Structural characterization of the hemophore HasAp from Pseudomonas aeruginosa: NMR spectroscopy reveals protein-protein interactions between Holo-HasAp and hemoglobin

Abstract

Pseudomonas aeruginosa secretes a 205 residue long hemophore (full-length HasAp) that is subsequently cleaved at the C'-terminal domain to produce mainly a 184 residue long truncated HasAp that scavenges heme [Letoffé, S., Redeker, V., and Wandersman, C. (1998) Mol. Microbiol. 28, 1223-1234]. HasAp has been characterized by X-ray crystallography and in solution by NMR spectroscopy. The X-ray crystal structure of truncated HasAp revealed a polypeptide alphabeta fold and a ferriheme coordinated axially by His32 and Tyr75, with the side chain of His83 poised to accept a hydrogen bond from the Tyr75 phenolic acid group. NMR investigations conducted with full-length HasAp showed that the carboxyl-terminal tail (21 residues) is disordered and conformationally flexible. NMR spectroscopic investigations aimed at studying a complex between apo-HasAp and human methemoglobin were stymied by the rapid heme capture by the hemophore. In an effort to circumvent this problem NMR spectroscopy was used to monitor the titration of 15N-labeled holo-HasAp with hemoglobin. These studies allowed identification of a specific area on the surface of truncated HasAp, encompassing the axial ligand His32 loop that serves as a transient site of interaction with hemoglobin. These findings are discussed in the context of a putative encounter complex between apo-HasAp and hemoglobin that leads to efficient hemoglobin-heme capture by the hemophore. Similar experiments conducted with full-length 15N-labeled HasAp and hemoglobin revealed a transient interaction site in full-length HasAp similar to that observed in the truncated hemophore. The spectral perturbations observed while investigating these interactions, however, are weaker than those observed for the interactions between hemoglobin and truncated HasAp, suggesting that the disordered tail in the full-length HasAp must be proteolyzed in the extracellular milieu to make HasAp a more efficient hemophore.

Figure 1

Alignment of amino acid sequences from HasAs and HasAp showing identical residues in bold face and residues involved in the coordination of the heme iron, including the “auxiliary” H83, which is thought to accept phenolic acid proton from Y75 (HasAp numbering), in red. The arrow indicates the length of truncated HasAp (full-length minus 21 amino acids) utilized in these investigations.

Figure 2

(A) The fold of HasAp harbors four α-helices and eight anti-parallel β-strands. The axial ligands, His32 and Tyr75 are located in two extended loops above and below the heme. The side chain of His83 is located within hydrogen bonding distance of the Tyr75 phenolic acid proton. (B) Presentation of the heme and its coordination sphere in HasAp at 1.7 Å resolution depicting hydrogen bonding interactions (2.2 – 3.2 Å; black dotted lines).

Figure 3

Electron density for the heme in molecules A and B, contoured at 1σ, derived from a 2F_o-1F_c Fourier synthesis after the last refinement cycle. The electron density from a F_o-1F_c Fourier synthesis omitting the heme molecules is presented in the supplemental material (Figure S3). The data is consistent with heme binding in only one orientation in HasAp.

Figure 4

(A) Per residue plot of B-factors of the C_α atoms in monomer A (black line) and monomer B (red line). The average B-factor of all main chain atoms is 15.9 Ų for both monomers (dashed horizontal line). (B) Stereo representation of molecule A of the HasAp crystal structure color coded according to thermal factors, with blue representing 8 Ų and red 51 Ų. The heme and its axial ligands His32 and Tyr75 are shown in yellow.

Figure 5

¹H-¹⁵N HSQC spectrum of (A) [U-¹⁵N]-truncated HasAp in phosphate buffer (µ = 0.1, pH 7.0 and 32 °C, obtained in a Varian Unity Inova 600 spectrometer. (B) [U-¹⁵N]-full-length HasAp under identical conditions. The cross-peaks within the boxes correspond to resonances originating from the C’-terminal tail in the full-length protein.

Figure 6

¹H-¹⁵N HSQC spectra of [¹⁵N-Val]-truncated-HasAp acquired with standard conditions (A) 128 (t1)×1632 (t2) complex points; spectral width, 2.4 kHz (t1)×9.60 kHz (t2); acquisition time, 85 ms; recycle delay, 1 s; 32 scans per increment. (B) fast recycling conditions; 128 (t1)×1050 (t2) complex points; spectral width, 3.6 kHz (t1)×15 kHz (t2); acquisition time, 35 ms; recycle delay, 50 ms; 256 scans per increment. (C) Strip plots showing heteronuclear correlations of residues Gln36 and Val37 in: HNCA, a) and e); HN(CO)CA, b) and f); CBCA(CO)NH, c) and g); and HNCACB, d) and h) spectra obtained with [U-¹³C, U-¹⁵N]-truncated-HasAp. Blue lines are used to highlight H^N to C_i^α and C^α _i-1 correlations in HNCA strips, whereas red lines illustrate correlations between H^N of both Gln 36 and Val 37 to the C^α and C^β cross-peaks of Gln 36.

Figure 7

A view of the structure of HasAp where residues exhibiting double cross-peaks in the ¹H-¹⁵N HSQC spectrum of [U-¹⁵N]-truncated HasAp acquired with fast repetition conditions have been highlighted in blue. The heme is shown in red.

Figure 8

(A) A portion of the ¹H-¹⁵N HSQC spectrum obtained from [U-¹⁵N]-full-length HasAp. The slices were obtained at the cross-peak corresponding to Leu201 (δ(¹H) = 8.13 ppm and δ(¹⁵N) = 122.8 ppm). The intensity of this N-H hydrogen cross-peak is significantly larger than the intensity of two other N-H hydrogens attached to nitrogens resonating at the same ¹⁵N frequency. These are the protons at 9.34 and 9.46 ppm. (B) A ¹H-¹H 2D slice obtained from a 3D NOESY HSQC (t_mix = 110 ms) spectrum of full-length HasAp. The slice was obtained at the ¹⁵N frequency of Leu201 and shows the few NOEs detected for the corresponding cross-peak at δ(¹H) = 8.13 ppm. (C) Same as B, except that the cross-peak corresponding to Tyr56 at δ(¹H) = 9.34 ppm shows several NOEs, as is typical of cross-peaks in structured proteins.

Figure 9

Per residue plot of cross-peak perturbations associated with the titration of (A) truncated- [U-¹⁵N]-HasAp and (B) full-length-[U-¹⁵N]-HasAp with met-Hb. The weighted values of chemical shift perturbations were obtained from $\Delta {\delta }_{\mathit{\text{weighted}}}=\sqrt{[{\left(\Delta \delta N/5\right)}^2+\Delta \delta H^2]/2}$. The average of all chemical shift perturbations is 0.011 in (A) and 0.0067 in (B) The two plots are presented with the same vertical scale two illustrate that the chemical shift perturbations are larger in the case of truncated HasAP. The negative peaks correspond to residues whose resonances disappear in the course of the titration.

Figure 10

(A) A view of the structure of truncated holo-HasAp where residues whose cross-peaks exhibit weighted chemical shift perturbations larger than three times the average of all chemical shifts (0.011) upon titration with met-Hb are highlighted in magenta and residues whose corresponding cross-peaks disappear upon addition of met-Hb are highlighted in blue. His32 is shown in yellow. (B) Surface representation of the view shown in (A). (C) A view of the structure of apo-HasAs (PDB ID 1YBJ) (45), where residues highlighted in magenta are equivalent to residues in holo-HasAp whose corresponding cross-peaks are affected by chemical shift perturbations and residues highlighted in blue are equivalent to those in holo- HasAp whose cross-peaks disappear upon titration with met-Hb. (D) Surface representation of the view shown in (C).