Fig 1.
Detection of formaldehyde in ALTAR urine samples.
A: Average formaldehyde detected across participants that had provided ≥ 5 samples during ALTAR. The grey and pink polygons indicate the confidence intervals calculated from the data. The green background indicates the period of treatment during ALTAR B. Heatmap of all data to show the variability in formaldehyde detection. Black squares are missing samples. All samples used correlate to routine 0-, 3-, 6-, 9-, 12-, 15-, and 18-month sampling and does not include UTI samples. For plotting purposes data is presented in days.
Fig 2.
Detection of formaldehyde resistance (FDHR) in sequenced ALTAR E. coli isolates.
Colours of tree branches represent the major clades from E. coli. Common sequence types (ST) defined by the Achtman MLST scheme are shown for reference. The branch tips are coloured to represent swab (green) or urine (orange) isolates. The heatmap is a binary representation of where growth was observed at increasing concentrations of formaldehyde. All strains were screened n = 1 and any FDHR phenotype confirmed in a minimum of n = 3 growth assays (Fig A and B in S1 Text).
Fig 3.
Case profiles of the four MH participants (defined here as A., B., C. and D.) identified to carry FDHR E. coli isolates.
The schematic below each graph represents the timeline based on the x-axis of the plot for when E. coli was isolated, with the shape defining the FDH phenotype (See key shown in A.). The colour of the objects, in the timeline schematics, reflects the isolates genotype based on the Achtman MLST scheme. Next to each plot is a clade view, taken from the phylogenetic tree in Fig 2, to show the phylogenetic relatedness of specific isolates, the same shape and coloured objects, as used in the timelines, are included in these trees to indicate the specific isolates. The line graphs show the detected formaldehyde from the urine samples, in A. & C. reduced formaldehyde detection coincided with the identification of FDHR E. coli isolates.
Table 1.
FrmR variants isolated during ALTAR.
Fig 4.
Characterisation of 3 frmR variants introduced into CFT073 by CRISPR/Cas gene editing.
A. Growth assays of each variant compared to CFT073 in the noted concentrations of formaldehyde. B. Expression analysis for frmA and frmB of each variant compared to CFT073 in uninduced conditions. Included here is a scarless deletion of frmR (∆frmR) for comparison. C. Detoxification rates determined during growth assay when each variant was challenged from Time 0 with 1 mM formaldehyde. All data represents a minimum of n = 3 biological independent repeats.
Fig 5.
Comparison of identified HPLC peak areas from analysis of 175 urine ALTAR samples.
A. scatter plot of all identified component peak areas > 0.07 demonstrating elevated levels in MH users. B. Component violin plots comparing known components between arms.
Fig 6.
Impact of pH on the growth of frmR+ and ∆frmR strains when exposed to methenamine in artificial urine.
Data shown is the average of 3 independent repeats for all conditions. Formaldehyde concentration detected in the cultures is plotted on the primary y-axis with culture density represented by OD600 values on the secondary y-axis. Data is colour coordinated with respect to the presence of absence of methenamine in the cultures (see key in top left graph).