Role of the K2 Capsule in Escherichia coli Urinary Tract Infection and Serum Resistance

Bacterial strains, plasmids, and urine samples

Characteristics of bacterial strains and plasmids used in this study are summarized in table 1. Bacteria were stored and cultured as described elsewhere [5]. Approval, informed consent, and procedures for obtaining and storing pooled human urine samples have been described elsewhere [13].

Construction of a CFT073 capsule deletion mutant

To construct a mutant of the K2 capsule–specific locus (ksl[k2]) of strain CFT073, the 5 genes of region 2 (figure 1) were deleted and replaced with the chloramphenicol resistance gene by use of primer pair Donne-934 and Donne-978 (table 2) and the lambda red recombinase system [16], with modifications as described elsewhere [20]. Chloramphenicol-resistant colonies were tested by polymerase chain reaction (PCR) analysis performed using primers Donne-936 and Donne-979 (table 2), for replacement of region 2 with the chloramphenicol resistance marker. The chloramphenicol resistance marker was eliminated by expression of the FLP recombinase, as described elsewhere [17, 20], and confirmed by PCR performed using primers Donne-936 and Donne-979 (table 2). The resulting strain was named “UMD141.”

Identification of kps cosmid clones and subcloning of the kps_K2 gene cluster

A genomic library of CFT073 chromosomal DNA [15] was systematically screened by PCR performed using primers Donne-930 and Donne-931, as described elsewhere [20], to determine the presence of the kps gene cluster encoding all 3 regions involved in K2 capsule export and biosynthesis. A DNA fragment that contained ksl(k2)ABCDE, region 2 of the kps cluster, was amplified from cosmid clone 9-A1 by PCR performed using primers Donne-936 and Donne-979 and the Pfx DNA polymerase (Invitrogen), in accordance with the manufacturer’s instructions. The PCR product was cloned into the TOPO blunt vector (Invitrogen) to create pXLW29 excised with EcoRI, and it was ligated to EcoRI-digested pWKS30 to create pXLW36.

Capsule staining and electron microscopy

Bacteria grown overnight in human urine under static conditions were stained with ferritin (1 g/mL; Sigma) and were examined using a JEOL 1200 EX transmission electron microscope, as described elsewhere [10].

Experimental infections in mice

Urovirulence was assessed using a CBA mouse model of ascending UTI, as described elsewhere [10, 14]. In brief, CBA mice were transurethrally challenged with 10⁸ cfu/mouse. After 48 h, urine was collected. The bladder and kidneys were removed, weighed, and homogenized, and dilutions were plated on selective media by use of a spiral plater. For cochallenge infections, mice were inoculated with a mixture of 5 × 10⁷ cfu of either the CFT073^rif wild-type strain and 5 × 10⁷ of the nalidixic acid–resistant mutant UMD141 (for a total of 10⁸ cfu/inoculum) or of the mutant UMD141 strain containing the control plasmid pWKS130 and the complementing plasmid pXLW36. Because the lower limit of detection was 10² cfu, samples yielding no colonies were scored as having this value. A competitive index was calculated for each mutant, and statistical analysis was performed as described elsewhere [21]. Details are provided in the Appendix, which appears only in the electronic version of the Journal.

Polymorphonuclear leukocyte (PMNL) and serum preparations and opsonophagocytosis assay

PMNLs and serum were isolated, as described elsewhere [22], from heparinized (50 U/mL) venous blood samples obtained from 1 healthy donor in all tests. In each experiment, serum samples were obtained from the same donor, under conditions that preserved complement activity, and the final concentration was 10% vol/vol.

The opsonophagocytosis assay was performed as described elsewhere [22], with minor modifications. In brief, bacteria (grown in human urine to maximize capsule expression; 2.5 × 10⁴ cfu/well), normal human serum (10%), and Hanks’ balanced salt solution with Ca²⁺ and Mg²⁺ supplemented with 0.1% bovine serum albumin were incubated in 96-well round-bottom microtiter plates. After 30 min at room temperature with shaking, PMNLs (2.5 × 10⁵ cells/well) were added, and the reaction was allowed to proceed for another 90 min. Samples (10 μL) were obtained from each well immediately after the addition of PMNLs (time 0) and at the end of the reaction period. Counts of viable bacteria were obtained by plating serial dilutions on Luria-Bertoni agar with appropriate antibiotics. The percentage of surviving bacteria was calculated according to the following formula: (number of colony-forming units at 90 min/number of colony-forming units at 0 min) × 100%. Control reactions consisted of wells with bacteria and serum without PMNLs or bacteria alone. In addition, all reactions were also performed with serum that had been heat inactivated at 56°C for 30 min.

Genomic organization of the capsular operon of ExPEC strain CFT073

The capsule locus of ExPEC strain CFT073 contains 13 open-reading frames [12] and is similar to that of other group 2 capsule loci [7] (figure 1). The overall guanine-cytosine content (41.45%) is significantly lower than that of the K12 E. coli genome (50.80%), suggesting a foreign origin. Genes within regions 1 and 3 of the kps_K2 cluster, which encode enzymes required for the transport of capsule components, are highly homologous to those of other group 2 capsule loci. In contrast, the genes of region 2, designated “ksl(k2)ABCDE,” are predicted to be responsible for the enzymatic synthesis of the polysaccharide and share no similarity with other E. coli sequences. Interestingly, the products of the ksl(k2)A and ksl(k2)C genes are similar to proteins involved in the synthesis of teichoic acid in gram-positive organisms. The major features of the kps_K2 genes and their putative functions are summarized in table 3, which appears only in the electronic edition of the Journal.

Construction of a kps_K2 mutant impaired in capsule synthesis

Previous research [10] has demonstrated that CFT073 mutants with transposon insertions within kps genes had a reduced ability to compete with the wild-type strain to colonize the murine urinary tract. To provide conclusive evidence for the role of capsule in CFT073 virulence, we constructed a nonpolar mutant (strain UMD141) in which the entire biosynthetic region of the capsule locus (ksl[k2]ABCDE) was deleted. The growth rate of the mutant in human urine, as measured by determination of the bacterial count, was equivalent to that of the wild type (figure 2, which appears only in the electronic version of the Journal). In Luria broth, the mutant exhibited a slightly slower rate of increase in the optical density measured at 600 nm; this finding is consistent with the smaller size of the mutant, which results from the absence of capsule. Thin sections of wild-type CFT073 and mutant UMD141 strains grown in human urine were prepared and examined by electron microscopy for capsule production. As shown in figure 3, no capsule could be detected in strain UMD141, whereas complementation of strain UMD141 with plasmid pXLW36 containing ksl(k2)ABCDE restored capsule expression. These results indicate that genes within region 2 are required for the production of K2 capsule.

Role of the K2 capsule in the pathogenesis of ascending UTI

To assess the role of K2 capsule in colonization of the murine urinary tract, wild-type CFT073^rif and ksl(k2)ABCDE mutant UMD141 strains were compared by transurethral inoculation. When administered alone, mutant strain UMD141 was able to colonize the urine, the bladder, and the kidneys at levels comparable to and not significantly different from those of wild-type strain CFT073 (data not shown). Thus, capsule does not appear to play a major role in colonization of the murine urinary tract, in experiments in which bacteria are inoculated separately.

For a more sensitive test of the role of capsule in colonization of the mouse urinary tract, we performed competitive-colonization experiments in which the wild-type strain CFT073 and the ksl(k2)ABCDE deletion mutant strain UMD141 were coinoculated in a 1:1 ratio. We found that the wild-type strain significantly outcompeted the mutant strain at 48 h in urine (P = .003) and the kidneys (P = .001). The difference in bladder colonization was not significant (P = .078) (figure 4A). These results suggest that the ability to synthesize K2 capsule confers a competitive advantage for the wild-type strain. To confirm that this advantage is the result of capsular synthesis, we next performed competitive-colonization experiments comparing the mutant UMD141 with a complemented strain. Because the complemented strain has a plasmid, which might impair its fitness, we transformed UMD141 with the empty plasmid vector pWKS130. As shown in figure 4B, the complemented strain UMD141 (pXLW36) was recovered at significantly higher levels than was the mutant strain UMD141 (pWKS130) in the bladder (P = .049) and urine (P = 0.002). In the kidneys, the levels of colonization were not significantly different (P = .475). These competitive-challenge experiments provide definitive evidence that synthesis of K2 capsule by CFT073 confers an advantage for survival within the murine urinary tract.

Association between K2 capsule and protection from serum

To assess the role of the CFT073 K2 capsule in the development of resistance to killing by complement and phagocytes, we compared the ability of wild-type, capsule-deficient mutant, or complemented mutant strains to survive in the presence of human serum and PMNLs. As shown in figure 5, there were no significant differences in the survival of the wild-type and the complemented mutant strains, under any of the conditions tested. However, significantly fewer colony-forming units of the ksl(k2)ABCDE mutant strain UMD141 (pWKS30) were recovered in the presence of human serum, compared with either the complemented strain UMD141 (pXLW36) (P = .048) or the wild-type strain CFT073 (pWKS30) (P = .002). No significant differences in survival among the strains were observed after incubation in the presence of heat-inactivated serum with or without PMNLs. However, even though all 3 strains were efficiently killed in the presence of serum plus PMNLs, the ksl(k2)ABCDE mutant strain was more susceptible to killing than was the wild-type strain (P = .013). On the basis of these results, we conclude that the capsule of ExPEC strain CFT073 is required for serum resistance.