Figure 1.
UPEC attachment in Bowman's capsule.
(A) Schematic showing nephron structure. (B) Dynamic imaging of LT004 infused directly into Bowman's capsule. (2 h) Bacteria (green, arrow) conformally lining to Bowman's capsule epithelia of a glomerulus (G) and in the early proximal tubule segment (S1). Proximal tubule epithelium is outlined by endocytosed 10 kDa cascade-blue labeled dextran (blue). Blood plasma is labeled with 500 kDa rhodamine dextran (red). Hoechst 33342 labels cell nuclei (cyan). (4–5 h) Bacterial colonization within Bowman's capsule and the proximal tubule is accompanied by reduced peritubular capillary blood flow (4 h, arrow). (8 h) The glomerular capillary loops succumb to infection and the blood stops flowing (G). Scale bar = 30 µm.
Figure 2.
Expression, binding and infection characteristics of P and Type 1 isogenic strains.
(A) qRT-PCR analysis of micro-dissected tissues 8 h after infusion with LT004 or PBS. Bars show cycle of threshold for detection of gfp+ (white), papA_2 (black) and fimA (grey) transcript. n/d = not detectable. Each group is from an individual animal (B) Relative agglutination of human O type red blood cells, indicating PapG mediated agglutination (black) and yeast cell agglutination in the absence of mannose, indicating FimH mediated agglutination (white) for indicated strains using LB medium as control. (C) Bacterial adhesion (counted per 40 cells) to A498 human renal epithelial cells. * P = 0.001, ** P = 0.045 Error bars in (a, c) = standard deviation. (D) Number of bacteria in the kidney, shown by CFU/g tissue, following a 4-day ascending infection.
Figure 3.
Progression of infection in live animals.
Multiphoton imaging of renal tissue after infusion of indicated strains (green). Injected proximal tubules are outlined by a co-injected 10 kDa cascade-blue dextran (blue); blood plasma is labeled with 500 kDa rhodamine dextran (red) and cell nuclei are stained with Hoechst 33342 (cyan). (A) Wild type LT004 bacteria (green, arrow) (n = 12) can be seen colonizing the infected tubule (blue outline) within 2 h. As bacteria multiplied, shutdown of the peritubular capillaries was observed by a loss of the red plasma marker within surrounding capillaries (arrow, 6 h). (B) The UPEC strain ARD41, which lacks PapG mediated attachment, showed compromised colonization kinetics with few bacteria visible before 8 h (arrow). At 10 h, bacteria colonized the tubule lumen and signs of vascular dysfunction appeared (arrow) (n = 12). (C) The E. coli K-12 strain ARD42 (n = 5) showed a delayed colonization, with few bacteria colonizing the tubule at 2 h (arrow and insert, which is shown as Figure 3F). Colonization at 7 h was less as compared to LT004 (arrow). (D) Expression of P fimbriae in strain ARD43 restores colonization kinetics. Arrow at 2 h identifies bacterial colonization. At 6.5 h the beginning of vascular dysfunction can be seen by the lack of black shadows typical of flowing erythrocytes (arrow) (n = 4). (E) Lack of FimH mediated attachment in UPEC strain ARD40 shows that bacteria struggle to maintain themselves in the center of the tubule lumen, with less bacteria in the center of the tubule (arrow 5 h) (n = 7) (F) 50 µm wide inset from Figure 3C showing bacterial colonization of the tubule. (G) shows x-z projection of panel 3E, ARD40 8 h, demonstrating the bacterial ‘tube’ structure. Images from 24 h show bacterial clearance and edema formation in all strains. Scale bars: 7–10 h = 30 µm, 24 h = 50 µm. All figures presented are representative images, displaying the characteristic colonization patterns for each experimental set.
Table 1.
Bacterial strains and oligonucleotides used in this study.
Figure 4.
(A) Crystal violet assay of biofilm formation from indicated strains. Visualization (top) and quantification (bottom) of biofilm at OD600. Error bars = standard deviation. * P = 0.001 ** P = 0.032 (B) Western blot of RpoS protein in indicated strains using MC1 as positive control.
Figure 5.
Infection affects renal filtration.
Glomerular filtration in non-infected (I) and infected (II) tubules 4 h (A,B) and 8 h (C, D) after LT004 infusion is visualized 7, 20, and 80 s after iv bolus infusion of 10 kDa dextran (red). These data sets are representatives from a single animal, the experiments were performed on at least 3 separate occasions. The dynamic aspect of renal filtration and clearance can be seen in Videos S2 and S3, from which these frames originate. (A) Efficient filtration of the non-infected nephron (I) is visualized by the appearance of the bright red dextran in the tubule lumen (20 s), followed by a drop in intensity (80 s) indicating clearance. A less dramatic intensity change is seen in the infected nephron, indicating less filtration. Epithelial dextran uptake (arrow, 80 s) indicates epithelial dysfunction in infected tubule. Healthy epithelia in non-infected tubule exclude the dextran (arrowhead, 80 s). (B) Quantification of the mean intensity of dextran in tubule lumens over 80 s. Dotted red line corresponds to luminal intensity in the non-infected tubule (I), black line corresponds to the infected tubule (II). (C–D) The absence of filtrated dextran in the infected tubule (II) demonstrates compromised filtration at 8 h. Tissue shows accumulation of some dextran from previous bolus infusions. Scale bars = 30 µm. (E) Enlarged 50 µm inset from 5A 7 s showing dextran leaking into epithelial cells (arrowhead) from the basal side of the cells.