EP4648780A1 - Potentiation of aminoglycosides through activation of carbohydrate transporters - Google Patents

Abstract

The invention relates to the potentiation of aminoglycosides through activation of carbohydrate transporters, in particular using nucleosides. The invention encompasses the combination of a nucleoside and an aminoglycoside antibiotic for treating a bacterial infection, wherein the nucleoside enhances the killing of the bacteria by the antibiotic, preferably wherein the nucleoside limits the emergence of antibiotic-resistant bacteria or re-sensitizes antibiotic-resistant bacteria to the killing by the antibiotic.

Description

POTENTIATION OF AMINOGLYCOSIDES THROUGH ACTIVATION OF CARBOHYDRATE TRANSPORTERS

FIELD OF THE INVENTION

The invention relates to the potentiation of aminoglycosides through activation of carbohydrate transporters, in particular using nucleosides. The invention encompasses the combination of a nucleoside and an aminoglycoside antibiotic for treating a bacterial infection, wherein the nucleoside enhances the killing of the bacteria by the antibiotic, in particular wherein the nucleoside limits the emergence of aminoglycoside-resistant bacteria and/or re-sensitizes aminoglycoside-resistant bacteria to the killing by the antibiotic.

BACKGROUND OF THE INVENTION

Antibiotics save many lives and play a major role in modern medicine. However, the use of any new antibiotic is followed by the spread of resistance in bacterial populations¹. As outlined by the World Health Organization (WHO), studies show a prevalence of Gram-negative bacteria regarding resistance ², mainly due to their double membrane barrier ³. A recent report associates 1,27 million death in 2019 directly due to bacterial AMR, with only 5 pathogens responsible for 95% of deaths, including Escherichia coli, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii and Pseudomonas aeruginosa, making AMR burden responsible for more death than malaria and VIH together⁴, highlighting the urgency to develop novel therapies. Among these pathogens, all but one (S. aureus) are Gram-negative bacteria.

Aminoglycosides (AGs) constitute a family of broad-spectrum antibiotics able to counteract the double membrane barriers of Gram-negative bacteria. They are used in the clinic to treat infections such as pneumonia, sepsis or urinary tract infections (UTI) when they are caused by various Gram- negative pathogens⁵, but AGs treatment are associated with ototoxicity⁶ and nephrotoxicity^7,8.

The entry of AGs in Gram-negative cells has been proposed to occur in three steps^9,10. The first step is an uptake phase considered as "passive", facilitated by the cationic properties of AGs⁹: they electrostatically interact with negatively charged components of the outer membrane (lipopolysaccharides, phospholipids), in an energy-independent manner^11,12. This leads to membrane disruption, followed by a "self-promoted" uptake in the periplasm¹³. Uptake by non-specific outer membrane porins OmpF and OmpC has also been monitored in E. coli¹⁴. During step two, an energy dependent phase allows AG entry from the periplasm into the cytoplasm. AG uptake in bacterial cells is thus described as dependent of proton motive force (PMF), which can be decreased by inhibitors of electron transport and oxidative phosphorylation^15,16. Finally, AGs target the ribosome, leading to translation slow-down¹⁷ and mistranslation^18,19. It is commonly accepted that mistranslated proteins, among which membrane proteins, alter the integrity of the membrane, allowing a final phase of uptake of more AGs and leading to more damages to cells. Aminoglycosides are molecules that show good efficacy but cause significant side effects in humans during the treatment of infections but also in prophylaxis²⁰. Nephrotoxicity and ototoxicity are observed in variable proportions according to the studies, varying on average between 3 and 15% of cases^21,22, although these number are difficult to determine because they depend on the patient: general state of health, presence of mutations (e.g. mutation in the mitochondrial gene MTRNR1 associated with aminoglycoside-induced deafness²³), and on the type of infection, in particular the duration of the treatment. Indeed, the appearance of these side effects is facilitated by long treatments²². Aminoglycosides, by virtue of their chemical properties, bind to negatively charged sites in the brush border cells of the proximal convoluted tubule of the kidney²⁴. They accumulate in the cell before being transferred to the vacuoles, where they cause excessive accumulation of phospholipids, alter lysosome membranes, and disrupt mitochondrial metabolism²⁵. The mean incidence of nephrotoxicity for gentamicin and tobramycin was estimated at 14 and 12%, respectively, and 9.4% for amikacin in a study combining data from 10,000 patients²¹. The onset of nephrotoxicity has been shown to be dependent on the area under the curve, hence the use of single doses spaced 12 hours apart^26,27. The use of aminoglycosides can cause a significant loss of high-frequency sound detection, dizziness, and even irreversible deafness²⁸. The longer the duration of treatment, the greater the side effects. For example, hearing loss has been observed in 4 to 15% of patients treated with 1 g/day of streptomycin for more than 7 days. Symptoms usually appear after a latency of 7 to 10 days²⁹. This ototoxicity is due to their high affinity for the hair cell membrane of the cochlea, resulting in damage first to the outer hair cells of the first to third row and then to the inner hair cells³⁰. Absorption by the organ of Corti located in the cochlea is rapid (saturation reached in 3h), with an effect on the hair cells depending on the concentration of aminoglycosides³¹. The residual concentration (Cmin) is predictive of toxicity. The dosage for long treatments must be performed after 48 hours of treatment. For amikacin, the toxicity threshold is between 2.5 and 5 µg/ml and between 0.5 and 1 µg/ml for tobramycin and gentamicin The maximum aminoglycoside concentration (Cmax/MIC) recommended is 8-10 times the Minimal Inhibitory Concentration (MIC). The international application WO 2020/227530 and the corresponding scientific publication (Yang et al., Cell, 2019, 177, 1649-1661) disclose a machine learning approach for identifying metabolic pathways involved in the lethality of 3 antibiotics (ampicillin (AMP, beta-lactam), ciprofloxacin (CIP, fluoroquinone) and gentamicin (GENT, aminoglycoside) using 206 metabolites from the Biolog phenotype microarrays (PMs 1-4, Bochner, FEMS Microbiol Rev, 2009, 33-191-205) that are included in the iJ0136 genome-scale model of E.coli metabolism. Model-guided machine learning predictions identified purine biosynthesis pathway as a new pathway involved in antibiotic lethality with share directionality between AMP and CIP and opposite directionality for GENT (Example 5, p79-80; Figure 3). The hypothesis was verified in wild-type E.coli by genetic deletion of enzymes involved in purine or pyrimidine biosynthesis, biochemical inhibition with purine biosynthesis enzyme inhibitors and biochemical supplementation with purine biosynthesis substrates (phosphoribosyl pyrophosphate (prpp), glutamine (gln); Example 6, p81; Figure 4A). Stimulation of purine biosynthesis pathway increased AMP and CIP lethality and decreased GENT lethality (Figure 4E); conversely inhibition of purine biosynthesis pathway decreased AMP and CIP lethality and increased GENT lethality (Figure 4B, 4C, 4D). Based on these predictions and observations, it was hypothesized that (i) purine supplementation would rescue antibiotic-induced purine depletion, reducing antibiotic lethality and (ii) pyrimidine supplementation such as with uracil would inhibit pyrimidine biosynthesis and promote purine biosynthesis activity via prpp accumulation, and consequently increase antibiotic lethality (Figure 5A). This hypothesis does not fit with the validated model for Gentamicin (GENT) which shows clearly that GENT has an opposite effect to AMP or CIP where GENT lethality is decreased by stimulation of the purine biosynthesis pathway (Figure 4E). Therefore, it is not surprising that for gentamicin no significant increase in antibiotic lethality was observed after 4h with all the tested pyrimidine (cytosine, thymine, uracil, uridine) (Figure 5B and 5C) indicating that pyrimidine does not improve the killing of bacteria by gentamicin or other aminoglycoside antibiotics in the test conditions. This was confirmed with the antibiotic half-maximum inhibitory concentration (IC50) measured after four hours of treatment using Biolog plates. No significant change in IC50 of Gentamicin compared to control was observed for all tested pyrimidines (PM1: negative control (1,31E-07); uridine (1,55E-07); PM3: negative control (9,98E-09); Cytidine (1,27E-08); Cytosine (1,14E-08); Thymine (1,20E-08); Thymidine (1,02E-08); Uracil (1,17E-08); Uridine (1,12E-08); a 2-fold change is considered significant). The inventors of the present application have reproduced these experiments and found no significant change in antibiotic lethality for all tested pyrimidine compounds in the test conditions (Figure 21 of present application). The application EP 3027213 discloses the use of a combination of an antibiotic and both an uridine compound and a pyruvate to limit the side effects of the antibiotic. It is disclosed that : (i) the antibiotic effect is solely based on the antibiotic compound and is not rendered by the co- administered uridine and/or pyruvate compounds and (ii) the presence or both uridine and pyruvate compounds is necessary to neutralize the side effects of the antibiotic compound. Consequently, there is a need in the art for compositions and methods to increase the effectiveness of Aminoglycosides against Gram-negative bacteria by increasing their uptake across bacterial membranes. The invention fulfills this need by allowing greater efficiency of AGs therapies, through an improved uptake in bacteria. BRIEF SUMMARY OF THE INVENTION The invention encompasses compositions, uses of these compositions for killing bacteria, and methods for killing bacteria. The invention encompasses the combination of a nucleoside together with an aminoglycoside antibiotic, wherein the nucleoside enhances the killing of the bacteria by the antibiotic, preferably wherein the nucleoside limits the emergence of bacteria resistant to the aminoglycoside antibiotic and/or re-sensitizes bacteria resistant to the aminoglycoside antibiotic to the killing by the antibiotic. In some embodiments, the nucleoside is selected from uridine, inosine, guanosine, cytidine, thymidine, or xanthosine. Preferably, the nucleoside is uridine. In some embodiments, the nucleoside is administered to the urinary tract, gastrointestinal tract, lungs, eyes, ears, nose, brain, heart, blood or skin. In some embodiments, the nucleoside is administered at a dose of 1-20 g/m². In some embodiments, the method is for the prevention or limitation of the emergence of antibiotic- resistant bacteria by killing the bacteria faster with a co-administration of the nucleoside together with the aminoglycoside antibiotic at concentrations already used in clinics. In some embodiments, the method is for killing antibiotic-resistant bacteria. Co-administration of the nucleoside together with the aminoglycoside antibiotic is able to kill antibiotic-resistant bacteria, whereas the same concentration of the aminoglycoside antibiotic used alone is ineffective. In some embodiments, the method is for decreasing of the toxicity of a reference treatment by a co- administration of the nucleoside together with the aminoglycoside antibiotic with the aminoglycoside being administrated at a lower concentration than the reference treatment. In some embodiments, the level of killing of the bacteria by the antibiotic is the same as the reference treatment. In some embodiments, the antibiotic is selected from tobramycin, gentamicin, and amikacin. In some embodiments, the bacteria are enterobacteria. In some embodiments, the bacteria are selected from Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, and Acinetobacter baumannii. In some embodiments, the bacteria causing the infection comprise bacteria resistant to the aminoglycoside antibiotic. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1A-C: CmtA and Crp are involved in tobramycin tolerance. A. Etest of E. coli cmtA and crp. Deletion of CmtA increases the MIC by 4-fold whereas deletion of Crp increases MIC by 15-fold, despite its growth defect. B. Growth curve of E. coli WT, Δ cmtA and Δ crp, in presence or not of tobramycin 0.4 µg/ml. C. Etests on E. coli carrying the empty vector (p0+) compared to the vector overexpressing cmtA (pcmtA+). The inducer was added on the medium. Figure 2A-C: A. Overexpression of 11 carbohydrates transporters sensitizes to AGs specifically. See also Table 3. Response to tobramycin (0.1 µg/ml), carbenicillin (3 µg/ml), ciprofloxacin (0.005 µg/ml) or gentamicin (0.08 µg/ml) of E. coli carrying the empty plasmid (p0) compared to the vector overexpressing carbohydrates transporters. Cultures of each strain were grown overnight. Susceptibility was assessed by serial dilutions (ND: non-diluted) and dropping 5 µl of each dilution on plates containing or not 0.1 µg/ml of tobramycin. Inducer was added in the media. A. Overexpression of 16 transporters sensitize to tobramycin. B. Overexpression of 5 transporters sensitize also to carbenicillin or ciprofloxacin. C. Overexpression of 11 transporters sensitize also to gentamicin (AG). Figure 3: CmtA is involved in the differential uptake of neocy5. Uptake of neocy5 evaluated by flow cytometry on E. coli ΔcmtA Δcrp compared to the WT strain and E. coli carrying a plasmid overexpressing cmtA compared to the strain carrying the empty vector (p0), expressed as fold change of mean fluorescence per cell (compared either to the WT for mutants or to the empty vector for overexpression). Figure 4: Expression of GFP under the dependence of cmtA promoter. Quantification of GFP fluorescence using flow cytometry depending on the strain (WT, ΔcmtA, Δcrp or Δcra) or the substrate added in the medium (supplemented or not with glucose or mannitol 0.5%), expressed as fold change of mean fluorescence per cell (compared either to the WT for mutants, or to the no substrate condition for carbon sources). Figure 5A-D: Uridine increases cmtA expression and decreases the MIC of tobramycin by enhancing the uptake. A. GFP expression from cmtA promoter depending on the condition (supplemented or not with glucose, uridine, ribose, uracile or mannitol 0.5%). Growth depending on the substrate is presented on the left by following OD_600nm. B. Quantification of GFP fluorescence using flow cytometry depending on the substrate added in the medium (supplemented or not with glucose, uridine, ribose, uracile or mannitol 0.5%) expressed as fold change of mean fluorescence per cell compared with the no substrate condition. C. Tobramycin MIC of E. coli WT determined by Etest, expressed in µg/ml, depending on the substrate added in the medium (glucose, ribose or uridine 0.5%). D. Uptake of neocy5 evaluated by flow cytometry on E. coli WT grown either with a supplementation of glucose (MIC 1.5 µg/ml) or uridine (MIC 0.1 µg/ml) 0.5%, expressed as fold change of mean fluorescence per cell compared to glucose. Figure 6A-B: GFP expression from fruA or btuB promoter depending on the substrate. A. Quantification of GFP fluorescence from the fruA promoter using flow cytometry depending on the substrate added in the media (supplemented or not with glucose, uridine, ribose, fructose or mannose 0.5%) expressed as fold change of mean fluorescence per cell. B. Quantification of GFP fluorescence from the btuB promoter using flow cytometry depending on the substrate added in the media (supplemented or not with glucose, uridine, ribose, fructose or mannose 0.5%) expressed as fold change of mean fluorescence per cell. Figure 7A-B: Uridine mediated AG susceptibility is not related to uridine uptake, catabolism or stress responses. A. Susceptibility to tobramycin of E. coli carrying the empty plasmid (p0) compared to the vector overexpressing nupG or nupC, two uridine transporters. Cultures of each strain were grown overnight. Susceptibility was assessed by serial dilutions and dropping 5 µl of each dilution on plates containing or not 0.1 µg/ml of tobramycin and uridine 0.5%. Inducer was added in the media. B. Quantification of GFP fluorescence in exponential phase from the P1rrnB using flow cytometry depending on the substrate added in the media (supplemented or not with glucose, uridine, ribose or maltose 0.5%) expressed as fold change of mean fluorescence per cell. Cultures in stationary phase (Stat) and cells treated with sub-MIC tobramycin (Tob, 0.06 µg/ml) were used as positive controls of stringent response activation. Figure 8A-B: Uridine induces fast killing and prevents the appearance of resistant mutants. A. Time kill curve of E. coli WT, growing in mid-exponential phase in liquid cultures supplemented or not with glucose, maltose or uridine 0.5%. Lethal treatment of tobramycin (10 µg/ml) was applied, and survival was assessed after 1, 2, 3, 4, 6 and 20 hours by plating and counting CFUs/ml. B. Photography of a plate presenting small and normal colonies after 20 hours of treatment with tobramycin and maltose.

Figure 9A-B: Uridine effect is PMF-dependent, but does not occur through a change of PMF. A. Uptake of Mitotracker Red evaluated by flow cytometry on E. coli ΔcmtA and Δcrp compared to the WT strain ; E. coli carrying a plasmid overexpressing cmtA compared to the strain carrying the empty vector (pO); and E. coli WT grown either with a supplementation of glucose or uridine 0.5%, expressed as fold change of mean fluorescence per cell. TCS (tetrachlorosalicylanilide) treatment was used as negative control. B. Survival to tobramycin 4 μg/ml treatment in urine synthetic media supplemented or not with uridine 0.5%, in the presence or not of 15 μM of the protonophore Carbonyl cyanide m- chlorophenyl hydrazine (CCCP). Survival was assessed by plating and counting CFU after 16 hours of treatment.

Figure 10A-B: Overexpression of the carbohydrate transporter MtlFGK in P. aeruginosa sensitizes to tobramycin by increasing the uptake. A. Etest on P. aeruginosa carrying the empty vector (p0+) compared to the vector overexpressing MtlFGK (pmtlFGK+). Inducer was added on the medium. B. Neocy5 uptake evaluated by flow cytometry on P. aeruginosa carrying a plasmid overexpressing MtlFGK (MtlFGK+) compared to the strain carrying the empty vector (p0+), expressed as fold change of mean fluorescence per cell.

Figure 11A-D: Uridine potentiates AGs in urine synthetic medium by enhancing uptake. A. Survival of E. coli after 20 hours of treatment with low-dose tobramycin (0.5 μg/ml) according to the addition of different concentrations of uridine (shown in %). Significance above each point represents the significance of the decrease in survival induced by the addition of the uridine percentage mentioned below, compared with the no uridine condition (0%), except for the comparison between 0.0315% and 1% as indicated. B. Liquid MIC test performed on E. coli, in synthetic urine medium by drop test as indicated in the method section, using concentration of the AG from 0 to 100 μg/ml, with or without addition of 0.5% uridine in the medium. C. Neocy5 uptake evaluated by flow cytometry on E. coli growing in synthetic urine medium supplemented or not with uridine 0.5%, expressed as fold change of mean fluorescence per cell compared with no supplementation. D. Time-kill curve of E. coli in urine synthetic medium, after addition of 4 μg/ml of tobramycin, 4 μg/ml of gentamicin or 8 μl/ml of amikacin in the medium supplemented or not with 0.031% or 0.0009% of uridine. Survival was assessed by plating and counting CFU/ml at time 0, 1, 2, 4, 6 and 24 hours after treatment. Detection limit: no colony on the pure culture plate.

Figure 12. Determination of the lowest efficient uridine concentration. Survival of E. coli after 20 hours of treatment with 4 μg/ml tobramycin as a function of concentrations of added uridine (shown in %). (MIC = 10 μg/ml, in synthetic urine). Significance above each point represents the significance of the decrease in survival induced by the addition of the uridine percentage compared with the no uridine condition (0%). For statistical significance calculations, one-way ANOVA was used. *** means p<0.001, * means p<0.05. Number of replicates: n=3.

Figure 13: Impact of nucleosides on the bactericidal effect of aminoglycosides, in synthetic urine. Survival of E. coli K12 strain after 20 hours of treatment with tobramycin 0.5 μg/ml (MIC=10), gentamicin 1 μg/ml (MIC=10), amikacin 6 μg/ml (MIC=50); or P. aeruginosa after treatment with tobramycin 10 μg/ml (MIC=15), supplemented or not with 0.031% of uridine, cytidine, adenosine, thymidine, or inosine. For statistical significance calculations, two way-ANOVA was used, with Bonferroni correction for multiple hypotheses testing. **** means p<0.0001, *** means p<0.001, ** means p<0.01, * means p<0.05, ns: non-significant. Number of replicates for each experiment: n=3. Number of replicates for each experiment: n=3. #: AG resistant strain. MICs are for synthetic urine.

Figure 14: Uridine potentiates tobramycin on Human Plasma Like Medium on E. coli K12. Survival of E. coli K12 after 20 hours of treatment with 1 μg/ml of tobramycin supplemented or not with 0.031% of uridine. The dotted line indicates the limit of detection (no colony on the pure culture plate).

Figure 15. Uridine potentiates AGs in human blood, with AG sensitive strain. Survival of E. coli CFT073 after 1 hour of treatment with 0.1 μg/ml of gentamicin supplemented or not with 0.05% of uridine in human blood. For statistical significance calculations, one-way ANOVA was used. *** means p<0.001. Number of replicates: n=3.

Figure 16. Uridine potentiates AGs in human blood, with AG (amikacin) resistant strain. Survival of E. coli 932 after 1 hour of treatment with 100 μg/ml of amikacin supplemented or not with 0.05% of uridine in human blood. For statistical significance calculations, one-way ANOVA was used. *** means p<0.001. Number of replicates: n=3. The right panel corresponds to the same AMI and AMI+U results as the left panel but in linear scale.

Figure 17: Uridine potentiates AGs on clinical E. coli strains in synthetic urine medium. Survival of susceptible and resistant E. coli strains after 20 hours of treatment with tobramycin, supplemented or not with 0.031% of uridine. Tobramycin concentration is indicated: 10 μg/ml for susceptible strains; 50 or 200 or 400 μg/ml for resistant strains, which are marked with For statistical significance calculations, two way-ANOVA was used. **** means p<0.0001, ** means p<0.01, ns: non-significant. Number of replicates for each experiment: n=3. Number of replicates for each experiment: n=3. #: AG resistant strain. Figure 18. Uridine potentiates AGs in urine in bladder infection model in mice. A. Uridine potentiates the UTI 89 strain in synthetic urine. B. in vivo. CFU/whole homogenized bladder in female C57BI/6 mice infected with the UTI 89 strain intravesically for 24 hours, then treated with PBS, gentamicin, or uridine and gentamicin for an additional 24 hours. For statistical significance calculations, one-way ANOVA was used. ** means p<0.01. ns: non-significant. Results are two pooled experiments with n=5-7/group.

Figure 19. Uridine induces fast killing and prevent the appearance of resistant mutants, in synthetic medium (tryptone). Survival of E. coli WT, growing in mid-exponential phase in liquid cultures supplemented or not with glucose, maltose or uridine 0.5%. MIC in tryptone = lpg/ml. Lethal treatment of tobramycin (10 μg/ml) was applied, and survival was assessed after 1, 2, 3, 4, 6 and 20 hours by plating and counting CFUs/ml. Geometric mean and geometric standard deviation on three biological replicates are represented. For statistical significance calculations, Welch's t-test was used. ** means p<0.01. Number of replicates: n=3 to 7

Figure 20. Neo-Cy5 uptake evaluated by flow cytometry on E. coli growing in synthetic urine medium supplemented or not with uridine 0.5%, expressed as fold change of mean fluorescence per cell compared with no supplementation. For statistical significance calculations, Welch's parametric t- test was used. * means p<0.05. ns: non-significant. Number of: replicates n=3.

Figure 21. Reproduction of the antibiotic lethality assay with pyrimidine supplementation disclosed in WO 2020/227530 and the corresponding scientific publication (Yang et al., Cell, 2019, 177, 1649- 1661). Methods from Yang et al. 2019 was reproduced as follow: E. coli K12 cells were grown overnight in MOPS minimal medium (MM) (Teknova) supplemented with 0.2% of glucose in triplicate. Cultures were diluted 500 fold in 10 ml of MOPS MM + glucose 0.2% in 125 ml baffled flasks. Culture were grown until an OD600nm of ~ 0.3 at 37°C with 300 rpm agitation, and then re-diluted to OD600nm = 0.1 in MOPS MM + glucose 0.2%. For antibiotic treatment, 1 ml of cultures was dispensed in 14 ml tubes and treated with the appropriate antibiotic (48 ng/mL of gentamicin), and biochemical supplementation (1 mM) where indicated. Samples were collected before and after 1, 2, 3 and 4 hours of treatment, serially diluted in PBS and plated for colony enumeration. Results are shown in CFU/ml before and after 1, 2, 3, and 4 hours of treatment. No significant change in gentamicin antibiotic lethality is observed for all tested pyrimidine compounds in the killing assay performed in the presence of glucose. DETAILED DESCRIPTION OF THE INVENTION The uptake of Aminoglycosides (AGs) by carbohydrate transporters, with particular focus on E. coli was investigated. The sugar transporters involved in AG uptake were determined, and it was shown that such uptake is not due to a change in proton motive force. This provides evidence for a previously undescribed mechanism of AGs uptake in Gram-negative bacteria. AGs, possibly because of their osidic structure, are substrates for redundant carbohydrates transporters in Gram-negative pathogens, allowing this molecule to hijack bacterial sugar transport systems. Uridine was identified as a substrate capable of increasing the number of these transporters under laboratory conditions, in rich media and in synthetic human urine. Increasing AG transporters in vivo using a boost with selected carbohydrates may be a way to exploit this new mechanism to potentiate AGs by reducing their effective doses in therapies, and the side effects associated with them. In addition to uridine, other nucleosides such as in particular cytidine, thymidine and inosine also potentiate aminoglycoside (AG) efficiency. The potentiation of aminoglycosides by nucleosides such as uridine is effective in various biological media including synthetic urine, synthetic plasma medium, human blood and in vivo, as shown in a mouse model of urinary tract infection. AG potentiation by nucleosides such as uridine is effective on various pathogens, including the major pathogens Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae and Acinetobacter baumannii. Furthermore, AG potentiation by nucleosides such as uridine is effective on clinical strains resistant to AG. In addition, the combination of AG and nucleoside such as uridine induces fast killing of the bacteria and prevents the selection of resistant mutants. The use of metabolites to resensitize to aminoglycosides is a strategy already employed in other studies ^32,33. The role of the addition of compounds, e.g. mannitol, allows a stimulation of PMF and thus of aminoglycoside entry in persister cells³³. In the present study, sensitive cells are being studied, and the addition of uridine does not alter the PMF indicating that the increase in aminoglycoside entry uses another pathway, although dependent on the presence of a proton gradient. Furthermore, the addition of mannitol has little or no effect on the MIC under the conditions that were used. In addition, the combination of aminoglycoside and nucleoside such as uridine induces fast killing and prevent the appearance of resistant mutants. In contrast, maltose-treated bacteria show a regrowth due to the selection of AG resistant mutants. The present study opens the way to the use of many molecules to potentiate aminoglycosides, with two distinct axes of improvement: (1) maintaining the same treatment efficacy with less aminoglycosides could allow for a reduction in the toxicity threshold, as well as a reduction in the doses of antibiotic ingested and therefore rejected; (2) allowing for a faster eradication of cells (shorter treatment, limitation of the appearance of resistance); (3) re-sensitization of aminoglycoside-resistant bacteria. This is an "old" class of antibiotics, yet it is being re-evaluated. Indeed, WHO predictions concerning the development of MDR bacteria indicate that the coming years will be critical, and each functional molecule will be important. The screening system employed to identify uridine uses the cmtA gene which showed increased induction in the presence of uridine. The use of biolog plates is a first step in the search for activating substrates, but a larger scale compound screen could identify other substrates that could be used in therapy, and why not even combined with each other. The administration of carbohydrates in humans suggests a reduced toxicity compared to other potentiators (e.g. n-butanol ³⁴ which is a primary alcohol).

Investigations were performed concerning MICs according to the carbon source in MH medium (traditionally used in diagnosis for MIC determination) then in 1% bactotryptone, 0.5% NaCI to avoid the presence of complex glucose (starch) in MH. When looking for an active compound, the MIC value induced by the addition of the compound was compared with that of glucose (which corresponds to a repression of the transporters). Indeed, the addition of uridine does not induce any difference in MIC in MH medium (MIC of tobramycin equal to 0.1 μg/ml with or without the addition of 0.5% of uridine) or in MOPS medium supplemented with 0.2 % glucose. Due to the repression of AG transporters in culture medium containing glucose shown for the first time in the present invention, no potentiation of gentamicin with uridine could be observed in the cited prior art (WO 2020/227530 and Yang et al., Cell, 2019, 177, 1649-1661) The unexpected effect of uridine on AG transporters could not be obvious in view of the cited prior art which discloses antibiotic potentiation by stimulation of purine biosynthesis using pyrimidine compounds (pyrimidine bases or nucleosides). Thus, the composition of the culture medium may induce different responses³⁵ call upon regulatory pathways that may antagonize the effect of the tested molecule, or simply not induce a response, which leads to reflect on the relevance of the culture medium used during the screening of molecules. The use of a synthetic urine medium allowed the observation of a difference in a condition with or without uridine, without the need to compare with glucose. Results were also verified in synthetic human plasma-like medium as well as in human blood and in vivo (mouse Urinary Tract Infection (UTI) model). Growth in human urine shows great variability and a recent study suggests the use of a synthetic medium to standardize the results obtained between laboratories ³⁶. The medium used here could provide a basis for future studies.

According to the invention, a nucleoside is used to enhance the killing of a bacteria by an aminoglycoside antibiotic. Accordingly, a combination of the aminoglycoside antibiotic and the nucleoside, is used for treating bacterial infections according to the invention. The combination of the invention which is a combination of (only) two compounds may thus be defined as a combination consisting of the aminoglycoside antibiotic and the nucleoside.

Methods for killing bacteria

The invention encompasses methods for killing bacteria. In one embodiment, the method comprises administering a nucleoside together with an aminoglycoside antibiotic. Preferably, the nucleoside enhances the killing of the bacteria by the antibiotic. More preferably, the nucleoside limits the emergence of antibiotic-resistant bacteria. More preferably, the nucleoside re-sensitizes antibioticresistant bacteria to the killing by the antibiotic.

The effect of the nucleoside on the killing of the bacteria by the antibiotic may be assessed by various assays that are well-known in the art and disclosed in the examples of the present application such as in particular a killing assay or MIC determinations. The effect of the nucleoside on the killing of the bacteria by the antibiotic may be : (i) the increase of the killing of antibiotic-susceptible bacteria, (ii) the reduction of the appearance (or selection) of antibiotic-resistance bacteria, and/or (iii) the killing of antibiotic-resistant bacteria as shown in the examples of the present application.

According to the invention, antibiotic-resistant bacteria include bacteria which are tolerant to an aminoglycoside antibiotic and bacteria which are genetically resistant to an aminoglycoside antibiotic.

In one embodiment, the method is for the prevention of the emergence of antibiotic-resistant bacteria by killing the bacteria faster with a co-administration of the nucleoside together with the aminoglycoside antibiotic. In preferred embodiments, the method is for the prevention of bacteria that are genetically resistant to the antibiotic.

In some embodiments, the method is for killing antibiotic-resistant bacteria. Co-administration of the nucleoside together with the aminoglycoside antibiotic is able to kill antibiotic-resistant bacteria, whereas the same concentration of the aminoglycoside antibiotic used alone is ineffective. In preferred embodiments, the method is for the killing of bacteria that are genetically resistant to the antibiotic.

In one embodiment, the method is for decreasing of the toxicity of a reference treatment by a co- administration of the nucleoside together with the aminoglycoside antibiotic with the aminoglycoside being administrated at a lower concentration than the reference treatment. Preferably, the administration of the nucleoside allows for a 2, 3, 4, 5, 6, 7, 8, 9, or 10-fold reduction in the amount of aminoglycoside antibiotic to achieve the same level of killing as without the nucleoside.

Preferably, the nucleoside and the aminoglycoside antibiotic are administered simultaneously. In some embodiments, the nucleoside is administered 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 16, 20, or 24 hours prior to or subsequent to the aminoglycoside antibiotic.

Nucleosides

Preferably, the nucleoside is uridine. In various embodiments, the nucleoside is selected from uridine, inosine, guanosine, cytidine, adenosine, thymidine, or xanthosine. Preferably, the nucleoside is not adenosine. Preferably, the nucleoside is selected from uridine, inosine, guanosine, cytidine, thymidine, or xanthosine.

In some embodiments, the nucleoside is administered at a dose of 0.5-20 g/m². In various embodiments, the dose of uridine is at least 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12 g/m². In various embodiments, the dose of uridine is up to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, or 12 g/m².

Antibiotics

Preferably, the aminoglycoside antibiotic is selected from paromomycin, amikacin, gentamicin, streptomycin, neomycin, and tobramycin, plazomicin, and kanamycin. Most preferably, the antibiotic is selected from tobramycin, gentamicin, and amikacin.

In various embodiments, the dose of the aminoglycoside antibiotic is from 0.3-30 mg/kg of body weight. In some embodiments, the dose of the aminoglycoside antibiotic is from 1.0 -7.0 mg/kg, 1.0- 5.0, or 1-2.5 mg/kg of body weight. Preferably, the dose of the aminoglycoside antibiotic is reduced to below 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1, 2, or 3 mg/kg of body weight when used with the nucleoside and can preferably still achieve the same level of killing as a higher dose.

In some embodiments, the aminoglycoside antibiotic is administered every 8, 12, 24, or 36 hours. In some embodiments, the aminoglycosides antibiotic in administrated in monodose. In various embodiments, the dose of the aminoglycoside antibiotic is 1 - 2.5 mg/kg/dose every 8 hours or every 12 hours. In various embodiments, the dose of the aminoglycoside antibiotic is 7 mg/kg every 24 hours or every 36 hours. In some embodiments, the dose of the aminoglycoside antibiotic is reduced to below 50% 40%, 30% or 20% of these doses when used with the nucleoside and can preferably still achieve the same level of killing as the higher dose. Preferred combinations include: Tobramycin and uridine; Gentamicin and uridine; Amikacin and uridine; Tobramycin and thymidine; Gentamicin and thymidine; Amikacin and thymidine; Tobramycin and cytidine; Gentamicin and cytidine; Amikacin and cytidine; Tobramycin and inosine; Gentamicin and inosine; Amikacin and inosine.

In some embodiments, when the nucleoside is uridine, the antibiotic is different from gentamicin.

More preferred combinations include: Tobramycin and uridine; Amikacin and uridine; Tobramycin and thymidine; Gentamicin and thymidine; Amikacin and thymidine; Tobramycin and cytidine; Gentamicin and cytidine; Amikacin and cytidine; Tobramycin and inosine; Gentamicin and inosine; Amikacin and inosine.

Bacteria

Preferably, the bacteria that are killed are gram-negative bacilli (aerobic or anaerobic), Staphylococci, or Mycobacterium tuberculosis.

In various embodiments, the bacteria are of the Enterobacteriaceae family, including Escherichia coli, Klebsiella pneumoniae and K. oxytoca, Enterobacter cloacae and E. aerogenes, Providencia spp., Proteus spp., Morganella spp., and Serratia spp. In various embodiments, the bacteria are Yersinia pestis or Francisella tularensis. In various embodiments, the bacteria are Staphylococcus aureus, including methicillin-resistant and vancomycin-intermediate and -resistant isolates, P. aeruginosa, or Acinetobacter baumannii. In various embodiments, the bacteria are of the Mycobacterium spp., including Mycobacterium tuberculosis, M. fortuitum, M. chelonae, and M. avium. In various embodiments, the bacteria are Streptococcus spp., including S. pneumoniae, S. pyogenes, S. gallolyticus, S. saprophyticus, S. agalactiae. In various embodiments, the bacteria are Vibrio cholerae.

Preferably, the bacteria are selected from Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, and Acinetobacter baumannii.

In some embodiments, the bacteria as disclosed herein comprise or consist of bacteria that are resistant to the aminoglycoside antibiotic, wherein the resistant bacteria may be tolerant or genetically resistant to the antibiotic; preferably wherein the resistant bacteria comprise genetically resistant bacteria.

In some preferred embodiments, a combination chosen from Tobramycin and uridine, Amikacin and uridine, Tobramycin and thymidine, Gentamicin and uridine, Gentamicin and thymidine, Amikacin and thymidine, Tobramycin and cytidine, Gentamicin and cytidine, or Amikacin and cytidine is used for killing E. coli, for instance in blood infections or urinary tract infections; preferably the combination is chosen from Tobramycin and uridine, Amikacin and uridine, Tobramycin and thymidine, Gentamicin and thymidine, Amikacin and thymidine, Tobramycin and cytidine, Gentamicin and cytidine, or Amikacin and cytidine . In some preferred embodiments, a combination chosen from Tobramycin and uridine, Tobramycin and adenosine, Tobramycin and thymidine, Gentamicin and uridine orTobramycin and inosine is used for killing P. aeruginosa, for instance during lung infections; preferably the combination is chosen from Tobramycin and uridine, Tobramycin and adenosine, Tobramycin and thymidine or Tobramycin and inosine.

Administration

In various embodiments, the nucleoside and/or antibiotic is administered systemically or to the urinary tract, gastrointestinal tract, lungs, eyes, ears, nose, brain, heart, blood or skin. Preferably, the nucleoside and/or antibiotic is administered systemically or to the lungs, eyes, ears, nose, or skin.

In various embodiments, the administration is through parenteral route such as subcutaneous (s.c.), intradermal (i.d.), intramuscular (i.m.), intraperitoneal (i.p.) or intravenous (i.v.) injection. In various, embodiments, the administration is through inhalation, oral, topical, ocular, rectal, or vaginal routes.

In various embodiments, the nucleoside and/or antibiotic is administered in one or multiple administration dose(s).

In various embodiments, the level of glucose in the bacterial environment is minimized. Preferably the patient does not have hyperglycemia or renal glucosuria,

The quantity to be administered (dosage) depends on the subject to be treated, including the condition of the patient, the state of the individual's immune system, the route of administration and the size of the host. Suitable dosage ranges can be determined by the skilled artisan and can be modified by one skilled in the art, depending on circumstances.

The invention encompasses different therapeutic applications. First of all, an application to the treatment of urinary tract infections. In synthetic urine medium, it has been shown that the concentration of uridine that induced the most important potentiating effect was 0.031% or 1.27 mM. Consumption of purine-rich beverages increases urinary uridine excretion from 0.21 μM to 0.23 μM, which is still far too low to reach the concentrations shown to be effective (0.031%) ³⁷ The simple ingestion of purine-enriched products could therefore not be sufficient to achieve a sufficient concentration of uridine in the bladder to observe a potentiating effect on aminoglycosides.

In animal models, in rats, intravenous administration of 0.5g/kg causes a peak of 13 mM uridine in plasma, but induces a decrease in blood pressure within 1 minute of injection³⁸. In rabbits, 7.7% of intravenously injected uridine is recovered in the urine compared to only 1% when taken orally. The plasma peak is about 25 μg/ml or 100 mM for 0.1 g/kg ³⁹. These values show a great variability in animal models that must be taken into account during pre-clinical studies.

In humans, the values for urinary excretion in particular seem to be more favorable. In order to counteract the toxicity of fluorouracil used in chemotherapy, uridine can be administered and studies have been conducted on the PK/PD and toxicity of the molecule. In a first study, uridine was administered as an intravenous infusion in doses ranging from 1 to 12 g/m² (it is estimated that an average human has a body surface area of 1.5 m²). The concentration of uridine in plasma thus increased from 1-8 μM to a peak of 2 mM after an infusion of 1 hour at 8-12 g/m². It was also shown that that 24% of the uridine was excreted in the urine within 24 h of treatment. Regarding toxicity, 1 out of 2 patients who received the highest dose of uridine shivered for 15 minutes 1 h after the end of the infusion, without fever or any neurological impairment. A second study involving prolonged or intermittent exposure to uridine also provides interesting data: prolonged exposure causes the appearance of fever, which is not the case when administration is spaced 3 h apart. Uridine is rapidly eliminated from plasma, from mM to μM between treatments. Urinary excretion has been estimated at 15-40% ⁴⁰.

Taking the low value of 15% uridine excretion in urine: if 12 g uridine (8 g/m²) is injected in a 1 h infusion, it can be estimated that 1.8 g will be excreted in urine. If a large volume of 2 liters in the bladder is considered, then 0.9 g/L or 0.09% of uridine can potentially be found in the urine. One can estimate that a dose of 0.031% induces a strong potentiating effect. These estimates are based on two very informative studies - with about 20 patients in all.

Oral administration of uridine is already practiced in the clinic. For this type of administration, the limitation of the dosage (between 8 and 12 g/m²) is caused by the occurrence of diarrhea. The dose of uridine in plasma following the oral ingestion can reach a value between 60 and 80 μM - 10X less than by intravenous route, associated with a very low urinary excretion (1%) ⁴⁰. Uridine triacetate granules are already marketed under the name Vistogard (https://www.vistogard.com/Professional/Data/Pre-Clinical). Ingestion of PN401, a drug composed of uridine, causes a peak of uridine at 200 μM in plasma for an intake of 9.9 g⁴¹, but no information on urinary excretion. The best approach to increasing concentrations therefore may be intravenous injection, as with the antibiotic. If infusion is better tolerated, a "shift" of treatments could be implemented, first by uridine infusion and then by injection of the aminoglycoside. Preliminary results still suggest that even the lowest dose of uridine tested (0.0009% or 0.036 mM) is effective but less than 0.031% when the inoculum is more important. This could be due to the consumption of uridine, added in the medium in limited quantity. Thus, a daily intake of uridine could be sufficient to see a potentiating effect.

Uridine could be used for the treatment of pyelonephritis (to be defined according to the accumulation in the kidneys), endocarditis, and also in the context of local treatments of eye infections, otitis, skin infections in which one would not be limited by the concentration of uridine since it could be administered locally. Recently, a tobramycin inhalation therapy has been developed to treat cystic fibrosis patients with S. aureus or P. aeruginosa infection⁴². Like mannitol ⁴³, or uridine triphosphate⁴⁴, uridine could be administered by inhalation to potentiate tobramycin on P. aeruginosa under conditions of pulmonary infection. Indeed, the too low dosage of aminoglycosides in nebulization is the cause of the failure of such treatments. Finally, uridine could be used in the case of blood infections, since the administration of a unique dose of aminoglycosides is preconized in patients hospitalized for sepsis, before further treatment with other antibiotics.

Compositions and uses thereof

The invention encompasses compositions for killing bacteria comprising a nucleoside and an aminoglycoside antibiotic, wherein the nucleoside enhances the killing of the bacteria by the antibiotic as disclosed herein. Preferably, wherein the nucleoside limits the emergence of antibioticresistant bacteria. Preferably, wherein the nucleoside re-sensitizes the antibiotic-resistant bacteria to the killing by the antibiotic.

In various embodiments the nucleoside is selected from uridine, inosine, guanosine, cytidine, adenosine, thymidine, or xanthosine. In some embodiments the nucleoside is selected from uridine, inosine, guanosine, cytidine, thymidine, or xanthosine. Preferably, the nucleoside is uridine.

Preferably, the aminoglycoside antibiotic is selected from paromomycin, amikacin, gentamicin, streptomycin, neomycin, tobramycin, plazomicin, apramycin and kanamycin. Most preferably, the antibiotic is selected from tobramycin, gentamicin, and amikacin.

In various embodiments, the nucleoside is not adenosine and/or the aminoglycoside antibiotic is not gentamicin.

The invention encompasses the use of a nucleoside to enhance the killing of the bacteria by an aminoglycoside antibiotic.

In various embodiments the nucleoside is selected from uridine, inosine, guanosine, cytidine, adenosine, thymidine, or xanthosine. In some embodiments, the nucleoside is selected from uridine, inosine, guanosine, cytidine, thymidine, or xanthosine. Preferably, the nucleoside is uridine. Preferably, the aminoglycoside antibiotic is selected from paromomycin, amikacin, gentamicin, streptomycin, neomycin, tobramycin, plazomicin, apramycin and kanamycin. Most preferably, the antibiotic is selected from tobramycin, gentamicin, and amikacin.

In various embodiments, the nucleoside is not adenosine and/or the aminoglycoside antibiotic is not gentamicin.

In various embodiments, the administration occurs without the co-administration of glucose. Thus, in some embodiments, the glucose levels are minimimized.

EXAMPLES

1. MATERIAL AND METHODS

1.1 Strains, plasmids and primers

The strains, plasmids and primers used in the study are presented in Table 1.

TABLE 1: STRAINS, PLASMIDS, AND PRIMERS

STRAINS

Strains Number Genotype of interest Origin

Escherichia coli K12 subs. MG1655 C349 WT Lab collection

Vibrio cholerae N16961 7805 WT Lab collection

Pseudomonas aeruginosa PAO1 N065 WT Lab collection

Klebsiella pneumoniae NTUH K2044 N105 WT Liver abscess, Wu et al 2009

Acinetobacter baumannii ATCC 19606 1276 WT Lab collection

Clinical strains:

E. coli 886 rectal

E. coli 932 aac(6')-lb3, aac(6')lb-cr, aadAl, sull_2, sul2_2, dfrAl rectal

E. coli 1193 blaCTX-M-14_l, blaTEM-lB_l, blaOXA-l_l, ant(3")-la_l, aac(3)-lld_l, aadA5_l, sull_2, dfrA17_l, aph(3')-la_l, mph(A)_l, mph(A)_2, tet(A)_4, catAl_l cutaneous

E. coli 1195 blaCTX-M-l_6, blaTEM-lB_l, aadA2_2, sull_2, sul2_2, dfrA12_l, aph(3")- lb_5;aph(3')-la_l; aph(6)-ld_l, mph(A)_l; mph(A)_2, tet(A)_4 feces

E. coli 1215 ant(3")-la_l, aac(6')-llc_l, sull_2, tet(A)_4 cutaneous

E. coli 1236 blaTEM-lB_l, blaSHV-2_2, aadA5_l, sull_2, sul2_2, sul2_2 cutaneous E. coli 1238 ant(2")-la_18; ant(3")-la_l, sull_2, sul2_2, aph(3")-lb_5; aph(6)-ld_l feces

E. coli Ec019 blaCTX-M-14, mdfA_l) anal

E. coli Ec068 mdfA_l) colostomy

Deletions in E.coli : Δcrp Δera P940 This study ΔcmtA 0056 This study ΔfruA N812 This study ΔchbC 0066 This study ΔlamB N810 This study ΔmalE N811 This study ΔfrwB N813 This study ΔfrvB N814 This study ΔptsG N815 This study ΔtreB N816 This study ΔmalX N818 This study ΔmtIA N819 This study ΔmglA N820 This study ΔmngA N821 This study ΔypdG N822 This study ΔsgcA N823 This study ΔsrlE N824 This study ΔascF N825 This study ΔagaW 0064 This study ΔbgIF 0067 This study ΔsgcC 0068 This study ΔmanY 0069 This study ΔgalP 0070 This study ΔgatC 0071 This study ΔnagE 0072 This study ΔglvC 0073 This study Δudk R060 This study Δudp R058 This study

GFP fusions:

MG1655 + psclOl pcmtA-GFP 0537 This study MG1655 Δcrp + psclOl pcmtA-GFP 0538 This study

MH1655 Δera psclOl pcmtA-GFP Q115 This study

MG1655 ΔcmtA + psclOl pcmtA-GFP P315 This study

MG1655 + psclOl-btuBGFP R051 This study

MG1655 + psclOl plrrnB-GFP R921 This study

MG1655 + psclOl pfruA-GFP P115 This study

MG1655 big colonies R796, R797, R798 fusA 1779T>G , rpIL 122_127delTAGCTG This study

MG1655 small colony R799 fusA 2015C> T, rpIL 122_127delTAGCTG This study

MG1655 small colony R800 fusA 2011C> T This study

Overexpressions P. aeruginosa:

P. aeruginosa + pO P845 This study

P. aeruginosa pSEVA-238 mtlFGK+ P847 This study

P. aeruginosa pSEVA-238 gtsB+ P848 This study

P. aeruginosa pSEVA-238 oprB+ P849 This study

P. aeruginosa pSEVA-238 PA2291 R434 This study

P. aeruginosa pSEVA-238 fruA+ S298 This study

Overexpression A. baumannii:

A. baumannii pSEVA-238 p0 R978 This study

A. baumannii pSEVA-238 fruA+ R980 This study

Overexpression E. coir.

MG1655 + pSEVA-238 p0 0897 This study pSEVA-238 cmtAB+ P134 This study pSEVA-238 fruBKA+ P314 This study pSEVA-238 malEFG+ P132 This study pSEVA-238 manXYZ+ P313 This study pSEVA-238 lamB+ P311 This study pSEVA-238 frwBC+ P715 This study pSEVA-238 mtlA+ P716 This study pSEVA-238 malX+ P717 This study pSEVA-238 mngA+ P718 This study pSEVA-238 glpTQ+ P719 This study pSEVA-238 bglF+ P720 This study pSEVA-238 ypdGH+ P721 This study pSEVA-238 bglH+ P722 This study pSEVA-238 treB+ P723 This study pSEVA-238 galP+ Q144 This study pSEVA-238 gatABC+ Q145 This study pSEVA-238 chiP+ Q146 This study pSEVA-238 chbCBA+ Q147 This study pSEVA-238 srlAEB+ Q148 This study pSEVA-238 acsF+ Q149 This study pSEVA-238 xylEFG+ Q169 This study pSEVA-238 nupC+ Q872 This study pSEVA-238 nupG+ Q873 This study pSEVA-238 frvAB+ Q168 This study pSEVA-238 ptsG+ Q257 This study

AcmtA + pSEVA238 cmtA+ Q322 This study

MG1655 ΔcmtA + pSEVA238 p0 Q321 This study

PLASMIDS

Plasmid Number Construction pSEVA-238 M027 psclOl 0896 Overexpression for plasmids for E. coli: pSEVA-238 cmtAB+ P134 PCR on gDNA using primers ML 290/291. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 fruBKA+ P314 PCR on gDNA using primers ML 295/296. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 malEFG+ P132 PCR on gDNA using primers ML 299/300. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 manXYZ+ P313 PCR on gDNA using primers ML 301/302. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 lamB+ P311 PCR on gDNA using primers ML 297/298. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 frwBC+ P715 PCR on gDNA using primers ML 311/312. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 mtlA+ P716 PCR on gDNA using primers ML 303/304. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 malX+ P717 PCR on gDNA using primers ML 315/316. Ligation into pSEVA-238 between Kpnl and Xbal restriction sites. pSEVA-238 mngA+ P718 PCR on gDNA using primers ML 307/308. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 glpTQ+ P719 PCR on gDNA using primers ML 317/318. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 bglF+ P720 PCR on gDNA using primers ML 323/324. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 ypdGH+ P721 PCR on gDNA using primers ML 309/310. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 bglH+ P722 PCR on gDNA using primers ML 321/322. Ligation into pSEVA-238 between Kpnl and Xbal restriction sites. pSEVA-238 treB+ P723 PCR on gDNA using primers ML 313/314. Ligation into pSEVA-238 between Kpnl and Xbal restriction sites. pSEVA-238 galP+ Q144 PCR on gDNA using primers ML 339/349. Ligation into pSEVA-238 between Xbal and Ptsl restriction sites. pSEVA-238 gatABC+ Q145 PCR on gDNA using primers ML 343/344. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 chiP+ Q146 PCR on gDNA using primers ML 333/334. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 chbCBA+ Q147 PCR on gDNA using primers ML 331/332. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 srlAEB+ Q148 PCR on gDNA using primers ML 337/338. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 acsF+ Q149 PCR on gDNA using primers ML 335/336. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 xylEFG+ Q169 PCR on gDNA using primers ML 325/326. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 nupC+ Q872 PCR on gDNA using primers ML 349/350. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 nupG+ Q873 PCR on gDNA using primers ML 351/352. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 frvAB+ Q168 PCR on gDNA using primers ML 327/328. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites. pSEVA-238 ptsG+ Q257 PCR on gDNA using primers ML 345/346. Ligation into pSEVA-238 between Ecorl and Xbal restriction sites.

Overexpression plasmids for P. aeruginosa: pSEVA-238 mtlFGK+ P828 PCR on gDNA using primers ML 365/366. Ligation into pSEVA-238 between BamHI and Xbal restriction sites. pSEVA-238 gtsB+ P829 PCR on gDNA using primers ML 355/356. Ligation into pSEVA-238 between EcoRI and Xbal restriction sites. pSEVA-238 oprB+ P831 PCR on gDNA using primers ML 359/360. Ligation into pSEVA-238 between EcoRI and Xbal restriction sites. pSEVA-238 PA2291 R433 PCR on gDNA using primers ML 414/415. Ligation into pSEVA-238 between EcoRI and Xbal restriction sites. pSEVA-238 fruA+ R963 PCR on gDNA using primers ML 363/364. Ligation into pSEVA-238 between EcoRI and Xbal restriction sites.

Overexpression plasmids for A. baumannii pSEVA-238 fruA+ R962 PCR on gDNA using primers ML 431/432. Ligation into pSEVA-238 between EcoRI and Xbal restriction sites.

Fusions with GFP: psclOl-pcmtAGFP 0501 PCR on gDNA using primers ML 252/253. Ligation into pTOPO-TA. Digestion with EcoRI and subcloning into psclOl. psclOl-plrrnBGFP R692 Fruchard et al., 2022 (tgt) psclOl-pfruAGFP P115 PCR on gDNA using primers ML 261/262. Ligation into pTOPO-TA. Digestion with EcoRI and subcloning into psclOl. psclOl-pbtuBGFP R051 PCR on gDNA using primers ML 411/412. Ligation into pTOPO-TA. Digestion with EcoRI and subcloning into psclOl.

PRIMERS

1.2 Growth conditions

MH medium was used to determine the MIC of the deletion or overexpression strains. MOPS Rich (Teknova EZ rich defined medium) was used for the screening assay to avoid intrinsic fluorescence of other media, and for the neocy5 assay. For the study of the specific-substrate responses, the substrate was added to the medium containing 1% bactotryptone and 0.5% NaCI. To mimic urinary tract infection, the urine synthetic medium was prepared according to ⁴⁵. To mimic the human plasma, the HPLM medium formulated to resemble the natural cellular environment found in the body, mimicking the metabolic profile of human plasma (Human Plasma Like Medium, Thermofischer Scientific) was used. The five species were grown at 37°C with shaking (100-150 rotations per minutes). 1.3 Genes deletions

All E. coli strains used in this work are derivatives of E. coli MG1655 and were constructed by transduction using Keio knockouts strains. Kanamycin resistance cassette aph was removed using FLP/FRT system ⁴⁶.

1.4 Overexpression of transporters

Overexpressions were performed by cloning the genes of interest into the pSEVA-238 vector ⁴⁷ under the dependence of the Pm promoter, using 1 mM of sodium benzoate as inducer ⁴⁸. The primers and restriction sites associated used for cloning are listed in Table 1.

1.5 MIC evaluation

Etest: Stationary phase cultures in MH were diluted 20X in PBS except for the Δcrp strain, 300 μl were plated on the appropriate medium: MH for genes deletions; MH supplemented with kanamycin and sodium benzoate for maintenance and induction of the plasmid in overexpression strains; 1% amino acid, 0.5% NaCI and 0.5% of substrate (eg. Glucose, ribose, uridine ...) to assess the impact of carbon sources. Plates were then dried for 10 minutes. Etests (Biomerieux) were placed on the plates and incubated overnight at 37°C.

Liquid cultures: MICs were determined by the microtiter broth dilution method with an initial inoculum size of 5.10⁵ CFUs/ml. In MH medium, the MIC was interpreted as the lowest antibiotic concentration preventing visible growth. In urine synthetic medium, growth being difficult to interpret with naked eye, 5μl of pure culture of each antibiotic dilution were plated and MIC was interpreted as the lowest concentration preventing growth.

1.6 RNA-seq

Cultures of E. coli were diluted 1000X and grown in triplicate in MH supplemented or not with 0.1 μg/ml of tobramycin, corresponding to 25% of the MIC in liquid cultures to an OD_600nm of 0.4.

Cultures of E. coli, P. aeruginosa, K. pneumoniae and A. baummannii were diluted 1000X grown in triplicate in MH supplemented or not with 0.5% of uridine to an OD_600nm of 0.4.

All RNAs were purified with the RNAeasy mini kit (Qiagen) according to manufacturer instruction. Briefly, 4 ml of RNA-protect (Qiagen) reagent were added on 2 ml of bacterial cultures during 5 minutes. After centrifugation, the pellets were conserved at -80°C until extraction. Protocol 2 of the RNAprotect Bacteria Reagent Handbook were performed, with addition of a proteinase K digestion step, such as described in the protocol 4. Quality of RNA was controlled using the Bioanalyzer. Count data were analyzed using R and the Bioconductor package DESeq2. Data were normalized with DESeq2. Raw p-values were adjusted for multiple testing according to the Benjamini and Hochberg procedure and genes with an adjusted p-value lower than 0.005 were considered differentially expressed.

1.7 Neocy5 uptake assay

Quantification of fluorescent neomycin (Neo-cy5) uptake was performed as described ⁴⁹. Neo-cy5 is an aminoglycoside (neomycin) coupled to the fluorophore Cy5 that retained activity and mode of uptake in Gram-negative ⁵⁰. Overnight cultures were diluted 100X in rich MOPS (Teknova EZ rich defined medium). When the bacterial cultures reached an OD₆₀₀ of 0.25, they were treated with 0.4 μM of Cy5 labeled Neomycin for 15 minutes at 37°C under aluminium foil. For the assays with different substrate, cultures were washed once with PBS before treatment. 20 μl of each treated culture were then used for flow cytometry, diluted in 200 μl of PBS before reading fluorescence. Flow cytometry experiments were performed as described ⁵¹. For each experiment, 50000 events were counted on the Miltenyi MACSquant device.

1.8 Evaluation of PMF

Quantification of PMF was performed using the Mitotracker Red CMXRos dye (Invitrogen)⁵², in parallel with the neocy5 uptake assay, using the same bacterial cultures. 50 μl of each culture were mixed with 60 μl of PBS. Tetrachlorosalicylanilide TCS (Thermofischer), a protonophore, was used as a negative control with a 500 μM treatment applied for 10 minutes at room temperature. Then, 25 nM of Mitotracker Red were added to each sample and let at room temperature for 15 minutes under aluminium foil. 20 μl of the treated culture were then used for flow cytometry, diluted in 200 μl of PBS before reading fluorescence. Flow cytometry was performed as described above.

1.9 Carbon sources screening assay

The GFPmut3 was fused to promoter of interest ⁵³ and cloned into a plasmid pSClOl.For the first screening, overnight cultures of the strain carrying the screening system were diluted 200X in MOPS Rich (Teknova EZ rich defined medium) supplemented with carbenicillin for plasmid maintenance. The phenotype Microarray (Biolog) plates PM1PM2B and PM3Bwere used for substrate screening. Each well was filled with 100 μl of inoculated media and mixed by pipetting. Media were transferred to 96 well dark-bottom plates (Thermo Scientific). GFP fluorescence was followed on the Tecan Infinite 200 PRO (Life Science) at 37°C during 8h. Fluorescence induction by the substrate was calculated using the ratio fluorescence (t8h-t0h) over growth (t8h-t0h OD_600nm). For flow cytometry quantification, overnight cultures of strain carrying the screening system were diluted 200X in rich MOPS (Teknova EZ rich defined medium) supplemented with carbenicillin for plasmid maintenance and grown overnight, and the substrate tested at 0.5%. Fluorescence was read on 5 μl of cultures diluted in 200 μl of PBS.

1.10 Stringent response

The P1rrnB-gfp fusion was constructed using gfp ASV, and cloned into a plasmid pSClOl. Cultures were grown overnight (positive control) or diluted 100X until reaching an OD₆₀₀nm of 0.4, in MH medium supplemented with carbenicillin for plasmid maintenance and 0.5% of the mentioned substrate or 0.1 μg/ml of tobramycin. Fluorescence was read by flow cytometry, on 20 μl of cultures diluted in 200 μl of PBS.

1.11 Killing assay

On bactotryptone medium: Overnight cultures were diluted 1000X in 25 ml of medium containing 1% bactotryptone and 0.5% NaCI, supplemented or not with glucose, maltose or uridine 0.5%. Cultures were grown to an OD_600nm of 0.3-0.4, and a 5 ml aliquot was treated with a lethal concentration of tobramycin (10 μg/ml). Cultures were plated at 0 h (to), 1, 2, 4, 6 and 20 hours after treatment, and survival was calculated by counting CFUs/ml after treatment divided by the initial number of CFUs/ml (to).

On synthetic urine medium: Overnight cultures in MH was diluted 10-fold in PBS. Approximately 5.10⁶ or 5.10⁷ CFUs/ml were diluted in 200 μl of urine-like medium containing or not uridine, cytidine, thymidine or inosine, and treated with AGs, in a 96-well plate incubated at 37°C with 100-120 rpm. Cultures were plated at 0 h (to), 1, 2, 4, 6 and 24 hours after treatment or only 20 hours of treatment, and survival was calculated by counting CFUs/ml after treatment divided by the initial number of CFUs/ml (tO).

1.12 Dose-response assay

Two fold dilutions of uridine in urine synthetic media were prepared into lines of a 96-well plate, from 2% to 0.0018%. The last column was only filled with synthetic urine medium without uridine. Then, 100 μl of a solution containing 2X concentration of the tested antibiotic was added, thus diluted uridine and antibiotics by two and increasing the total volume per well at 200 pl. The last line was only filled with urine synthetic media without antibiotic. Each well was then inoculated with approximately 5.10⁶ CFUs from MH stationary phase culture, and incubated 20 hours at 37°C with agitation (100 rpm). CFUs were count by plating before and after treatment on MH medium. 1.13 Whole genome sequencing gDNA was extracted from 500 μl of overnight cultures in MH, using the Blood and Tissue Extraction Kit (Qiagen) according to the manufacturer instructions. The presence of variants (Single Nucleotide Polymorphism) was analyzed with SnpEff 5.0⁵⁴.

1.14 Urinary Tract Infection model

Urinary tract infection was induced in 6-7 week old female C57BL/6J mice from Charles River, France as previously described [101], Briefly, the human UPEC cystitis isolate UTI89, engineered to express the fluorescent protein RFP and antibiotic resistance to kanamycin (UPEC-RFP) [101] was grown statically in Luria-Bertani (LB) broth for 18 hours at 37°C in the presence of kanamycin (50 μg/ml). Cultures were adjusted to 2 x10⁸ CFU/mL in PBS and 50 μL (10⁷ CFU/mouse) was delivered via catheter directly into the bladder of mice anesthetized by intraperitoneal injection of 100 mg/kg ketamine and 5 mg/kg xylazine. At 24 hours post-infection, mice were treated with 100 μL PBS, 0.2 mg/kg gentamicin or 0.2 mg/kg gentamicin + 0.5 g/kg uridine by retro-orbital intravenous injection. Mice were sacrificed at 48 hours post-infection (24 hours post treatment) by cervical dislocation after isoflurane inhalation. To calculate CFU, bladders were aseptically removed and homogenized in 1 ml of PBS. Serial dilutions were plated on LB agar plates with kanamycin. All animals used in this study had free access to standard laboratory chow and water at all times. Infections were conducted at Institut Cochin in accordance with approval of APAFIS #34290 by SC3 - CEEA34 - Universite de Paris Cite, at Institut Cochin, in application of the European Directive 2010/63 EU.

1.15 Human blood infection

This study was carried out by QIMA Life Science (1 bis rue des plantes 86160 GENCAY- France). Blood came from a 39-year-old male donor. E. coli CFT073 (ATCC® 700928™), or strain 932, were grown overnight in MH, and then diluted to an OD₆₀₀nm of 1 in PBS. Blood (500 pl) was inoculated with 2.10⁴ CFU/ml (2X), and treated with uridine (0.05%), gentamicin for the susceptible strain CFT073 (0.1 μg/ml, or amikacin for the resistant strain 932 (100 μg/ml), or the combination of uridine and the antibiotic (500 μl mix; prepared in blood, 2X). Blood containing or not the compounds and/or E. coli bacteria was then incubated for 1 hour at 37°C under agitation (150 rpm). Bacterial enumeration was carried out before (to) and after culture on the treated blood of each condition. For each condition, treated and untreated blood was plate onto two MH agar plates (100 μl and the pellet of the remaining volume of blood (900 pl)). 1.16 Statistical analysis

F-test was performed in order to determine whether the variances were equal or different between conditions. For conditions with equal variance, Student's t-test was used. For conditions with significantly different variances, Welch correction was applied. One-way ANOVA or two-way ANOVA were used for multiple comparisons. GraphPad Prism was used to determine the statistical differences between groups. **** means p<0.0001, *** means p<0.001, ** means p<0.01, * means p<0.05. Number of replicates for each experiment was 3<n<7. Means and standard deviations for growth curves and survival rate, means and geometric means for logarithmic values were calculated using GraphPad Prism.

2. RESULTS

2.1 cmtA deletion decreases susceptibility to AGs

Deletion mutants of carbohydrates transporters in E. coli were constructed and the impact of a single deletion on tobramycin susceptibility was tested. For multi-component systems, the protein located at the membrane was deleted. For the 30 mutants tested (Table 3), the response to tobramycin was assessed using serial dilution assays and Etests (Table 3). One of those deletions, CmtA, showed a phenotype of decreased susceptibility to tobramycin: the deletion strain showed a 4-fold increase in the Minimal Inhibitory Concentration (MIC) than the wild type (WT) strain (Figure 1A), and decreased susceptibility in the presence of tobramycin at 4xMIC (Figure IB). CmtA has 52% similarity to MtlA⁵⁵, and is annotated as a cryptic mannitol E 11 C PTS enzyme. CmtA has the ability to complement mannitol transport in a ΔmtIA strain only when expressed under a heterologous promoter⁵⁶. An increase in MIC of ΔcmtA with other AGs (kanamycin, gentamicin) was also observed, but not with antibiotics from other families, such as trimethoprim, ciprofloxacin, amoxicillin or chloramphenicol (Table 1). Susceptibility of ΔcmtA was also unchanged for spectinomycin, an aminocyclitol antibiotic with a chemical structure close to that of AGs, but without aminosugars or glycosidic bonds (Table 1). The decreased susceptibility of AcmtA was thus specific to AGs. The other transporter deletion mutants showed either no impact on the serial dilution test or a slight increase in the MIC of tobramycin (Table 3), which is consistent with specificity towards AGs.

Deletion of Crp, the primary regulator of carbon catabolic repression and transcriptional activator of non-preferential sugar transporters⁵⁷, has previously been linked to streptomycin tolerance⁵⁸. The Acrp strain showed a 10-fold higher MIC than the WT strain despite a significant growth defect (Figure 1A), and no susceptibility to a 0.4 μg/ml tobramycin treatment (Figure 1B). The effect of each transport systems was tested by overexpressing them in trans. 2.2 Overexpression of various carbohydrate transporters increase susceptibility to AGs

First, the cmtAB PTS transporter gene was cloned on a plasmid under the control of an inducible promoter to assess its response to tobramycin. While deletion of this transporter increased the AGs concentrations required to kill bacteria (increased MIC), its overexpression increased susceptibility to tobramycin and gentamicin, and not to spectinomycin (Figure 1C and Table 2). The response was specific to AGs, since there was no effect on ciprofloxacin, chloramphenicol, trimethoprim and amoxicillin used as negative controls (Table 2).

Next, the 23 carbohydrate transporters that were identified were cloned on the inducible plasmid. Survival by serial dilution were first determined on media containing tobramycin, gentamicin, chloramphenicol and carbenicillin. The MIC was then evaluated for strains with reduced tolerance to AGs (Table 3). Of the overexpressed carbohydrate transporters, 16 showed increased susceptibility to tobramycin upon overexpression, compared to the empty vector (Figure 2A and Table 3). Growth of the frwBC+ strain was more strongly impaired than the empty plasmid control in serial dilution tobramycin containing-media (Figure 2A) but did not exhibit a lower MIC to tobramycin than pO. Overexpression of the lamB porin at the outer membrane showed no response to AGs, nor did chiP (chitin porin), ptsG (glucose PTS), xylEFG (Major Facilitator Superfamily (MFS) protein of xylose, galP (MFS protein of galactose), frvAB (fructose-like PTS) and mtlA (Mannitol PTS) (Figure 2A). These increased susceptibility profiles were AGs-specific for 11 transporters: cmtAB, chbCB (chitobiose PTS), srIEAB (glucitol PTS), ascF (cellobiose/arbutine PTS), malEFG (maltose ABC transporter), fruBKA (fructose PTS), frwBC (fructose-like PTS/ mngA (fructose-like PTS), ypdGH (fructose-like PTS bgIF (Pglucoside PTS), and malX (maltose PTS) (Figure 2C). Strains overexpressing manXYZ (mannose PTS/ treB (trehalose PTS/ bgIH (beta glucoside porin), gatABC (galacticol PTS) and glpTQ (glycerol-3- phosphate permease already known to favor uptake of fluoroquinolone⁵⁹) showed a phenotype of susceptibility to AGs , but also to ciprofloxacin or carbenicillin (Figure 2B).

Table 2: MIC (μg/ml) to different antibiotic determined by Etest 2.3 Carbohydrate transporters promote AG uptake.

To verify whether the effect of carbohydrate transporters expression was related to differential uptake, the intracellular uptake of Neo-cy5 was monitored, a fluorescent AG synthesized for uptake studies in bacteria, which carries the properties of AGs for uptake, mode of action and activity against Gram-negative bacteria ⁵⁰, as previously performed⁴⁹. It was found that cmtA deletion decreased the amount of fluorescence inside the cell after treatment, meaning that the uptake of AG is decreased (Figure 3). The Δcrp mutant showed a greater decrease of fluorescence, consistent with the increased MIC value (Figure 1 and Figure 3A). As a corollary, it was next tested whether overexpression of CmtAB leads to increased entry of Neo-cy5. No difference was observed between the WT strain carrying pO and the WT strain carrying the pcmtAB+ overexpression vector during the 15 minutes treatment (not shown). This may be due to the fact that the membrane is already saturated with CmtA during this observation window. Thus, the AcmtA strain carrying the empty plasmid or overexpressing CmtA was tested, which confirmed that Neo-cy5 uptake was increased in the presence of CmtA (Figure 3).

Together, these observations indicate that redundant transporters are involved in AGs uptake, and that increasing the expression of these transporters could potentiate AGs therapeutically.

2.4 CmtA as a screening tool to look for activators of AG transporters

The transcriptome of growing E. coli was studied with or without tobramycin. Transcriptome analysis under sub-MIC (25% of the MIC) tobramycin treatment in MH medium showed no induction of the carbohydrates transporters. Some transporters were even repressed upon addition of tobramycin such as mglA (galactose permease), the PTS malX, manXY, nagE, or rbsC (ribose permease) (Table 4).

In order to define the conditions favoring AG uptake by sugar transporters, a screening system was constructed based on the fusion between the promoter of the cmtAB operon and gfp, carried by a plasmid. In the mutant deleted for crp, fluorescence was lost, confirming that Crp is a positive regulator of this transport system (Figure 4). In the strain lacking cmtA, fluorescence showed a 2-fold increase, suggesting a potential positive feedback control (Figure 4). Among the regulators related to sugar utilization, the Cra protein has also been described as a major player: this protein is involved in the repression of genes subject to catabolite activation such as fruBKA or mtlADR ⁶⁰. Deletion of cra also increased fluorescence by a factor of 2. As a corollary, addition of glucose, which leads to catabolic repression of non-glucose transporters and increases the AG MIC^{61 62 63} (Table 3), strongly decreases fluorescence. Mannitol, as proposed ⁵⁶ was not able to increase cmtA expression (Figure 4). These observations validate the use of cmtA as a tool to screen for conditions inducing of the expression of the transporter. Table 3 : MIC (μg/ml) determined by Etest

2.5 Uridine is able to activate expression of CtmA

In order to search for carbohydrate sources which could activate expression from the cmtA promoter, two Biolog 96 well plates (PMl and PM2B) were used containing 180 carbon sources from the Biolog Phenotype Microarray system. Those plates were inoculated with medium containing WT E. coli carrying the PcmtA-gfp plasmid. Growth at OD_600nm and GFP production were monitored during 8 hours (Table 5). Using the fluorescence (t8h-t0h) over growth (t8h-t0h OD_600nm) ratio, uridine was identified as the strongest activator of PcmtA-gfp (ratio of 9959) (Table 5), followed by bromo- succinic acid and inosine. Glucose was confirmed as one of the compounds with the lowest activation of cmtA (ratio 698). Nucleosides were tested in the PM3B: uridine (ratio of 7601) was again followed by inosine (ratio of 6235), guanosine (ratio of 5669), cytidine (ratio of 5177), adenosine (ratio of 5154), thymidine (ratio of 4152), xanthosine (ratio of 3907) and finally glucose as control (ratio of 839). The above data obtained from Biolog plates was confirmed, this time using standard media and solutions and in triplicate, on a plate reader (Fig. 5A) and flow cytometry (Figure 5B). Again, 6-fold increase in fluorescence production was measured when the media was supplemented with uridine, compared to a condition without additional carbohydrates.

Whether uridine was able to activate the promoter of another carbohydrate transporter involved in AGs uptake was next tested. fruA was chosen, as it is a PTS system present in both enterobacteria and pseudomonales, annotated as a fructose-specific transporter. The pfruA-GFP construction also showed increased fluorescence when uridine was added to the medium, suggesting that uridine is able to activate transcription from more than one sugar transporter. As a negative control to rule out any pleiotropic effect of uridine on gene expression in the experimental conditions, the promoter of the vitamin B12 transporter btuB, was fused to GFP: pbtuB-GFP construction did not show increased GFP expression when the medium was supplemented with uridine (Figure 6).

2.6 Uridine as carbon source decreases the MIC to aminoglycosides by increasing uptake

Because uridine is a substrate that strongly activates the promoter of cmtA, the effects of uridine on the MIC of several antibiotics in E. coli were tested. If uridine is able to increase the expression of AGs transporters, a lower dose of AGs would be needed to kill bacteria, resulting in a decrease of the MIC.

All MICs measured on different substrates in this study are presented in Table 3. Regarding AGs, uridine decreased the MIC of tobramycin and gentamicin by 10-fold, compared to glucose. As expected from the AG specificity of these transporters, this phenotype is not observed with amoxicillin, nor chloramphenicol (Table 3).

Uridine is composed of a sugar moiety, ribose, and a nucleotide moiety, uracile. Interestingly, the 10- fold decrease in MIC promoted by uridine addition in the medium was not observed upon addition of ribose (MIC at 1 μg/ml for ribose versus 1.5 μg/ml for glucose, Figure 5C and Table 3).

To test whether the potentiating effect of uridine was dependent on Crp, the MIC in the presence of glucose or uridine was tested in the Δcrp mutant. The MIC of the Δcrp strain in the presence of glucose was increased to 6 μg/ml and did not change in the presence of uridine. In the absence of crp, the potentiating effect of uridine on AGs was thus abolished, (6 μg/ml in Δcrp versus 0.1 μg/ml in the WT strain). These results support the hypothesis that the carbohydrate transporters stimulated by uridine are Crp-dependent. Similar to uridine, cytidine, adenosine and guanosine also led to susceptibility to AGs (MIC 0.1 μg/ml) compared to glucose supplementation, but not inosine (MIC 1 μg/ml) (Table 3). To test whether the effect of uridine on susceptibility to AGs was related to increased AG uptake, Neo-cy5 uptake was monitored in cells treated with uridine or glucose. Addition of uridine to the medium increased fluorescence from 1,92 with glucose to 2,82 with uridine, during 15 minutes of sub-MIC treatment (Figure 5D).

2.7 Uridine mediated AG susceptibility is not related to uptake through uridine transporters, uridine catabolism or stress responses

It was shown that uridine induces the expression of at least two carbohydrate transporters (cmtA and fruA), which leads to increased AG uptake and susceptibility. In order to test whether the uptake and utilization of uridine itself is responsible of these effects, it was asked whether nucleoside transporters might be also involved in AGs uptake. nupG and nupC were overexpressed, which are able to transport uridine and other nucleosides⁶⁴. None of these overexpressions had any impact on tobramycin susceptibility (Table 3), showing that AGs are not able to enter through nucleoside transporters. Moreover, addition of uridine to the medium when nucleoside transporters were overexpressed, resulting in an increase in the amount of uridine inside the cells, did not induce an increase in susceptibility to tobramycin (Figure 7A). This suggests that uridine uptake by itself is not responsible for the increased susceptibility to tobramycin.

In order to test whether catabolism of uridine was necessary for the AG potentiation phenotype, udk (degradation of uridine into uracile and ribose-l-phosphate) and udp (degradation of uridine into uridine monophosphate by reduction of guanosine triphophaste) were deleted. Their deletion had no impact on the MIC of tobramycin (Table 3). Together with the nucleoside transporter overexpression results, these data suggest that uridine utilization is not necessary for the AG potentiating effect of uridine. Moreover, the fact that other nucleosides can also decrease the MIC to AGs suggests that the phenotype is not linked to uridine utilization pathways.

Finally, it was considered whether the effect of uridine could be linked with the stringent response. Previous studies have shown that the stringent response may contribute to streptomycin uptake⁶⁵ and AGs susceptibility^66,67 ppGpp synthesis by SpoT has also been shown to be modulated by carbon sources⁶⁸. To test whether uridine supplementation could increase or reduce the stringent response, a reporter plasmid was used containing the rmB 16S ribosomal RNA promoter 1 fused to the GFP, where a decrease in fluorescent corresponds to the induction of the stringent response. Tobramycin sub-MIC^69,70 and stationary phase cultures were used as controls that induce the stringent response. Treatment with various carbon sources, including uridine, did not induce the stringent response in the experimental conditions (Figure 7B). Together, these results support an effect of uridine at level of carbohydrate transporters.

2.8 Uridine induces rapid death and prevents the appearance of resistant mutants

The impact of carbon sources was evaluated in a time kill curve experiment on a media containing only tryptone and the substrate, in which bacteria were grown to the mid-exponential phase and treated with a bactericidal dose of tobramycin. The addition of glucose prevented the bactericidal effect, despite the fact that tobramycin was added at a dose 10 times the MIC (Figure 8). Addition of maltose (MIC 0.4 μg/ml) or uridine (MIC 0.1 μg/ml) increased bactericidal effect at different levels (Figure 8A), of which maltose was the less effective. Regrowth was observed for maltose after 20 hours of treatment. Since the number of surviving bacteria after 6 hours of treatment was lower than after 20 hours of treatment, the growing cells were probably suppressor mutants which appeared after 6 hours of treatment. The uridine supplementation was most efficient because it did not promote the regrowth, since all survivors had the same MIC as the WT strain.

Among the suppressor mutants surviving in maltose, two populations could be distinguished: small colonies and normal colonies (Figure 8B), in the same proportion. Sequencing of genomic DNA of these resistant mutants (MIC around 1 μg/ml) showed the appearance of mutations in the fusA (elongation factor G) and rpIL (50S ribosomal protein L7/L12) gene, both linked with the translation process targeted by AGs. The fast killing induced by uridine could avoid selection of these mutations in the population.

2.9 Uridine does not induce changes in PMF but its potentiating effect requires membrane potential.

Changes in AG resistance have been previously linked to changes in membrane potential ⁷¹. To address whether changes in AG uptake was related to variations of PMF, on the WT, the ΔcmtA, the Δcrp and the p0/pcmtA+ strains a treatment was applied with the probe Mitotracker Red whose accumulation depends on the membrane potential ^49,52. No change in PMF was observed in either the deletion strains or cmtAB overexpression strain (Figure 9), showing that carbohydrate transporters promote AG uptake without affecting the PMF. Because the impact of metabolites on AGs uptake has previously been associated with changes in PMF, it was essential to test whether uridine supplementation affects the PMF. Addition of 0.5% uridine or glucose in the medium did not modify the PMF (Figure 9A). Thus, the positive effect of uridine on AG uptake is not due to an increase of membrane potential.

However, the fact that uridine did not increase the PMF does not mean that the effect of uridine on AG uptake is independent of PMF. To address this question, the impact of uridine addition on the MIC of tobramycin in the presence or not of carbonyl cyanide m-chlorophenyl hydrazine (CCCP). CCCP is a protonophore which disrupts the gradient of protons, and thus the PMF. Addition of uridine in the presence of 4 μg/ml of tobramycin prevents bacterial growth while the absence of uridine enables bacteria to growth up to IO^-7 UFC/ml (Figure 9B). Addition of 15 μg/ml of CCCP decreased strongly the efficiency of uridine as potentiator of AGs. The presence of the proton gradient allowed the observation of the potentiating effect of uridine.

2.10 Ags uptake by carbohydrate transporters and the effect of uridine are conserved among Gram negative pathogens.

Since Ags uptake by carbohydrates transporters is present in E. coli, it was considered whether the Ags uptake by carbohydrates transporters is also conserved in pseudomonales, such as Pseudomonas aeruginosa, and Acinetobacter baumannii, Gram-negative pathogens involved in resistant infections. Five P. aeruginosa transporters involved in the uptake of glucose, mannose, maltose and fructose were overexpressed. Overexpression of any of those five transporters increased susceptibility to tobramycin and gentamicin (Figure 10A and Table 3), and not to ciprofloxacin or chloramphenicol (Table 3). Overexpression of gtsB,fruA and mtlFGK decreased the MIC of tobramycin by 10-fold (MIC 0.4 μg/ml vs 4 μg/ml for the empty vector), overexpression of oprB and a homologous gene showed a 2,5-fold decrease (the MIC to amoxicillin was not tested due to the strain's resistance). Neocy5 assay on the strain overexpressing mtlFGK confirmed its increased uptake of AGs compared to the empty vector (about 1.3-fold after 15 minutes of sub-MIC treatment with neocy5) (Figure 10B). For A. baumannii, overexpression of the PTS fruA decreased the MIC of tobramycin to <0.064 compared to 0.1 with the empty vector.

Moreover, the addition of uridine decreased the MIC of tobramycin compared to glucose for V. cholerae (10-fold) and ESKAPE pathogens K. pneumoniae (4-fold), P. aeruginosa (4-fold) and A. baumannii (8-fold) in MH medium (Table 3). Involvement of carbohydrates transporters in AG uptake is thus shared among Gram-negative bacteria, and not limited to a single genera. The potentiating effects of carbohydrate substrates seem to be shared in these pathogens.

2.11 Uridine potentiates AG efficiency in synthetic urine

AG exhibit a concentration-dependent killing ⁷² meaning that aggressive dosing of AG increase the probability of treatment success ⁷³. Therefore, the use of uridine to boost the uptake of AGs appears as a solution to increase the effective doses in bacteria without increasing the toxicity for the patient.

Gram-negative bacteria are frequently involved in urinary tract infections. To test this hypothesis, a range of uridine concentrations were tested for their effect on AG susceptibility on a synthetic human urine medium ⁴⁵ to mimic bacterial growth conditions in UTL From overnight cultures, 5.10⁶ CFU/ml of bacteria were treated with low-dose tobramycin and survival was assessed after 24 hours of treatment, depending on the presence or not of uridine (2-fold sequential dilutions of uridine from a concentration of 1% to 0%) (Figure 11A). 0.5 μg/ml of tobramycin was used, and it was observed that 0.031% of uridine induced the largest effect and confirm the tobramycin potentiation by uridine addition in a synthetic urine medium. Moreover, the potentiating effect increased between uridine at 0.0009% and 0.031%, and then decreased for uridine at 0.031% to 1% (Figure 11A). This could indicate a competition for the transporter between uridine and tobramycin, in a situation in which uridine is in large excess compared to the AG and thus prevents AG entry through the transporter. Of particular significance, an uridine concentration of 0.031% is reachable in the human body ⁷⁴, making this molecule a promising adjuvant to improve AG treatment.

Liquid MIC was performed with other AGs: streptomycin, neomycin and apramycin: uridine addition decreased the MIC for the three AGs (Figure 11B).

A Neo-cy5 assay confirmed an approximately 2.5-fold higher uptake at the single cell level when cells grew in uridine-supplemented media (Figure 11C).

Finally, concentrations of 0.0009% (0,036 mM) or 0.031% (1.27 mM) of uridine were used to observe the time-kill curve in the presence of tobramycin or gentamicin (4 μg/ml) or amakacin (8 μg/ml), the three most commonly used clinical AGs (Figure 11D). The addition of uridine resulted in rapid eradication of the bacteria, with a large drop of survival after 24 hours, compared with a decrease of only one to two logs when uridine was not present in the medium for all three AGs. The concentration of AGs used here is lower than the concentration found in therapy in the bladder, which explains the low killing without uridine, but highlights the effect of the addition of uridine (example: in human, administration of 1 mg/kg gentamicin, urinary concentration between 113 and 423 μg/ml after 1 hours of treatment, 12 to 271 μg/ml after 2 hours ⁷⁵).

The lowest efficient uridine concentration was determined as 0.031% (Figure 12).

2.12 In addition to uridine, other nucleosides also potentiate AG efficiency in synthetic urine, in E. coli and P. aeruginosa

Then, the efficiency of other nucleosides (cytidine, thymidine, adenosine and inosine) as potentiators of the aminoglycosides tobramycin, gentamicin, amikacin on E. coli was compared. In synthetic urine, cytidine and thymidine also showed a significant potentiating effect (Figure 13). P. aeruginosa is also an opportunistic pathogen causing severe UTI infections. The efficiency of nucleosides with tobramycin was also tested on P. aeruginosa. Uridine, adenosine, thymidine and inosine efficiently enhance killing of P. aeruginosa by tobramycin in synthetic urine.

2.13 Uridine potentiation of aminoglycosides is effective in synthetic plasma medium and in human blood

In order to determine whether uridine could be used to treat blood infections such as sepsis, similar killing experiments were performed on a plasma-like medium (HPLM for Human Plasma Like Medium). E. coli K12 or E. coli strain CFT073 [100] (a pyelonephritis strain isolated from a patient with sepsis) were treated with tobramycin, and the survival after 20 hours depending on the addition of 0.031% of uridine was assessed. It was observed that uridine decreased survival at 1 μg/ml of tobramycin, although this medium contains other sugars (Figure 14).

This result was also confirmed using human blood. In these experiments, human blood was inoculated with 1.10⁴ CFU of aminoglycoside susceptible E. coli strain CFT073 [100], or amikacin resistant strain E. coli 932. For the susceptible strain CFT073, gentamicin (0.1 μg/ml) treatment for 1 hour resulted in killing of 30% of the initial population, while combination of gentamicin with uridine killed 50% of the population (Figure 15). For the resistant strain, 932 was treated with 100 μg/ml of amikacin (Figure 16), addition of uridine decreased the survival of the strain after one hour of treatment, in a statistically significant way.

These results support the hypothesis that uridine could potentiate AGs in therapy.

2.14 Uridine potentiation of tobramycin is effective on clinical E. coli strains

Clinical E. coli strains were next used to assess the potentiating effect of uridine addition (0.031% as determined before) combined to tobramycin on the same synthetic urine medium. Synthetic urine medium was used to assess the AG-potentiating action of uridine, on clinical strains of E. coli, isolated from a long-stay hospital (Table 6) as well as uropathogenic UPEC strains from the NILS collection (Natural Isolate with Low Subcultures)[99], for which the complete genome sequences are available. Table 6: Clinical strains

Strains that do not carry any known tobramycin resistance genes were first tested, hence are susceptible to aminoglycosides: strain 886 and NILS 9, 10, 23, 24, 29, 31, 47, 49 and 78 with no resistance genes (but different level of tolerance); strains Ec019 and Ec068 with the MdfA efflux pump associated with a 2-3 fold increase in AG MICs [76] ; and strain 1236, which harbors resistance genes against various other antibiotics (R-lactams, sulfonamides, spectinomycin, streptomycin). In all strains, the addition of uridine (at 0.031 %) resulted in strongly decreased survival after 20 hours lethal tobramycin treatment (10 to 50 μg/ml) (Figure 17). AG resistant clinical strains with aminoglycoside resistance (modification enzymes) were also evaluated (Figure 17). With tobramycin concentrations of 200 and 400 μg/ml, which are typically ineffective against these strains, the addition of uridine showed promising results. For the strains 1193 and 1195, uridine addition decreased the survival at 200 μg/ml. For strains 1215 and 932, and NILS 55 and 64, the effect of uridine was observed at 400 μg/ml of tobramycin treatment. Aminoglycosides concentration reachable in the urine has been shown to be up to 423 μg/ml after 1 hour of treatment with 1 mg/kg of gentamicin⁷⁵, while doses used in therapy nowadays are in the order of 3 to 8 mg/kg.

This data indicates that uridine could be used in combination with AGs to treat certain infections due to resistant E. coli. The benefit of uridine treatment has been observed for the evaluated resistant strains with a tobramycin concentration superior or equal to 200 μg/ml.

These results support the hypothesis that uridine could potentiate AGs in therapy against resistant strains. 2.15 Uridine potentiation of aminoglycosides is effective in vivo

A mouse UTI model was used to determine whether the effect of uridine could also be detected in vivo. First, the strain UTI 89 [101] was treated with gentamicin in synthetic urine in order to verify if the potentiation by uridine was valid in the strain, which was the case (Figure 18). Six-week-old female C57BI/6 mice were infected transurethrally with 10⁷ CFU of E. coli UTI89 strain and treated 24 hours later with gentamicin (0.2 mg/kg) combined or not with 0.5 g/kg of uridine. Bacterial survival was assessed after 24 hours of treatment, i.e. 48 hours post-infection. While gentamicin treatment did not have a significant effect compared to the PBS control, the combination with uridine significantly decreased the number of bacteria in the bladder (Figure 18).

2.16 Uridine induces rapid death and prevents the selection of resistant mutants

To evaluate the impact of different carbon sources on the effectiveness of tobramycin, the kinetics of cell death was quantified by measuring survival rates during several times points after lethal tobramycin treatment. Supplementation with maltose or uridine enhanced the bactericidal effect of tobramycin, with maltose showing the least effectiveness. Notably, in the maltose-treated bacteria, regrowth was observed after 20 hours of treatment, due to the selection of fusA or rpIL resistant mutants, known to be involved in AG resistance mechanisms [102, 103], Conversely, tobramycin showed the highest efficacy on uridine-treated bacteria (Figure 19). The rapid bactericidal effect induced by uridine supplementation may have prevented the selection of these mutations in the bacterial population.

2.17 Uridine supplementation decreases the MIC of aminoglycosides by increasing their uptake

As uridine enhanced the expression of carbohydrate/AG transporters, its addition in growth media would be expected to increase bacterial AG uptake, resulting in a lower dose of AGs needed to kill bacteria, i.e. reduced MIC. MIC values of various antibiotics were measured in E. coli, in the presence of different carbohydrate sources (Table 7). In the case of AGs, supplementation with uridine decreased the MIC of tobramycin and gentamicin by 10-fold compared to glucose as a baseline (Figure 5C). No increase of susceptibility was observed to amoxicillin or chloramphenicol (Table 7). Table 7. Effect of different carbohydrates on the MIC (μg/ml) of different antibiotics.

Determined by Etest in MH medium for all antibiotics except for neomycin (microtiter broth dilution method). (Tob: Tobramycin; Kan: kanamycin; Gen: gentamicin; Neo: neomycin; Spec: spectinomycin; Cip: ciprofloxacin; Trim: trimethoprim; Amox: amoxicillin; Chlo: chloramphenicol). AG: aminoglycoside. R: plasmid resistance. means the condition was not tested.

Unlike with uridine, no significant decrease in MIC was observed upon supplementation with uridine's sugar moiety, ribose (Figure 20 and Table 7).

Supplementation with uridine also decreased the MIC of tobramycin compared to glucose for V. cholerae (10-fold) and ESKAPE pathogens K. pneumoniae (4-fold), P. aeruginosa (4-fold) and A. baumannii (8-fold) (Table 7). Thus, the involvement of carbohydrate transporters in AG uptake is a shared characteristic among Gram-negative bacteria, extending beyond a single genus, and the potentiating effects of uridine were present in all the pathogens examined.

Similar to uridine, supplementation with cytidine, adenosine, and thymidine also resulted in increased susceptibility to AGs (MIC of 0.1-0.4 μg/ml) compared to glucose supplementation, whereas inosine did not (MIC of 1 μg/ml) (Table 7). Guanosine was poorly soluble.

The fact that uridine increases susceptibility to AGs through increasing AG uptake was confirmed using Neo-Cy5 fluorescence per cell, which varied from 1.92 with glucose to 2.82 with uridine after a 15-minute sub-MIC tobramycin treatment (Figure 20). Altogether, these data support that uridine supplementation leads to a higher AG uptake through carbohydrate transporters, on growing cells. Thus, uridine can be an effective potentiator of aminoglycosides on actively dividing cells, on susceptible as well as tolerant or resistant strains. Importantly, uridine can re-sensitize resistant strains.

2.18 Discussion

AG uptake in Gram-negative cells has been evaluated for several decades, and highlighted a major role for PMF ^{9 15}-⁷⁷. The mechanism of AG uptake has been thus previously proposed to depend mainly on membrane potential. Here an active mechanism of uptake for AGs in Gram-negative bacteria through carbohydrate transporters has been identified and elucidated.

Overexpression of transporters enabled to identify at least 11 transport systems, in E. coli, involved in an AG-specific response: mainly PTS systems for different substrates and the ABC transporter MalEFG. Such diversity of involved transporters, and the weak phenotype conferred by single deletions, may explain why this mechanism has not been previously identified using high-throughput techniques such as Tn-seq in P. aeruginosa ⁷⁸ A. baumannii ⁷⁹ or in V. cholerae ⁸⁰. Conversely, respiratory chain mutants for example, are more easily identifiable because of the impact of a single deletion on AGs uptake ⁸¹.

This is not the first report of antibiotic uptake by the PTS⁸² but an active mechanism had never been described for AGs. The effect of individual transporters on susceptibility to AGs is modest comparatively to genetic resistances, but could play a major role if many transporters are expressed at the same time (e.g. through addition of a substrate, such as uridine). Taken together, these data support the hypothesis of an equilibrium between redundant systems able to transport a range of carbohydrates, and AGs. Improving therapies based on carbohydrate transport pathways appears to be an efficient way to avoid selection for resistance, both because single deletion would have little or no impact on bacterial fitness, and because mutations in a large range of transporters would be deleterious for bacterial growth.

Aminoglycoside molecules, as their name indicates, consist of a core structure composed of two or more amino sugars linked via glycosidic links to an aminocyclitol ¹⁰. The presence of the sugar pattern could explain the fact that these molecules are able to be recognized as substrate of bacterial carbohydrates transporters. This is supported by the fact that spectinomycin, with a similar structure, but without the sugar moiety, did not induce the same response. Similar observations were made regarding the uptake of imipenem in P. aeruginosa, capable of being carried by the amino acid porin OprD2 : the authors also noted that imipenem shared structural similarities with amino acids which could explain the recognition of the transporter ⁸³.

While assessing the role of diverse carbohydrate transporters on AG uptake in E. coli, the study of single transporter deletion strains led to identification of ΔcmtA, associated with a 4-fold increase of tobramycin MIC.

Carbon sources and metabolites have been described to impact AGs susceptibility through modulation of PMF in P. aeruginosa ^84,85, persister cells of E. coli, K. pneumoniae, Salmonella typhimurium ^32,33, Salmonella spp, E. coli and S. aureus⁸⁶ and A. baumannii ⁸⁷. However, the effect of carbohydrate transporters is not due to a PMF enhancement, but to AG uptake.

The fact that the PMF remains unchanged during carbohydrate treatment does not mean that this mechanism is independent of PMF, only that the mechanism does not rely on an increased membrane potential. Indeed, it was shown that the effect of uridine is conditioned by the functionality of proton motive force. Regarding PTS, NADH is required for the phosphorylation of EIA of E. coli ^89,90 suggesting a link between the redox state of the cell (and thus the electron transport chain) and the functional activity of the PTS. ABC transporters are also powered by ATP ⁹¹, which could be partially produced by the PMF. The need for PMF in AG uptake could be linked, at least in part, to the functionality of these transporters. These findings constitute an important element to understand the bacterial uptake of AGs.

The role of Crp and cAMP on streptomycin uptake has already been observed, and a mechanism of AGs penetration through polyamine transport system was proposed⁵⁸. A previous study predicted that Crp was a regulator of cmtA and several sugar transporters⁹². It was confirmed here the involvement of Crp as a regulator of cmtA and a player in AG susceptibility in E. coli. However, an indirect effect of Crp cannot be excluded. This also does not exclude Crp/cAMP independent regulation of AG uptake, notably with the cAMP independent action of Cra. The AG uptake mechanism by carbohydrate transporters is shared with Pseudomonadales and other Gram-negative bacteria, since carbohydrate transporters overexpression also enhances AG killing in P. aeruginosa and A. baumannii. Catabolic repression and carbon sources utilization and preferences in Pseudomonadales are different than what it is described in enterobacteria such as E. coll⁹³, and could imply a different regulation concerning AG entry.

AGs have been classified as critically important antimicrobials for human medicine by the WHO ⁹⁴, making the search for improvements of such treatments relevant ⁹⁵. Using a screening tool based on the expression of cmtA, uridine was identified among 188 substrates as the strongest activator of cmtA transcription and enhancer of killing of E. coll by AGs. Since AGs show toxicity towards the host, uridine and AGs co-treatment could be a way to improve treatment by decreasing the efficient AG doses and thus the adverse effects associated with AG therapies. For example, the use of mannitol on K. pneumoniae was shown to allow gentamicin to became effective below the toxic threshold, and protect the kidneys from AGs toxicity ³².

Treatment with high doses of uridine (up to 12g/m², mean body surface for a human is 1,7 m²) induces very limited side effects: no side effect under 10 g/m², and a fifteen minutes shivering without fever for higher doses, which could allow usage of uridine in human therapy⁷⁴. Moreover, a 2 mM peak (0.5 g/L) is reached after uridine injection in the plasma (0.05%), and 15-40% of the injected doses is excreted in the urine ⁹⁶. These doses correspond to a concentration of uridine that were shown to be efficient to potentiate AGs on a synthetic urine medium.

Mimicking bacterial growth conditions in the bladder in UTI, the killing effect of tobramycin, gentamicin and amikacin was enhanced, and was shown to be due to an increased AG uptake. Appearance of resistance to antibiotics is described in 5% of the treatment ⁹⁷ , which is propose to be due to insufficient antibiotics dosage ^9S. Utilization of uridine accelerates the bactericidal kinetics and decreases the risk of spontaneous resistance formation in laboratory conditions, as observed with the absence of spontaneous mutants in the presence of uridine compared to maltose (4-fold higher MIC). More rapidly killing bacterial populations may thus prevent or limit AG resistance formation and dissemination and relapsing infections. Uridine co-treatment with AGs could be relevant in the case of UTI, such as cystitis, but maybe also pyelonephritis, otitis or eyes infections as doses could be easily modulated and administered, as well as pulmonary infections combined with inhaled tobramycin for example ⁴². Moreover, although glucose is present in the blood, a potentiating effect of uridine in a medium mimicking the plasma (including the sugars composition) was observed, which could be relevant for the treatment of blood infection (sepsis). 2.19 Conclusions

This study has identified carbohydrate transporters as players in a new mechanism, which can tune Aminoglycosides (AG) uptake in various Gram-negative bacteria.

AG resistance is usually associated either with genomic AG-modifying factors, which inactivate the AG molecules, or genetic mutations. These mutations can either impact the target of AGs (the ribosome), or more frequently, their uptake via decreased PMF [102, 104, 105], The reason why mutations in carbohydrate transporters have not been previously linked with AG resistance may be explained by the diverse array of transporters capable of transporting AGs. Through systematic overexpression of sugar transporters, this study has identified at least 11 transport systems in E. coll that are specifically involved in AG uptake. These systems include phosphotransferase systems (PTS) for various sugars (chitobiose, glucitol, cellobiose/arbutine, fructose and fructose-like, |3-glucoside and maltose), and the ABC transporter MalEFG for maltose.

Simultaneous activation of multiple transporters can have a significant effect on sensitization to AGs. It is shown here that such upregulation can be achieved by supplementing uridine, the most potent activator of cmtA transcription among 188 substrates tested. Uridine potentiated the killing effect of the clinical AGs tobramycin, gentamicin, and amikacin in synthetic urine, a standardized medium that mimics the bacterial growth conditions in UTI. This enhancement was due to an increased uptake of AGs by the bacteria.

The uptake of AGs through carbohydrate transporters appears to be a shared property among Pseudomonadales [106], and overexpression of carbohydrate transporters also enhanced the killing of P. aeruginosa and A. baumannii by AGs. Therefore, improving treatments by exploiting carbohydrate/AGs transport represents a promising strategy to combat antibiotic resistance, considering the potential synergistic effects of AG and sugars.

In clinical settings, the emergence of AG resistance has been detected in 5.5% of treated cases [97], and inadequate antibiotic dosing has been identified as a primary cause of resistance [98], Uridine supplementation, by accelerating the kinetics of bacterial death, may reduce the selection of AG resistance, as observed in synthetic urine, and potentially limit the occurrence of recurrent infections. Co-treatment of uridine with AGs may be particularly beneficial in the case of UTI but may also be applicable to otitis, or eye infections, where the administered doses can be easily adjusted due to the mode of drug delivery. For pulmonary infections, combining uridine with inhaled tobramycin, for example, could improve outcomes [42], Treatment with high doses of uridine (up to 10 g/m²) induces no side effects [107] in human, and constitutes a promising adjuvant to AG. The World Health Organization has classified AGs as critically important antimicrobials for human medicine [108], underscoring the significance of seeking improvements in AG treatments [95], The concentrations of AGs used in these experiments were lower than those typically achieved in the bladder during therapy. In humans, the administration of 1 mg/kg of gentamicin leads to urinary concentrations ranging from 113 to 423 μg/ml after 1 hour of treatment, and 12 to 271 μg/ml after 2 hours [75], Despite this, the addition of uridine still demonstrated a substantial effect, suggesting that the use of uridine for UTI could theoretically be feasible.

The use of uridine as adjuvant to AGs could be a potent approach to enhance treatment outcomes either by reducing the required AG dosage and mitigating the associated adverse effects, or by limiting the appearance of AG resistance and even re-sensitizing AG resistant bacteria. Treatment with high doses of uridine (up to 10 g/m²) induces no adverse effects in humans [107], AGs exhibit a concentration-dependent killing [72], and are, thus, more effective in treating bacterial infections when higher doses are administered [73], Uridine offers a solution to enhance AG uptake, allowing for increased effective doses in bacteria without escalating toxicity for the patient.

Ċ

Table 4: Transcriptome analysis of growing E. coli in MH medium with or without sub-MIC tobramycin (25% of the MIC).

Log2fc indicates the ratio untreated over sub-MIC TOB.

REFERENCES

1. Clatworthy, A. E., Pierson, E. & Hung, D. T. Targeting virulence: a new paradigm for antimicrobial therapy. Nat. Chem. Biol. 3, 8 (2007).

2. Cassini, A. et al. Attributable deaths and disability-adjusted life-years caused by infections with antibiotic-resistant bacteria in the EU and the European Economic Area in 2015: a populationlevel modelling analysis. Lancet Infect. Dis. 19, 56-66 (2019).

3. Miller, S. I. Antibiotic Resistance and Regulation of the Gram-Negative Bacterial Outer Membrane Barrier by Host Innate Immune Molecules. mBio 7, e01541-16 (2016).

4. Murray, C. J. et al. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. The Lancet 399, 629-655 (2022).

5. Avent, M. L., Rogers, B. A., Cheng, A. C. & Paterson, D. L. Current use of aminoglycosides: indications, pharmacokinetics and monitoring for toxicity: Aminoglycosides: review and monitoring. Intern. Med. J. 41, 441-449 (2011).

6. Selimoglu, E. Aminoglycoside-induced ototoxicity. Curr. Pharm. Des. 13, 119-126 (2007).

7. Martinez-Salgado, C., Lopez-Hernandez, F. J. & Lopez-Novoa, J. M. Glomerular nephrotoxicity of aminoglycosides. Toxicol. Appl. Pharmacol. 223, 86-98 (2007).

8. Lopez-Novoa, J. M., Quiros, Y., Vicente, L., Morales, A. I. & Lopez-Hernandez, F. J. New insights into the mechanism of aminoglycoside nephrotoxicity: an integrative point of view. Kidney Int. 79, 33-45 (2011).

9. Taber, H. W., Mueller, J. P., Miller, P. F. & Arrow, A. S. Bacterial uptake of aminoglycoside antibiotics. Microbiol. Rev. 51, 439-457 (1987).

10. Krause, K. M., Serio, A. W., Kane, T. R. & Connolly, L. E. Aminoglycosides: An Overview. Cold Spring Harb. Perspect. Med. 6, a027029 (2016).

11. Martin, N. L. & Beveridge, T. J. Gentamicin interaction with Pseudomonas aeruginosa cell envelope. Antimicrob. Agents Chemother. 29, 1079-1087 (1986).

12. Bryan, L. E. & Van DenElzen, H. M. Streptomycin Accumulation in Susceptible and Resistant Strains of Escherichia coll and Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 9, 928- 938 (1976).

13. Vaara, M. Agents That Increase the Permeability of the Outer Membrane. MICROBIOL REV 56, 17 (1992).

14. Bafna, J. A. et al. Kanamycin Uptake into Escherichia coli Is Facilitated by OmpF and OmpC Porin Channels Located in the Outer Membrane. ACS Infect. Dis. 6, 1855-1865 (2020).

15. Fraimow, H. S., Greenman, J. B., Leviton, I. M., Dougherty, T. J. & Miller, M. H. Tobramycin uptake in Escherichia coli is driven by either electrical potential or ATP. J. Bacterial. 173, 2800-2808 (1991). 16. Muir, M. E., Van Heeswyck, R. S. & Wallace, B. J. Effect of Growth Rate on Streptomycin Accumulation by Escherichia coli and Bacillus megaterium. Microbiology 130, 2015-2022 (1984).

17. Aguirre Rivera, J. et al. Real-time measurements of aminoglycoside effects on protein synthesis in live cells. Proc. Natl. Acad. Sci. 118, e2013315118 (2021).

18. Davies, J., Gilbert, W. & Gorini, L. Streptomycin, suppression, and the code. Proc. Natl. Acad. Sci. U. S. A. 51, 883-890 (1964).

19. Davis, B. D., Chen, L. L. & Tai, P. C. Misread protein creates membrane channels: an essential step in the bactericidal action of aminoglycosides. Proc. Natl. Acad. Sci. 83, 6164-6168 (1986).

20. Wilson, A. P. R., Sturridge, M. F. & Treasure, T. Aminoglycoside toxicity following antibiotic prophylaxis in cardiac surgery. J. Antimicrob. Chemother. 26, 713-720 (1990).

21. Kahlmeter, G. & Dahlager, J. I. Aminoglycoside toxicity - a review of clinical studies published between 1975 and 1982. J. Antimicrob. Chemother. 13 Suppl A, 9-22 (1984).

22. Craig, W. A. Optimizing Aminoglycoside Use. Crit. Care Clin. 27, 107-121 (2011).

23. Foster, J. & Tekin, M. Aminoglycoside induced ototoxicity associated with mitochondrial DNA mutations. Egypt. J. Med. Hum. Genet. 17, 287-293 (2016).

24. Molitoris, B. A. Cell biology of aminoglycoside nephrotoxicity: newer aspects. Curr. Opin. Nephrol. Hypertens. 6, 384-388 (1997).

25. Baliga, R., Ueda, N., Walker, P. D. & Shah, S. V. Oxidant mechanisms in toxic acute renal failure. Am. J. Kidney Dis. Off. J. Natl. Kidney Found. 29, 465-477 (1997).

26. Rybak, M. J. et al. Prospective evaluation of the effect of an aminoglycoside dosing regimen on rates of observed nephrotoxicity and ototoxicity. Antimicrob. Agents Chemother. 43, 1549-1555 (1999).

27. Drusano, G. L., Bhavnani, S. M., Bertino, J. S., Nafziger, A. N. & Louie, A. Back to the Future: Using Aminoglycosides Again and How to Dose Them Optimally. Clin. Infect. Dis. 45, 753-760 (2007).

28. Haute Autorite de Sante, Direction de I'Evaluation et de I'Acces a I'lnnovation. Comission de la transparence, avis du 15 Decembre 2021 : amikacine B Braun. (2021).

29. Willems, P. et al. Prevention and modulation of aminoglycoside ototoxicity. Mol. Med. Rep. (2008) doi:10.3892/mmr.l.l.3.

30. Tran Ba Huy, P. Ototoxicite des aminoglycosides, medecine/sciences 4, 36-41 (1988).

31. Hiel, H. et al. Gentamicin uptake by cochlear hair cells precedes hearing impairment during chronic treatment. Audiol. Off. Organ Int. Soc. Audiol. 32, 78-87 (1993).

32. Rosenberg, C. R., Fang, X. & Allison, K. R. Potentiating aminoglycoside antibiotics to reduce their toxic side effects. PLoS ONE 15, e0237948 (2020).

33. Allison, K. R., Brynildsen, M. P. & Collins, J. J. Metabolite-enabled eradication of bacterial persisters by aminoglycosides. Nature 473, 216-220 (2011). 34. Lv, B. et al. n-Butanol Potentiates Subinhibitory Aminoglycosides against Bacterial Persisters and Multidrug-Resistant MRSA by Rapidly Enhancing Antibiotic Uptake. ACS Infect. Dis. (2022) doi:10.1021/acsinfecdis.lc00559.

35. Berryhill, B. A. et al. What's the matter with MICs: The contribution of nutrients and limiting resources to the pharmacodynamics of antibiotics and bacteria. http://biorxiv.org/lookup/doi/10.1101/2022.09.30.510422 (2022) doi:10.1101/2022.09.30.510422.

36. Hogins, J. et al. Bacterial Growth of Uropathogenic Escherichia coli in Pooled Urine Is Much Higher than Predicted from the Average Growth in Individual Urine Samples. Microbiol. Spectr. e02016-22 (2022) doi:10.1128/spectrum.02016-22.

37. Yamamoto, T. et al. Effect of purine-free low-malt liquor (happo-shu) on the plasma concentrations and urinary excretion of purine bases and uridine-comparison between purine- free and regular happo-shu. Harm. Metab. Res. Harm. Stoffwechselforschung Harm. Metab. 36, 231-237 (2004).

38. Yilmaz, M. S. et al. Hypotensive effects of intravenously administered uridine and cytidine in conscious rats: Involvement of adenosine receptors. Eur. J. Pharmacol. 584, 125-136 (2008).

39. Kim, E. & Kang, W. Pharmacokinetics of Uridine Following Ocular, Oral and Intravenous Administration in Rabbits. Biomol. Ther. 21, 170-172 (2013).

40. Van Groeningen, G. J., Peters, G. J., Nadal, J. C., Laurensse, E. & Pinedo, H. M. Clinical and Pharmacologic Study of Orally Administrated Uridine. JNCI J. Natl. Cancer Inst. 83, 437-442 (1991).

41. Kelsen, D. P. et al. Phase I trial of PN401, an oral prodrug of uridine, to prevent toxicity from fluorouracil in patients with advanced cancer. J. Clin. Oncol. 15, 1511-1517 (1997).

42. Smith, S., Rowbotham, N. J. & Regan, K. H. Inhaled anti-pseudomonal antibiotics for long-term therapy in cystic fibrosis. Cochrane Database Syst. Rev. (2018) doi:10.1002/14651858.CD001021.pub3.

43. Bilton, D. et al. Pooled analysis of two large randomised phase III inhaled mannitol studies in cystic fibrosis. J. Cyst. Fibros. Off. J. Eur. Cyst. Fibros. Soc. 12, 367-376 (2013).

44. Noone, P. G. et al. Effect of aerosolized uridine-5'-triphosphate on airway clearance with cough in patients with primary ciliary dyskinesia. Am. J. Respir. Crit. Care Med. 160, 144-149 (1999).

45. Sarigul, N., Korkmaz, F. & Kurultak, I. A New Artificial Urine Protocol to Better Imitate Human Urine. Sci. Rep. 9, 20159 (2019).

46. Zhu, X.-D. & Sadowski, P. D. Cleavage-dependent Ligation by the FLP Recombinase: Characterization of a mutant Flp protein with an alteration in a a catalytic amino acid. J. Biol. Chem. 270, 23044-23054 (1995). Silva-Rocha, R. et al. The Standard European Vector Architecture (SEVA): a coherent platform for the analysis and deployment of complex prokaryotic phenotypes. Nucleic Acids Res. 41, D666- D675 (2013). Kessler, B., Timmis, K. N. & de Lorenzo, V. The organization of the Pm promoter of the TOL plasmid reflects the structure of its cognate activator protein XylS. Mol. Gen. Genet. MGG 244, 596-605 (1994). Lang, M. et al. Sleeping ribosomes: Bacterial signaling triggers RaiA mediated persistence to aminoglycosides. IScience 24, 103128 (2021). Sabeti Azad, M. et al. Fluorescent Aminoglycoside Antibiotics and Methods for Accurately Monitoring Uptake by Bacteria. ACS Infect. Dis. 6, 1008-1017 (2020). Baharoglu, Z., Bikard, D. & Mazel, D. Conjugative DNA Transfer Induces the Bacterial SOS Response and Promotes Antibiotic Resistance Development through Integron Activation. PLoS Genet. 6, el001165 (2010). El Mortaji, L. et al. A peptide of a type I toxin-antitoxin system induces Helicobacter pylori morphological transformation from spiral shape to coccoids. Proc. Natl. Acad. Sci. 117, 31398- 31409 (2020). Cormack, B. P., Valdivia, R. H. & Falkow, S. FACS-optimized mutants of the green fluorescent protein (GFP). Gene 173, 33-38 (1996). Cingolani, P. et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w ¹¹¹⁸ ; iso-2; iso- 3. Fly (Austin) 6, 80-92 (2012). Grisafi, P. L. et al. Deletion mutants of the Escherichia coli K-12 mannitol permease: dissection of transport-phosphorylation, phospho-exchange, and mannitol-binding activities. J. Bacterial. 171, 2719-2727 (1989). Sprenger, G. A. Two open reading frames adjacent to the Escherichia coli K-12 transketolase (tkt) gene show high similarity to the mannitol phosphotransferase system enzymes from Escherichia coli and various Gram-positive bacteria. Biochim. Biophys. Acta BBA - Gen. Subj. 1158, 103-106 (1993). Bruckner, R. & Titgemeyer, F. Carbon catabolite repression in bacteria: choice of the carbon source and autoregulatory limitation of sugar utilization. FEMS Microbiol. Lett. 209, 141-148 (2002). Holtje, J.-V. Streptomycin Uptake via an Inducible Polyamine Transport System in Escherichia coli. Eur. J. Biochem. 86, 345-351 (1978). 59. Castaneda-Garcia, A., Rodn'guez-Rojas, A., Guelfo, J. R. & Blazquez, J. The glycerol-3-phosphate permease GIpT is the only fosfomycin transporter in Pseudomonas aeruginosa. J. Bacterial. 191, 6968-6974 (2009).

60. Saier, M. H. & Ramseier, T. M. The catabolite repressor/activator (Cra) protein of enteric bacteria. J. Bacterial. 178, 3411-3417 (1996).

61. Ferenci, T. Adaptation to life at micromolar nutrient levels: the regulation of Escherichia coli glucose transport by endoinduction and cAMP. FEMS Microbiol. Rev. 18, 301-317 (1996).

62. Perlman, R. L., De Crombrugghe, B. & Pastan, I. Cyclic AMP regulates Catabolite and Transient Repression in E. coli. Nature 223, 810-812 (1969).

63. Zubay, G., Schwartz, D. & Beckwith, J. Mechanism of Activation of Catabolite-Sensitive Genes: A Positive Control System. Proc. Natl. Acad. Sci. 66, 104-110 (1970).

64. Patching, S. G. et al. The nucleoside transport proteins, NupC and NupG, from Escherichia coli: specific structural motifs necessary for the binding of ligands. Org. Biomol. Chem. 3, 462 (2005).

65. Sakai, T. T. & Cohen, S. S. Interrelation Between Guanosine Tetraphosphate Accumulation, Ribonucleic Acid Synthesis, and Streptomycin Lethality in Escherichia coli CP78. Antimicrob. Agents Chemother. 7, 730-735 (1975).

66. Koskiniemi, S., Pranting, M., Gullberg, E., Nasvall, J. & Andersson, D. I. Activation of cryptic aminoglycoside resistance in Salmonella enterica: Activation of cryptic resistance. Mol. Microbiol. 80, 1464-1478 (2011).

67. Kim, N. et al. DksA Modulates Antimicrobial Susceptibility of Acinetobacter baumannii. Antibiot. Basel Switz. 10, (2021).

68. Park, Y.-H., Lee, C.-R., Choe, M. & Seok, Y.-J. HPr antagonizes the anti-o70 activity of Rsd in Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 110, 21142-21147 (2013).

69. Carvalho, A., Krin, E., Korlowski, C., Mazel, D. & Baharoglu, Z. Interplay between Sublethal Aminoglycosides and Quorum Sensing: Consequences on Survival in V. cholerae. Cells 10, 3227 (2021).

70. Fruchard, L. et al. Queuosine modification of tRNA-Tyrosine elicits translational reprogramming and enhances growth of Vibrio cholerae with aminoglycosides. http://biorxiv.org/lookup/doi/10.1101/2022.09.26.509455 (2022) doi:10.1101/2022.09.26.509455.

71. Miller, M. FL, Edberg, S. C., Mandel, L. J., Behar, C. F. & Steigbigel, N. H. Gentamicin Uptake in Wild-Type and Aminoglycoside-Resistant Small-Colony Mutants of Staphylococcus aureus. Antimicrob. Agents Chemother. 18, 722-729 (1980).

72. Reed, M. D. Optimal antibiotic dosing. The pharmacokinetic-pharmacodynamic interface. Postgrad. Med. 108, 17-24 (2000). 73. Kashuba, A. D., Nafziger, A. N., Drusano, G. L. & Bertino, J. S. Optimizing aminoglycoside therapy for nosocomial pneumonia caused by gram-negative bacteria. Antimicrob. Agents Chemother. 43, 623-629 (1999).

74. Albert Leyva et al. Phase I and Pharmacokinetic Studies of High-Dose Uridine Intended for Rescue from 5-Fluorouracil Toxicity. Cancer Res. 44, 5928-5933 (1984).

75. Wood, M. J. & Farrell, W. Comparison of Urinary Excretion of Tobramycin and Gentamicin in Adults. J. Infect. Dis. 134, S133-S138 (1976).

76. Edgar, R. & Bibi, E. MdfA, an Escherichia coli multidrug resistance protein with an extraordinarily broad spectrum of drug recognition. J. Bacterial. 179, 2274-2280 (1997).

77. Webster, C. M. et al. Proton motive force underpins respiration-mediated potentiation of aminoglycoside lethality in pathogenic Escherichia coli. Arch. Microbiol. 204, 120 (2022).

78. Cianciulli Sesso, A. et al. Gene Expression Profiling of Pseudomonas aeruginosa Upon Exposure to Colistin and Tobramycin. Front. Microbiol. 12, (2021).

79. Gallagher, L. A., Lee, S. A. & Manoil, C. Importance of Core Genome Functions for an Extreme Antibiotic Resistance Trait. mBio 8, e01655-17 (2017).

80. Babosan, A. et al. Non-essential tRNA and rRNA modifications impact the bacterial response to sub-MIC antibiotic stress, http://biorxiv.org/lookup/doi/10.1101/2022.02.06.479318 (2022) doi:10.1101/2022.02.06.479318.

81. Shan, Y., Lazinski, D., Rowe, S., Camilli, A. & Lewis, K. Genetic Basis of Persister Tolerance to Aminoglycosides in Escherichia coli. mBio 6, e00078-15 (2015).

82. Ammer, J., Brennenstuhl, M., Schindler, P., Holtje, J. V. & Zahner, H. Phosphorylation of streptozotocin during uptake via the phosphoenolpyruvate: sugar phosphotransferase system in Escherichia coli. Antimicrob. Agents Chemother. 16, 801-807 (1979).

83. Trias, J. & Nikaido, H. Protein D2 channel of the Pseudomonas aeruginosa outer membrane has a binding site for basic amino acids and peptides. J. Biol. Chem. 265, 15680-15684 (1990).

84. Martins, D. & Nguyen, D. Stimulating Central Carbon Metabolism to Re-sensitize Pseudomonas aeruginosa to Aminoglycosides. Cell Chem. Biol. 24, 122-124 (2017).

85. Crabbe, A. et al. Host metabolites stimulate the bacterial proton motive force to enhance the activity of aminoglycoside antibiotics. PLOS Pathog. 15, el007697 (2019).

86. Kitzenberg, D. A. et al. Adenosine Awakens Metabolism to Enhance Growth-Independent Killing of Tolerant and Persister Bacteria across Multiple Classes of Antibiotics. mBio 13, e00480-22 (2022).

87. Deng, W. et al. L-lysine potentiates aminoglycosides against Acinetobacter baumannii via regulation of proton motive force and antibiotics uptake. Emerg. Microbes Infect. 9, 639-650 (2020). 88. Herisse, M., Duverger, Y., Martin-Verstraete, I., Barras, F. & Ezraty, B. Silver potentiates aminoglycoside toxicity by enhancing their uptake. Mol. Microbiol. 105, 115-126 (2017).

89. Dannelly, H. K. & Roseman, S. Active site phosphorylation of enzyme I of the bacterial phosphotransferase system by an ATP-dependent kinase. J. Biol. Chem. 271, 15285-15291 (1996).

90. Dannelly, H. K. & Roseman, S. NAD+ and NADH regulate an ATP-dependent kinase that phosphorylates enzyme I of the Escherichia coli phosphotransferase system. Proc. Natl. Acad. Sci. U. S. A. 89, 11274-11276 (1992).

91. Schneider, E. ABC transporters catalyzing carbohydrate uptake. Res. Microbiol. 152, 303-310 (2001).

92. Shimada, T., Fujita, N., Yamamoto, K. & Ishihama, A. Novel Roles of cAMP Receptor Protein (CRP) in Regulation of Transport and Metabolism of Carbon Sources. PLoS ONE 6, e20081 (2011).

93. Sonnleitner, E., Abdou, L. & Haas, D. Small RNA as global regulator of carbon catabolite repression in Pseudomonas aeruginosa. Proc Natl Acad Sci USA 106, 21866-21871 (2009).

94. World Health Organization. Critically important antimicrobials for human medicine : 6th revision. https://www.who.int/publications-detail-redirect/9789241515528 (2018).

95. Serio, A. W., Keepers, T., Andrews, L. & Krause, K. M. Aminoglycoside Revival: Review of a Historically Important Class of Antimicrobials Undergoing Rejuvenation. EcoSal Plus 8, (2018).

96. van Groeningen, C. J., Leyva, A., Kraal, I., Peters, G. J. & Pinedo, H. M. Clinical and pharmacokinetic studies of prolonged administration of high-dose uridine intended for rescue from 5-FU toxicity. Cancer Treat. Rep. 70, 745-750 (1986).

97. Fish, D. N., Piscitelli, S. C. & Danziger, L. H. Development of Resistance During Antimicrobial Therapy: A Review of Antibiotic Classes and Patient Characteristics in 173 Studies. Pharmacother. J. Hum. Pharmacol. Drug Ther. 15, 279-291 (1995).

98. Abdelraouf, K., Linder, K. E., Nailor, M. D. & Nicolau, D. P. Predicting and preventing antimicrobial resistance utilizing pharmacodynamics: part II Gram-negative bacteria. Expert Opin. Drug Metab. Toxicol. 13, 705-714 (2017).

99. Bleibtreu, A., et al., The rpoS gene is predominantly inactivated during laboratory storage and undergoes source-sink evolution in Escherichia coli species. J Bacteriol, 2014. 196(24): p. 4276- 84.

100. Kao, J.S., et al., Pathogenicity island seguences of pyelonephritogenic Escherichia coli CFT073 are associated with virulent uropathogenic strains. Infect Immun, 1997. 65(7): p. 2812-20.

101. Mora-Bau, G., et al., Macrophages Subvert Adaptive Immunity to Urinary Tract Infection. PLoS Pathog, 2015. 11(7): p. el005044. 102. Bolard, A., P. Plesiat, and K. Jeannot, Mutations in GenefusAl as a Novel Mechanism of Aminoglycoside Resistance in Clinical Strains of Pseudomonas aeruginosa. Antimicrob Agents Chemother, 2018. 62(2). 103. Maisnier-Patin, S., et al., Compensatory evolution reveals functional interactions between ribosomal proteins S12, L14 and L19. J Mol Biol, 2007. 366(1): p. 207-15. 104. Thorbjarnardottir, S.H., R.A. Magnusdottir, and G. Eggertsson, Mutations determining generalized resistance to aminoglycoside antibiotics in Escherichia coli. Mol Gen Genet, 1978. 161(1): p. 89-98. 105. Hancock, R.E., Aminoglycoside uptake and mode of action-with special reference to streptomycin and gentamicin. II. Effects of aminoglycosides on cells. J Antimicrob Chemother, 1981.8(6): p. 429-45. 106. Pierle, S.A., et al., Identification of the active mechanism of aminoglycoside entry in <em>V. cholerae</em> through characterization ofsRNA <em>ctrR,</em> regulating carbohydrate utilization and transport. bioRxiv, 2023: p. 2023.07.19.549712. 107. Leyva, A., et al., Phase I and pharmacokinetic studies of high-dose uridine intended for rescue from 5-fluorouracil toxicity. Cancer Res, 1984. 44(12 Pt 1): p. 5928-33. 108. Organization, W.H., Critically important antimicrobials for human medicine : 6th revision. https://www.who.int/publications-detail-redirect/9789241515528. 2018.

Claims

1. A combination of a nucleoside and an aminoglycoside antibiotic for use in the treatment of a bacterial infection, wherein the nucleoside is administered together with the aminoglycoside antibiotic, and the nucleoside enhances the killing of the bacteria by the antibiotic.

2. The combination for use of claim 1, wherein the nucleoside limits the emergence of bacteria resistant to the aminoglycoside antibiotic.

3. The combination for use of claim 1, wherein the nucleoside re-sensitizes bacteria resistant to the aminoglycoside antibiotic to the killing by the antibiotic.

4. The combination for use of any one of claims 1 to 3, wherein the nucleoside is selected from uridine, inosine, guanosine, cytidine, thymidine, or xanthosine.

5. The combination for use of claim 4, wherein the nucleoside is uridine.

6. The combination for use of any one of claims 1-5, wherein the nucleoside is administered to the urinary tract, gastrointestinal tract, lungs, eyes, heart, brain, blood, ears, nose, or skin.

7. The combination for use of any one of claims 1-6, wherein the nucleoside is administered at a dose of 1-20 g/m².

8. The combination for use of claims 1-7, for the prevention of the emergence of antibioticresistant bacteria by killing the bacteria faster with a co-administration of the nucleoside together with the aminoglycoside antibiotic.

9. The combination for use of any one of claims 1-8, for decreasing of the toxicity of a reference treatment by a co-administration of the nucleoside together with the aminoglycoside antibiotic with the aminoglycoside being administrated at a lower concentration than the reference treatment.

10. The combination for use of claim 9, wherein the level of killing of the bacteria by the antibiotic is the same as the reference treatment.

11. The combination for use of any one of claims 1-10, wherein the antibiotic is selected from tobramycin, gentamicin, and amikacin.

12. The combination for use of any one of claims 1-11, wherein the bacteria causing the infection are enterobacteria.

13. The combination for use of any one of claims 1-12, wherein the bacteria causing the infection are selected from Escherichia coli, Pseudomonas aeruginosa, Klebsiella pneumoniae, and Acinetobacter baumannii.

14. The combination for use of any one of claims 1 and 3 to 13, wherein the bacteria causing the infection comprise bacteria resistant to the aminoglycoside antibiotic.

15. A composition for killing bacteria comprising a nucleoside and an aminoglycoside antibiotic, wherein the nucleoside enhances the killing of the bacteria by the antibiotic.

16. The composition of claim 15, wherein the nucleoside is selected from uridine, inosine, guanosine, cytidine, thymidine, or xanthosine.

17. The composition of claim 16, wherein, wherein the nucleoside is uridine.

18. The composition of any one of claims 15-17, wherein the antibiotic is selected from tobramycin, gentamicin, and amikacin.