Lipopolysaccharide structure modulates cationic biocide susceptibility and crystalline biofilm formation in Proteus mirabilis

Abstract

Chlorhexidine (CHD) is a cationic biocide used ubiquitously in healthcare settings. Proteus mirabilis, an important pathogen of the catheterized urinary tract, and isolates of this species are often described as "resistant" to CHD-containing products used for catheter infection control. To identify the mechanisms underlying reduced CHD susceptibility in P. mirabilis, we subjected the CHD tolerant clinical isolate RS47 to random transposon mutagenesis and screened for mutants with reduced CHD minimum inhibitory concentrations (MICs). One mutant recovered from these screens (designated RS47-2) exhibited ~ 8-fold reduction in CHD MIC. Complete genome sequencing of RS47-2 showed a single mini-Tn5 insert in the waaC gene involved in lipopolysaccharide (LPS) inner core biosynthesis. Phenotypic screening of RS47-2 revealed a significant increase in cell surface hydrophobicity and serum susceptibility compared to the wildtype, and confirmed defects in LPS production congruent with waaC inactivation. Disruption of waaC was also associated with increased susceptibility to a range of other cationic biocides but did not affect susceptibility to antibiotics tested. Complementation studies showed that repression of smvA efflux activity in RS47-2 further increased susceptibility to CHD and other cationic biocides, reducing CHD MICs to values comparable with the most CHD susceptible isolates characterized. The formation of crystalline biofilms and blockage of urethral catheters was also significantly attenuated in RS47-2. Taken together, these data show that aspects of LPS structure and upregulation of the smvA efflux system function in synergy to modulate susceptibility to CHD and other cationic biocides, and that LPS structure is also an important factor in P. mirabilis crystalline biofilm formation.

Keywords: AMR; Proteus mirabilis; antimicrobials; biocide; biocide resistance; biofilm; catheter associated urinary tract infection; lipopolysaccharide.

Figure 1

Impact of RS47-2 waaC disruption on lipopolysaccharide (LPS) structure. The effects of waaC disruption on LPS structure were assessed by comparison of RS47 WT and RS47-2 LPS profiles. (A) Illustration showing predicted impact of waaC inactivation on LPS structure in RS47-2. (B) Separation of LPS lysates on tris-glycine gels comparing core and O-antigen regions between RS47 WT and RS47-2.

Figure 2

Impact of LPS alterations on hydrophobicity and autoaggregation in RS47-2. (A) The hydrophobicity of Proteus mirabilis clinical isolate RS47 WT and RS47-2 was evaluated by measuring percentage (%) adhesion to hexadecane. (B) Autoaggregation of RS47 WT and RS47-2 cells after 2 h incubation at room temperature. All data represent the mean of 3 biological replicates. Error bars show standard error of the mean (SEM). ^*p ≤ 0.05. ^**p ≤ 0.01.

Figure 3

Impact of LPS alterations on survival in NHS. (A) Survival of WT RS47 in normal human serum (NHS). (B) Survival of RS47-2 in NHS. Heat inactivated serum (HI) was included as a control. Figures show the mean of three biological replicates. Error bars show standard error of the mean (SEM). Ordinary one-way ANOVA with Dunnett’s post hoc test were performed, comparing bacterial survival in NHS to the 0 % serum input control. ^*p ≤ 0.05. ^***p ≤ 0.001.

Figure 4

Impact of LPS alterations on motility. (A) Swimming ability of RS47 WT and RS47-2 in 0.3  % LB agar. Data shows distance migrated after 12 h incubation. (B) Swarming ability of RS47 WT and RS47-2. Data presented shows the diameter of swarm migration over 1.5  % LB agar from the central inoculum after 18 h incubation. A distance of zero mm indicates no expansion of the colony beyond the initial inoculum and no detectable swarming motility. All data represent the mean of five biological replicates and error bars show standard error of the mean (SEM). ^**p ≤ 0.01; ^****p ≤ 0.0001.

Figure 5

Impact of waaC disruption on crystalline biofilm formation. In vitro bladder models were used to simulate infection of the catheterized urinary tract. (A) Time taken for catheters to block in models inoculated with the RS47 WT or RS47-2 mutant. (B) pH of residual bladder urine at time of catheter blockage. (C) Number of viable cells in residual bladder urine at time of catheter blockage. Data represents the mean of a minimum of five biological replicates and error bars represent SEM. ^*p ≤ 0.05; ^**p ≤ 0.01.

Figure 6

Overview of mechanisms modulating CHD susceptibility in Proteus mirabilis. (A) Low susceptibility to CHD is conferred by LPS structure and smvA efflux in WT P. mirabilis clinical isolate RS47 (smvA on, waaC on). Positively charged CHD molecules are occluded by negative charges on the LPS while smvA mediated efflux provides further protection from CHD that enters the periplasmic space. (B) Intermediate susceptibility to CHD is conferred by variation in LPS structure or inactivation of smvA efflux activity, observed in RS47::pGEM-TsmvR (smvA off, waaC on) and RS47-2 (smvA on, waaC off) respectively. Intact LPS in RS47::pGEM-TsmvR occludes entry of CHD into the periplasmic space but subsequent protection from smvA efflux is absent. Truncation of LPS in RS47-2 increases penetration of CHD into periplasmic space, but some protection is still provided by smvA efflux. (C) High susceptibility to CHD is observed in RS47-2::pGEM-TsmvR where protection from LPS and smvA efflux are compromised. LPS truncation allows greater CHD penetration into periplasmic space and protection from smvA efflux is absent.