CXCL10 Acts as a Bifunctional Antimicrobial Molecule against Bacillus anthracis

Abstract

Bacillus anthracis is killed by the interferon-inducible, ELR(-) CXC chemokine CXCL10. Previous studies showed that disruption of the gene encoding FtsX, a conserved membrane component of the ATP-binding cassette transporter-like complex FtsE/X, resulted in resistance to CXCL10. FtsX exhibits some sequence similarity to the mammalian CXCL10 receptor, CXCR3, suggesting that the CXCL10 N-terminal region that interacts with CXCR3 may also interact with FtsX. A C-terminal truncated CXCL10 was tested to determine if the FtsX-dependent antimicrobial activity is associated with the CXCR3-interacting N terminus. The truncated CXCL10 exhibited antimicrobial activity against the B. anthracis parent strain but not the ΔftsX mutant, which supports a key role for the CXCL10 N terminus. Mutations in FtsE, the conserved ATP-binding protein of the FtsE/X complex, resulted in resistance to both CXCL10 and truncated CXCL10, indicating that both FtsX and FtsE are important. Higher concentrations of CXCL10 overcame the resistance of the ΔftsX mutant to CXCL10, suggesting an FtsX-independent killing mechanism, likely involving its C-terminal α-helix, which resembles a cationic antimicrobial peptide. Membrane depolarization studies revealed that CXCL10 disrupted membranes of the B. anthracis parent strain and the ΔftsX mutant, but only the parent strain underwent depolarization with truncated CXCL10. These findings suggest that CXCL10 is a bifunctional molecule that kills B. anthracis by two mechanisms. FtsE/X-dependent killing is mediated through an N-terminal portion of CXCL10 and is not reliant upon the C-terminal α-helix. The FtsE/X-independent mechanism involves membrane depolarization by CXCL10, likely because of its α-helix. These findings present a new paradigm for understanding mechanisms by which CXCL10 and related chemokines kill bacteria.

Importance: Chemokines are a class of molecules known for their chemoattractant properties but more recently have been shown to possess antimicrobial activity against a wide range of Gram-positive and Gram-negative bacterial pathogens. The mechanism(s) by which these chemokines kill bacteria is not well understood, but it is generally thought to be due to the conserved amphipathic C-terminal α-helix that resembles cationic antimicrobial peptides in charge and secondary structure. Our present study indicates that the interferon-inducible, ELR(-) chemokine CXCL10 kills the Gram-positive pathogen Bacillus anthracis through multiple molecular mechanisms. One mechanism is mediated by interaction of CXCL10 with the bacterial FtsE/X complex and does not require the presence of the CXCL10 C-terminal α-helix. The second mechanism is FtsE/X receptor independent and kills through membrane disruption due to the C-terminal α-helix. This study represents a new paradigm for understanding how chemokines exert an antimicrobial effect that may prove applicable to other bacterial species.

FIG 1

The CXCR3-similar region of FtsX is important in mediating CXCL10 antimicrobial activity against the B. anthracis parent strain. (A) A synthetic peptide consisting of the 27 aa (aa 54 to 80) of FtsX with similarity to the CXCR3-binding region (aa 9 to 35) of CXCL10 (35) and a scrambled control. (B) A peptide competition assay with the FtsX peptide in a 10:1 or 20:1 molar ratio with 0.46 µM CXCL10 resulted in statistically significant less killing of the B. anthracis parent strain than by CXCL10 alone. Incubation with the control peptide had no effect on killing. Bacterial viability was measured by alamarBlue reduction, and fluorescence is expressed as a percentage of that of the strain-specific untreated control. Data points represent mean values ± the standard errors of the means; n = 3 separate experiments using triplicate wells. **, P ≤ 0.001; ***, P ≤ 0.0001

FIG 2

Complete amino acid sequence and secondary-structure comparison of CXCL10 and CTTC. (A) Amino acid sequences of intact CXCL10 and synthetic protein CTTC (aa 1 to 54) lacking the C-terminal α-helix. The amino acids removed from intact CXCL10 are in bold. (B, C) CD showed a distinct α-helical pattern in the intact CXCL10 spectrum, as indicated by peaks at wavelengths of 190 and 210 nm; these peaks were absent from the CTTC spectrum. The calculated β-sheet contents of both the CXCL10 and CTTC spectra were consistent with the estimated percentage of β-sheets in the published CXCL10 crystal structure (45).

FIG 3

Antimicrobial activity of CTTC was retained, albeit less potently, against the B. anthracis parent strain but not against the B. anthracis ΔftsX mutant. (A) CTTC killed the B. anthracis parent strain but was less potent than intact CXCL10 at concentrations ranging from 0 to 2.8 µM. (B) CTTC at 0 to 2.8 µM exhibited no antimicrobial effect against the B. anthracis ΔftsX mutant, with a statistically significant difference observed between intact CXCL10 and CTTC. Bacterial viability was measured by alamarBlue reduction, and fluorescence is expressed as a percentage of that of the strain-specific untreated control. Data points represent mean values ± the standard errors of the means; n = 3 separate experiments using triplicate wells in each experiment. **, P ≤ 0.001; ***, P < 0.0001

FIG 4

The B. anthracis ftsE(K123A/D481N) mutant was resistant to CXCL10 and CTTC, in contrast to the parent strain. (A) The B. anthracis ftsE(K123A/D481N) mutant (designated as "ftsE" in this figure) had a slightly longer lag phase prior to initiating log-phase growth than the B. anthracis parent strain, a growth characteristic similar to that observed in the B. anthracis ΔftsX mutant (35). (B) The B. anthracis ftsE(K123A/D481N) mutant was relatively resistant to CXCL10 and resistant to CTTC at all of the concentrations tested, from 0 to 2.8 µM. Results were similar to those obtained with the B. anthracis ΔftsX mutant. A statistically significant difference between CXCL10 and CTTC was observed at concentrations of ≥0.9 µM. (C) The B. anthracis ftsE(K123A/D481N) mutant transformed with IPTG-inducible plasmid pUTE973 (empty vector control) exhibited resistance to 0.9 µM CXCL10 in the absence or presence of 25 µM IPTG. Transformation of the B. anthracis ftsE(K123A/D481N) mutant with IPTG-inducible plasmid pUVA424 (ftsE complementation vector) resulted in restoration of susceptibility to 0.9 µM CXCL10 only upon treatment with 25 µM IPTG to induce gene expression. (D) The B. anthracis ftsE(K123A/D481N) mutant transformed with plasmid pUTE973 exhibited resistance to 2.3 µM CTTC in the absence or presence of 25 µM IPTG. The B. anthracis ftsE(K123A/D481N) mutant transformed with pUVA424 exhibited susceptibility to 2.3 µM CTTC only upon the addition of 25 µM control. Data points represent mean values ± the standard errors of the means; n = 3 separate experiments using IPTG to induce gene expression. Bacterial viability was measured by alamarBlue reduction, and fluorescence is expressed as a percentage of that of the strain-specific untreated triplicate wells in each experiment. *, P ≤ 0.01; ***, P ≤ 0.0001; ****, P < 0.0001.

FIG 5

Membrane depolarization occurs after exposure of B. anthracis to CXCL10. (A) B. anthracis parent strain membrane depolarization occurs after exposure to CXCL10 or CTTC. (B) B. anthracis ΔftsX mutant membrane depolarization occurs upon exposure to CXCL10 but not upon exposure to CTTC. The positive control (LL37 at 25 µM) induced membrane depolarization. Exposure to the negative control (CCL5 at 2.8 µM) did not cause membrane depolarization (n = 4 or 5 separate experiments). (C) Killing of the bacteria by CXCL10 or CTTC was correlated with the membrane depolarization assay results obtained under the same conditions (n = 4 or 5 separate experiments). Numbers of CFU per milliliter are plotted on a log scale. (D) Treatment of the B. anthracis ΔftsX mutant with CXCL10 or CTTC at the concentrations indicated results in less killing than treatment of the parent strain shown in panel A (n = 4 or 5 separate experiments). Incubation with the positive control (LL37) resulted in less survival of both strains than of the untreated control strain. The negative control (CCL5) was not bactericidal (n = 4 or 5 separate experiments). Statistical analysis was performed by two-way ANOVA. ns, not significant; *, P ≤ 0.01; **, P ≤ 0.001; ***, P ≤ 0.0001; ****, P < 0.0001.

FIG 6

Hypothetical model of the bifunctional antimicrobial activity of CXCL10 against B. anthracis. (A) FtsE/X-mediated antimicrobial activity of CXCL10. CXCL10 elicits an antimicrobial effect against B. anthracis via an FtsE/X-dependent pathway through the N-terminal portion and/or other regions of CXCL10, independently of the C-terminal α-helix. Cell death results through lysis. (B) FtsE/X-independent antimicrobial activity of CXCL10. CXCL10 interacts with B. anthracis in an FtsX-independent manner to cause bacterial killing, likely through interaction or insertion of the C-terminal α-helix into the bacterial membrane, resulting in cell lysis. While the mechanistic pathways are different, both appear to lead to the same result of cell lysis.