The role of bacterial size, shape and surface in macrophage engulfment of uropathogenic E. coli cells

Abstract

Uropathogenic Escherichia coli (UPEC) can undergo extensive filamentation in the host during acute urinary tract infections (UTIs). It has been hypothesised that this morphological plasticity allows bacteria to avoid host immune responses such as macrophage engulfment. However, it is still unclear what properties of filaments are important in macrophage-bacteria interactions. The aim of this work was to investigate the contribution of bacterial biophysical parameters, such as cell size and shape, and physiological parameters, such as cell surface and the environment, to macrophage engulfment efficiency. Viable, reversible filaments of known lengths and volumes were produced in the UPEC strain UTI89 using a variety of methods, including exposure to cell-wall targeting antibiotics, genetic manipulation and isolation from an in vitro human bladder cell model. Quantification of the engulfment ability of macrophages using gentamicin-protection assays and fluorescence microscopy demonstrated that the ability of filaments to avoid macrophage engulfment is dependent on a combination of size (length and volume), shape, cell surface and external environmental factors. UTI89 filamentation and macrophage engulfment efficiency were also found to occur independently of the SOS-inducible filamentation genes, sulA and ymfM in both in vivo and in vitro models of infection. Compared to filaments formed via antibiotic inhibition of division, the infection-derived filaments were preferentially targeted by macrophages. With several strains of UPEC now resistant to current antibiotics, our work identifies the importance of bacterial physiological and morphological states during infection.

Fig 1. THP-1 macrophages engulf UTI89 filaments significantly less than rods.

UTI89 was treated with either cephalexin (rods: untreated, LEX rods: 2.5 μg/ml, LEX filaments: 10 μg/ml) or ciprofloxacin (rods: untreated, CIP rods: 3.75 ng/ml, CIP filaments: 15 ng/ml). UTI89/pLau80 was treated with 0.2% glucose (FtsZ (plasmid repressed) rods) or 0.2% arabinose to induce expression of ftsZ-yfp (FtsZ-overproduction filaments). (A) Bacterial lengths as determined by phase contrast microscopy with n = 110–426. (B-D) THP-1 macrophages were infected (MOI 10) for 1 hour and bacterial loads were assessed 2-hours post infection in a gentamicin-protection assay. Data are the averages of 4 independent experiments with error bars representing the SEM. * indicates p <0.05, ** p <0.01, **** p <0.0001, determined by one-way ANOVA with multiple comparisons or Welch’s t-test.

Fig 2. In a heterogeneous population of rods and filaments, rods are engulfed by THP-1 macrophages preferentially over filaments.

UTI89 and UTI89/pGI5 (msfGFP) were untreated (rods) or exposed to 10 μg/ml cephalexin (LEX filaments). (A) rods and filaments were combined to create one population (Mixed bacteria) of 24% rods and 76% filaments. (B) THP-1 macrophages were infected (MOI 10) for 1 hour and intracellular bacterial loads were assessed 2-hours post infection in a gentamicin-protection assay. (C-D) Macrophages were infected with UTI89/pGI5 (msfGFP) (cyan) Mixed bacteria at MOI 10, fixed after 60 minutes and stained with 1X CellMask Orange (red). Images were acquired using a DeltaVision Elite microscope with the 40X dry NA 0.60 objective. (C) Percentage of bacterial populations fully engulfed (number of bacteria fully engulfed/total number of bacteria counted X 100) is normalised and presented as per macrophage to allow for comparisons despite variations in number of macrophages counted from microscopy images. (D) Representative image of infected macrophage with arrow indicating an internalised rod bacterium. Orthogonal xz and yz views highlight internalisation of rod bacterium. Scale bar = 20 μm. Data are the averages of 4 independent experiments (A and B) or 2 independent experiments (C) with error bars representing the SEM. * indicates p < 0.05, ** p < 0.01, determined by one-way ANOVA with multiple comparisons or Welch’s t-test.

Fig 3. THP-1 macrophage engulfment of UTI89 decreases exponentially as bacterial length increases.

(A-B) UTI89 was treated with 10 μg/ml cephalexin for increasing periods of time to produce filaments of increasing average length. (C-D) UTI89/pLau80 was treated with 0.2% (v/v) arabinose for increasing periods of time to produce FtsZ filaments of increasing average length. (A-D) THP-1 macrophages were infected (MOI 10) for 1 hour and bacterial loads were assessed 2-hours post infection in a gentamicin-protection assay. Data are the averages of 4 independent experiments with vertical error bars representing the SEM of intracellular bacteria and horizontal error bars (B and D) represent SEM of bacterial length, n = 105–250. * indicates p < 0.05, ** p < 0.01, *** p < 0.001, determined by one-way ANOVA with multiple comparisons. R squared values generated from linear and nonlinear regression analyses, which were plotted as straight or curved lines, respectively, on B and D.

Fig 4. THP-1 macrophages have bacterial size and shape preferences during engulfment.

UTI89 cells were untreated (rods) or treated with 10 μg/ml cephalexin (filaments), 10 μg/ml mecillinam (Spheres), and grown for 3 additional hours with 10 μg/ml mecillinam and cephalexin (Big spheres) or with 10 μg/ml cephalexin (Long filaments). (A) Phase contrast images were acquired using a Zeiss Axioplan 2 microscope with the 100X oil immersion NA 1.4 objective. Images are representative from 1 experiment. Scale bar = 5 μm. Table of average cell volumes with standard deviation (SD) as measured by a Coulter counter, n = 36927–281142. (B) THP-1 macrophages were infected (MOI 10) for 1 hour and bacterial loads were assessed 2 hours post infection in a gentamicin-protection assay. Data are the averages of 3 independent experiments with error bars representing the SEM. * indicates p < 0.05, determined by one-way ANOVA with multiple comparisons.

Fig 5. Blocking mannose binding with methyl α-D-mannopyranoside reduces THP-1 macrophage engulfment of UTI89 rods and filaments.

UTI89 was untreated (rods) or treated with 10 μg/ml cephalexin (filaments). Both populations were also treated with 3% methyl α-D-mannopyranoside (rods + α-D-MP and filaments + α-D-MP). THP-1 macrophages were infected (MOI 10) for 1 hour and bacterial loads were assessed 2 hours post infection in a gentamicin-protection assay. Data are the averages of 3 independent experiments with error bars representing the SEM. * indicates p < 0.05, ** p < 0.01, determined by one-way ANOVA with multiple comparisons.

Fig 6. The growth environment of bacteria contributes to engulfment dynamics.

UTI89/pGI5 (msfGFP) were isolated from an in vitro human bladder model (bladder bacteria; ~86% filaments) or treated with 10 μg/ml cephalexin (filaments) or untreated (rods) and grown in LB medium. (A) THP-1 macrophages were infected (MOI 10) for 1 hour and bacterial loads were assessed 2-hours post infection in a gentamicin-protection assay. (B-D) Macrophages were infected with bladder bacteria (cyan) at MOI 10, fixed after 60 minutes and stained with 1X CellMask Orange (red). Images were acquired using a DeltaVision Elite microscope with the 40X dry NA 0.60 objective. Internalisation of bacteria was determined using a FIJI macro on 3D image data. (B) Percentage of bacterial populations fully engulfed (number of bacteria fully engulfed/total number of bacteria counted) is normalised and presented as per macrophage to allow for comparisons despite variations in number of macrophages counted from microscopy images. (C) Percentage of bacterial populations partially engulfed per macrophage as determined by microscopy presented as in B. (D) Representative image of infected macrophages with white box indicating bacteria fully engulfed. Orthogonal xz and yz views highlight internalisation of rod bacterium. (E-F) THP-1 macrophage engulfment of UTI89 filaments at the single-cell level. X-axis shows engulfment with partial engulfment being < 0 and full engulfment > 0, as determined by nearest neighbour distance calculations. (E) LEX Filaments, (F) bladder model bacterial population. Scale bar = 20 μm. Data are the averages of 3 independent experiments with error bars representing the SEM. * indicates p < 0.05, ** p < 0.01 and ns indicates p > 0.05, determined by one-way ANOVA with multiple comparisons.

Fig 7. UTI89 mutants lacking both sulA and ymfM retain the ability to form filaments and replicate engulfment dynamics of wild-type UTI89.

(A-B) Six hours post-infection, with (A) UTI89/pMAN01 (GFP) or (B) UTI89ΔsulAΔymfM/pMAN01 (GFP), C3H/HeN mouse bladders (magenta) were bisected and splayed on a silicone pad, fixed and imaged by an Olympus FV1000MPE microscope with a 20X objective (NA 0.75). Each bacterial strain was tested on a group of 4 mice with images representative from different mice. Scale bar = 50 μm. Representative images at higher magnification are shown for UTI89/pMAN01 (GFP) and UTI89ΔsulAΔymfM/pMAN01 (GFP) in (C). Scale bar = 20 μm.

Fig 8. UTI89 mutants lacking sulA and ymfM retain the ability to form filaments and replicate engulfment dynamics of wild-type UTI89.

(A-D) UTI89/pGI5 (msfGFP) or UTI89ΔsulAΔymfM/pGI5 (msfGFP) were isolated from an in vitro human bladder model (bladder bacteria or ΔsulAΔymfM bladder bacteria respectively). (A) THP-1 macrophages were infected (MOI 10) for 1 hour and bacterial loads were assessed 2-hours post infection in a gentamicin-protection assay. (B-D) Macrophages were infected with bladder bacteria or ΔsulAΔymfM bladder bacteria (cyan) at MOI 10, fixed after 60 minutes and stained with 1X CellMask Orange (red). Images were acquired using a DeltaVision Elite microscope with the 40X dry NA 0.60 objective. Internalisation of bacteria was determined using a FIJI macro on 3D image data. (B) Percentage of bacterial populations fully engulfed (number of bacteria fully engulfed/total number of bacteria counted) is normalised and presented as per macrophage to allow for comparisons despite variations in number of macrophages counted from microscopy images. (C) Percentage of bacterial populations partially engulfed per macrophage as determined by microscopy presented as in D. (D) Representative image of infected macrophages with white boxes indicating bacteria fully engulfed, and yellow arrows indicating extracellular bacteria. Orthogonal xz and yz views highlight internalisation of rod bacterium. Scale bar = 20 μm. Data are the averages of 3 independent experiments with error bars representing the SEM. * indicates p < 0.05, ** p < 0.01 and ns indicates p > 0.05, determined by one-way ANOVA with multiple comparisons.