Complement protective epitopes and CD55-microtubule complexes facilitate the invasion and intracellular persistence of uropathogenic Escherichia coli

Abstract

Background: Escherichia coli-bearing Dr-adhesins (Dr+ E. coli) cause chronic pyelonephritis in pregnant women and animal models. This chronic renal infection correlates with the capacity of bacteria to invade epithelial cells expressing CD55. The mechanism of infection remains unknown.

Methods: CD55 amino acids in the vicinity of binding pocket-Ser155 for Dr-adhesin were mutated to alanine and subjected to temporal gentamicin-invasion/gentamicin-survival assay in Chinese hamster ovary cells. CD55/microtubule (MT) responses were studied using confocal/electron microscopy, and 3-dimensional structure analysis.

Results: Mutant analysis revealed that complement-protective CD55-Ser165 and CD55-Phe154 epitopes control E. coli invasion by coregulating CD55-MT complex expression. Single-point CD55 mutations changed E. coli to either a minimally invasive (Ser165Ala) or a hypervirulent pathogen (Phe154Ala). Thus, single amino acid modifications with no impact on CD55 structure and bacterial attachment can have a profound impact on E. coli virulence. While CD55-Ser165Ala decreased E. coli invasion and led to dormant intracellular persistence, intracellular E. coli in CD55-Phe154Ala developed elongated forms (multiplying within vacuoles), upregulated CD55-MT complexes, acquired CD55 coat, and escaped phagolysosomal fusion.

Conclusions: E. coli target complement-protective CD55 epitopes for invasion and exploit CD55-MT complexes to escape phagolysosomal fusion, leading to a nondestructive parasitism that allows bacteria to persist intracellularly.

Keywords: CD55; Dr+ E. coli; intracellular persistence; invasion; microtubules.

Figure 1.

CD55 SCR3 mutants differ in Escherichia coli invasion and survival capacities. A, Location of mutated amino acids on CD55's SCR2 and SCR3 domains. This figure was obtained using Cn3D (PubMed structure). The yellow labels represent amino acids mutated for the study. B, Results from a 3-hour Dr+ E. coli invasion assay using Chinese hamster ovary (CHO) cells expressing wild-type (WT) and mutated human CD55; the CD55(−) control represents group I mutants with no binding or invasion characteristics. Other mutant group representatives (all selected mutants are alanine substitutions) include group II (G159), group III (S165), group IV (F123), and group V (F148 and F154). Results are shown as means ± standard deviation (SD; n = 3). Asterisk (*) indicates significance (P ≤ .05) vs CD55(+) WT control. C, Results from temporal invasion assays with invasion stopped at earlier time points (15, 30, 45, or 60 minutes). D, Results from invasion assays stopped at 15 or 30 minutes; data are reported using expanded scales for both time points. E, Invasion results from assays stopped at 90 minutes and 180 minutes. F, Survival assay comparing WT, mutant, or CD55(−) CHO control cells that had no significant change in the number of bacteria within CHO cells between 30 and 120 minutes. G, Mutant F154A survival assay data supporting killing followed by multiplication of bacteria over time. H, Mutant F148A survival assay data supporting multiplication followed by killing of bacteria over time. Results shown are mean ± SD for 3 independent experiments.

Figure 2.

Transmission electron microscopy analysis of Escherichia coli invasion in wild-type (WT)-CD55 and mutants F148A and F154A. A, Dr+ E. coli 2-hour invasion results of Chinese hamster ovary (CHO) cells containing WT-CD55 showing a tight vacuole containing a single bacterium. B, Dr+ E. coli 2-hour invasion results of CHO cells containing CD55 mutant F154A showing a microcolony containing 4 bacteria in a tight vacuole. The arrows indicate the tightly apposing limiting membranes of 4 individual vacuoles fusing to form the big vacuole. C, Following a 2-hour invasion assay, a CD55 mutant F148A example showing a dividing bacterium within a single vacuole. D, Mutant F148A showing a larger colony with several bacteria, each within its own vacuole (arrows point to individual bacteria within their own vacuoles). E, Mutant F148A showing a loose vacuole containing 4 bacteria and some debris after a 2-hour invasion. F, Mutant F154A showing a loose vacuole containing 2 bacteria and some bacterial debris after a 2-hour invasion (arrows point to the bacterial debris). Bar = 500 nm.

Figure 3.

Dr+ Escherichia coli interact with microtubules (MT) and MT and CD55 follow the same expression patterns in wild-type (WT) and mutant CD55 expressing Chinese hamster ovary (CHO) cells. A, Immunofluorescence confocal microscopy images showing invasion of Dr+ E. coli in CHO cells expressing WT-CD55, mutated CD55, or cells lacking CD55(−). Red = MT and green = bacteria. The panels indicate projected images of internal Z stacks (60 × magnification) of 0.5 μm each. Mutants F148A and F154A showed an increased invasion capacity compared with WT-CD55, while mutant S165A behaved like the negative control CD55(−). White arrows in WT-CD55 panels indicate a single bacterium. Arrows in mutants F148A and F154A indicate larger microcolonies of bacteria. Insets in 3-hour images show bacteria apparently tethered to, or in very close association with MT. B, Quantitation of MT fluorescence intensity (MtRelFl) at 3 hours for WT-CD55, mutants, and negative controls. Values represent the mean ± standard deviation (SD) obtained from at least 20 cells. All images were taken at the same intensity (set at 5) and gain (set at 3) with a pixel dwell time of 1.1 μs/pixel. C, Graphic representation of the densitometric analysis of immunoblots showing changes in CD55 protein expression upon infection of WT and mutant CD55 expressing CHO cells with Dr+ E. coli. Quantitation and analysis of bands were performed using Odyssey software. β-actin was used as a normalization control, and the results are expressed relative to the level of β-actin. Asterisk (*) indicates significance, P ≤ .05, vs CD55+ controls. Values are mean ± SD (n = 3).

Figure 4.

Dr+ Escherichia coli bind to a complex including microtubules (MT) and CD55 and develop a CD55-mutant–dependent morphological plasticity. A, Wild-type (WT)-CD55–containing Chinese hamster ovary (CHO) cell assays only. Taken from a 3-hour E. coli invasion assay. Examples show a colocalization overlay, followed by E. coli (blue), CD55 (green), and MT (red) immunofluorescences alone. Arrows indicate the position of a single E. coli that is colocalized with CD55 and MT, as an example. B, WT-CD55–containing CHO cell assays only. From a 3-hour invasion assay that shows colocalization of CD55 and MT yielding a yellow overlay. Arrows pinpoint various spots and indicate colocalization of CD55 and MT. C, Four comparative overlays. A comparison of E. coli 3-hour invasion assays using WT-CD55 or selected CD55 mutants (S165A, F148A, and F154A). Individual figures show the differences in the number of bacteria invading the cells and also the different forms of E. coli with respect to size and extent of CD55 coating (white arrows, bacteria that were heavily coated with CD55; yellow arrows, bacteria that were lightly coated with CD55; and fuchsia arrows, bacteria that had an undetectable CD55 coating). Alexa Fluor 633–labeled bacteria (blue) were used to infect cells in these experiments. A–C are projected images of Z stacks (60 × magnification), 0.5 μm each.

Figure 5.

CD55 and microtubules (MT) form molecular complexes. A, In situ proximity ligation assay (PLA) showing green spots (from fluorescein isothiocyanate–labeled probe) indicate the close association of CD55 and MT as seen by immunofluorescence (60 × magnification) in wild-type (WT)-CD55 cells. B, PLA showing WT-CD55 cells after nocodazole (NDZ) treatment (5 µg/µL for 1 hour). C, CD55 (−) Chinese hamster ovary cell patterns. Diamidino-2-phenylindole (blue) was used to stain the nuclei.

Figure 6.

Dr+ Escherichia coli escape phagolysosomal fusion in WT-CD55 Chinese hamster ovary (CHO) cells; modeled structures of CD55 SCR3 generated using MPACK showing CD55 SCR3 and Ser165 mutants. A, Wild-type (WT)-CD55 expressing CHO cells transfected with EGFP-LC3 were infected with Alexa Fluor 633–labeled Dr+ E. coli (red) for 3 hours. White arrows in the merged inset image show that bacteria do not colocalize with LC3. The panels are projected images of Z stacks (60 × magnification) 0.5 μm each. B, The negatively charged surface in Ser165 and Asp181 of WT-CD55 is indicated by arrow 1, and the surface of Dr-adhesin binding site is indicated by arrow 2. This figure was generated using MOLecule analysis and MOLecule display (MOLMOL). Other key residues (D181, K161, and R139) are also noted. Areas in red have a negative potential of −0.5 units and those in blue have a positive potential of +0.5 units. C, A modeled structure of a CD55 Ser165Ala mutant that binds to Dr+ E. coli as well as WT-CD55 but has a significantly reduced capability of promoting the invasion of bacteria. D, A modeled structure of a CD55 Ser165Leu mutant that neither binds to Dr+ E. coli nor promotes its invasion capabilities.