Lipopolysaccharide Domains Modulate Urovirulence

Abstract

Uropathogenic Escherichia coli (UPEC) accounts for 80 to 90% of urinary tract infections (UTI), and the increasing rate of antibiotic resistance among UPEC isolates reinforces the need for vaccines to prevent UTIs and recurrent infections. Previous studies have shown that UPEC isolate NU14 suppresses proinflammatory NF-κB-dependent cytokines (D. J. Klumpp, A. C. Weiser, S. Sengupta, S. G. Forrestal, R. A. Batler, and A. J. Schaeffer, Infect Immun 69:6689-6695, 2001, http://dx.doi.org/10.1128/IAI.69.11.6689-6695.2001; B. K. Billips, A. J. Schaeffer, and D. J. Klumpp, Infect Immun 76:3891-3900, 2008, http://dx.doi.org/10.1128/IAI.00069-08). However, modification of lipopolysaccharide (LPS) structure by deleting the O-antigen ligase gene (waaL) enhanced proinflammatory cytokine secretion. Vaccination with the ΔwaaL mutant diminished NU14 reservoirs and protected against subsequent infections. Therefore, we hypothesized that LPS structural determinants shape immune responses. We evaluated the contribution of LPS domains to urovirulence corresponding to the inner core (waaP, waaY, and rfaQ), outer core (rfaG), and O-antigen (waaL, wzzE, and wzyE). Deletion of waaP, waaY, and rfaG attenuated adherence to urothelial cells in vitro In a murine UTI model, the ΔrfaG mutant had the most severe defect in colonization. The mutation of rfaG, waaL, wzzE, and wzyE resulted in an inability to form reservoirs in mouse bladders. Infection with the LPS mutant panel resulted in various levels of urinary myeloperoxidase. Since the ΔwaaL mutant promoted Th₁-associated adaptive responses in previous studies (B. K. Billips, R. E. Yaggie, J. P. Cashy, A. J. Schaeffer, and D. J. Klumpp, J Infect Dis 200:263-272, 2009, http://dx.doi.org/10.1086/599839), we assessed NU14 for Th₂-associated cytokines. We found NU14 infection stimulated TLR4-dependent bladder interleukin-33 (IL-33) production. Inoculation with rfaG, waaL, wzzE, and wzyE mutants showed decreased IL-33 production. We quantified antigen-specific antibodies after infection and found significantly increased IgE and IgG1 in ΔwaaP mutant-infected mice. Our studies show LPS structural constituents mediate multiple aspects of the UPEC life cycle, including the ability to acutely colonize bladders, form reservoirs, and evoke innate and adaptive immune responses.

FIG 1

Deletion of LPS biosynthetic genes modulates LPS structure. (A) Structure of LPS, with mutants depicted with arrows. Mutation of the inner core (waaP, waaY, and rfaQ), outer core (rfaG), and O-antigen (O-Ag) (wzzE, wzyE, and waaL). (B) Silver-stained SDS-PAGE of purified LPS of NU14, waaL, waaP, waaY, rfaG, rfaQ, wzzE, and wzyE strains. Lower-molecular-weight species pertain to the core-lipid A region; higher-molecular-weight species pertain to O-antigen.

FIG 2

LPS structure modulates bacterial growth and outer membrane functions. (A) Bacterial growth of NU14, ΔwaaL, ΔwaaP, ΔwaaY, ΔrfaG, ΔrfaQ, ΔwzzE, and ΔwzyE strains in LB medium. ΔrfaG and ΔwaaL mutants showed a significant decrease in bacterial growth in LB medium. (B) Bacterial susceptibility to 1 μM EDTA. ΔrfaG and ΔwaaL mutants showed a significant decrease in bacterial growth in LB medium containing 1 μM EDTA. (C) Bacterial susceptibility to 1% SDS. ΔwaaL, ΔwaaP, ΔwaaY, ΔrfaG, ΔrfaQ, and ΔwzzE mutants showed a significant decrease in bacterial growth in LB medium containing 1% SDS. Data represent means ± standard errors of the means; this experiment was performed in duplicate and repeated three times. Statistical analysis was determined by one-way analysis of variance followed by Dunnett's multiple-comparison test; the asterisks (*, P < 0.05; **, P < 0.01; ***, P < 0.001) indicate statistically significant differences between NU14 and the LPS mutants.

FIG 3

LPS structure modulates bacterial adherence and invasion. (A) Cell adherence to PD07i cells was determined at 2 hpi. ΔwaaP, ΔwaaY, and ΔrfaG mutants had severe defects in the ability to adhere to bladder cells. (B) Intracellular CFU were determined after gentamicin treatment. ΔwaaP, ΔwaaY, ΔrfaQ, and ΔrfaG mutants were severely attenuated in their ability to invade bladder cells, while the ΔwaaL mutant had a significant increase in the number of intracellular CFU. Data represent means ± standard errors of the means; each experiment was performed three times in triplicate. Statistical analysis was determined by one-way analysis of variance followed by Dunnett's multiple-comparison test; the asterisks (*, P < 0.05; **, P < 0.01; ***, P < 0.001) indicate statistically significant differences between NU14 and the LPS mutants.

FIG 4

LPS structure modulates acute bladder colonization and bladder reservoirs. (A) Acute bladder colonization. Female C57BL/6 mice were instilled via transurethral catheterization with 2 × 10⁸ CFU of NU14, ΔwaaL, ΔwaaP, ΔwaaY, ΔrfaG, ΔrfaQ, ΔwzzE, and ΔwzyE strains. ΔwaaP, ΔwaaY, ΔrfaQ, ΔrfaG, ΔwaaL, and ΔwzyE mutants were severely attenuated in their ability to colonize bladders at 24 hpi. (B) Colonization at 2 weeks postinfection likely to indicate reservoirs. Mice infected with ΔrfaG, ΔwaaL, ΔwzzE, and ΔwzyE mutants had sterile bladders at 2 weeks postinoculation. Data represent means ± standard errors of the means; each experiment was performed at least twice. Statistical analysis was determined by one-way analysis of variance followed by Dunnett's multiple-comparison test; the asterisks (*, P < 0.05; **, P < 0.01; ***, P < 0.001) indicate statistically significant differences between NU14 and the LPS mutants.

FIG 5

O-antigen suppresses urinary MPO. Female C57BL/6 mice were instilled via transurethral catheterization with PBS or 2 × 10⁸ CFU of NU14, ΔwaaL, ΔwaaP, ΔwaaY, ΔrfaG, ΔrfaQ, ΔwzzE, and ΔwzyE strains, and urine was collected at 6 hpi. Mice infected with NU14, ΔrfaQ, ΔwaaL, ΔwzzE, and ΔwzyE strains had increased urinary MPO compared to PBS-instilled mice. There were no differences between NU14-infected mice and the LPS mutant-infected mice (n = 8). Data represent means ± standard errors of the means; each experiment was performed at least twice. Statistical analysis was determined by one-way analysis of variance followed by Dunnett's multiple-comparison test; the asterisks (*, P < 0.05; **, P < 0.01; ***, P < 0.001) indicate statistically significant differences between results for PBS- and LPS mutant-treated mice.

FIG 6

UPEC induces IL-33 secretion in a TLR4-dependent manner from bladder urothelial cells. (A) Female C57BL/6 mice were instilled via transurethral catheterization with PBS or 2 × 10⁸ CFU of NU14, and bladders were harvested at 12, 24, 48, and 72 h postinoculation (n = 5). IL-33 production was increased at 12 hpi, peaked at 24 hpi, and decreased by 72 hpi. Statistical analysis was determined by one-way analysis of variance followed by Dunnett's multiple-comparison test; the asterisks (*, P < 0.05; **, P < 0.01; ***, P < 0.001) indicate statistically significant differences between PBS- and NU14-instilled mice. (B) Female C57BL/6 and TLR4^−/− mice were instilled via transurethral catheterization with PBS or 2 × 10⁸ CFU of NU14, and bladders were harvested at 24 hpi. TLR4^−/− mice infected with NU14 did not result in significant IL-33 production. Statistical analysis was determined by one-way analysis of variance followed by Tukey's multiple-comparison test; the asterisks (**, P < 0.01; ***, P < 0.001) indicate statistically significant differences between WT C57B6 mice instilled with PBS, WT C57B6 mice inoculated with NU14, and TLR4 ^−/− NU14-inoculated mice (n = 5). (C) Bladder IL-33 staining. (i) Negative control of C57BL/6 bladder section stained for IL-33, with all antibodies carried out except IL-33. (ii) PBS-instilled C57BL/6 bladder section stained for IL-33 shows minimal staining. (iii) NU14-infected C57BL/6 bladder shows increased IL-33 staining in the urothelium (arrows). (iv) NU14-infected TLR4^−/− mouse bladder shows minimal IL-33 staining. (D) O-antigen induces bladder IL-33 production. Female C57BL/6 mice were instilled via transurethral catheterization with PBS or 2 × 10⁸ CFU of NU14, ΔwaaL, ΔwaaP, ΔwaaY, ΔrfaG, ΔrfaQ, ΔwzzE, ΔwzyE strains, and bladders were harvested at 24 hpi. Mice infected with ΔrfaG, ΔwaaL, ΔwzzE, and ΔwzyE mutants showed no significant induction in IL-33 production compared to NU14-infected mice. Statistical analysis was determined by one-way analysis of variance followed by Dunnett's multiple-comparison test; the asterisks (*, P < 0.05; **, P < 0.01) indicate statistically significant differences between WT NU14 and the LPS mutants.

FIG 7

Antigen-specific antibody responses were determined by ELISA. Female C57BL/6 mice were instilled via transurethral catheterization with PBS or 2 × 10⁸ CFU of NU14-OVA, ΔwaaL-OVA, ΔwaaP-OVA, ΔwaaY-OVA, ΔrfaG-OVA, ΔrfaQ-OVA, ΔwzzE-OVA, and ΔwzyE-OVA strains on day 1 and again on day 14. The mice were challenged with NU14 on day 28, and serum was collected 24 h after UPEC challenge. (A) Mice infected with ΔwaaP and ΔwaaY mutants showed significant OVA-specific IgG1 production compared to PBS-instilled mice (n = 15). There were no significant increases compared to NU14. (B) Mice infected with the ΔwaaP mutant showed significant OVA-specific IgE production compared to PBS-instilled mice (n = 5). There were no significant increases compared to NU14. (C) No significant differences in OVA-specific IgG2c between groups (n = 15). Data represent means ± standard errors of the means; each experiment was performed at least twice. Statistical analysis was determined by one-way analysis of variance followed by Dunnett's multiple-comparison test; the asterisks (*, P < 0.05; **, P < 0.01) indicate statistically significant differences between PBS and LPS mutant treatment.