletion of the hypothetical protein region upstream of the fruRBA operon abolishes fructose-stimulated growth in F. nucleatum. This phenotype cannot be restored by complementation, suggesting possible misannotation of this region. Growth was measured as OD600 after 12 h in the presence of 20 mM fructose or TSPC alone. (C) Sequence analysis of the fruRBA promoter region with annotated regulatory elements. The predicted CarR binding site shaded in red is located within a region encoding a hypothetical 45-amino-acid protein. This hypothetical protein’s coding sequence overlaps with the fruR gene. Potential FruR binding sites are highlighted in green. (D) Electrophoretic mobility shift assay demonstrating CarR binding to the intact fruR promoter. Purified CarR protein was incubated with a Cy5-labeled fruR promoter probe in the presence of a 10-fold excess of non-specific competitor DNA (poly[dI-dC]) and 25 ng of the labeled probe. As indicated, competitive binding was assessed by pre-incubating the CarR mixture with a 100-fold excess of unlabeled probes before adding the labeled probes. (E) Schematic of two PfruR promoter versions: wild type and a variant with a mutated CarR binding site both used to drive luciferase expression as shown in panel F. P1 represents the wild-type CarR binding site within the fruR promoter, while P2 contains distinct mutations within the binding site. (F) Schematic of plasmid constructs for luciferase reporter assays showing the replacement of mCherry in pCWU6 with luciferase driven by the wild-type or mutated fruR promoter. (G) Promoter activity of PfruR and its mutant variant fused to luciferase in WT, ∆carR, and ∆carS backgrounds. Cultures were sampled at an OD₆₀₀ of 0.6 for luciferase assays. Luciferase activity (RLU) was normalized to cell density (OD600), with data representing the mean ± standard deviation from three biological replicates.

Fig 5

Validation of ChIP-seq findings by EMSA, confirming CarR’s regulatory role in F. nucleatum metabolic pathways through direct promoter binding. (A) Alignment of ChIP-enriched regions from five target genes identifies a conserved CarR-binding motif, underlined and highlighted in black, demonstrating sequence specificity in CarR binding. (B, C, D, E) Combined ChIP-seq and EMSA analyses reveal CarR binding at the megL, gluD, and kal promoters while showing no binding at the fap2 promoter, which is used as a negative control. Top panels: ChIP-seq coverage tracks from two independent experiments (ChIP sample in blue and input sample in red) exhibit consistent enrichment at the promoter regions of megL, gluD, and kal mapped to the F. nucleatum ATCC 23726 genome. Data are visualized using the Integrated Genome Viewer. Bottom panels: Electrophoretic mobility shift assay results confirm CarR binding to Cy5-labeled promoter probes (megL, gluD, and kal) in a concentration-dependent manner. A 10-fold excess of non-specific competitor DNA was included to validate specificity, and competitive binding was further confirmed by pre-incubation with a 100-fold excess of unlabeled probes. The fap2 promoter is a negative control, demonstrating no interaction with CarR.

Fig 6

Coaggregation as a signal sensed by the CarRS two-component system in F. nucleatum. (A) Diagram of the radABCD operon structure, with the radB gene highlighted in a box to indicate the target detection region for qRT-PCR. (B) Schematic of the experimental design illustrating F. nucleatum single strain and three co-culture conditions with A. oris MG-1 used to investigate coaggregation signaling. Co-culture conditions include: (i) simple co-culture mix; (ii) high-density co-culture; and (iii) coaggregation clump formation. Detailed descriptions of each condition are provided in the Methods section. (C) qRT-PCR analysis of radABCD operon expression across the different experimental conditions, with RNA extracted from each indicated experimental setup. radB expression levels were normalized to the gyrB gene, and the experiments were validated in two independent runs, each conducted in triplicate. The results are shown as means ± standard deviation (n = 3). *, P < 005, **, P < 0.01, n.s, not significant (Student’s t test). (D) Coaggregation assay showing that RadD expression in the ∆carS mutant restores coaggregation ability to levels similar to those of the wild-type strain in coaggregation with A. oris MG-1. Representative images from three independent experiments with the indicated strains.

Fig 7

Model of CarS-mediated sensing of RadD-dependent coaggregation and regulation of radD expression. (A) In wild-type F. nucleatum, RadD interacts with an unknown receptor on A. oris, initiating coaggregation. This coaggregation generates a signal that activates CarS, leading to CarR-mediated transcriptional regulation of radD. RadD expression forms a positive feedback loop, enhancing coaggregation. See main text for details. (B) In the absence of coaggregation, CarS remains inactive, leaving CarR in its non-phosphorylated state. Non-phosphorylated CarR is a repressor of the radABCD operon, leading to reduced radD expression. (C) Coaggregation still occurs without CarS (∆carS pradD), but the signal cannot be relayed to CarR. As a result, CarR remains non-phosphorylated and unable to activate radD transcription, leading to repression of radD expression despite coaggregation.