All the E. faecalis used herein were derived from E. faecalis OG1RF (ATCC 47077). The OG1RF∆ccfA strain was constructed in our lab as outlined in Yang et al. (2023). Supplementary Text S1 presents the details. The bacteria were grown in the brain heart infusion (BHI) medium (Coolaber Science and Technology, China) at 37°C and 150 rpm agitation.

First, OG1RF cultures in the logarithmic growth phase were initiated by inoculating them into the BHI medium (1:1,000). Following that, cells were grown for 4 h and then exposed to different LVF concentrations. Herein, LVF concentrations were MIC-based. Supplementary Text S2 describes antibiotic type selection and MIC measurement. Bacteria were treated with various antibiotic concentrations, followed by sample collection at different time intervals for enumeration. Subsequently, the cultures were gradient-diluted with a Phosphate-Buffered Solution (PBS) and then grown on BHI agar for counting.

First, OG1RF cultures in the logarithmic growth phase were inoculated into the BHI medium (1:1000) and then cultured at 37°C. At specific culture time points (4, 4.5, 5, 5.5, and 6 h), a sample was removed from the culture medium and challenged with LVF at a final concentration of 20 mg/L. The mixed cultures were then incubated at 37°C for 8 h. Subsequently, the samples were diluted with PBS before inoculating their appropriate dilutions on BHI agar to count the surviving bacteria. Simultaneously, the samples’ bacterial concentrations before LVF treatment was determined. The results were presented as a colony-forming unit (CFU)/mL, and the persister frequency (f) was determined using the following Formula (1):

where N2 represents the number of bacteria in the culture after antibiotic treatment (CFU/mL), and N1 is the number of bacteria in the culture before antibiotic treatment (CFU/mL).

First, OG1RF cultures in the logarithmic growth phase were inoculated into the BHI medium (1:1,000) and then supplemented with different concentrations of the pheromone cCF10 (8, 10, 12, 14, 16, and 20 ng/mL). After incubating the cultures for 3 h, the bacteria were collected through centrifugation at 8,000 rpm for 3 min. The collected material was re-suspended in the BHI medium, and then exposed with LVF at a final concentration of 20 mg/L before determining the persistence rate as earlier mentioned. We detected the persistence rate of OG1RF∆ccfA to further verify the effect of cCF10 on the formation of E. faecalis persistence. We also examined the effects of different cCF10 concentrations (1, 10, and 20 ng/mL) on the persistence rate of OG1RF∆ccfA. Notably, Kingsray Biotechnology Co., Ltd. (Nanjing, China) synthesized the peptide pheromone used herein, the 7-amino acid sex pheromone cCF10 (C-clumping-inducing peptide, amino acid sequence LVTLVFV; Dunny and Leonard, 1997). To confirm the specificity of cCF10, a site-mutated sequence called cCF10-F (sequence = LVFLVTV) was employed and cCF10-F were synthesized by GenScript (China). The procured pheromones cCF10 and cCF10-F were dissolved in acetonitrile (Maclean’s, China), respectively, and then stored at −20°C in the dark.

We extracted the extracellular pheromone per the methodology described by Zhou et al. (2023). Briefly, the collected samples were placed on ice for 20 min and then centrifuged at 6,000 rpm for 10 min at 4°C. The supernatant was then filtered using a 0.22 μm filter (SLGP033RB Millipore, Massachusetts, United States) and mixed with ammonium hydroxide solution and acetonitrile (both from Shanghai Macklin Biochemical Technology Co. Ltd., China) at a ratio of 8:1:1. The resulting mixtures were vortexed at 1,400 rpm for 15 min at room temperature (RT) and then centrifuged for 15 min at 25,000 rpm (ST16R; ThermoFisher Scientific, Massachusetts, United States). Following that, thorough mixing was done with the supernatant and an equivalent amount of an aqueous solution of 10% ammonium hydroxide. The cCF10 elution process was as follows. First, the extraction column (Sep-Pak C18 WAT054945 Waters, Massachusetts, United States) was activated by sequentially flowing 5 mL acetonitrile through it, followed by 5 mL water at a rate of one drop/s. A volume of 30 mL of test solution was passed through the extraction column at a speed of one drop per 3 s. Subsequently, to eliminate water-soluble impurities from the column bed matrix material within the extraction column or those that have been absorbed during activation or sample loading, 2 mL of water was introduced as an eluent. This step was followed by flushing with an aqueous acetonitrile solution (30%), also amounting to 2 mL, to ensure the impurities were properly removed. Finally, the collected eluate was sent to Science Compass (Zhejiang, China) for Liquid Chromatography Mass Spectrometry (LCMS) testing. Supplementary Text S4 presents the detailed detection procedure. Preparation of Scanning Electron Microscope (SEM) samples.

The cell pellet was gently re-suspended in a pre-cooled 2.5% glutaraldehyde fixative for 24 h at 4°C. The fixative was then removed from the sample via centrifugation at 6000 rpm for 5 min, followed by dehydration using an ethanol gradient approach. Ultimately, the bacteria underwent lyophilization with a FDU1200 EYELA freeze-dryer (Tokyo, Japan) and were then imaged using a Sigma 300 SEM (Zeiss, Germany).

Firstly, E. faecalis OG1RF normal bacteria, persisters, and cCF10 treatment persisters were collected as previously described. The bacterial cultures were first collected via centrifugation at 8,000 rpm for 3 min. The collected residues were then cooled in liquid nitrogen for 15 min and stored at −80°C before being sent to Allwegene Technology Co. Ltd. (Beijing, China) for transcriptome sequencing. Supplementary Text S5 describes the specific sequencing methods employed. The transcriptome data (project number PRJCA026007) has been deposited in the China National center for Bioinformation. The URL is https://ngdc.cncb.ac.cn/bioproject/browse/PRJCA026007.

The OG1RF strain was cultured in a six-well plate, and any unadhered bacteria were gently washed off with PBS. The plates with adherent biofilms were then air-dried at RT and fixed with methanol (1 mL/well) for 15 min. The methanol was discarded, and the plates were air-dried again. A 1% crystal violet staining solution (1 mL/well) was then added to stain the biofilms for 1 h. Subsequently, the PBS was used to carefully wash away the crystal violet staining solution until it became colorless, after which the plates were left to air dry. The OD₅₇₀ was then determined using a Multifunctional Fluorescent Enzyme Labeler (Spectra Max M5, United States).

First, the bacterial cultures were centrifugally washed with PBS at 6,000 rpm for 5 min and then subjected to ATP measurement per the instructions in the ATP Assay Kit (Beyotime Biotechnology Co. Ltd., China). Chemiluminescence measurements were recorded using MFEL (SpectraMax M5, United States), and the data were normalized to the total amount of protein measured per the instructions in the BCA Protein Concentration Measurement Kit (Beyotime Biotechnology Co. Ltd., China).

The samples were transferred to ice for immediate cooling, and total RNA extraction was performed per the instructions in the Gram-Positive Microbes RNA Isolation Kit (Beibei Biotechnology Co, China). Subsequently, reverse transcription polymerase chain reaction (RT-qPCR) was conducted to convert total RNA into cDNA using the first strand cDNA synthesis kit (Tiangen, China) and random primers. Subsequently, the Power Up SYBR Master Mix (ThermoFisher, United States) was used for real-time PCR analysis, along with the CFX96 Real-Time System (Bio-Rad Laboratories Inc., Hercules, United States). The absolute quantification method was utilized for determining the mRNA levels of specific genes related to pheromones, such as those encoding pheromones and responding to them. The 16S rRNA gene was utilized as an internal reference for normalization purposes. The primer sequences for RT-qPCR analysis can be found in Supplementary Table S1, and were designed with the assistance of DNASTAR.

The experiments were independently repeated a minimum of three times. Data analysis was conducted with SPSS 25 software (IBM, Armonk, NY). All data were presented as mean ± standard deviation (SD) and were analyzed using the independent samples t-test or one-way analysis of variance followed by the student–Newman–Keuls test. Results or differences with p < 0.05 were considered statistically significant.

Herein, we first established a persistent bacteria model and subjected E. faecalis strains to a series of persister analyses to examine the effect of pheromone cCF10 on the formation of E. faecalis persister cells (Supplementary Figures S1, S2). According to the “biphasic killing” curve, bacterial populations decreased sharply after antibiotic exposure and plateaued at 8 h (Supplementary Figure S2B). Furthermore, the number of surviving bacteria was almost stable when antibiotic concentrations exceeded 20 mg/L, indicating that the surviving bacteria were persisters (Balaban et al., 2004; Lewis, 2007; Kint et al., 2012; Harms et al., 2016). Based on SEM images (Supplementary Figure S2C), normal-growing E. faecalis cells exhibited a relatively round and full morphology, whereas LVF-exposed cells shriveled, with a large amount of biofilm attached to their surfaces, further indicating that the surviving bacteria were persisters. Extracellular polymers (EPS) is mainly a number of polymer substances, such as polysaccharides, proteins and nucleic acids. Our results showed that more polysaccharides, proteins and nucleic acids were produced around the persisters compared to normal bacteria (Supplementary Figure S3). Unlike dead bacteria, the persistent bacteria were able to resuscitate after the removal of antibiotic pressure. To illustrate this point, we added electron microscope images of the persistent bacteria at different resuscitation times (Supplementary Figure S4). We observed changes in the morphology of bacteria that had been screened for antibiotics when added to fresh media. We found that after 1 h of resuscitation, most of the bacteria in the visual field returned to normal bacterial form. In addition, we also compared the resuscitation curves of the persisters and the sterilization curves of the recovered persisters. We obtained the same growth curve as normal bacteria (Supplementary Figure S5A). Then we exposed the resuscitated to 20 mg/L of levofloxacin hydrochloride for 12 h. We found that the screened bacteria were still sensitive to levofloxacin hydrochloride (Supplementary Figure S5B). Our results indicated that the bacteria we screened were persisters rather than resistant bacteria.

Although bacterial growth was not affected when E. faecalis OGIRF was exposed to 10 and 12 ng/mL cCF10 concentrations (Supplementary Figure S6), the persister rate decreased from 0.109 to 0.050 and 0.047% (Figure 1A). This finding indicated that cCF10 could inhibit the formation of E. faecalis persisters. Interestingly, the persister rate of E. faecalis increased to 0.201%, 0.211%, and 0.205% when pheromone concentrations reached 14, 18, and 20 ng/mL, respectively. It could be attributed to the fact that the excessively high pheromone concentrations exceeded the regulatory range of E. faecalis, interfering with the physiological metabolism of the bacterium.

We also detected the persister rate of a mutant OG1RF∆ccfA strain to verify the effect of cCF10 on the formation of E. faecalis persistence. According to the results, the persister rate of the mutant strain increased to 0.31%. On the other hand, adding exogenous cCF10 at 10 and 20 ng/mL concentrations reduced the persister rate to 0.064% and 0.043%, respectively (Figure 1B), further indicating that the pheromone cCF10 could inhibit the persistence of E. faecalis within a certain range. To confirm the specificity of cCF10, we employed a site-mutated sequence called cCF10-F. Our results indicated that unlike cCF10, cCF10-F did not impact the formation of OG1RF persisters (Supplementary Figure S7). Based on SEM images (Figure 1C), normal-growing OG1RF∆ccfA cells exhibited a rounded and plump morphology, whereas LVF-exposed cells had a shriveled appearance, indicating that the surviving bacteria were persisters.

Herein, the persisters were screened with 20 mg/L LVF at different culture times to evaluate the changes in persister rate during the OG1RF growth stages. The bacteria reached the logarithmic phase at 4 h, at which the persister rate was 0.11%, and plateaued at 6 h at which the persister rate was 7.68%. Furthermore, the persister generation rate gradually increased with the growth of OG1RF, exhibiting 0.25%, 0.72%, and 2.48% persister rates after growth for 4.5, 5, and 5.5 h, respectively (Figure 2A). To explore the potential mechanisms, we evaluated the biofilm content of OG1RF during the growth process. The results (Figure 2B) revealed OG1RF biofilm accumulation during the growth process (the OD₅₇₀ changed from 0.204 to 1.289). Notably, although there was a concomitant increase in pheromone content as the bacterial population proliferated (Figure 2C), this heightened pheromone concentration did not cause a decline in the persister rate. Furthermore, the biofilm exerted its primary influence during this period. Additionally, we measured biofilm content in E. faecalis to investigate the potential impact of pheromones on biofilm formation. According to the results, bacteria accumulated a large amount of biofilm after persistence (OD₅₇₀ increased from 0.442 to 0.600 to 0.601), and cCF10 addition did not alter biofilm accumulation (Supplementary Figure S8). To support this idea, we added the results of different concentrations (8, 10, 12, 14, 18, and 20 ng/mL) of cCF10 on OG1RF biofilm formation. Our results suggested that pheromones do not affect the formation of bacterial biofilms compared to control group (Supplementary Figure S9).

Herein, we examined transcript level changes to further determine the molecular mechanisms of E. faecalis persistence and the mechanism by which cCF10 inhibits OG1RF persistence. We described the process of transcriptomics analysis in detail in Supplementary Text S5. Persisters had 2,065 differential genes compared to normal bacteria (Figure 3A). On the other hand, the cCF10 treatment group had 231 differential genes compared to the persisters group (Figure 3B). The top 20 enriched pathways are shown in order of q-value from smallest to largest. DEGs were mainly concentrated in metabolic pathways in normal and persisters groups (Figure 3C). On the other hand, in the persisters and cCF10 treatment groups, the results showed that the DEGs were mainly enriched in ABC transfer system (Figure 3D). Furthermore, in the persisters group, the expression levels of genes involved in DNA replication (Figure 3E) and ATP synthesis (Figure 3F) were decreased compared to normal bacteria. Among genes involved in DNA replication, danB2 exhibited the most significant changes (Figure 3E), with reductions to 0.89 and 0.90 times that of the control group in the persister and cCF10 treatment groups, respectively. On the other hand, among the genes associated with ATP synthesis, atpB (Figure 3F) had the most significant alterations, with reductions to 0.79-fold and 0.80-fold that of the control group in the persister and cCF10 treatment groups, respectively. In addition, to further verify whether cCF10 could alter the energy metabolism of OG1RF and prevent bacterial persistence, we focused on changes in gene expression levels associated with glycolysis and tricarboxylic acid (TCA) cycling. Clearly, the expression levels of genes related to metabolism decreased after the bacteria entered the persistence state. In cCF10 treatment group, the expression levels of genes (gatC, gatA, gatB, galK, nifJ, lpd, aceF, and pdhA) were all increased. Of these, galk showed the most significant change with a 1.07-fold increase compared to the persisters group (Figures 3K,L). Furthermore, phoU expression (Figure 3G) in the persister and cCF10 treatment groups were 0.89-fold and 0.90-fold that of the control group, respectively. These findings indicate a notable phoU downregulation following bacterial persistence. Furthermore, recA, which is associated with SOS response, was upregulated by 1.19-fold (in the persisters group) and 1.17-fold (in the cCF10 treatment group) relative to the control group (Figure 3H). The Opp2 system serves as the conduit for cCF10 secretion and uptake in OG1RF (Segawa et al., 2021). In the persisters group, the expression levels of genes involved in Opp2 system were decreased compared to normal bacteria (Figure 3I). However, no changes in Opp2 system gene expression were found in cCF10 treatment group compared to persisters (Figure 3I). In order to further verify this result, we carried out laboratory verification in 3.4. Our results have shown that the accumulation of biofilm played a crucial role in the generation of E. faecalis persisters. Although transcriptomic results showed that the expression of several genes (gelE, ccpA, eno, can and OG1RF_12096) associated with biofilm formation was mostly reduced in the persistent bacteria compared with normal bacteria (Figure 3J). However, no changes in biofilm formation genes expression were found in cCF10 treatment group compared to persisters. In addition, we added different concentrations of cCF10 during the growth of OG1RF in order to further verify the effect of cCF10 on OG1RF biofilm formation. Consistently, the addition of cCF10 did not alter the formation of OG1RF biofilms (Supplementary Figure S9).

The Opp system serves as the conduit for pheromone secretion and uptake (Segawa et al., 2021). Segawa et al. (2021) discovered that pheromone uptake in E. faecalis OG1RF was primarily dependent on the Opp2 system. The OppA lipoprotein is a well-conserved protein known for its ability to bind peptides. It interacts with an ABC transporter system consisting of two channel-forming proteins (OppB and OppC) and two membrane-associated ATPases (OppD and OppF; Segawa et al., 2021). Herein, we found that the expression of genes related to the Opp2 system decreased after bacteria entered the persistence phase (Figure 4), indicating that the pheromone-binding ability of E. faecalis persisters was reduced and that the uptake pathway was partially blocked. However, opp2A expression (Figure 4A) increased by 1.61-fold (10 ng/mL of cCF10) and 1.11-fold (14 ng/mL of cCF10) relative to the persisters group, suggesting that the cCF10-binding ability was improved. Similar changes were observed in opp2B, opp2C, opp2D, and opp2F. The expression of opp2B (Figure 4B) increased, respectively, by 1.25-fold (10 ng/mL of cCF10) and 1.50-fold (14 ng/mL of cCF10) relative to the persisters group, suggesting that the cCF10-binding ability was improved. The expression of opp2C (Figure 4C) increased, respectively, by 1.90-fold (10 ng/mL of cCF10) and 1.62-fold (14 ng/mL of cCF10) relative to the persisters group, suggesting that the cCF10-binding ability was improved. The expression of opp2D (Figure 4D) increased, respectively, by 1.53-fold (10 ng/mL of cCF10) and 1.66-fold (14 ng/mL of cCF10) relative to the persisters group, suggesting that the cCF10-binding ability was improved. The expression of opp2F (Figure 4E) increased, respectively, by 2.43-fold (10 ng/mL of cCF10) and 3.11-fold (14 ng/mL of cCF10) relative to the persisters group, suggesting that the cCF10-binding ability was improved. These findings suggested that cCF10 could induce the Opp2 system and enter bacterial cells.

(p)ppGpp is a signaling molecule that bacteria produce when they encounter challenging conditions, such as exposure to antibiotics. The build-up of (p)ppGpp can ultimately interfere with the movement of protons and inhibit ATP synthesis, leading to bacterial dormancy (Shang et al., 2020). Alterations in (p)ppGpp production negatively affect bacterial stress survival and virulence in E. faecalis, and (p)ppGpp directly inhibits the activity of enzymes involved in GTP biosynthesis (Gaca et al., 2013, 2015). According to research, RelA (synthase) and SpoT (hydrolase) maintain a steady state of (p)ppGpp (Hobbs and Boraston, 2019). At the same time, phoU was reported to affect the intracellular level of (p)ppGpp (Shang et al., 2020). Herein, relA (Figure 5A) and spoT (Figure 5B) expression decreased after the formation of persisters, and no significant changes were observed in the cCF10 treatment group. Furthermore, phoU expression (Figure 5C) in the cCF10 treatment group was 4.5 × 10⁻⁵ copies/16 s RNA, a 3.21-fold increase compared to the persisters group. These findings indicate that cCF10 could inhibit persistence by suppressing (p)ppGpp production in E. faecalis.

Repressed metabolic activity is a distinctive feature of bacterial persistence (Amato et al., 2014). Herein, When OG1RF entered the persistence state, ATP level decreased from 1.91 μmol/ mg protein to 0.20 μmol/ mg protein. We also observed that compared to the persisters group, 10 ng/mL cCF10 treatment group significantly promoted ATP synthesis (about 1.16-fold). However, the ATP content decreased by ~0.91 times relative to that of the persisters group after adding 14 ng/mL of cCF10 (Figure 6A). Additionally, we assessed the expression levels of genes involved in energy metabolism, including ATP synthesis and DNA replication to further investigate the effect of cCF10 in preventing the formation of persister cells. As expected, atpB (Figure 6B) and atpD (Figure 6C) were downregulated after persistence (~0.073-fold and 0.069-fold of normal bacteria, respectively). After cCF10 exposure, atpB and atpD were upregulated by 2.82-fold and 1.29-fold, respectively, compared to the persisters group, further implying that cCF10 could inhibit persistence by increasing ATP levels. Similarly, danB (Figure 6D), danE (Figure 6E), and recG (Figure 6F) expression levels in persister cells were 0.15, 0.18, and 0.11 times those in normal bacteria, respectively. Notably, cCF10 upregulated danE (Figure 6E) from 7.9 × 10⁻⁶ to 1.3 × 10⁻⁵ copies/16 s RNA (~1.6-fold increase) compared to persisters. These findings illustrated that cCF10 reduces the generation of persister cells by maintaining the metabolic activity of E. faecalis OG1RF.