Nonnutritive sweeteners can promote the dissemination of antibiotic resistance through conjugative gene transfer

Nonnutritive sweeteners promote conjugative transfer

To test the effect of nonnutritive sweeteners on the conjugative transfer of ARGs, both intra- and intergenus-transfer experiments (model I) were first conducted, in which the bacteria were exposed to various concentrations of four commonly used nonnutritive sweeteners (SAC, SUC, ASP, and ACE-K) for 8 h at room temperature. Notably, in both mating systems, the whole concentration range (from 0.03 to 300 mg/L) of three sweeteners (SUC, ASP, and ACE-K) caused a significant concentration-dependent increase (p = 0.00017 ~ 0.047, Fig. S1a, b); Pearson correlation analysis was shown in Table S3 in conjugative transfer compared to the control (Fig. 1a, b). The intragenus (donor Escherichia coli K-12 LE392 and recipient E. coli K-12 MG1655) spontaneous conjugative transfer frequency was (1.9 ± 0.2) × 10⁻⁴ transconjugants per recipient cell (Fig. S2). However, the conjugative transfer frequencies were significantly enhanced by the sweeteners SUC, ASP, and ACE-K at 0.3 mg/L or above. For example, SUC, ASP, and ACE-K at 300 mg/L enhanced the conjugative frequencies by 1.5- (p = 0.00027), 4.1- (p = 0.000000089), and 3.4-fold (p = 0.0000020), respectively (Fig. 1a). In contrast, SAC did not significantly change the conjugative transfer frequency in the conjugation system (p = 0.200 ~ 0.670, Fig. 1a). For intergenus conjugation (donor E. coli K-12 LE392 and recipient Pseudomonas alloputida), all sweeteners at concentrations of 3 mg/L or higher (except for SAC) were seen to promote the conjugative transfer of the donor RP4 plasmid to recipients of different genera (p = 0.000047 ~0.042, Fig. 1b). SUC, ASP, and ACE-K at 300 mg/L caused a great increase in conjugative transfer, by 2.6- (p = 0.0000020), 4.1- (p = 0.000036), and 4.2-fold (p = 0.000019), respectively (Fig. 1b). It should be noted that the enhanced transfer frequencies were associated with the increased number of colonies on selective transconjugant plates, rather than decreased recipient numbers (Fig. S3).

To test whether conjugative transfer would be promoted by the tested compounds under clinical conditions, we used an IncA/C plasmid pMS6198A, which contains multiple clinically relevant resistance genes, in a uropathogenic E. coli strain isolated from a patient who suffered from a urinary tract infection [30] to conduct a conjugation experiment (model II, donor E. coli K-12 MG1655 and recipient E. coli J53) in liquid culture at 37 °C. Similarly, a significant concentration-dependent increase (p = 0.00017 ~ 0.047, Fig. S4 and Table S3) in pMS6198A transfer was also observed upon exposure to all four sweeteners (Fig. 1c and Fig. S5). This is different from the result in which SAC failed to enhance both intra- and intergenus conjugative transfer of the RP4 plasmid under environmental conditions.

The successful transfer of donor plasmids (RP4 and pMS6198A) to recipients was confirmed by a series of analyses. The MIC levels for transconjugants (recipient with transferred plasmid) were comparable or even higher than those for their corresponding parents that were resistant to all antibiotics (Fig. S6). In addition, plasmid extraction and PCR amplification of related resistance genes showed that all the transconjugants contained the same type of plasmid as the donor did and harbored the corresponding ARGs (Fig. S7). Thus, profiling of all the transconjugants confirmed that the donor plasmid was transferred to the recipient cells and conferred antibiotic resistance.

To understand whether transconjugants were capable of transferring ARGs to other candidates, we performed a reverse conjugation experiment (model III, donor P. alloputida containing the RP4 plasmid and recipient E. coli K-12 MG1655; Methods and Text S1) for 8 h at room temperature. Interestingly, this reverse conjugative transfer occurred and it was enhanced by all nonnutritive sweeteners (p = 0.000033 ~ 0.036, Fig. 1d; Fig. S8 and S9). Specifically, similar to the result of pMS6198A transfer, SAC significantly increased RP4 plasmid transfer (p = 0.036). Thus, the transferred RP4 plasmid in transconjugants maintained its mobility and could be transferred to other potential recipients. Collectively, these results demonstrate that the tested four sweeteners promoted plasmid-mediated conjugation in both nonpathogenic and pathogenic strains.

Real-time visualization of conjugative plasmid transfer

The spread of ARGs by conjugation is a dynamic process. However, clear real-time visualization of the cellular dynamics of conjugative transfer has yet to be obtained. Specifically, little is known about whether nonnutritive sweeteners could speed up the transfer of conjugative plasmids at the single-cell level. To achieve this, we developed an experimental system that is capable of real-time visualization of RP4 plasmid transfer by conjugation at the single-cell level using a fluorescence labeling technique (Methods and Text S2). This method could be useful or inspirational to other ecological studies, such as evolution of antibiotic resistance at a community-wide level. Time-lapse microscopy images showed that in the control group (without the addition of any sweeteners), the RP4 plasmid was successfully transferred to the recipient after 5 min of contact (Fig. 2a; Movie S1). In contrast, under exposure to four sweeteners, the plasmid acquisition process was accelerated, and successful transfer was observed within 5 min (Fig. 2b; Movie S2). Moreover, the number of transconjugants within 150 min were significantly increased under exposure to the tested sweeteners compared to the control (Fig. 2c–g). This was also mathematically supported by single-cell quantitative analysis, which showed higher transfer rates (r) under exposure to the tested sweeteners than in the control (Fig. 2h). A 6.0-, 5.0-, 5.0-, and 7.0-fold increase of transfer rate was induced by SAC, SUC, ASP, and ACE-K, respectively. The fitted maximum number of transconjugants (N_m) showed a more than 2.0- to 4.0-fold increase under treatment with these sweeteners. These results further confirm that nonnutritive sweeteners promote plasmid-mediated conjugation at the single-cell level.

Nonnutritive sweeteners enhance ROS production and the SOS response

It has been reported that nonnutritive sweeteners can cause DNA damage [26] and induce oxidative stress in bacteria [32]. Thus, we measured ROS production in cells exposed to the tested sweeteners. Notably, in the strains treated with the sweeteners (except for SAC), concentration-dependent increase in ROS was detected (Fig. 3a–c; Fig. S10). For example, under treatment with the sweeteners (except SAC) at 300 mg/L, the relative fold change in ROS production in the donor E. coli K-12 LE392 increased 1.5-fold (p = 0.0010 ~ 0.0070, Fig. 3a). This increase was also observed when both the recipients E. coli K-12 MG1655 and P. alloputida were treated with these sweeteners (Fig. 3b, c). In contrast, exposure to the sweetener SAC did not result in any increase in ROS production in these donor and recipient strains. To further verify the promotion of ROS production by these compounds, thiourea, a scavenger of oxygen-free radicals [33, 34], was added to the mating system together with nonnutritive sweeteners. As expected, significant decrease in ROS production was detected in the three strains treated with SUC, ASP, and ACE-K after thiourea was added (Fig. S11).

To determine whether the increased production of ROS induced by these sweeteners promotes conjugative transfer, thiourea was added to both the intra- and intergenus conjugative systems, together with the sweeteners. Compared to the group without thiourea in both conjugative systems, addition of the scavenger caused no change in the conjugative transfer frequency in the SAC-treated group but significantly decreased the transfer frequency in the SUC-, ASP-, and ACE-K-treated groups (p = 0.000033 ~0.035, Fig. 3d, e). These results suggest that ROS production plays a significant role in the plasmid-mediated conjugation process.

Under oxidative stress caused by the tested sweeteners, bacterial defense may include changed expression of ROS-related genes (or ROS detoxification) as well as changed expression of SOS response-related genes. To track the expression of genes related to ROS production and the SOS response regulated by the tested sweeteners, the mRNA transcription level of genes from both the donor E. coli K-12 LE392 and recipient P. alloputida was measured. The transcripts for genes encoding enzymes involved in ROS detoxification in the donor were increased under exposure to the tested sweeteners (Fig. 3f; Table S4). For example, the gene encoding superoxide dismutase (sodA) exhibited a 1.3- (p = 0.032), 1.8- (p = 0.0052), and 1.8-fold (p = 0.0023) increase in expression in the donor treated with SUC, ASP, and ACE-K, respectively. Accordingly, an SOS response was detected in the donor after exposure to these sweeteners (Fig. 3f; Table S5). Increased transcription was detected for the genes sulA (1.3- to 1.6-fold change, p = 0.0040 ~ 0.036) and umuC (1.6- to 2.1-fold change, p = 0.014 ~ 0.040) after treatment with the three sweeteners. We also found increased expression of ROS detoxification- and SOS response-related genes in recipients treated with SUC, ASP, and ACE-K (Fig. 3g; Tables S6 and S7). In contrast, SAC treatment did not result in significant regulation of the genes mentioned above in the donor or recipient.

It is apparent that SUC, ASP, and ACE-K increased ROS production in both intra- and intergenus conjugative systems, whereas SAC did not induce any change in ROS production in the strains tested. We propose that nonnutritive sweeteners induced ROS overproduction and increased ROS detoxification, and correspondingly stimulated oxidative stress responses (the SOS response) in the strains, eventually contributing to the promotion of intra- and intergenus conjugative transfer.

Nonnutritive sweeteners increased cell membrane permeability

It is worth noting that addition of the scavenger thiourea did not completely reverse the increase in the frequency of conjugative transfer caused by sweeteners (i.e., SUC, ASP, and ACE-K). For example, after addition of thiourea to the mating system that was exposed to 300 mg/L ACE-K, the frequency of conjugative intragenus transfer was (3.5 ± 0.4) × 10^–4 transconjugants per recipient cell, which was still significantly (p = 0.0090) higher than the value for the control ((2.0 ± 0.3) × 10^–4 transconjugants per recipient cell, Fig. 3d). This indicated that ROS production is not the only mechanism for conjugation.

Bacterial cell membranes act as a barrier against horizontal transfer of ARGs [11], and the permeability of this barrier could play a significant role in the transport of substances into or outside of cells. Therefore, we tested whether nonnutritive sweeteners could change the cell membrane permeability. To examine this concept, we treated both the donor E. coli LE392 and the two recipients (E. coli K-12 MG1655 and P. alloputida) with different concentrations of the four sweeteners. After 2 h of treatment, the cell membrane permeability was increased by up to 3.7-fold (p = 0.000047 ~ 0.0019) in the donor by all the sweeteners (300 mg/L) except SAC (Fig. 4a; Fig. S12). This increase was also observed in the two recipients treated with all sweeteners (Fig. 4b, c), exhibiting an increase of up to 1.5- and 1.6-fold, respectively (p = 0.0010 ~ 0.028), under exposure to all tested sweeteners at 300 mg/L. In addition, we found that SAC, together with the other three sweeteners, significantly increased the permeability of another donor strain, E. coli K-12 MG1655 (p = 0.00024 ~ 0.0082, Fig. S13a). Similar results were also found for the permeability of the recipient strain E. coli J53 (p = 0.000088 ~ 0.0074, Fig. S13b).

Our results showed that the SAC failed to enhance ARG transfer from E. coli LE392 to either E. coli K-12 MG1655 and P. alloputida but enhanced conjugative transfer between E. coli K-12 MG1655 and E. coli J53, as well as between the strain P. alloputida containing the RP4 plasmid and the strain E. coli K-12 MG1655 (i.e., reverse conjugation). This raises the question of whether the permeability of the donor or recipient (or both) is important for conjugative transfer. To this end, we performed different pre-exposure to separately increase the cell membrane permeability of the donor (E. coli LE392) and two recipients (E. coli K-12 MG1655 and P. alloputida) before mixing for conjugation experiments (Text S3; Table S8). We treated each strain with 3 mg/L SUC, which significantly increased the permeability of the donor and recipient (Fig. 4a–c). The results showed that there was no significant change in conjugation frequency when the recipients were pre-exposed to SUC (p = 0.078 ~ 0.19, Fig. 4d, e; Fig. S14). This indicates that the increased permeability of the recipient did not guarantee enhanced conjugative transfer. In contrast, both intra- and intergenus conjugation frequencies were significantly (p = 0.00000000034 ~ 0.040) promoted when the permeability of the donor was significantly increased. Overall, these results indicate that cell membrane permeability is positively mediated by nonnutritive sweeteners in mating systems and that donor permeability plays a more critical role than recipient permeability in the conjugation process.

We tested whether genes that mediate membrane permeability at the molecular level were upregulated in the donor and recipient (P. alloputida) after treatment with the tested sweeteners at 3 mg/L. Genes related to cell membrane permeability in both donor and recipient cells were upregulated after exposure to the tested sweeteners (Fig. 4f, g). For instance, fecA, which encodes an outer membrane transporter, and ybgE, which encodes an inner membrane protein, were significantly upregulated (1.5- to 7.0-fold (p = 0.0020 ~0.037) and up to 2.7-fold (p = 0.0060 ~0.044), respectively, Table S9) in the donor cells treated with SUC, ASP, and ACE-K (Fig. 4f). In contrast, SAC did not induce significant changes in the expression of cell membrane-related genes in the donor. These findings are consistent with the results showing that all sweeteners, except SAC, increased the membrane permeability of the donor. Nevertheless, cell membrane-related genes in the SAC-treated recipient showed increased expression, as seen in the other sweetener-treated recipient groups (Fig. 4g; Table S10).

These results confirm that the tested sweeteners increased the expression of cell membrane-related genes, and then increased the cell membrane permeability of the donors or both parents to promote the conjugative transfer of plasmids between strains.

Nonnutritive sweeteners upregulate the expression of conjugation-related genes on the RP4 plasmid

The process of conjugative transfer mediated by the RP4 plasmid requires the regulation of global regulatory genes and conjugative transfer-related genes [35, 36]. After treatment with 3 mg/L SUC, ASP, and ACE-K, we found that the core global regulatory gene korC was repressed in comparison with the control group. Accordingly, the conjugative transfer regulator gene traG (which connects the relaxosome and mating pair formation complex) and DNA transfer and replication (Dtr) genes traC1 and traC2 exhibited increased expression compared with the mRNA expression detected in the control group (Fig. 5). During the conjugation process, the plasmid is thought to be transferred via a pilin bridge between donor and recipient cells [35]. The expression levels of pilin formation-related genes, including traA, traB, traF, and trap, were also upregulated in the sweetener-treated groups (Fig. 5). These genes also contribute to the formation of the mating pair formation (Mpf) system. For example, compared to the control, mRNA expression of gene traB (encoding a conjugative transfer protein) was significantly upregulated by more than 1.5- (p = 0.0046), 1.4- (p = 0.047), and 2.1-fold (p = 0.000050) under exposure to SUC, ASP, and ACE-K, respectively (Fig. 5a; Table S11). Together, these results reveal that these sweeteners (except for SAC) cause increased mRNA expression levels of genes related to RP4 plasmid transfer and replication (Dtr genes) and then promote the formation of the pilus channel in the Mpf system to allow RP4 plasmid transfer from the donor to the recipient.

The conjugation process also requires direct cell-to-cell contact. This can be achieved by the periplasmic pilin of donor cells. In the donor E. coli K-12 strain, pili operons have been found to play roles in cell adhesion and colonization [37]. In the present study, the expression of genes in these adhesion-relevant operons was upregulated. For example, the expression of gene ecpA was upregulated by 1.8- (p = 0.035), 4.2- (p = 0.0070), and 2.1-fold (p = 0.033) after treatment with SUC, ASP, and ACE-K, respectively (Fig. 5b; Table S12).

Source: Ecology - nature.com