LPS O Antigen Plays a Key Role in Klebsiella pneumoniae Capsule Retention

Abstract

Despite the importance of encapsulation in bacterial pathogenesis, the biochemical mechanisms and forces that underpin retention of capsule by encapsulated bacteria are poorly understood. In Gram-negative bacteria, there may be interactions between lipopolysaccharide (LPS) core and capsule polymers, between capsule polymers with retained acyl carriers and the outer membrane, and in some bacteria, between the capsule polymers and Wzi, an outer membrane protein lectin. Our transposon studies in Klebsiella pneumoniae B5055 identified additional genes that, when insertionally inactivated, resulted in reduced encapsulation. Inactivation of the gene waaL, which encodes the ligase responsible for attaching the repeated O antigen of LPS to the LPS core, resulted in a significant reduction in capsule retention, measured by atomic force microscopy. This reduction in encapsulation was associated with increased sensitivity to human serum and decreased virulence in a murine model of respiratory infection and, paradoxically, with increased biofilm formation. The capsule in the WaaL mutant was physically smaller than that of the Wzi mutant of K. pneumoniae B5055. These results suggest that interactions between surface carbohydrate polymers may enhance encapsulation, a key phenotype in bacterial virulence, and provide another target for the development of antimicrobials that may avoid resistance issues associated with growth inhibition. IMPORTANCE Bacterial capsules, typically comprised of complex sugars, enable pathogens to avoid key host responses to infection, including phagocytosis. These capsules are synthesized within the bacteria, exported through the outer envelope, and then secured to the external surface of the organism by a force or forces that are incompletely described. This study shows that in the important hospital pathogen Klebsiella pneumoniae, the polysaccharide capsule is retained by interactions with other surface sugars, especially the repeated sugar molecule of the LPS molecule in Gram-negative bacteria known as "O antigen." This O antigen is joined to the LPS molecule by ligation, and loss of the enzyme responsible for ligation, a protein called WaaL, results in reduced encapsulation. Since capsules are essential to the virulence of many pathogens, WaaL might provide a target for new antimicrobial development, critical to the control of pathogens like K. pneumoniae that have become highly drug resistant.

Keywords: Klebsiella; LPS; O antigen; capsule; encapsulation; retention; virulence.

FIG 1

Capsule of Klebsiella pneumoniae B5055. (A) Measurements of single B5055 cells were made by atomic force microscopy to calculate capsule thickness. The dashed red line represents the fit of the curve to Hooke’s law, and blue circles represent individual measurements. (B) Representation of the capsule secretion machinery in K. pneumoniae, based on references and . The secretion pore, composed of Wza, Wzb, and Wzc subunits, spans the inner membrane (IM) and outer membrane (OM). The beta-barrel outer membrane protein Wzi is depicted in proximity to the pore. For scale, the distance from the inside surface of the outer membrane to the peptidoglycan layer (PG) is approximately 10 nm, compared with the capsule thickness, which is >250 nm. (C) Colonies of K. pneumoniae B5055 grown on LB agar can be drawn into strings through contact with a bacteriological loop. These string lengths (i.e., at the point of breakage) were measured directly using a ruler. (D) Maneval stain of the encapsulated isolate B5055. (E) Eleven insertion mutants were identified within the cps gene cluster, predominantly within the wza and wzb genes that encode the capsule secretion pore subunits Wza and Wzb. The insertions within the CPS cluster identified in the capsule-deficient screen are denoted by green arrowheads, with the number of independent insertions shown above each arrowhead.

FIG 2

Transposon insertions into LPS-associated genes may have reduced capsules. (A) Eight capsule mutants were found to have transposon insertions within lipopolysaccharide biosynthesis genes of B5055; seven were within the waa operon. The insertions identified in the capsule-deficient screen are denoted by green arrowheads. (B) The sensitivity of various K. pneumoniae B5055 strains to human serum killing was evaluated in an assay where bacteria were exposed to normal human serum (NS) or heat-treated human serum (HTS) for 90 min. The viable counts were expressed as the fold change ratio from the starting inoculum for each isolate after 90 min. The values represent the mean and standard error of the mean (SEM) from 3 biological replicates and were analyzed using a two-way analysis of variance (ANOVA) with Bonferroni’s posttest adjustment for multiple testing. *, P < 0.05; **, P < 0.01; ns, not significant. (C) Uronic acid was measured in the pellets of K. pneumoniae B5505 (WT) and in selected mutants. Uronic acid was estimated in bacterial pellets of mid-log samples, and the histograms are the mean and SEM from 3 biological replicates. (D) O1 LPS structure of B5055 inferred from reference . The glycosyltransferases that add onto each of the sugars are denoted. The LPS insertions that affected encapsulation are shown in red. The urinate-containing sugars are shown in green. For orientation, the LPS is embedded in the membrane through lipid A (RHS), and the O antigen is added to the LHS.

FIG 3

Loss of O antigen is associated with capsule loss. (A) The string test length was used to examine the relative contributions of WaaL and LPS O antigen to capsule retention. The strains tested were B5055 (WT), the capsule mutant (B5055 Δwzb-c; capsule-deficient “smooth LPS”), the defined WaaL mutant (B5055 ΔwaaL), and the complemented defined WaaL mutant (B5055 ΔwaaL-C′) with pACYC184 expressing WaaL from the Tet^r promoter. The values represent the mean and SEM from 3 biological replicates. (B) Silver stain and Western blotting were used to study the LPS phenotype in each strain. LPS preparations from K. pneumoniae strains were separated by Tricine-SDS-PAGE on 15% gels and visualized by a modified silver staining method or transferred to nitrocellulose for incubation with monoclonal antibody I12, specific for the O1 LPS of K. pneumoniae (60). The Western blot of the WaaL mutant (B5055 ΔwaaL) showed the presence of polymerized O antigen in the WaaL mutant, albeit of reduced intensity. This strain was negative by silver stain. Complementation of the WaaL mutant with pACYC184 expressing WaaL from the Tet^r promoter (B5055ΔwaaL-C′) restored LPS, producing silver-stained and Western blot patterns that were similar to those in the wild-type B5055 and mutant (B5055 Δwzb-c) LPS profiles. (C) Flow cytometry of difference mutants of K. pneumoniae B5055 stained with antibody I12 to determine O-antigen surface expression. The 5 strains tested were B5055 (WT), B5055 with wzb-c and waaF deleted (B5055Δwzb-cΔwaaF), the defined WaaL deletion mutant (B5055 ΔwaaL), and the mutant B5055 ΔwaaL-CV complemented with the empty complementation vector (pACYC184) or with pACYC184 expressing WaaL from the Tet^r promoter (B5055 ΔwaaL-C′). (D) Capsule thickness of the WaaL mutant (B5055 ΔwaaL) and the complemented WaaL mutant (B5055 ΔwaaL-C′) was measured by AFM. Cells were imaged in contact mode at a scan rate of 1 Hz at room temperature. The cell apex was probed during force measurements (82). The histograms represent the mean and SEM from 30 to 40 bacteria for each mutant for capsule thickness. (E and F) Groups of 10 to 12 C57BL/6 mice were inoculated intranasally with 5 × 10⁴ cells of K. pneumoniae B5055 or the various mutants and monitored for weight loss for 3 days (E). (F) After 3 days, the animals were killed and the viable count of K. pneumoniae estimated by plating of serially diluted lung homogenates onto LB agar. Groups were compared using a one-way ANOVA.

FIG 4

Contribution of WaaL and Wzi to capsule retention in K. pneumoniae. To examine the relative contributions of Wzi and LPS O antigen to capsule retention, the wild-type B5055 strain, single mutant strains (Δwzb-c, Δwzi, ΔwaaL, and complemented ΔwaaL) (C and E), and double mutant strain (ΔwaaL Δwzi) were analyzed for (A) string test length, (B) uronic acid levels associated with the bacterial pellets, and (C) survival in unheated human serum for 90 min, as previously described. The mutants were also tested for (D) formation of static biofilms. In panels A to C, the histograms represent the mean and SEM from 3 biological replicates, and the different strains were compared using a one-way ANOVA. For biofilm measurements (D), B5055 and the deletion mutants were tested with (group B) and without (group A) transformation with a plasmid expressing MrkH (59). MrkH is a cyclic-di-GMP-sensitive transcriptional activator of Mrk fimbria expression that has been naturally deleted from B5055, resulting in poor biofilm formation. The bacteria were added to the wells under optimized conditions and left for 24 h, and the wells were stained with crystal violet. The binding of crystal violet was determined using acetic acid solubilization, and absorbance was measured at 595 nm in a spectrophotometer. The mean and SEM from each sample (3 biological replicates, 5 technical replicates of each) were compared with those of the parent WT strain by one-way ANOVA (E). The strains were also tested for their sensitivity to a K2-specific bacteriophage, RAD2, in a plaque assay (39). The number and type of plaques reflect the sensitivity of each of the mutants to RAD2.

FIG 5

Amended model for capsule retention by K. pneumoniae. The association between the O antigen in the LPS and the polysaccharide capsule occurs through unresolved forces and possibly simple entanglement of the fibrils. The O antigen is shown in red and the core oligosaccharide in blue. Some capsule types may use an acylated carrier (pink) to link the capsule fibril to the bacterial surface. The ligation of LPS O antigen is required for maximal capsule retention, and Wzi plays a minor role in capsule retention, possibly through its properties as a lectin (57).