Aerobactin mediates virulence and accounts for increased siderophore production under iron-limiting conditions by hypervirulent (hypermucoviscous) Klebsiella pneumoniae

Abstract

Hypervirulent (hypermucoviscous) Klebsiella pneumoniae (hvKP) strains are an emerging variant of "classical" K. pneumoniae (cKP) that cause organ and life-threatening infection in healthy individuals. An understanding of hvKP-specific virulence mechanisms that enabled evolution from cKP is limited. Observations by our group and previously published molecular epidemiologic data led us to hypothesize that hvKP strains produced more siderophores than cKP strains and that this trait enhanced hvKP virulence. Quantitative analysis of 12 hvKP strains in iron-poor minimal medium or human ascites fluid showed a significant and distinguishing 6- to 10-fold increase in siderophore production compared to that for 14 cKP strains. Surprisingly, high-pressure liquid chromatography (HPLC)-mass spectrometry and characterization of the hvKP strains hvKP1, A1142, and A1365 and their isogenic aerobactin-deficient (ΔiucA) derivatives established that aerobactin accounted for the overwhelming majority of increased siderophore production and that this was not due to gene copy number. Further, aerobactin was the primary factor in conditioned medium that enhanced the growth/survival of hvKP1 in human ascites fluid. Importantly the ex vivo growth/survival of hvKP1 ΔiucA was significantly less than that of hvKP1 in human ascites fluid, and the survival of outbred CD1 mice challenged subcutaneously or intraperitoneally with hvKP1 was significantly less than that of mice challenged with hvKP1 ΔiucA. The lowest subcutaneous and intraperitoneal challenge inocula of 3 × 10(2) and 3.2 × 10(1) CFU, respectively, resulted in 100% mortality, demonstrating the virulence of hvKP1 and its ability to cause infection at a low dose. These data strongly support that aerobactin accounts for increased siderophore production in hvKP compared to cKP (a potential defining trait) and is an important virulence factor.

FIG 1

Growth/survival of hypervirulent Klebsiella pneumoniae (hvKP) strains grown in human ascites fluid supplemented with 0% and 5% homologous, conditioned ascites fluid. The growth/survival of six hvKP strains was assessed at 0, 3, 6, and 24 h in 100% human ascites fluid supplemented with 0% or 5% homologous conditioned ascites fluid. Both low (between 10³ and 10⁴ CFU) and high (between 10⁷ and 10⁸ CFU) starting titers were evaluated. (A) NTUH-K2044. (B) N7205. (C) A4528. (D) A1365. (E) A9534. (F) A1142. Supplementation with conditioned, homologous ascites fluid (5% final concentration) resulted in a significant increase in growth/survival for all strains at the low starting titer compared to 0% supplementation (*, P < 0.05/2). Data are means ± SEM; n = 3 or 4.

FIG 2

Quantitative measurement of siderophores (SP) in 12 hvKP and 14 cKP strains grown in iron-poor minimal medium (MM), human ascites fluid, and Luria-Bertani broth (LB). Quantitative SP measurements were performed on bacterium-free supernatants harvested after overnight growth. A, MM. B, human ascites fluid. C, LB. The median SP concentration for hvKP strains grown in MM and human ascites fluid, but not LB, was significantly greater than that for cKP strains (*, P < 0.05/3). Each symbol represents the mean concentration from 3 independent conditioned media measured in duplicate for a single strain.

FIG 3

HPLC-mass spectrometry analysis of hvKP1-generated conditioned medium. (A) A representative HPLC profile of conditioned medium generated from the growth of hvKP1 in human ascites fluid. The fractions were designated 1 to 5. (B) Fraction 5 was further separated by HPLC into fractions 5a to 5e. (C) Mass spectrometric analysis was performed on fraction 5c and established that the species present was aerobactin. (D) The biological activities of fractions 5a to 5e were assessed by assessing the growth of hvKP1 in human ascites fluid supplemented with each fraction (5% final concentration). Supplementation with fractions 5c and 5d resulted in a significant increase in growth and/or survival for hvKP1 compared to 0% supplementation (*, P < 0.05/5). Data are means ± SEM; n = 3.

FIG 4

Quantitative measurement of siderophores (SP) in hvKP1, hvKP1 ΔiucA, A1142, and A1142 ΔiucA grown in iron-poor MM, human ascites fluid, and LB broth. Quantitative SP measurements were performed on bacterium-free supernatants harvested after overnight growth. 1, MM; 2, human ascites fluid; 3, LB. (A) hvKP1 and hvKP1 ΔiucA. (B) A1142 and A1142 ΔiucA. (C) A1365 and A1365 ΔiucA. The median SP concentrations for hvKP1, A1142, and A1365 compared to hvKP1 ΔiucA, A1142 Δiuc, and A1365 ΔiucA when grown in MM and for hvKP1 and A1142 compared to hvKP1 ΔiucA and A1142 Δiuc when grown in ascites fluid were significantly greater (*, P < 0.05/3). There was a trend for the median SP concentration of A1365 compared to A1365 ΔiucA when grown in ascites fluid (#, P < 0.05 but >0.05/3). Each symbol represents the mean concentration from an independent conditioned medium measured at least in duplicate for each strain.

FIG 5

The gene copy numbers for iucA, entH, iroB, and irp2 are similar. Quantitative PCR was performed on 4 independent concentrations of chromosomal DNA. The calculated threshold (cycles) correlates inversely with gene copy number. Black, iucA; red, entH; blue, iroB; yellow, irp2; purple, 23S RNA. *, P < 0.05/4 for 23S RNA compared to iucA, entH, iroB, and irp2 for all DNA concentrations. #, P > 0.05/4 but P < 0.1 for entH compared to iucA, iroB, and irp2 for 5,000 pg. *, P < 0.05/4 for iucA compared to entH and irp2 for 50 pg. n = 3 for all measurements.

FIG 6

Growth/survival of hvKP1 and hvKP1 ΔiucA in human ascites fluid supplemented with purified aerobactin or conditioned medium generated by hvKP1 or hvKP1 ΔiucA. The growth/survival of hvKP1 and hvKP1 ΔiucA was assessed at 0, 3, 6, and 24 h in 100% human ascites fluid. (A) Growth/survival of hvKP1 in ascites fluid that was not supplemented (0%) or was supplemented with 10% conditioned ascites fluid generated by hvKP1 (aerobactin replete) or hvKP1 ΔiucA (aerobactin deficient) or purified aerobactin (460 nM; 260 μg/ml). *, P < 0.05/3. (B) Growth/survival of hvKP1 ΔiucA in ascites fluid that was not supplemented (0%) or was supplemented with 10% of conditioned ascites fluid generated by hvKP1 (aerobactin replete) or hvKP1 ΔiucA (aerobactin deficient) or purified aerobactin (460 nM; 260 μg/ml). *, P < 0.05/3. Data are means ± SEM; n = 4 to 6.

FIG 7

Growth/survival of hvKP1 and hvKP1 ΔiucA in human ascites fluid. The growth/survival of hvKP1, hvKP1/pFUS2, hvKP1 ΔiucA, hvKP1 ΔiucA/pFUS2, and hvKP1 ΔiucA/pFUS2[iucA-D] was assessed in 100% human ascites fluid. (A) A midlevel starting inoculum of approximately 1 × 10⁶ CFU/ml was used, and growth/survival was assessed at 0, 3, 6, and 24 h. The growth/survival of hvKP1 and hvKP1/pFUS2 was significantly increased compared to that of hvKP1 ΔiucA and hvKP1 ΔiucA/pFUS2, respectively. *, P < 0.05/2. Data are means ± SEM (n = 5 to 8). (B) A low starting inoculum of <1 × 10⁵ CFU/ml was used, and growth/survival was assessed at 0, 12, 24, 48, 72, and 96 h. The growth/survival of hvKP1 was significantly increased compared to that of hvKP1 ΔiucA. *, P < 0.05. Data are means ± SEM from 3 independent experiments with n = 3 for each experiment.

FIG 8

Survival of outbred CD1 mice after subcutaneous (s.c.) or intraperitoneal (i.p.) challenge with hvKP1 and hvKP1 ΔiucA. (A and B) Animals were challenged s.c. with 3.0 × 10² (n = 5), 3.5 × 10³ (n = 10), 3.5 × 10⁴ (n = 10), or 3.5 × 10⁵ (n = 10) CFU of hvKP1 (A) or with 2.8 × 10² (n = 5), 3.6 × 10³ (n = 10), 3.6 × 10⁴ (n = 10), or 3.6 × 10⁵ (n = 10) CFU of hvKP1 ΔiucA (B). (C and D) Animals were challenged i.p. with 3.2 × 10¹ (n = 5), 3.2 × 10² (n = 5), 3.2 × 10³ (n = 5), or 3.5 × 10⁴ (n = 5) CFU of hvKP1 (C) or with 2.4 × 10¹ (n = 5), 2.4 × 10² (n = 5), 2.4 × 10³ (n = 5), or 2.4 × 10⁴ (n = 5) CFU of hvKP1 ΔiucA (D). Strains were grown overnight in LB medium. An in extremis state or death was scored as nonsurvival. *, P < 0.05 for hvKP1 compared to hvKP1 ΔiucA.