Artificial Sweeteners Induce Bacterial Drug Resistance and Modulate Gene Expression

Introduction

Artificial sweeteners are a class of food additives that are produced by humans using chemical methods. Because they are high in sweetness and mostly non-caloric, they are considered an alternative to sugar in a wide variety of foods.¹ Currently, natural sweeteners fall far short of the demand for their use because of cost and commercialization, and artificial sweeteners continue to be major sweeteners in the food industry.² Its consumption has increased significantly over the past few years and its demand in the global consumer market is very strong.^3,4 Artificial sweeteners were once considered the best option for obese people and those with diabetes, and some children’s foods have started to promote artificial sweeteners over natural ones.⁵ Non-caloric artificial sweeteners have long been assumed to be metabolically inert, similar to natural sweeteners, with no significant side effects.⁶ However, a growing body of evidence suggests that long-term consumption of artificial sweeteners may affect the immune system, leading to an increased risk of type 2 diabetes, cardiovascular disease, and even cancer.^7–10 Therefore, whether artificial sweeteners are safe alternatives to natural sweeteners requires further study.

Approximately 700,000 people worldwide die annually from infections caused by drug-resistant bacteria. It is estimated that ten million people will die from AMR infections by 2050, if no action is taken now.¹¹ The emergence and spread of antibiotic resistance genes (ARGs) are usually attributed to the misuse or overuse of antibiotics.¹² These resistance genes can also be transferred and expressed in the gut microbiome, further increasing the risk of drug resistance.¹³ Some commonly used artificial sweeteners such as saccharin and acesulfame potassium also contain an antibiotic-like sulfonyl group, which is a functional group of sulfonamides. Artificial sweeteners are more likely to cause oxidative stress than natural sweeteners, which typically leads to bacterial resistance to antibiotics.^14,15 Overall, there has not been much research on the effects of artificial sweeteners on bacterial resistance and their associated genes, especially on the differences between artificial and natural sweeteners at daily doses.

Recent evidence indicates that artificial sweeteners can reshape gut microbial composition and function through mechanisms beyond simple metabolic inertness, potentially creating selective pressure for antimicrobial resistance (AMR) evolution.¹⁶ The gut microbiome represents a crucial interface where sweetener-induced metabolic shifts may drive bacterial adaptation strategies, including enhanced efflux pump activity and oxidative stress tolerance, thereby facilitating AMR development.^16,17 Therefore, we selected five commonly used artificial sweeteners (saccharin (SAC), cyclamate (CYC), sucralose (SUC), aspartame (ASP), and acesulfame potassium (ACE-K)) and two strains (E. coli and B. subtilis) for validation, covering both gram-negative and gram-positive bacteria. We first investigated whether each artificial sweetener had a significant antibacterial effect and changes in bacterial susceptibility to antibiotics and ROS production before and after the addition of artificial sweeteners. Bacterial models were subsequently constructed and transcriptome analysis was performed to investigate differences in the expression of genes associated with resistance, oxidative stress, and virulence in bacteria treated with different artificial sweeteners and glucose (GLC). This article also discusses the safety risks of using artificial sweeteners in daily life as substitutes for natural sweeteners.

Materials and Methods

Bacterial Strains and Chemicals

Gram-negative E. coli ATCC 25922 and Gram-positive B. subtilis ATCC 6633 were used in this study. E. coli and B. subtilis are representative commensal bacteria in the gut. Both E. coli and B. subtilis were susceptible to ampicillin (Amp). Five analytical chemicals of artificial sweeteners (SAC, CYC, SUC, ASP, and ACE-K) were ordered from Altascientific Technology Co., Ltd. (Tianjin, China) and dissolved in Milli-Q water as stock solutions for further use. Antibiotics were purchased from the National Institutes for Food and Drug Control (Beijing, China), dissolved in Milli-Q water at an initial concentration of 1024 μg/mL, and filtered before use.

Growth Curve Determination

A single bacterial colony of E. coli or B. subtilis was inoculated aseptically into LB broth supplemented with artificial sweeteners and GLC at various concentrations (SAC, 2500 mg/L; CYC, 2000 mg/L; ASP, 5000 mg/L; ACE-K, 5000 mg/L; SUC, 1800 mg/L; GLC, 5000 mg/L), and allowed to grow for up to 48 h. Growth was recorded as absorbance at 600 nm (OD600) using a Multimode plate reader.

Minimum Inhibitory Concentrations Determination

According to the Clinical and Laboratory Standards Institute (CLSI, 2019) standard, the MIC is defined as the minimum concentration of the antibiotic ampicillin (Amp) and sweetener (CYC, SAC, SUC, ASP, and ACE-K) at which no bacteria (E. coli and B. subtilis) are visibly growing. Antibiotics were prepared using 2-fold serial dilutions in LB broth, starting from 1024 μg/mL down to 0.5 μg/mL, following CLSI 2019 guidelines. For each artificial sweetener, 2-fold serial dilutions were similarly prepared in LB broth, with the CAC-defined maximum concentration serving as the starting point. Initially, each bacterial strain was incubated overnight in LB medium and washed thrice with LB broth. The cell pellets were resuspended in fresh LB medium and uploaded to a 96-well plate at approximately 10⁵ CFU/mL. Each well consisted of a total volume of 200 μL containing various concentrations of artificial sweeteners or antibiotics. The plate was incubated overnight at 37 °C and was scanned using a multimode plate reader (Molecular Devices, USA) at a wavelength of 600 nm (OD600). Each trial was repeated three times. LB culture medium with antibiotics was used as a positive control and culture medium without antibiotics (LB broth-only wells) was used as a negative control in all experiments.

Measurement of ROS Production

To determine whether artificial sweeteners influence oxidative stress, intracellular ROS concentrations were measured using a 2′,7′-dichlorofluorescein diacetate (DCFH-DA) cellular ROS detection assay kit (Solarbio, Beijing, China). Initially, the cell suspensions (initial concentration of approximately 10⁶ CFU/mL in PBS solution) were incubated with DCFH-DA at 10 μM for 30 min at 37 °C in the dark. Sweeteners were added and incubated at room temperature in the dark for 2 h. The suspensions were analyzed using a fluorescence microplate (Molecular Devices, USA) at excitation and emission wavelengths of 488 and 525 nm, respectively. Both positive (3% hydrogen peroxide, final concentration) and negative (Milli-Q water) suspensions were used as controls for the ROS analysis.

Transcriptome Sequencing

Both bacterial strains were cultured at 10⁶ CFU/mL in the LB medium. They were then treated with artificial sweeteners at CAC-defined concentrations (SAC: 2500 mg/L; CYC: 2000 mg/L; ASP:5000 mg/L; ACE-K:5000 mg/L; SUC:1800 mg/L; GLC: 5000 mg/L) and incubated at 37 °C for 2 h (the exposure time selected to capture early transcriptional stress responses before growth phase alterations). Samples were prepared with three biological replicates per condition. Bacterial cells were collected by centrifugation (8000 × g, 3 min), rapidly frozen in liquid nitrogen, and stored at −80°C. RNA quality was assessed using Agilent 2100 Bioanalyzer (RIN > 7.0), and libraries were constructed using the TruSeq RNA Sample Preparation Kit. Sequencing was performed on an Illumina NovaSeq platform with paired-end 150 bp reads. After identification by 16S rRNA sequencing, samples were subjected to RNA-seq analysis. Library construction was performed once sequencing requirements were met. An Agilent 2100 Bioanalyzer was used to measure the library size and concentration for RNA-seq. The procedures for sample determination, sequencing, and analysis were conducted using PersonalBio. Sequencing was performed using an Illumina platform.

Bioinformatics Analysis

The gene read count was calculated using HTSeq (v0.9.1) as the original expression level of the gene. To make the gene expression levels comparable across samples, Fragments Per Kilobase of transcript per Million mapped reads (FPKM) were used for normalization for visualization purposes. In transcriptome sequencing, genes with an FPKM >1 are generally considered to be expressed. For exploratory data presentation, fold-change was calculated as FPKM treatment/FPKM control, representing the relative level of gene upregulation or downregulation. However, for rigorous statistical identification of differentially expressed genes (DEGs), DESeq2 (v1.34.0) was employed to account for library size variations and RNA composition bias. Significantly differentially expressed genes were defined as those meeting two criteria: |log2FoldChange| >1 and Benjamini-Hochberg adjusted p-value (padj) <0.05. Genes with log2FoldChange >0 were designated as upregulated, and those with log2FoldChange <0 were designated as downregulated.

Statistical Analysis

All data are presented as mean ± standard deviation (SD), and the results were analyzed using analysis of variance (ANOVA) and independent-sample t-test methods, with the Benjamini–Hochberg correction. The corrected p-values were less than 0.05 and were statistically significant.

Results

Effect of Artificial Sweeteners on Bacterial Growth Curves

The growth of E. coli and B. subtilis was measured every 6–12 hours after exposure to various concentrations of artificial sweeteners (CYC, SAC, SUC, ASP, and ACE-K) for 48 h. In E. coli, growth was not significantly affected by the different artificial sweetener groups compared with the control group supplemented with GLC over the measured time, except for the SAC group (Figure 1A). Similarly, there was no significant effect of artificial sweeteners on the growth of B. subtilis at any time point compared with that in the GLC group (Figure 1B). In addition, the growth rates of E. coli and B. subtilis slowed or even decreased after 24 h of exposure to all sweeteners. However, E. coli and B. subtilis continued to grow after 24 h of exposure to GLC.

Figure 1 Growth curves of (A) E. coli and (B) B. subtilis following artificial sweetener treatment (n=3). PCA of artificial sweetener and glucose gene expression profiles in (C) E. coli and (D) B. subtilis. Data were presented as mean ± standard error mean. Significant differences between individual sweetener-treated groups and the control (GLC) were tested with Independent-sample t test, *p < 0.05.

Minimum Inhibitory Concentration of Artificial Sweeteners

The maximum concentration of each sweetener used in this study was determined by the CAC, which is the maximum limit for additives in food and beverages (SAC, 2500 mg/L; CYC, 2000 mg/L; ASP, 5000 mg/L; ACE-K, 5000 mg/L; SUC, 1800 mg/L). Of these, the maximum dose of ASP recommended by the CAC was 10,000 mg/L, but it was experimentally found that ASP could not be completely dissolved at this concentration, which was reduced to 5000 mg/L to avoid compromising the results of the experiment. In this study, the MIC of each sweetener against both strains were higher than the maximum concentration in the experiment, which proved that the antibacterial properties of artificial sweeteners at the daily usage concentrations were not obvious (Table 1). Notably, both strains were sensitive to Amp (CLSI 2019) before exposure to artificial sweeteners, with MIC values of 2μg/mL and 1μg/mL, respectively. However, 48 h after exposure to each artificial sweetener, the MIC values for Amp increased for both strains, indicating reduced antibiotic susceptibility (ranging from 2-fold to 4-fold increases) rather than stable clinical resistance, whereas the MIC values for Amp did not change for either strain after exposure to GLC (Table 2). These modest effect sizes are consistent with adaptive tolerance rather than the development of stable resistance.

Table 1 MIC (mg/Ml) Data of 5 Artificial Sweeteners to 2 Bacterial Based on Antibiotic Susceptibility Testing

Table 2 MIC (μg/Ml) Data of Amp After the Treatment of Artificial Sweeteners to 2 Bacteria Based on Antibiotic Susceptibility Testing

Expression of Drug Resistance-Related Genes

To investigate early transcriptional stress responses and avoid confounding by growth-phase alterations, transcriptome analysis was performed on E. coli and B. subtilis after 2 h exposure to each artificial sweetener. Differentially expressed genes (DEGs) were identified using DESeq2 with |log2FoldChange| >1 and Benjamini–Hochberg adjusted p-value <0.05. Based on gene expression values of each sample PCA principal component analysis, the result shows that the largest principal component of E. coli (PC1:68.5%) and B. subtilis (PC1:61.5%) was better able to distinguish between the artificial sweetener and GLC groups, while the artificial sweetener group clustered together (Figure 1C and D).

To test whether artificial sweeteners influenced the expression of resistance-related genes, we compared the expression of genes highly associated with resistance in two species of bacteria, most of which were efflux pump genes. They are involved in the active efflux of drugs and enhance bacterial survival under antibiotic pressure. The results showed that in E. coli and B. subtilis, the expression of resistance-related genes was mostly upregulated in the sweetener-treated group compared to the control GLC group. Especially in E. coli, this phenomenon is particularly pronounced in E. coli because of the abundance of efflux pump genes (Figure 2). For example, DESeq2 analysis revealed 1.6-, 2.5-, and 2-fold increases in acrA, acrB and tolC were detected in CYC-treated E. coli compared with the control, whereas in E. coli treated with SAC, 1.7-, 5.2-, and 8.2-fold increases in ompA, ompC, and ompF, respectively, were detected in E. coli treated with SUC, 2.0 and 2.3-fold increases in evgA and mdtJ were detected, respectively. In E. coli treated with ACE-K, 9.3-fold and 3.7-fold increases in msbA and proV, respectively, were detected. A 2.7-fold increase in arcA expression was detected in ASP-treated E. coli (Tables S1 and S2).

Figure 2 Heatmap showing the expression of resistance-associated genes in E. coli and B. subtilis following artificial sweetener treatment.

Effects of Artificial Sweeteners on ROS and Expression of Related Genes

As intracellular reactive signaling molecules associated with various bacterial physiological activities.^18,19 Once exposed to adverse conditions, reactive oxygen species can accumulate in bacterial cells. Excessive ROS can lead to oxidative stress and cell damage, which are also mechanisms of action of common antimicrobial drugs.²⁰ Therefore, we measured the ROS production in E. coli and B. subtilis treated with artificial sweeteners for 2 h. We found a significant increase in intracellular ROS levels for both bacterial species in the treated group compared to the control group, and this was positively correlated with concentration (Figure 3A and C). For example, at 300 mg/L, CYC, ASP, ACE-K, and SUC induced 1.53-, 1.58-, 1.62- and 1.58-times higher ROS production in E. coli, respectively, whereas at 1000 mg/L, ROS production increased 1.69-, 1.74-, 1.79 and 1.72 times, respectively. However, the SAC group shows 1.32 times increase at the highest concentration. The antibiotic Amp group also showed significantly increased intracellular ROS production in both the bacterial species. In contrast, GLC treatment did not increase ROS production in either the bacterial cell line or significantly decreased ROS production in E. coli.

Figure 3 Generation of reactive oxygen species (ROS) induced by treatments with artificial sweeteners at different concentrations. Effects of artificial sweeteners (n=3, ANOVA, p < 0.05) on intracellular ROS production in E. coli (A). Heatmap depicting the expression profiles of oxidative stress-associated genes in E. coli (B). Effects of artificial sweeteners (n=3, ANOVA, p < 0.05) on intracellular ROS production in B. subtilis (C). Heatmap showing the expression profiles of oxidative stress-associated genes in B. subtilis (D). Significant differences between individual sweetener-treated groups and the control (0 mg/L of sweeteners) were assessed using an independent-sample t-test. Statistical significance is indicated by symbols (* p < 0.05 and # p < 0.01).

Regarding gene expression at the 2 h exposure timepoint, we focused on changes in the expression of genes associated with oxidative stress in both bacteria after each sweetener was used. These genes were generally highly expressed in the artificial sweetener group compared to the GLC control group, and this was more pronounced in E. coli (Figure 3B and D). For example, in SAC treated E. coli, 5.3-, 4.8-, and 5.5-fold increases in recR, oxyR and yfgD, respectively, while the CYC-treated group showed 14.4-, 4.7-, and 4.7-fold increases in soxS, hupA and oxyR, respectively, while the SUC-treated group showed 9.4-, 4.4-, and 4.4-fold increases in soxS and lexA, respectively. 18.1-, 5.1-, and 4.3-fold increases in soxS, recR and tpx were detected in ACE-K-treated cells. 12.6, 2.2 and 2.1-fold increases in soxS, katG and ruvB were detected in the ASP treated group, respectively (Tables S3 and S4). Taken together, these results further confirmed that artificial sweeteners are more likely than GLC to promote ROS overproduction in bacterial cells at early exposure timepoints (2 h).

Effect of Artificial Sweeteners on the Expression of Virulence-Related Genes

To assess early transcriptional changes in bacterial pathogenicity at 2 h post-exposure, we focused on the expression of virulence factor-related genes in two bacterial strains treated with artificial sweeteners. Similar to the above results, in E. coli, virulence factor-related genes were upregulated in almost all artificial sweetener-treated groups compared with the GLC controls (Figure 4). The expression of the rstA gene increased 13-fold in all five artificial sweetener-treated groups compared with that in the control group, and the expression of the rstA gene was up to 16.7-fold higher in the CYC group. In avian pathogenic E. coli (APEC) and uropathogenic E. coli (UPEC), deletion of the rstA gene significantly reduces the virulence of the bacteria.²¹ In addition, the expression of slyD increased more than 5-fold in all five artificial sweetener-treated groups (Table S5). In Helicobacter pylori, SlyD has been identified as a new virulence factor that promotes intestinal metaplasia of the gastric mucosa, a precancerous lesion in gastric cancer, by regulating signaling pathways in host cells.²²

Figure 4 Heatmap showing the expression of virulence-associated genes in E. coli and B. subtilis following sweetener treatment.

Similar results were obtained for B. subtilis (Figure 4). Notably, the expression of several genes involved in sporulation was significantly upregulated in some of the artificial sweetener-treated groups compared to the controls, with spo0A, spo0B and spoIIAA all of which are key regulators of sporulation.²³ In particular, the expression of spoIIAA was upregulated by 4.4-, 8.1-, 4.9-, 9.0-, and 2.2-fold in the SAC, CYC, SUC, ACE-K, and ASP treatment groups, respectively, compared to the control group (Table S6). This suggests that the addition of artificial sweeteners causes a degree of stress to bacteria, which may promote spore formation.

Discussion

Although artificial sweeteners have been one of the most important achievements of the food industry to date, their impact on human health is controversial.²⁴ A recent study found that high intake of sucralose reduced the activity of T cells in mice, which had some effect on the immune system.²⁵ It is essential to note that the first digestive organ that artificial sweeteners contact after entering the body is the gastrointestinal tract, which is in contact with the intestinal flora for 18–36 h and may affect the functions of many bacteria, including intestinal probiotics.²⁶ Recent evidence indicates that artificial sweeteners, once considered metabolically inert, function as active microbiota-modulating agents that can reshape gut microbial composition and function, thereby creating selective pressure for antimicrobial resistance (AMR) evolution beyond simple caloric replacement.¹⁶ The gut microbiome serves as both a reservoir for resistance genes and a potential therapeutic target for metabolic interventions, where sweetener-induced oxidative stress and efflux pump activation may inadvertently drive bacterial adaptation strategies.¹⁷ Studies have shown that exposure to these sweeteners increases cell envelope permeability, upregulates genes encoding DNA uptake and translocation machinery, and may affect the balance of the gut microbial community via quorum sensing inhibition.^27,28 Therefore, we hypothesized that at CAC-permitted dietary concentrations, artificial sweeteners induce oxidative stress, leading to upregulation of efflux pump expression and subsequent reduction in antibiotic susceptibility, without exerting significant bactericidal effects.

Our study focused on whether artificial sweeteners affect the resistance of gut bacteria and the expression of related genes. CAC allowed eight types of artificial sweeteners, including alitame, aspartame, cyclamate, neotame, sucralose, advantame, saccharin, and acesulfame potassium, and established the corresponding upper limits for each sweetener. In our study, the results showed that under the experimental concentrations applied, except for SAC, which had a significant impact on the growth of E. coli, the artificial sweeteners exhibited no significant inhibitory effects on the growth of the two bacterial strains tested. This finding is consistent with previously reported results.²⁹ In this study, the MIC values of all artificial sweeteners exceeded the maximum experimental concentration. Nonetheless, after 48 h of exposure to artificial sweeteners, bacterial susceptibility to antibiotics decreased (adaptive tolerance), whereas in the GLC group, antibiotic sensitivity was maintained. Similar findings have been reported previously, suggesting that certain artificial sweeteners may induce the evolution of bacterial antibiotic resistance. The MIC values of the four artificial sweeteners used in a previous study against the three strains also exceeded the maximum experimental concentration of 1500 mg/L.¹⁴

However, it is crucial to note that these effects are concentration-dependent, distinguishing between environmental/lethal concentrations and sublethal daily intake levels. At higher concentrations (beyond CAC limits), artificial sweeteners may exhibit bactericidal properties or potentiate antibiotic activity through membrane disruption, as reported in recent studies.¹⁷ This paradoxical effect underscores the translational relevance of our findings: while artificial sweeteners may increase antibiotic sensitivity at high concentrations, their effects at daily exposure levels (CAC-permitted) appear to be the opposite, inducing adaptive tolerance through efflux pump upregulation and ROS-mediated stress responses rather than stable clinical resistance.

Other studies have demonstrated that ACE-K and SAC exhibit stronger inhibitory effects on bacterial strains,^30,31 ACE-K has a strong inhibitory effect on A. baumannii and P. aeruginosa but has no effect on S. aureus, showing that the antibacterial efficacy of artificial sweeteners varies significantly among different bacteria. The concentration of artificial sweeteners is also a key factor determining their inhibitory effects on bacterial growth. Similar to our results, previous studies have shown that low concentrations of sucralose (28.7 mM=11.4 mg/mL, with a concentration of 1.8 mg/mL in our experiment) had no inhibitory effect on all six bacterial strains. 55.7 mM sucralose showed a slight inhibitory effect on the isolates, with a significant difference between the two isolates.³² Notably, these findings also revealed that sublethal concentrations of ACE-K enhanced the sensitivity of A. baumannii to multiple antibiotics, particularly carbapenems. In contrast, our experimental findings showed a different trend. This discrepancy may stem from differences in the concentrations of artificial sweeteners used in our experiment compared to those used in previous studies. Specifically, our study focused on the concentrations of artificial sweeteners recommended daily by CAC, which are far below the lethal levels for bacteria. At these concentrations, only the bacterial efflux pump gene was activated, reducing antibiotic sensitivity, whereas damage to the bacterial cell membrane was minimal and insufficient to enhance the penetration of carbapenem drugs. These experimental results suggest that, while artificial sweeteners may increase antibiotic sensitivity at high concentrations, their effects on daily exposure levels may be the opposite. Additionally, our experiments corroborated previous findings¹⁴ showing that ROS production in bacteria correlated positively with artificial sweetener concentration. Although the concentration range in our study was broad, the overall trends remained consistent.

Therefore, we speculated that daily intake of artificial sweeteners within this concentration range is more likely to reduce antibiotic susceptibility compared to GLC. This hypothesis was supported by gene expression analysis at 2 h post-exposure, in which genes associated with antibiotic resistance in E. coli and B. subtilis were significantly upregulated in the artificial sweetener group compared with those in the GLC group. Among them, the AcrAB-TolC complex can significantly reduce intracellular antibiotic accumulation by efflux of drug molecules, and high expression of this efflux pump is one of the important mechanisms of bacterial adaptation and survival under antibiotic pressure.³³ These metabolic and stress-related pathways induced by sweeteners represent relevant targets in combination therapy strategies to overcome reduced susceptibility.¹⁷ The excessive production of ROS may activate the bacterial stress response mechanism and then upregulate the expression of the AcrAB-TolC efflux pump,³⁴ consistent with our hypothesis linking ROS induction to reduced antibiotic susceptibility via efflux pump activation. We observed that arcA expression in the sweetener group was more than 2.5 times higher in the artificial sweetener group than in the GLC group. Upregulation of arcA reduces ROS production by regulating metabolic pathways and oxidative stress responses, thereby enhancing bacterial tolerance to oxidative stress.³⁵

In addition, msbA is highly expressed in all artificial sweeteners (seven times higher than GLC), and the MsbA protein is an ATP-binding cassette (ABC) transporter in gram-negative bacteria, such as E. coli which is responsible for transporting lipooligosaccharides (LOS), the precursor of lipopolysaccharide (LPS), from the cytoplasm to the periplasmic side of the inner membrane. This process is critical for LPS biosynthesis, as LPS is a major component of the outer membrane of gram-negative bacteria and provides a barrier against antibiotics and environmental stresses.³⁶ The expression of the ompA, ompC, and ompF, which encode extracellular proteins that play important roles in bacterial extracellular permeability, resistance, and virulence, respectively, is upregulated in almost all artificial sweeteners. These proteins enhance bacterial resistance by regulating the permeability of the outer membrane and affecting the entry of antibiotics. Therefore, these proteins play an indispensable role in biofilm formation and stabilization, which is one of the most influential mechanisms of bacterial resistance.³⁷

Notably, in E. coli, stpA was significantly downregulated in all artificial sweetener groups compared to that in the GLC group. StpA, a homolog of H-NS, can activate the CRISPR-Cas system in E. coli to defend it from natural transformations. This defence mechanism can prevent the integration of foreign DNA, including resistance plasmids, and thus reduce the spread of resistance genes.³⁸ In addition, the expression of feoA and feoB, which are key components of the ferrous ion transport system (Feo system) in gram-negative bacteria, was downregulated at 2 h post-exposure. The Feo system plays an important role in bacterial oxidative stress protection and virulence, and its loss of function may lead to an increased tolerance of bacteria to oxidative stress and certain antibiotics.³⁹ While this pattern suggests potential implications for ferroptosis-like mechanisms, we acknowledge that without direct lipid peroxidation assays (eg, malondialdehyde measurement), iron-dependent cell death validation, or ferroptosis inhibitor experiments, this interpretation remains speculative and requires further investigation.

The expression of related genes is also upregulated in B. subtilis. For example, murAA and murAB, which play key roles in bacterial peptidoglycan biosynthesis, were significantly upregulated in almost all artificial sweetener groups. In Enterococcus faecalis, the gram-positive bacterium MurAB regulates the stability of MurAA. This regulatory mechanism is activated under cell wall stress (such as cephalosporin exposure), leading to the accumulation of MurAA. Thus, the enhanced bacterial resistance to cephalosporins.⁴⁰

Similar results were observed for ROS gene expression, such as a significant upregulation of rlmN gene expression (both more than seven times higher than that in the GLC group). Recent studies have shown that the rlmN acts as a stress sensor that directly senses reactive ROS and triggers a bacterial stress response. Upregulation of rlmN gene may lead to the production of stress response proteins, improve tolerance to oxidative stress, and thus affect the sensitivity of bacteria to drugs.⁴¹ In B. subtilis, changes in ROS in the SAC and CYC groups were not as pronounced as those in the other artificial sweetener groups. It is suggested that other artificial sweeteners are stronger than SAC in increasing bacterial oxygen-containing free radicals, especially at low concentrations; however, this difference decreases at higher concentrations.⁴²

These results indicate that artificial sweeteners can exert pressure on bacterial survival, and this pressure stimulates bacterial stress response and activation of efflux pumps, which is key to the formation of tolerant cells in bacteria and a universal mechanism observed in response to antibiotics.^43,44 No such phenomena were observed in the GLC group.

Notably, the expression of rstA, which encodes RstA protein, an important regulator of iron uptake in E. coli, was significantly elevated in the artificial sweetener groups. Together, rstA and rstB form a two-component system (RstA/RstB) that regulates the expression of iron-uptake-related genes by sensing the concentration of iron in the environment. Studies have shown that deletion of rstA reduces the ability of E. coli to take up iron and affects the expression of virulence factors.²¹ The upregulation of virulence-associated genes, including rstA and slyD, aligns with the concept that resistance and virulence are frequently co-regulated traits in bacterial adaptation, suggesting that artificial sweeteners may inadvertently promote a more virulent phenotype alongside reduced antibiotic susceptibility.⁴⁵ Iron overload is a key factor in ferroptosis. Recent studies have shown that high concentrations of sucralose induce cell ferroptosis.⁴⁶ Based on the current experimental results, we found for the first time that some artificial sweeteners affect the expression of genes responsible for iron ion absorption in bacteria, and their ability to induce ferroptosis needs to be further validated in follow-up studies.

In addition, our recent study revealed that probiotics in commercially available probiotic preparations are commonly resistant to multiple antibiotics⁴⁷ and that some of these probiotic preparations are supplemented with artificial sweeteners to improve taste; however, whether there is a causal relationship requires further research.

Several limitations should be acknowledged. The 2 h exposure captures early transcriptional responses but does not reflect long-term resistance development. The observed MIC increases (2–4 fold) represent adaptive tolerance rather than stable clinical resistance. Furthermore, single-species bacterial cultures cannot recapitulate the complex ecological interactions of the gut microbiome, limiting extrapolation to human microbiota. In vivo validation using complex models is essential. Finally, the ferroptosis hypothesis requires validation through lipid peroxidation assays.

Conclusion

The results indicate that the MIC values of five commonly employed artificial sweeteners (saccharin, cyclamate, sucralose, aspartame, and acesulfame potassium) for E. coli and B. subtilis exceeded the CAC-determined daily limit concentration. Except for SAC, the artificial sweeteners did not significantly affect the growth of E. coli or B. subtilis within 48 h. However, at this limiting concentration, they were associated with reduced antibiotic susceptibility and increased ROS production at 2 h post-exposure. Transcriptome analysis at 2 h revealed that the artificial sweetener group generally had upregulated genes associated with resistance mechanisms, oxidative stress, and virulence compared to the GLC group. This transcriptional response is consistent with the induction of adaptive tolerance rather than stable clinical resistance. In addition, artificial sweeteners affected the expression of iron uptake-related genes in E. coli, though the implications for ferroptosis remain speculative and require validation through lipid peroxidation assays.

These findings derive from in vitro single-species bacterial cultures and should be interpreted with caution, as they cannot fully recapitulate the complexity of the human gut microbiome. While our results highlight the need for careful consideration of artificial sweetener safety profiles at permitted daily intake levels, they do not replace clinical or in vivo studies. Future research should include longitudinal in vivo studies using complex microbiome models to assess whether chronic low-dose exposure selects for stable resistance, and to validate the ferroptosis hypothesis through iron metabolism studies. Considering the widespread use of artificial sweeteners in food and medicine worldwide, their potential impact on bacterial antibiotic susceptibility warrants further in vivo investigation under physiologically relevant conditions.