Distinct Morphological Fates of Uropathogenic Escherichia coli Intracellular Bacterial Communities: Dependency on Urine Composition and pH

Abstract

Uropathogenic Escherichia coli (UPEC) is the leading cause of urinary tract infections. These bacteria undertake a multistage infection cycle involving invasion of and proliferation within urinary tract epithelial cells, leading to the rupture of the host cell and dispersal of the bacteria, some of which have a highly filamentous morphology. Here, we established a microfluidics-based model of UPEC infection of immortalized human bladder epithelial cells that recapitulates the main stages of bacterial morphological changes during the acute infection cycle in vivo and allows the development and fate of individual cells to be monitored in real time by fluorescence microscopy. The UPEC-infected bladder cells remained alive and mobile in nonconfluent monolayers during the development of intracellular bacterial communities (IBCs). Switching from a flow of growth medium to human urine resulted in immobilization of both uninfected and infected bladder cells. Some IBCs continued to develop and then released many highly filamentous bacteria via an extrusion-like process, whereas other IBCs showed strong UPEC proliferation, and yet no filamentation was detected. The filamentation response was dependent on the weak acidity of human urine and required component(s) in a low molecular-mass (<3,000 Da) fraction from a mildly dehydrated donor. The developmental fate for bacteria therefore appears to be controlled by multiple factors that act at the level of the whole IBC, suggesting that variable local environments or stochastic differentiation pathways influence IBC developmental fates during infection.

Keywords: UPEC; infection model; microfluidics; morphological differentiation; stress response; urinary tract infection; urine.

FIG 1

Microfluidic model for the intracellular UPEC and release stages of UTI. Diagrammatic representation of the CellASIC Onix microfluidics plate showing the contents of each well that were sequentially pumped through the main chamber during the defined stages of the infection. The lower panels show phase-contrast and fluorescence microscopy images of the infection chamber surface at the end of the intracellular growth stage (29 h postinfection), showing a well-developed IBC, and at the end of the dispersal phase (49 h postinfection), showing filamentous UTI89/pGI5 in among BECs. Scale bars, 5 μm.

FIG 2

PD07i bladder cells infected with UTI89/pGI5 are mobile. Time-lapse microscopy showing 3-h intervals during the 9- to 29-h stage of infection, with a flow of EpiLife+Gm, was performed. Phase-contrast and GFP channel overlays (left) with SYTOX Orange channel (right) are shown for the indicated time points after switching to Gm-containing medium. Infected bladder cells (black, white, and blue arrowheads, which track individual BECs) appeared at different positions at each time point. SYTOX Orange staining indicated that most infected and uninfected bladder cells were not permeable, whereas permeable, assigned-dead bladder cells are indicated by red arrowheads; these occasionally drifted detached from the surface during the movie. Scale bars, 20 μm (60× oil objective). Refer to Movie S1 in the supplemental material for the full set of time-lapse images.

FIG 3

Bladder cells become immobile after exposure to urine. Microscopy images were taken 20 min before urine exposure, immediately after switching to urine (0 h), and 10 h into the urine exposure. Dead bladder cells have detached from the monolayer and taken up SYTOX Orange stain (e.g., white arrowheads), but the staining disappears soon after urine exposure (within 10 to 20 min). Scale bars, 20 μm (40× objective). All GFP exposures = 50 ms; all SYTOX exposures = 100 ms. Refer to Movie S2 for the full set of time-lapse images.

FIG 4

Filamentous bacteria emerging from an infected bladder cell. (A and B) Phase-contrast (A) and GFP fluorescence (B) time-lapse microscopy images, showing an IBC developing from within an infected bladder cell during the 20-h urine exposure. The bacteria overwhelmed the bladder cell, causing it to rupture and release filamentous bacteria. Scale bars, 10 μm (40× objective). All GFP exposures = 50 ms. Refer to Movie S3 for the full set of time-lapse images.

FIG 5

Short bacteria emerging from an infected bladder cell. (A and B) Phase-contrast (A) and GFP fluorescence (B) time-lapse microscopy showing an IBC developing from within an infected bladder cell during the 20-h urine exposure. The bacteria overwhelmed the bladder cell, causing it to rupture and release many short bacteria and no detected filaments. Scale bars, 10 μm, 40× objective. All GFP exposures = 50 ms. Refer to Movie S4 for the full set of time-lapse images.

FIG 6

Urine pH controls UPEC filamentation. Phase-contrast microscopy and flow cytometry of UTI89/pGI5, harvested from the flow chamber infection model after day 2, with exposure to urine under different pH conditions, were performed on various samples: control (normal human urine, red, top left), acidified urine (blue, top right), neutralized urine (amber, bottom left), and pH-readjusted urine (green, bottom right). Scale bars, 5 μm. The lower panel shows flow cytometry frequency distributions, normalized to the sample mode, indicating the fraction of cells below or above a side-scatter area (SSC-A) cutoff value, defined as the SSC-A below which 99% of the cells are from a UTI89/pGI5 LB medium mid-log-phase culture. Sample colors correspond to the image frames above.

FIG 7

Synthetic human urine does not induce UPEC filamentation. Phase-contrast microscopy and flow cytometry of UTI89/pGI5, harvested from the flow chamber infection model after day 2, with exposure to human urine (red) or synthetic human urine (SHU) (blue), were performed. Scale bars, 5 μm. Flow cytometry of infection effluent with human urine shows a left-hand peak with a large right-hand shoulder indicating a mixed population of short and filamentous bacteria (12.2% filamentous, according to the cutoff); the bacteria exposed to SHU are represented by a narrow left-hand peak indicating a population of short bacteria of a similar length (99% short).

FIG 8

The urine small-molecule fraction supports robust UPEC filamentation. Phase-contrast microscopy and flow cytometry of UTI89/pGI5, harvested from the flow chamber infection model after day 2, with exposure to whole human urine (red) or small molecule human urine (<3000-Da fraction, blue), were performed. Scale bars, 5 μm. Flow cytometry showed very similar curves for both urine types, a left-hand peak with a large right-hand shoulder indicating a mixed population of short and filamentous bacteria; 12.2% of bacteria for whole urine and 12.7% of small molecule urine were filamentous.