Bacterial strain and growth media
Artificial urine (AU) with a physiological nitrate concentration of 20 mM was prepared following the procedures outlined in ref. 9 and was sterilised using a syringe filter (Corning, nylon membrane, diam. 25 mm and pore size 0.2 µm). AU with a concentration of 1.5 × 108 CFU/mL of E.coli strain 25922 or strain 35218 was achieved by comparing absorbance to a 0.5 McFarland standard. This solution was diluted to a final concentration of 1.5 × 103 CFU/mL, aligning with the typical bacterial count observed in uncomplicated UTI10.
Growth conditions and antimicrobial killing
E.coli cultures were cultivated in 1 mL volumes within 96-deep well plates with non-oxygen permeable lids to create anaerobic conditions. All samples were incubated in a single batch, using two 96-deep well plates which were filled with triplicates of E.coli strain 35218 and 25922 for each time point with a known starting concentration of 1.5 × 103 CFU /mL. 164 mg/mL sterilised amoxicillin, a concentration high enough to inhibit the growth of high bacterial count cultures, was introduced to samples after 15 h of growth. To minimise exposure to oxygen, the non-oxygen permeable lid was only removed from sample triplicates that were analysed at the timepoint of interest.
At each time-point, samples of both bacterial strains were removed from the 96-well plate and transferred to Eppendorf tubes. A range of dilutions were immediately plated on Mueller-Hinton agar plates for CFU determination. The Eppendorf tubes containing the samples were stored at −18 °C to measure nitrite concentrations within a 72 h timeframe. All nitrite measurements were performed simultaneously at the end of the in-vitro study, alongside a series of AU samples spiked with known nitrite concentrations to generate a calibration curve. All experiments were conducted in triplicates.
Growth determination
For quantifying CFU, the specimen was diluted (10−2, 10−3, 10−4, 10−5, 10−6) and subsequently plated on Mueller-Hinton agar plates. Following 24 h of incubation, CFUs were counted manually.
Nitrite and creatinine measurement
Nitrite was assessed using the Griess method11 with a detection range up to 100 μM and a limit of detection (LOD), defined according to the IUPAC as 3.3 times standard deviation of the intercept divided by the slope of the calibration12, of 5.9 μM. Briefly, 1 mg/mL N-(1-naphthyl)ethylenediamine dihydrochloride was mixed with 10 mg/ mL sulfanilic acid in 5% (w/w) orthophosphoric acid to form the Griess reagent.
AU samples were centrifuged for 5 min at 2000 rpm. 40 μL of the supernatant was mixed with 10 μL of Griess reagent and shaken for 15 min before reading the absorbance at 548 nm. Creatinine was measured using the Jaffe method as described in ref. 13. 100 μM of a working reagent consisting of equal volumes of 17.5 mmol/L picric acid and 0.29 mol/L sodium hydroxide were mixed with 10 μM of the sample of interest. Absorbance was read in 96-well plates in a FLUOstar Omega plate reader after 90 s of incubation. The assay has an upper detection limit of 20 μM and a LOD of 2.18 μM according the IUPAC definition12.
Urine collection of patients with and without UTI
Urine samples of twenty-five female patients with uncomplicated UTI (confirmed growth of E. coli > 105 CFU/mL, white cell count (WCC) > 50–100, no epithelial cells) and 25 female patients without UTI (no suspected infection, no bacterial growth, WCC < 50, no epithelial cells) were stored in boric acid at 4 °C until further analysis. Boric acid preserves bacterial counts and has no known effect on nitrite detection using dipstick tests14. At collection, the samples were routinely analysed for growth of organisms, white cell count and epithelial cells at Northwest London Pathology, Imperial College Healthcare NHS Trust. Bacterial growth determination, nitrite and creatinine measurement was conducted 4 days after collection at Hammersmith Hospital. Regional Ethics Committee approval was obtained prior to the start of study and is summarised in IRAS Project ID 162013, COREC Application 06/Q0406/20: The detection of microbial products and effects on patients in infection. The study protocol allows use of clinical samples routinely submitted to a Diagnostic laboratory as part of patient care, and where written consent is not normally obtained.
Statistical analysis
Statistical analyses were performed using Spearman rank correlation and Mann-Whitney U-tests. Spearman rank correlation was used to assess the monotonic relationship between nitrite levels and CFU counts for each strain under different conditions and in UTI patients. The Spearman correlation S was calculated, and a p-value of less than 0.05 was considered statistically significant. Spearman correlation was used instead of a Pearson correlation as a normal distribution of values was not provided.
Mann-Whitney U-test was used to compare nitrite concentrations and CFU counts between treated and untreated groups for each strain after addition of antimicrobial drug, as well as to evaluate differences in nitrite levels between UTI and non-UTI patient groups. Statistical significance was defined as p < 0.05 for all tests. Mann-Whitney U-test is used to determine if there is a significant difference between two groups, without assuming any distribution of values.
Statistical analyses were conducted using Python with the following packages: pandas, numpy, and scipy for statistical tests, alongside matplotlib and seaborn for data visualization.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
