The iron-siderophore transporter FhuA is the receptor for the antimicrobial peptide microcin J25: role of the microcin Val11-Pro16 beta-hairpin region in the recognition mechanism

Abstract

The role of the outer-membrane iron transporter FhuA as a potential receptor for the antimicrobial peptide MccJ25 (microcin J25) was studied through a series of in vivo and in vitro experiments. The requirement for both FhuA and the inner-membrane TonB-ExbB-ExbD complex was demonstrated by antibacterial assays using complementation of an fhuA(-) strain and by using isogenic strains mutated in genes encoding the protein complex respectively. In addition, MccJ25 was shown to block phage T5 infection of Escherichia coli, in vivo, by inhibiting phage adhesion, which suggested that MccJ25 prevents the interaction between the phage and its receptor FhuA. This in vivo activity was confirmed in vitro, as MccJ25 inhibited phage T5 DNA ejection triggered by purified FhuA. Direct interaction of MccJ25 with FhuA was demonstrated for the first time by size-exclusion chromatography and isothermal titration calorimetry. MccJ25 bound to FhuA with a 2:1 stoichiometry and a K(d) of 1.2 microM. Taken together, our results demonstrate that FhuA is the receptor for MccJ25 and that the ligand-receptor interaction may occur in the absence of other components of the bacterial membrane. Finally, both differential scanning calorimetry and antimicrobial assays showed that MccJ25 binding involves external loops of FhuA. Unlike native MccJ25, a thermolysin-cleaved MccJ25 variant was unable to bind to FhuA and failed to prevent phage T5 infection of E. coli. Therefore the Val11-Pro16 beta-hairpin region of MccJ25, which is disrupted upon cleavage by thermolysin, is required for microcin recognition.

Figure 1. Structural comparison of native MccJ25 (PDB entry 1Q71) and thermolysin-cleaved t-MccJ25 (PDB entry 1S7P)

The Figure illustrates the disruption of the MccJ25 Val¹⁰–Pro¹⁶ β-hairpin upon cleavage by thermolysin. Structures are represented with black backbones and grey side chains using MOLMOL software [38]. Thick covalent bonds were used for the peptide backbones as well as the Glu⁸ side chain engaged in the ring. The Val¹⁰–Phe¹¹ bond targeted by thermolysin is indicated by an arrow. Selected residues are numbered.

Figure 2. Inhibitory effects of Mcc25 and t-MccJ25 on infection by phage T5

E. coli F was incubated with MccJ25 (A) or t-MccJ25 (B) at concentrations of 0.5 μM (△), 1 μM (▲), 5 μM (□) and 10 μM (◆), or with solvent only (○). After addition of phage T5, the culture turbidity (Abs) was monitored over 120 min at 620 nm. A control was performed in the absence of phage T5 (×). Data are representative of four independent experiments.

Figure 3. Inhibition of phage T5 adsorption

Phage adsorption was first measured in vivo by incubating E. coli W3110 for 10 min with 0.1–10 μM MccJ25 (■) or t-MccJ25 (○) before addition of phage T5. The amount of adsorbed phage was determined after a 15 min incubation with phage T5 at MOI=10. Results are expressed as means±S.E.M. for three independent experiments. In vitro, phage adsorption to FhuA was evaluated by measuring the microcin-induced inhibition of DNA ejection. FhuA (3 nM) was incubated for 10 min with 0.1, 0.2, 0.8, 1.6 or 3.1 μM MccJ25 (*) or solvent only in the presence of the fluorescent probe YO-PRO-1. The reaction was initiated by phage T5 addition. The inhibitory activity of 3.1 μM t-MccJ25 (black star) was compared with that of MccJ25 at the same concentration. The fluorescence signal indicative of phage DNA release was measured over 10 min (λ_excitation 490 nm; λ_emission 509 nm). The rate of fluorescence increase (V_f), which is directly proportional to the number of phage bound to FhuA, was measured in order to calculate the percentage of adsorbed phage. Data are representative of three independent experiments.

Figure 4. MccJ25–FhuA interaction in vitro

³H-labelled MccJ25 (4 nmol) was incubated for 10 min in the presence of 4 nmol (B) or 16 nmol (C) of FhuA. In a control experiment (A), incubation was performed in FhuA buffer only. Samples were analysed by gel-permeation chromatography on a Superose 12 HR 10/30 column. Absorbance was monitored at 226 nm (black line). Radioactivity was measured in every collected fraction by liquid scintillation counting (grey line).

Figure 5. Microcalorimetry analysis

(A) Microcalorimetric titration isotherm for the binding of MccJ25 to FhuA. Data were obtained at 25 °C using an automated sequence of 28 injections of 500 μM MccJ25 from a 300 μl syringe into a reaction cell containing 26.5 μM FhuA. The volume of each injection was 10 μl, and injections were made at 1 min intervals. Raw data were treated with ORIGIN® software and the values are plotted against molar ratio. Each point corresponds to the heat generated by the reaction upon each injection. The curve fit to the data (solid line) obtained by the ORIGIN® software yields values for K_d and stoichiometry (see the Experimental section). (B) DSC endotherm of the MccJ25–FhuA complex. Baseline-corrected DSC thermograms of FhuA alone (13 μM) (solid line) and the MccJ25–FhuA complex formed by complete saturation of FhuA with MccJ25 (broken line) were recorded in the same buffer as for ITC experiments with a heating rate of 1 K/min.

Figure 6. Partial alignment of FhuA amino acid sequences from MccJ25-susceptible (E.c., S.p.) and MccJ25-tolerant (S.t., P.a.) strains

Alignment was performed using the Multalign software [39]. Extracellular loops L4–L8 of the E. coli FhuA protein [12] are in black boxes, together with the corresponding sequence in other Enterobacteriaceae. Abbreviations and sources for the sequences are: E.c, E. coli K12 W3110 (Swiss-Prot; accession no. P0671); S.p., S. enterica Paratyphi SL369 (TrEMBL; accession no O86903); S.t., S. enterica Typhimurium LT2 (TrEMBL; accession no O86925); P.a., P. agglomerans K4 (TrEMBL; accession no O86924).