Discovery, characterization and in vivo activity of pyocin SD2, a protein antibiotic from Pseudomonas aeruginosa

Abstract

Increasing rates of antibiotic resistance among Gram-negative pathogens such as Pseudomonas aeruginosa means alternative approaches to antibiotic development are urgently required. Pyocins, produced by P. aeruginosa for intraspecies competition, are highly potent protein antibiotics known to actively translocate across the outer membrane of P. aeruginosa. Understanding and exploiting the mechanisms by which pyocins target, penetrate and kill P. aeruginosa is a promising approach to antibiotic development. In this work we show the therapeutic potential of a newly identified tRNase pyocin, pyocin SD2, by demonstrating its activity in vivo in a murine model of P. aeruginosa lung infection. In addition, we propose a mechanism of cell targeting and translocation for pyocin SD2 across the P. aeruginosa outer membrane. Pyocin SD2 is concentrated at the cell surface, via binding to the common polysaccharide antigen (CPA) of P. aeruginosa lipopolysaccharide (LPS), from where it can efficiently locate its outer membrane receptor FpvAI. This strategy of utilizing both the CPA and a protein receptor for cell targeting is common among pyocins as we show that pyocins S2, S5 and SD3 also bind to the CPA. Additional data indicate a key role for an unstructured N-terminal region of pyocin SD2 in the subsequent translocation of the pyocin into the cell. These results greatly improve our understanding of how pyocins target and translocate across the outer membrane of P. aeruginosa. This knowledge could be useful for the development of novel anti-pseudomonal therapeutics and will also support the development of pyocin SD2 as a therapeutic in its own right.

Keywords: Pseudomonas aeruginosa; antibiotics; bacteriocins; common polysaccharide antigen; outer membrane; pyocins.

Figure 1. Discovery and purification of pyocins SD1, SD2 and SD3

(a) Comparison of the amino acid % identity for pyocins SD1, SD2 and SD3, with pyocins S1, S2, S3 and colicin D. The cytotoxic domains of pyocins SD1, SD2 and SD3 are homologous with colicin D and the N-terminal domains are homologous with the respective pyocins. Pyocins SD1 and SD2 are nearly identical in their cytotoxic domains and immunity proteins. (b) SDS/PAGE gel (4–12%) of purified pyocins SD1 (63 kDa), SD2 (72 kDa) and SD3 (79 kDa) and their immunity proteins (10 kDa) post purification. (c) Pyocin S2(1–209) (1 mg·ml⁻¹) is not active against CF18. (d) Inhibition of pyocin S2 killing (1 μg·ml⁻¹) by pyocin S2(1–209) (1 mg·ml⁻¹).

Figure 2. Sequence alignment and secondary structure predictions for pyocin SD2

(a) Previously proposed domain architecture of pyocin S2. (b) Predicted secondary structure features of pyocin SD2 using PSIPRED-software [32]. The first 50 residues of pyocin SD2 have little regular secondary structure and are rich in proline and glycine residues. (c) Newly proposed domain architecture of pyocin SD2.

Figure 3. Small angle X-ray scattering model of pyocin SD2

(a) Ab initio model of pyocin SD2 computed with DAMMIF and averaged with DAMAVER. (b) Overlay of the experimentally determined pyocin SD2 SAXS curve (black points) with the fit of the ab initio model (purple line), produces a good fit (χ=1.158). (c) Pair-distance distribution plots from experimental scattering data for pyocin SD2 suggest that this protein is an elongated multi-domain protein in solution. D_max=215 Å. (d) Guinier plot of scattering data indicates that pyocin SD2 is monomeric in solution. Radius of gyration is 54.4 Å.

Figure 4. Pyocin SD2 utilizes the outer membrane protein FpvAI

Ten microlitres of purified pyocin was spotted on to a growing lawn of cells. Clear zones indicate cell death. (a) Inhibition of growth of P. aeruginosa PAO1 by pyocin SD2 (1 mg·ml⁻¹). (b) Pyocin SD2 (1 mg·ml⁻¹) is not active against PAO1ΔfpvAI. (c) Pyocin SD2 killing (1 mg·ml⁻¹) is restored against PAO1ΔfpvAI::fpvAI. (d) Pyocin S2 (2 mg·ml⁻¹) is not active against PAO1. (e) Inhibition of pyocin SD2 killing (1 mg·ml⁻¹) by pyocin S2 (2 mg·ml⁻¹) against PAO1. (f) Pyocin SD2 (1 mg·ml⁻¹) is not active against PAO1ΔfpvAI::fpvAI_TBM. (g) P. aeruginosa PAO1 was grown in LB broth with shaking at 37°C to OD₆₀₀=0.2. The culture was then divided into four separate flasks. Pyocin SD2 was added to a final concentration of 6.3 μg·ml⁻¹ (square), 63 μg·ml⁻¹ (triangle) and 630 μg·ml⁻¹ (star) to three separate flasks. Growth was continued with shaking at 37°C and the OD₆₀₀ of the untreated (circle) and pyocin-treated cultures were monitored at 20 min intervals. Bars represent mean ± S.E.M. (n=2). The highest concentration of pyocin SD2 used (630 μg·ml⁻¹) resulted in 13.7×10⁶ pyocin SD2 molecules per cell. Killing of PAO1 by pyocin SD2 was seen at all three pyocin concentrations.

Figure 5. The translocation domain is at the N-terminus of pyocins S2/SD2

Ten microlitres of purified pyocin was spotted on to a growing lawn of cells. Clear zones indicate cell death. (a) Inhibition of growth of P. aeruginosa CF18 by pyocin S2 (0.1 μg·ml⁻¹). (b) Pyocin S2Δ17 (750 μg·ml⁻¹) is not active against P. aeruginosa CF18. (c) Inhibition of pyocin S2 killing (0.1 μg·ml⁻¹) by pyocin S2Δ17 (750 μg·ml⁻¹) against P. aeruginosa CF18. (d) Inhibition of pyocin SD2 killing (100 μg·ml⁻¹) by pyocin S2Δ17 (750 μg·ml⁻¹) against P. aeruginosa PAO1.

Figure 6. Pyocin SD2 requires the CPA for efficient killing

(a) Inhibition of growth of P. aeruginosa PAO1 and P. aeruginosa PAO1wzt by pyocin SD2 (1–5: 313 μg·ml⁻¹, 2-fold dilutions) as shown by a soft agar overlay spot-test. Ten microlitres of purified pyocin SD2 was spotted on to a growing lawn of cells. Clear zones indicate cell death. (b) Representative growth curves for P. aeruginosa strains PAO1 and PAO1wzt grown in LB broth for 180 min with and without the addition of 3 μM pyocin SD2. OD₆₀₀ measured 0, 30, 60, 90 and 180 min after pyocin treatment. Error bars represent the standard error mean between replicate samples (n=3). PAO1 (circle), PAO1wzt (up triangle), PAO1+pyocin SD2 (square) and PAO1wzt+pyocin SD2 (down triangle). (c) Percent killing for one of the replicates in (b). Ten microlitres of cells (10-fold serial dilutions) were spotted on LB agar plates at 60, 90 and 180 min after pyocin SD2 treatment, incubated at 37°C for 16 h and CFU determined. % killing determined by comparison to untreated controls.

Figure 7. Pyocins SD2 and S2 bind to the CPA of P. aeruginosa LPS

(a) ITC binding isotherm of pyocin SD2 (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg ml⁻¹) from wild-type P. aeruginosa PAO1. Saturable heats were observed indicative of an interaction. Data were fitted to a single-binding site model that yielded a K_d=0.22±0.02 μM. Reactions were performed in 0.2 M sodium phosphate buffer, pH 7.5 at 30°C using a MicroCal iTC200. (b) ITC isotherm of pyocin SD2* (111 μM) titrated into isolated LPS-derived polysaccharide (1 mg·ml⁻¹) from PAO1wzt. No saturable binding isotherm was observed. (c) ITC binding isotherm of pyocin S2 (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from wild-type P. aeruginosa PAO1. Data were fitted to a single-binding site model that yielded a K_d=0.62±0.04 μM. Reactions were performed in 0.2 M sodium phosphate buffer, pH 7.5 at 30°C using a MicroCal iTC200. (d) ITC isotherm of pyocin S2* (111 μM) titrated into isolated LPS-derived polysaccharide (1 mg·ml⁻¹) from PAO1wzt. No saturable binding isotherm was observed. Errors reported are the error for the single site binding model fit. *Reactions were performed in 0.2 M sodium phosphate buffer, pH 7.5 at 25°C using a MicroCal VP-ITC.

Figure 8. The helical domain of pyocins S2/SD2 is involved in CPA-binding

(a) Schematic of the protein variants used in this study. (b) ITC binding isotherm of pyocin S2Δ318 (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from wild-type P. aeruginosa PAO1. No saturable binding isotherm was observed. (c) ITC binding isotherm of pyocin S2 (1–209) (300 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from wild-type P. aeruginosa PAO1. No saturable binding isotherm was observed. (d) ITC binding isotherm of pyocin SD2Δ216 (888 μM) titrated into isolated LPS-derived polysaccharide (1.5 mg·ml⁻¹) from wild-type P. aeruginosa PAO1. Saturable heats were observed indicative of a weak interaction. (e) ITC binding isotherm of pyocin SD2Δ216 (300 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from wild-type P. aeruginosa PAO1. Heats were observed indicative of a weak interaction. Data were fitted to a single-binding site model, with N set to 0.1, that yielded a K_d=23.3±0.59 μM. (f) ITC isotherm of pyocin SD2Δ216 (300 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from PAO1wzt. No saturable binding isotherm was observed. Reactions were performed in 0.2 M sodium phosphate buffer, pH 7.5 at 30°C using a MicroCal iTC200. Errors reported are the error of the single site binding model fit. (g) Far-UV circular CD spectra were scanned from 250 to 195 nm three times at 25°C for pyocin S2 (green line), pyocin S2(1–209) (blue line) and pyocin S2Δ318 (purple line). All proteins have helical content demonstrating folding. (h) CD spectra for pyocin SD2 (orange line) and truncated pyocin SD2Δ216 (light blue line). Both proteins have helical content demonstrating folding. Proteins in 0.2 M sodium phosphate buffer, pH 7.5 at 25°C were diluted to a concentration of 15–150 μg·ml⁻¹ and CD spectra were recorded on a J-815 spectropolarimeter (Jasco) in a 0.1 cm quartz cuvette. The final spectra were represented by molar ellipticity, Δε (mdeg·M⁻¹·cm⁻¹).

Figure 9. CPA-binding ability of pyocins SD1, SD3, S5 and AP41

(a) ITC binding isotherm of pyocin S5 (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from wild-type P. aeruginosa PAO1. Strong, saturable heats were observed indicative of a strong interaction. Data were fitted to a single-binding site model that yielded a K_d=0.44±0.01 μM. (b) ITC isotherm of pyocin S5 (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from PAO1wzt. No saturable binding isotherm was observed. (c) ITC isotherm of pyocin SD1 (300 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from PAO1. No saturable binding isotherm was observed. (d) ITC binding isotherm of pyocin SD3 (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from wild-type P. aeruginosa PAO1. Saturable heats were observed indicative of an interaction. Data were fitted to a single-binding site model that yielded a K_d=1.1±0.04 μM. (e) ITC isotherm of pyocin SD3 (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg·ml⁻¹) from PAO1wzt. No saturable binding isotherm was observed. (f) ITC binding isotherm of pyocin AP41* (150 μM) titrated into isolated LPS-derived polysaccharide (3 mg ml⁻¹) from wild-type P. aeruginosa PAO1. No saturable binding isotherm was observed. Errors reported are the error of the single site binding model fit. *Reactions were performed in 0.2 M sodium phosphate buffer, pH 7.5 at 25°C using a MicroCal VP-ITC.

Figure 10. Pyocin binding and translocation model

(a) Newly proposed domain architecture of pyocin SD2. (b) Proposed model of pyocin binding and translocation. Pyocin SD2 binds to the CPA on the cell surface of P. aeruginosa, via the receptor and CPA-binding domain, which orientates the N-terminus close to the TonB-dependent outer membrane protein FpvAI. The intrinsically unstructured translocation domain is then positioned to thread through the FpvAI pore and interact with translocation machinery. The interaction of TonB with FpvAI is important for pyocin SD2 translocation across the outer membrane.