1. Introduction
Antimicrobial resistance (AMR) poses a major global health threat and has attracted growing public attention [
1]. The widespread use of antibiotics has facilitated the spread of antibiotic resistance genes (ARGs) across animals, humans, and the environment [
2]. Horizontal gene transfer (HGT) is the primary mechanism of ARG transmission [
3], occurring through conjugation, transformation, and transduction. Among these, plasmid-mediated conjugation plays a particularly prominent role [
4]. Common plasmid types, such as IncI2, IncX4, and IncX3, mediate the transfer of the colistin (CS) resistance gene
mcr-1, the carbapenem resistance gene
blaNDM-1, and the tigecycline resistance gene
tet(X4), respectively. This promotes the spread of resistant bacteria and compromises the efficacy of last-resort antibiotics [[
5], [
6], [
7], [
8], [
9], [
10]].
The mammalian intestine hosts ∼100 trillion bacteria—collectively known as the intestinal microbiota—which serve as a major reservoir for ARGs [[
11], [
12], [
13]]. Consequently, developing novel strategies to curb AMR plasmid transfer in the gastrointestinal (GI) tract is essential. Conjugation inhibitors represent a promising strategy for limiting ARG dissemination. For example, Fe
2O
3@MoS
2 particles have been shown to block the conjugative transfer of the RP4 plasmid [
14]. Similarly, isothiocyanates reportedly inhibit the transfer of IncN, IncI2, IncP, and IncW plasmid types [
15]. However, these compounds exhibit toxicity and limited
in vivo efficacy [[
16], [
17], [
18]]. Therefore, safer and more effective conjugation inhibitors are urgently needed to address the public health risks posed by AMR plasmids. Cinnamic acid (CA; 3-phenyl-2-acrylic acid) is an organic compound naturally present in plants such as
Cinnamomum cassia (Chinese cinnamon) and
Panax ginseng, as well as in fruits, whole grains, vegetables, and honey [
19,
20]. In adults, daily intake of CA can reach up to 211 mg∙d
−1 [
21], and may exceed 599 mg∙d
−1 in some cases [
22]. Growing evidence indicates that CA exhibits a range of biological activities, including antioxidant [
23], anti-inflammatory [
24], and antithrombotic effects [
25]. However, its potential influence on plasmid conjugation has not been previously reported.
In this study, we demonstrate that CA inhibits the conjugation of various Inc-type plasmids. Conjugation inhibition within bacterial populations was assessed using a novel plasmid-tracing system developed in our laboratory [
26] and further validated
in vivo in a mouse model. Additionally, a comprehensive genome-wide transcriptomic analysis was performed to explore the underlying mechanisms of CA-mediated inhibition. Oral toxicity tests were conducted to assess the biocompatibility of CA. To our knowledge, this is the first study to report that CA can inhibit the conjugative transfer of ARGs.
2. Materials and methods
2.1. Bacterial strains and reagents
The donor strains selected for this study included
Escherichia coli (
E. coli) MG1655 carrying an IncP plasmid (RP4) encoding the ampicillin (Amp), kanamycin (Km), and tetracycline (Tc) resistance genes [
27];
E. coli CSZ4 harboring an IncX4 plasmid (pCSZ4) with CS resistance gene
mcr-1;
E. coli S110 containing an IncFII plasmid (pS110) with tigecycline resistance gene
tet(X4
);
Salmonella typhimurium (
S. typhimurium) LS3479 carrying an HI2 plasmid (pLS3479) with
mcr-1; and
S. typhimurium 14E1050 carrying an IncI2 plasmid (p14E1050) with
mcr-1. The recipient strain,
E. coli MG1655S, was genetically modified and chromosomally tagged with a
rpsL gene variant conferring streptomycin (Str) resistance [
26].
Strains were cultured in Luria-Bertani (LB) broth or on LB agar supplemented with the following antibiotics: Amp (100 mg∙L−1), Km (30 mg∙L−1), Tc (25 mg∙L−1), Str (2000 mg∙L−1), CS (2 mg∙L−1), apramycin (APR, 30 mg∙L−1), rifampicin (RIF, 50 mg∙L−1), and meropenem (MEM, 1 mg∙L−1). The MQ-RP4-sfGFP strain, used to track plasmid transfer, was constructed as described in Text S1 in Appendix A.
2.2. Determination of the minimum inhibitory concentration (MIC) and growth curves
MICs of the tested strains were determined by the broth dilution method according to Clinical and Laboratory Standards Institute (CLSI) guidelines. E. coli ATCC 25922 was used as the quality control strain, and each isolate was tested in triplicate.
For growth curve analysis, overnight cultures were diluted in fresh LB broth to an optical density at 600 nm (OD600) of 0.2. CA was added to final concentrations of 0, 50, 100, and 200 mg∙L−1. The cultures were incubated at 37 °C, and OD600 was measured hourly over a 12-h period. All experiments were performed in triplicate.
2.3. In-vitro conjugation experiments
Overnight cultures of donor and recipient strains were sub cultured in LB broth at 37 °C for 4 h to reach the logarithmic phase. Cultures were then adjusted to an OD
600 = 0.3 (10
8 CFU∙mL
−1) using phosphate-buffered saline (PBS) or LB broth. Donor and recipient strains were mixed at a 1:1 ratio (1 mL total volume) with final CA concentrations of 0 (with ethyl alcohol as the solvent), 50, 100, or 200 mg∙L
−1. After 12 h of static incubation at 37 °C, mating mixtures were plated on LB agar containing appropriate antibiotics to enumerate donors, recipients, and transconjugants. For the pre-exposure assay, donor and recipient strains were exposed to 200 mg∙L
−1 CA for 4 h, washed twice with PBS, and used in conjugation assays without further CA treatment [
28].
2.4. Fluorescence microscopy of bacteria
Bacterial cells were first washed with PBS and centrifuged at 6000 r∙min−1. A 1 µL aliquot of the suspension was carefully placed on a microscope slide. Imaging was performed at 50× magnification using a Leica TCS SP8 laser scanning confocal microscope (LSCM). The tetramethyl rhodamine isothiocyanate (TRITC) and fluorescein isothiocyanate (FITC) channels were used to visualize bacterial morphology and spatial distribution.
2.5. Animal
Specific pathogen-free, 6-week-old female C57BL/6 mice were housed at the Laboratory Animal Center of South China Agricultural University (Guangzhou, China). All animal care procedures followed the guidelines of the American Association for Accreditation of Laboratory Animal Care (2023C082) and complied with the Guide for the Care and Use of Laboratory Animals. Mice were acclimated for 7 d prior to experimentation.
2.6. Ex-vivo conjugation experiments
Live bacteria from mice were isolated using a previously established protocol [
26]. The donor strain MQ-RP4-
sfGFP was cultured in LB broth at 37 °C for 4 h to reach the logarithmic phase. Donor and recipient (gut) bacteria were washed twice with PBS and adjusted to an OD
600 = 0.5. A 10 µL mixture of donor and recipient cells was spotted onto brain heart infusion (BHI) agar plates supplemented with 0, 50, or 200 mg∙L
−1 CA. Plates were incubated aerobically at 37 °C for 12 h. Cells were then harvested by scraping and resuspended in 1 mL PBS. All samples were diluted in PBS and filtered through 0.2-μm filters to optimize sorting. PBS, donor, and recipient bacterial populations were used as gating controls. All samples were analyzed using a CytoFLEX flow cytometer (Beckman Coulter, USA) equipped with blue (488 nm) and yellow (561 nm) lasers. For each sample, data were collected for at least 500 000 bacterial cells, gated based on green fluorescent protein (GFP) and mCherry fluorescence. Gating involved three sequential steps: Gate 1 discriminated bacterial cells based on size using side scatter (SSC-A) and volume scatter (VSSC PB450-A); Gate 2 isolated singlet cells using high forward scatter (VSSC PB450-H) and volume scatter (VSSC PB450-A); and Gate 3 identified green-fluorescing cells (lower left quadrant) and red-fluorescing cells (upper right quadrant). Transconjugants were quantified by counting green-fluorescing cells. The number of recipient cells was calculated as the total bacterial count minus the number of mCherry-positive cells. Flow cytometry data were analyzed using CytExpert SRT software (Beckman Coulter), and fluorescent populations were compared using one-way analysis of variance (ANOVA).
2.7. 16S ribosomal RNA (rRNA) sequencing and analysis
16S rRNA sequencing was performed following established protocols from a previous study [
26]. Sequences that met quality criteria and shared ≥ 97% similarity were clustered into a single operational taxonomic unit (OTU) using USEARCH (v10.0). Taxonomic annotation of operational taxonomic units (OTUs)/amplicon sequence variants (ASVs) was conducted using the Naive Bayes classifier in QIIME2 [
29], with the SILVA database (release 138.1) [
30] as the reference and a confidence threshold of 70%. The Shannon index was calculated to assess species diversity within each sample. Beta diversity was evaluated by principal coordinate analysis (PCoA) to examine inter-sample diversity. One-way analysis of variance was used to compare bacterial abundance and diversity.
2.8. In-vivo conjugation experiments
All mice were pretreated with streptomycin (5 g∙L
−1) in drinking water for four days to eliminate colonization resistance. Bacterial populations were assessed using eosin methylene blue (EMB) agar to evaluate colonization resistance [
31]. Mice were then administered a mixture of donor
S. typhimurium 14E1050 pIncI2-
mcr-1 (or
E. coli MG1655 harboring pRP4) and recipient
E. coli MG1655S at a concentration of 10
8 CFU∙mL
−1. After 0.5 h, each mouse received 0.2 mL of CA at 10, 100, or 200 mg∙kg
−1∙d
−1 (
n = 8 per group). The control group received 0.5% sodium carboxymethyl cellulose (CMC-Na) as the solvent. Fecal samples were collected, homogenized by vortexing, and plated on selective EMB agar containing 2 mg∙L
−1 CS for donors, 2 g∙L
−1 Str for recipients, or both CS and Str for transconjugants. Conjugation frequency was calculated as the number of transconjugants divided by the total number of recipients.
2.9. RNA extraction and real-time polymerase chain reaction (PCR)
E. coli MG1655/pRP4 was grown to the stationary phase and resuspended in PBS to an OD600 of 0.5. CA was then added to the donor cells at final concentrations of 0, 50, or 200 mg∙L−1, and the cultures were incubated for 4 h. Total RNA was extracted using the Easy RNA Extraction Kit (Omega Bio-Tek, USA), and reverse transcribed into complementary DNA (cDNA) using the HiScript II Q RT SuperMix for qPCR (Vazyme, China). Gene expression levels were quantified by real-time PCR using primers listed in Table S1 in Appendix A.
2.10. Measurement of permeability of cell membrane
N-Phenyl-1-naphthylamine (NPN) fluorescent dye was used at a final concentration of 10 μmol∙L
−1 to evaluate cell membrane permeability [
32]. The donor strain was treated with CA as described above. Fluorescence was measured using a microplate reader (excitation: 350 nm; emission: 420 nm). Each experiment included three biological and three technical replicates.
2.11. RNA sequencing and bioinformatics analysis
Overnight cultures of
E. coli MG1655/pRP4 were diluted 1:100 into fresh LB broth and incubated at 37 °C for 4 h. The cultures were then adjusted to 10
8 CFU∙mL
−1 in PBS supplemented with 0, 50, or 200 mg∙L
−1 CA and incubated for 4 h before rapid freezing in liquid nitrogen. Samples were sent to Novogene (China) for RNA sequencing. Gene expression was quantified using the fragments per kilobase of transcript per million mapped reads (FPKM) method. Differential expression analysis was performed using DESeq2 and R packages [
33]. Functional annotation was carried out using the KEGG database.
2.12. Measurement of intracellular adenosine triphosphate (ATP) and nicotinamide adenine dinucleotide (NAD+)/reduced nicotinamide adenine dinucleotide (NADH)
Intracellular ATP and NAD
+/NADH levels in donor bacteria treated with varying CA concentrations (0, 50, and 200 mg∙L
−1) were quantified using the ATP Assay Kit and NAD
+/NADH Assay Kit (Beyotime, China), following the manufacturer’s protocols [
34]. Donor cultures were grown overnight, diluted 1:100 in fresh LB medium, and incubated at 37 °C for 3 h. Bacterial cells were harvested by centrifugation (5 500 r∙min
−1, 5 min, 37 °C), resuspended in PBS to an OD
600 of 0.5, and treated with CA for 4 h. After treatment, cells were centrifuged again (12 000 r∙min
−1, 5 min, 4 °C), and assays were performed as per the manufacturer's instructions. Measurements were taken using a multimode reader (Model 1600801W; PerkinElmer, USA). All experiments were conducted in triplicate with three biological replicates.
2.13. Iodonitrotetrazolium chloride (INT) reduction assay
INT is reduced to insoluble formazan (INF) by components of the bacterial respiratory chain, making it a useful marker for assessing bacterial respiratory activity [
34]. In this study, an overnight donor culture was diluted 1:100 in fresh LB medium and incubated at 37 °C for 4 h. Bacterial cells were then harvested by centrifugation at 5 500 r∙min
−1 for 5 min at 28 °C and resuspended in PBS to an OD
600 of 0.3 (10
8 CFU∙mL
−1), then kept on ice. CA (0, 50, or 200 mg∙L
−1), 1 mmol∙L
−1 INT, and 1 mL of the bacterial suspension were mixed and incubated for 4 or 12 h. After incubation, 200 µL of the mixture was transferred to a 96-well plate, and absorbance at 490 nm was measured using a multimode reader (Model 1600801W, PerkinElmer). All experiments included three biological replicates.
2.14. Measurement of membrane potential (ΔΨ) and transmembrane proton gradient (ΔpH)
To assess the effects of CA on the proton motive force (PMF)—comprising Δ
Ψ and ΔpH—
E. coli MG1655/pRP4 was treated with CA, and the fluorescent dyes 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF-AM) and 3,3-dipropylthiadicarbocyanine iodide (DiSC
3(5)) were used to measure ΔpH and Δ
Ψ, respectively, following established protocols [
35]. Overnight bacterial cultures were diluted 1:100 in fresh LB medium, incubated at 37 °C for 4 h, and treated with CA (0, 50, or 200 mg∙L
−1). Cells were then harvested at 5 500 r∙min
−1 for 5 min at room temperature and resuspended in PBS to an OD
600 of 0.5. The suspension was incubated with BCECF-AM (20 µmol∙L
−1) and DiSC
3(5) (1 µmol∙L
−1) and transferred to a 96-well plate for equilibration. Each condition was tested in triplicate.
2.15. Swarming motility assay
Swarming motility assays were following a previously described method [
36]. LB broth plates containing 0.5% agar were supplemented with CA at concentrations of 0, 50, or 200 mg∙L
−1. A 10 µL aliquot of donor bacteria (adjusted to an OD
600 of 0.3) was inoculated at the center of each plate. After incubation at 37 °C for 48 h, inhibition zones were photographed and measured.
2.16. In-vivo biocompatibility assay
To evaluate the biocompatibility of CA, mice were orally administered 0.2 mL of CA at a dose of 200 mg∙kg
−1, the highest concentration used in the animal experiments. Mortality was monitored daily for one week. Mice were randomly assigned to two groups (
n = 6 per group) and treated via oral gavage for one week: the CA group (200 mg∙kg
−1) and the control group, which received 0.5% CMC-Na. Body weight and general health status were recorded daily throughout the study. At the end of the treatment period, the mice were euthanized, and liver, spleen, colon, kidney, stomach, and cecum tissues were collected and fixed overnight in 4% PBS-buffered paraformaldehyde. The tissues were then embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) for histological analysis. Stained sections were examined using a Nikon TE2000U optical microscope (Japan) [
31]. Stool samples were collected at the end of the treatment period for microbiota analysis. DNA was extracted, and 16S rRNA sequencing was conducted to assess the composition and diversity of the gut microbiota.
2.17. Statistical analysis
Statistical analysis was performed using GraphPad Prism (v9.1.1). All data were obtained from at least three independent experiments and presented as means ± standard deviation (SD). p-values were obtained by unpaired t-test, one-way ANOVA, or Mann-Whitney U test. Significance levels were determined at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).
3. Results and discussion
3.1. CA inhibits the conjugation of various Inc-type plasmids
Since its identification in the 1970s, the pRP4 plasmid has been a model for studying plasmid transfer [
37]. Plasmids from the IncI2, IncX4, and IncHI2 incompatibility groups account for over 90% of
mcr-1-harboring plasmids, often carrying multiple resistance genes for aminoglycosides, β-lactams, fluoroquinolones, sulfonamides, Tcs, and trimethoprim [
38]. IncFII plasmids are particularly prevalent and drive the global spread of antimicrobial resistance [
39], underscoring the need for effective clinical control measures. Inc-type plasmids are also commonly used to study conjugation inhibitors [
40,
41]. In this study, plasmids representing the IncP (pRP4), IncI2 (p14E1050), IncHI2 (pLS3479), IncX4 (pCSZ4), and IncFII (pS110) incompatibility groups were selected to examine the impact of CA on plasmid conjugation.
The conjugation frequency of the pRP4 plasmid was first evaluated at different CA concentrations (0, 50, 100, and 200 mg∙L−1) to determine its effect on donor and recipient strain growth. As shown in Table S2 and Fig. S1 in Appendix A, all CA concentrations used in the conjugation experiments were below the MIC, and none had a significant impact on the growth of either donor or recipient strains.
The baseline transfer frequency of the pRP4 plasmid was approximately 4.5 × 10
−3 transconjugants per recipient cell (Tc/R). CA significantly reduced this frequency in a concentration-dependent manner: by 13.3-fold at 50 mg∙L
−1 (
p < 0.0001), 20.6-fold at 100 mg∙L
−1 (
p < 0.0001), and 24.5-fold at 200 mg∙L
−1 (
p < 0.0001), compared to the control. CA also exhibited a concentration-dependent inhibitory effect on the conjugation frequencies of various clinically relevant resistance plasmids (
p < 0.05; Figs. 1(a)-(e)). For the IncI2 plasmid, the baseline conjugation frequency was 0.34, consistent with previously reported [
6,
42]. CA reduced the frequency by 3.6-fold at 50 mg∙L
−1 (
p = 0.105), 24.0-fold at 100 mg∙L
−1 (
p < 0.05), and 661.6-fold at 200 mg∙L
−1 (
p < 0.05). Similarly, 200 mg∙L
−1 CA significantly inhibited the conjugation frequencies of IncX4, IncHI2, and IncFII plasmids by approximately 238.5-fold (
p < 0.05), 23.1-fold (
p < 0.0001), and 199.0-fold (
p < 0.001), respectively.
Previous studies have shown that dihydroartemisinin (DHA) inhibited the conjugation of IncI2 and IncX4 plasmids by 180- and 160-fold, respectively, at a concentration of 200 mg∙L
−1 [
34]. However, CA exhibited even greater efficacy under the same conditions, reducing conjugation by 661.6- and 238.4-fold, respectively, indicating substantially higher inhibitory activity compared to DHA. These findings demonstrate that CA effectively inhibits the conjugation of various Inc-type plasmids
in vitro (
Fig. 1(f)). However, this study is limited to a subset of Inc-type plasmids. Further research is needed to evaluate the efficacy of CA against a broader range of plasmid types to validate its potential as a conjugation inhibitor.
3.2. CA inhibits plasmid conjugation within the intestinal microbiome
The horizontal transfer of plasmid-encoded ARGs between commensal and pathogenic bacteria in the GI tract is well established, highlighting the need to limit their dissemination within the gut microbiome [
43]. To date, the inhibitory effects of CA on ARG transfer within the gut microbiome have not been comprehensively investigated. To address this, we developed an
ex vivo model using microbial communities isolated from mouse feces, which were subsequently co-incubated with a donor strain under laboratory conditions for a conjugation assay [
44].
The donor strain, MQ-RP4-
sfGFP, harbors the pRP4 plasmid carrying the
sfGFP gene and a non-transferable ColE1 plasmid (pMQ). The pMQ encodes the mCherry fluorescent protein and LacIq, which represses the
sfGFP promoter (pTrc) [
26,
45,
46]. Consequently, the donor strain fluoresces red (RED
+), while transconjugants fluoresce green (GFP
+). This fluorescence-based reporter system allows for clear identification and quantification of transconjugants via flow cytometry or confocal microscopy (Figs. 2(a)-(c)).
The spontaneous conjugation frequency of the pRP4 plasmid from the donor to fecal bacteria was 8.58 × 10
−4, similar to the transfer frequency observed in soil bacteria but significantly lower than that in activated sludge [
47,
48]. The proportion of GFP
+ fluorescent cells significantly decreased in a concentration-dependent manner following exposure to CA. Specifically, 50 and 200 mg∙L
−1 CA reduced the proportion by 2.33-fold (
p < 0.01) and 4.69-fold (
p < 0.01), respectively, compared to the baseline (
Fig. 2(d)). To verify that CA did not affect sfGFP expression or protein folding, we exposed
sfGFP-carrying
E. coli to 50 and 200 mg∙L
−1 CA for 4 h. Flow cytometry analysis showed no significant differences in GFP fluorescence intensity between the groups (
p > 0.05; Fig. S6 in Appendix A). These results suggest that CA does not directly impact sfGFP expression in transconjugants but inhibits plasmid DNA transfer within gut bacterial communities, thereby reducing the proportion of GFP
+ cells.
Given CA’s ability to block conjugation in the gut microbiome, we investigated which commensal bacterial genera are most affected. We performed cell sorting and 16S rRNA gene amplicon sequencing of transconjugant communities, comparing samples treated with and without CA. The pRP4 plasmid was transferable to seven genera across the transconjugant communities:
Enterobacter,
Bacillus,
Enterococcus,
Lysinibacillus,
Escherichia-Shigella,
Paucibacter, and
Caenimonas (
Fig. 2(e)). These findings are consistent with previous studies [
49,
50].
Among these genera,
Enterobacter emerged as the predominant recipient, indicating a higher degree of plasmid permissiveness among closely related phylogenetic species [
50,
51]. Both CA-treated and control samples shared the same genera—
Enterobacter,
Bacillus,
Enterococcus,
Lysinibacillus,
Escherichia-Shigella,
Paucibacter, and
Caenimonas (Fig. S7in Appendix A). However, their relative abundances differed. For example,
Bacillus represented 35% of OTUs in control samples but decreased to 20% and 31% in samples treated with 50 and 200 mg∙L
−1 CA, respectively. Genera within Firmicutes, such as
Bacillus and
Enterococcus, accounted for 32%-37% of control samples and 13%-43% of CA-treated transconjugants.
Caenimonas, a conditional pathogen, represented 0.0014% of OTUs in the control group but was absent in the 200 mg∙L
−1 CA group (Fig. S7). These findings suggest that CA effectively inhibits the spread of the pRP4 plasmid within the gut microbiome, particularly by targeting the conditional pathogen
Caenimonas.
3.3. CA inhibits plasmid conjugation in vivo
To assess the effects of CA on plasmid conjugation within a mammalian host, the GI tract was selected as the site for bacterial interaction. Colonization resistance limits the establishment of both donor and recipient strains and the subsequent conjugation process [
52,
53]. Previous studies have shown that administering 5 g∙L
−1 streptomycin disrupts colonization resistance, facilitating stable conjugation [
52,
54]. Therefore, streptomycin was used to overcome this barrier and evaluate CA’s impact on
in vivo conjugation. As expected, administering 5 g∙L
−1 streptomycin significantly reduced endogenous enterobacterial populations in feces by more than 10 000-fold (
p < 0.05; Fig. S8 in Appendix A).
The IncP (pRP4) plasmid is widely used as a standard model for studying plasmid conjugative transfer
in vitro [[
55], [
56], [
57]]. In our previous research, the
mcr-1-harboring IncI2 (p14E1050) plasmid in the clinical isolate
S. typhimurium 14E1050 demonstrated a higher conjugation frequency, posing a significant clinical risk [
26]. Therefore,
E. coli MG1655/pRP4 and
S. typhimurium 14E1050/p14E1050 were selected as primary donor for experimental studies. Conjugation events were assessed by quantifying CFU∙g
−1 of feces collected daily over a two-week period (
Fig. 3(a)). Colonization levels of recipient strains (
E. coli MG1655S) were similar across groups (
Fig. 3(b)). However, compared to
E. coli MG1655/pRP4,
S. typhimurium 14E1050/p14E1050 exhibited higher colonization levels and a longer duration (
Fig. 3(c)). Additionally, the IncI2 plasmid p14E1050 demonstrated a significantly higher conjugation frequency than the pRP4 plasmid
in vivo (
Fig. 3(d)).
Consistent with our findings, previous studies have reported similar
in vivo inhibitory effects of conjugation inhibitors using IncI2 plasmids as models [
34,
58]. Given its well-documented higher conjugative activity in the mammalian gut,
S. typhimurium 14E1050/p14E1050 was chosen for further investigation. A schematic of the animal experiment is shown in
Fig. 3(e).
The spontaneous conjugation frequency was determined to be 1.3, aligning with previously reported values [
59]. Upon exposure to varying CA concentrations (0, 10, 100, and 200 mg∙kg
−1), a dose-dependent reduction in conjugation frequency was observed, aligning with results from
in vitro and
ex vivo studies. Specifically, 10 mg∙kg
−1 CA reduced the conjugation frequency of the IncI2 plasmid by twofold
in vivo, while 200 mg∙kg
−1 CA resulted in an eightfold reduction (0.1572,
p < 0.05; Figs. 3(f)-(h)). A previous study reported that treatment with chelerythrine or DHA led to a two-fold reduction in conjugation frequency, suggesting that CA exhibits comparable or superior inhibitory effects in
in vivo models [
34,
58]. Therefore, these findings demonstrate that CA significantly inhibits plasmid conjugation across
in vitro,
ex vivo, and
in vivo conditions, thereby highlighting its potential as a promising strategy to mitigate the spread of antimicrobial ARGs and enhance the effectiveness of antibiotic therapies.
3.4. CA suppresses the expression of conjugation-related genes
To investigate how CA influences conjugation, we conducted a pre-exposure experiment. The results indicated that pre-exposure of recipient bacteria to CA did not significantly affect conjugation frequency. In contrast, pre-exposure of donor bacteria led to a significant reduction in conjugation frequency (
p < 0.01). Interestingly, when both donor and recipient bacteria were treated with CA, the suppressive effect was less pronounced than with donor exposure alone (
Fig. 4(a)). These findings suggest that the inhibitory effect of CA is primarily associated with the donor bacteria (
p < 0.01,
Fig. 4(a)). To elucidate the molecular mechanisms underlying this inhibition, we examined factors that may impair conjugation in donor cells.
A reduction in plasmid copy number within donor bacteria can hinder the process of conjugation [
59]. Therefore, we quantified the pRP4 copy number using real-time PCR. As shown in
Fig. 4(b), CA treatment did not significantly alter plasmid copy numbers.
Bacterial outer membrane proteins (OMPs) are essential for the transport of solutes, peptides, proteins, and nucleic acids, and they play a critical role in conjugation [[
60], [
61], [
62], [
63], [
64]]. Specifically, OmpA, OmpC, and OmpF contribute to pore formation, membrane transport, and the establishment of effective mating pairs during conjugation [[
65], [
66], [
67]]. Among these OMPs, only
OmpA expression was significantly reduced (by 1.5- to 2.0-fold,
p < 0.05) in a dose-dependent manner following CA exposure (
Fig. 4(c)).
To further assess the effect of CA on membrane integrity, we used the NPN fluorescence assay, which measures membrane permeability by detecting fluorescence upon interaction with hydrophobic regions of the phospholipid bilayer [
68]. CA treatment increased membrane permeability by 1.81-fold at 200 mg∙L
−1 (
p < 0.0001,
Fig. 4(d)). Similarly, recipient bacteria exposed to CA exhibited significantly increased outer membrane permeability (
p < 0.0001, Fig. S9 in Appendix A). Increased membrane permeability is known to facilitate plasmid uptake by recipient cells [
35], potentially explaining the partially reduced inhibitory effect observed when both donor and recipient bacteria were treated with CA.
Conjugation requires the formation of mating bridges between donor and recipient cells, a process mediated by the mating pair formation (Mpf) and DNA transfer and replication (Dtr) system genes [
32,
69]. To assess whether CA affects these systems, we examined the expression of key Mpf genes (
trbBp and
traL) and Dtr genes (
trfAp,
trfA, and
traJ). Exposure to 200 mg∙L
−1 CA significantly downregulated all tested genes (
Fig. 4(e)). Specifically,
trbBp and
traL were downregulated by 3.33-fold (
p < 0.05) and 2.70-fold (
p < 0.05), respectively, indicating impaired formation of mating bridges and channels.
Dtr genes facilitate plasmid transfer by regulating relaxase activity and initiating replication [
35,
70], which encodes the traJ protein and promotes transcription of additional
tra (transfer) genes [
71]. Following treatment with 200 mg∙L
−1 CA,
traJ expression decreased by 2.5-fold (
Fig. 4(e)). Collectively, these findings collectively suggest that CA disrupts the function of the pRP4-encoded transfer machinery, providing a mechanistic explanation for its inhibitory effect on conjugation.
3.5. CA limits energy supply for conjugation
To investigate the underlying causes of reduced conjugation-related gene expression following CA exposure, we conducted a comprehensive transcriptomic analysis of donor bacteria treated with high-dose (200 mg∙L−1) and low-dose (50 mg∙L−1) CA. Differentially expressed genes (DEGs) were identified by pairwise transcriptome comparisons between CA-treated and control samples.
As shown in Figs. S10 in Appendix A, relative to the control, 223 genes (136 upregulated and 87 downregulated) and 1299 genes (646 upregulated and 653 downregulated) were significantly differentially expressed (|log2(fold change)| > 1 and p-adjust (padj) < 0.05) in donors treated with 50 and 200 mg∙L−1 CA, respectively. This indicates a dose-dependent increase in the number of DEGs following CA exposure. These DEGs were subsequently analyzed through Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis.
A total of 22 dominant functions were identified based on expression patterns in the 50 mg∙L−1 CA group, compared to 31 functions in the 200 mg∙L−1 CA group. Relative to the control group, the top 20 enriched KEGG signaling pathways in the 50 mg∙L−1 CA group included geraniol degradation, degradation of aromatic compounds, phenylalanine metabolism, fatty acid degradation, valine, leucine, and isoleucine degradation, tyrosine metabolism, fatty acid metabolism, arginine and proline metabolism, oxidative phosphorylation, butanoate metabolism, the tricarboxylic acid cycle (TCA) cycle, alanine, aspartate and glutamate metabolism, microbial metabolism in diverse environments, pyruvate metabolism, carbon metabolism, biosynthesis of antibiotics, biosynthesis of secondary metabolites, and general metabolic pathways.
Fatty acid degradation, tyrosine metabolism, and branched-chain amino acid degradation (valine, leucine, and isoleucine), along with tryptophan metabolism, are fundamental to core metabolic processes [
72]. Altered intracellular pools of amino acids such as tryptophan and methionine play critical roles in regulating the conjugative transfer apparatus [
73]. These findings suggest that CA may influence plasmid conjugation by modulating central metabolic pathways. Among these, the TCA cycle is particularly important, as it drives both energy production and amino acid biosynthesis [
74]. As shown in
Figs. 5(a) and
(b), twenty metabolic pathways—including the TCA cycle—were commonly affected in both the low-dose and high-dose CA groups.
KEGG enrichment analysis revealed that several regulated genes were associated with the TCA cycle, a key pathway responsible for generating reducing equivalents [
75]. The expression levels of TCA cycle-related genes such as
aacA,
aacD,
ackA, and
eno were significantly downregulated (Fig. S11 in Appendix A), indicating a potential disruption in cellular respiration. The TCA cycle generates NADH, which is subsequently utilized for ATP synthesis through the PMF and the electron transport chain (ETC) [[
76], [
77], [
78]]. Consequently, the NAD
+/NADH ratios were measured (
Fig. 5(c)), revealing a marked decrease following CA exposure, further suggesting TCA cycle disruption. Additionally, the expression of ETC-associated genes (
nuoE,
cydAB, and
atpABEF) was significantly downregulated in response to CA treatment (Fig. S11).
The ETC activity was assessed using INT reduction to INF as a marker, reflecting respiratory chain dehydrogenase activity [
34]. After 4 h of exposure to CA, ETC activity in the donor strain showed a slight, non-significant decrease compared to the control (
p > 0.05). However, extending exposure to 12 h resulted in a significant reduction in ETC activity, decreasing to 91.15% at 50 mg∙L
−1 (
p < 0.05) and to 84.43% at 200 mg∙L
−1 (
p < 0.01) (Fig. S12 in Appendix A).
PMF regulation involves complex compensatory mechanisms, including adjustments in intracellular pH and membrane potential (Δ
Ψ) [
76,
77]. Fluorescent probes BCECF-AM and DiSC
3(5) were used to assess PMF activity, following previously described methods [
34]. In the CA-treated group, BCECF-AM fluorescence consistently decreased compared to the control (
p < 0.0001;
Figs. 5(d) and
(e)), indicating a reduction in intracellular pH. Conversely, DiSC
3(5) fluorescence increased in a concentration-dependent manner (
p < 0.0001;
Figs. 5(f) and
(g)), suggesting changes in membrane potential. Together, these findings indicate that CA primarily disrupts ETC activity and PMF.
Conjugation is an ATP-dependent and energetically demanding process [[
78], [
79], [
80]]. As shown in
Fig. 5(h), ATP levels declined in a dose-dependent manner after 4 h of CA treatment. ATP is also critical for bacterial motility, including flagellar movement. Notably, CA inhibited the expression of cell adhesion genes (
dgcZ,
uspF,
dgcP, and
znuA) and flagellum assembly genes (
flgN and
ybjN) (Fig. S13 in Appendix A). Consequently, CA significantly impaired donor swarming motility (
Figs. 5(i) and
(j)). These results suggest that CA disrupts the TCA cycle, leading to PMF perturbation, reduced intracellular ATP levels, and decreased plasmid conjugation efficiency.
Consistent with these findings, previous studies have shown that PMF inhibitors such as DHA, melatonin, and antimicrobial peptides can inhibit conjugation [
34,
77,
78,
81]. Conversely, compounds targeting the type IV traffic adenosine triphosphatase (ATPase) TrwD (e.g., unsaturated fatty acids) or the TraE protein (e.g., 105055 and 239852) exhibit specificity toward certain Inc plasmids [
82], thereby limiting their broad-spectrum utility against antibiotic resistance. Additionally, Domenech et al. reported that PMF inhibitors such as triclosan, hydrochloride, and pimozide can also suppress transformation, another HGT mechanism of HGT [
83]. Therefore, PMF inhibitors like CA may offer broader-spectrum inhibition of HGT. Given the central role of energy metabolism in bacterial physiology, these disruptions may also increase bacterial susceptibility to antibiotics [
84], which warrants further investigation.
3.6. CA maintains optimal biosafety in vivo
Conjugation inhibitors function primarily as preventive agents rather than direct therapeutics and must be administered via the intestinal route in mammals to effectively target plasmid conjugation [
31,
85]. Therefore, a favorable biosafety profile is essential for their use through the digestive tract (
Fig. 6(a)). Although various conjugation inhibitors have shown efficacy, safety data remain limited. To address this, we conducted an
in vivo study to assess the biosafety of CA.
After seven days of intragastric administration of 200 mg∙kg
−1 CA (high dose, HCA), all mice in both the control and HCA groups survived without observable adverse effects. No significant differences in body weight were noted (
Fig. 6(b)). Given the tendency of environmental pollutants to accumulate in organs [
86], we performed histopathological analyses of the liver, spleen, kidney, stomach, colon, and cecum using H&E staining. No tissue damage or inflammation was observed in CA-treated mice (
Fig. 6(c)).
A healthy gut microbiota is essential for overall health [
87]. To assess the impact of CA on gut microbiota composition, we conducted 16S rRNA gene sequencing of fecal samples. Phylum-level analysis identified 21 phyla across all samples (
Fig. 6(d)). α-Diversity (Shannon index;
Fig. 6(e)) and β-diversity (principal component analysis [
88]) demonstrated no significant differences between control and HCA-treated groups (
Fig. 6(f)). The relative abundance of the ten dominant phyla—Desulfobacterota, Acidobacteriota, Chloroflexi, Actinobacteriota, Proteobacteria, Bacteroidota, Firmicutes, Bdellovibrionota, Campylobacterota, and Patescibacteria—also remained unchanged (
Figs. 6(g)-(p)). These results support the favorable biosafety profile of CA and its potential for
in vivo use as a conjugation inhibitor.
Consistent with our results, prior studies by Lee et al. [
89] reported that 200 mg∙kg
−1 of CA did not induce hepatic or renal toxicity in male C57BL/6 mice. Similarly, Liu et al. [
90] demonstrated low cytotoxicity of CA using an NRK-52E cell model. The compound’s widespread use in the food, health, cosmetic, and pharmaceutical industries—along with the absence of significant toxic side effects [
91]—further supports its favorable biosafety profile and underscores its potential for
in vivo applications as a safe conjugation inhibitor.
However, a limitation of this study is the lack of identification of CA’s active site. Given that CA serves as a key scaffold for the synthesis of various derivatives [
92], exploring its structural features may offer a promising approach to developing more effective conjugation inhibitors. This aspect warrants further exploration in future studies.
4. Conclusions
In summary, this study demonstrates that the food additive CA acts as a safe and effective broad-spectrum inhibitor of plasmid-mediated conjugation. This inhibition occurs through downregulation of the TCA cycle and ETC activity, disruption of the PMF, and a reduction in intracellular ATP levels, ultimately leading to a decreased frequency of plasmid transfer. These findings suggest that CA holds promise as a strategic intervention to curb the spread of ARGs and address the growing global threat of antibiotic-resistant infections.
CRediT authorship contribution statement
Gong Li: Conceptualization, Writing - review & editing, Supervision. Ang Gao: Writing - original draft, Visualization, Formal analysis, Data curation. Xin-Yi Lu: Data curation. Tian-Hong Zhou: Methodology, Data curation. Shi-Ying Zhou: Formal analysis, Data curation. Li-Juan Xia: Data curation. Lei Wan: Data curation. Yu-Zhang He: Data curation. Xin-Yi Chen: Data curation. Wen-Ying Guo: Software. Jia-Min Zheng: Methodology. Hao Ren: Supervision. Sheng-Qiu Tang: Supervision. Xiao-Ping Liao: Supervision, Resources. Liang Chen: Writing - review & editing, Supervision. Jian Sun: Writing - review & editing, Supervision, Project administration, Data curation, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was supported by the National key research and development program of China (2022YFD1800400), the National Natural Science Foundation of China (32402943 and 32273066), the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (32121004), the Guangdong Provincial Natural Science Foundation (2025A1515012412), the Local Innovative and Research Teams Project of Guangdong Pearl River Talents Program (2019BT02N054), the Guangdong Major Project of Basic and Applied Basic Research (2020B0301030007), the 111 Center (D20008), and the Specific university discipline construction project (2023B10564003).
Ethics approval
Animal experimentation was approved by the ethics Committee of the Laboratory Animal Center of South China Agricultural University (Guangzhou, China), approval number 2023C082.
Data availability statement
The high-throughput 16S RNA gene sequence data have been submitted to the NCBI under accession number PRJNA1178588 and PRJNA1177220 and accessible at the following link:
https://www.ncbi.nlm.nih.gov/search/all/?term=+PRJNA1178588 and
https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1177220/. The raw sequence reads and transcriptome sequencing data from this paper have been submitted to the NCBI under accession number PRJNA1118489 and are accessible at the following
https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA1118489.
Appendix A. Supplementary material
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2025.06.040.
PDF
(4265KB)
