Topography for independent binding of alpha-helical and PPII-helical ligands to a peroxisomal SH3 domain

Abstract

While the function of most small signaling domains is confined to binary ligand interactions, the peroxisomal Pex13p SH3 domain has the unique capacity of binding to two different ligands, Pex5p and Pex14p. We have used this domain as a model to decipher its structurally independent ligand binding sites. By the combined use of X-ray crystallography, NMR spectroscopy, and circular dichroism, we show that the two ligands bind in unrelated conformations to patches located at opposite surfaces of this SH3 domain. Mutations in the Pex13p SH3 domain that abolish interactions within the Pex13p-Pex5p interface specifically impair PTS1-dependent protein import into yeast peroxisomes.

Figure 1. Structure-Sequence Relations in the Pex13p SH3 Domain

The multisequence alignment includes presently known 29 SH3 domain sequences from S. cerevisiae. Except for the Pex13p SH3 domain sequence (top line), only the segments that belong to the canonical SH3 domain structure are included (Mayer, 2001). The locations of the secondary structural elements, as determined from the apo-Pexp13p SH3 domain crystal structure, are indicated above the alignment. The loops, connecting these elements, are labeled. Residues for which the accessible surface area dropped by more than 50% or whose backbone amides were considerably shifted (cf. Figures 4A and 4D) upon ligand binding are marked by “+”. Residues in the Pex14p binding site that are conserved among the SH3 domain sequences from S. cerevisiae are highlighted in magenta. None of the residues of the Pex5p binding site are conserved. Three residues (M324, P342, and Y361) forming a small structural cluster, absent in most other SH3 domains, are shown with orange background. Pex13p mutations (cf. Figures 5B and 6) are indicated on top.

Figure 2. Structures of the Pex13p SH3 Domain in the Absence and Presence of Peptide Ligands

(A) Ribbon representation of the crystal structure of the apo-Pexp13 SH3 domain. The segments covering the canonical SH3 domain (Mayer, 2001) and the terminal extensions are shown in blue and cyan colors, respectively. The side chains of a network of three residues (M324, P342, and Y366), structurally bridging the RT loop and the n-Src loop, are shown in orange ball-and-sticks. The side chains of the conserved residues forming the canonical PXXP binding site in SH3 domains (2–3) (Y315, F317, W349, P363, and Y366; cf. Figure 1) are shown in magenta ball-and-sticks. The side chain atoms are displayed in atom-specific colors (carbon, black; oxygen, red; nitrogen, blue; sulfur, yellow). (B) Ribbon of the Pex13p(SH3)-Pex14p peptide complex structure. The Pex14p peptide is shown in yellow ball-and-sticks. In addition to the conserved residues in the PXXP binding site of the SH3 domain (cf. Figure 2A), residues E320 and E325, which provide specific interactions with side chains of the Pex14p peptide, are shown. (C) Conformation of the Pex5p peptide. Secondary structure defining NOEs in the Pex5p peptide when bound to the Pex13p SH3 domain, indicating α-helical conformation for residues 204–214 (cf. Figure 3B). The secondary structure is evidenced by strong Hα (i),HN(i+3), Hα (i), Hβ(i+3), and sequential HN,HN NOE cross peaks (Wüthrich, 1986) in the filtered NOESY spectra. NOEs in the C-terminal region of the helix could not be unambiguously assigned due to signal overlap. (D) The α-helical region of the Pex5p peptide when bound to the Pex13p SH3 domain is shown as a cylinder. Positively charged, negatively charged, polar, and other residues are shown in blue, red, green, and gray shapes, respectively. Residues that are located on the front and back sides of the helix are in black and white circles, respectively.

Figure 3. Increased α-Helical Content of Pex5p upon Pex13p SH3 Domain Binding

(A) Conformation of the Pex5p peptide upon binding to the Pex13p SH3 domain by CD spectroscopy. The fraction of α-helical conformation of Pex5p peptide increases upon complex formation with Pex13p SH3 (curve in green color), whereas isolated Pex5p peptide remains mostly unstructured (curve in blue color). The curve in green represents the molar ellipticity difference between the spectrum of the complex (at a molar ratio of 2:1 Pex5p/Pex13p SH3) and the sum of the spectra of the free peptide and the SH3 domain. The distortion of the helical curve beyond 210 nm is due to the presence of aromatic residues in the sequence of Pex5p. (B) H_α secondary chemical shifts for residues 198–216 of the Pex5p peptide in the free (blue) and bound form (green). Upon ligand binding, the α-helical content of the Pex5p peptide is increased, and the helix is extended beyond Lys210 up to Glu214.

Figure 4. Pex14p (Magenta, [A]–[C]) and Pex5p (Green, [D]–[F]) Binding Sites on the Pex13p SH3 Domain

NMR chemical shift changes (Δδ, as defined in the Experimental Procedures) in the presence of saturated concentrations of the Pex14p and Pex5p ligands are shown in (A) and (D), respectively. The topography of the Pex14p and Pex5 binding sites on the Pex13p SH3 domain is shown in surface (B and E) and ribbon (C and F) representations, respectively. The orientation of the Pex14p binding site in (B) and (C) is as in Figure 2A. Surface (E) and ribbon (F) representations of the Pex5p binding site are in an orientation rotated by 180° around a vertical axis. Exposed side chains of residues for which large chemical shift perturbations were observed are shown in ball-and-stick presentation (C and F).

Figure 5. NMR Titration Experiments Demonstrating Effects of Pex5p Site Mutations of the Pex13p SH3 Domain on Binding to Pex5p and Its Independence from the Pex14p Binding Site

(A) Independent binding of the Pex14p and Pex5p ligands to the Pex13p SH3 domain. ¹H,¹⁵N correlation NMR spectra are shown for the uncomplexed (black) and the simultaneously Pex14p and Pex5p bound SH3 domain. Titration endpoints where the order of peptide addition was reversed are shown in blue (addition of Pex5p, then Pex14p) and orange (addition of Pex14p, then Pex5p). Chemical shift changes induced by Pex14p and Pex5p are indicated by pink and green colors, respectively. (Inset) The titration endpoints of M334 and I362 are independent of the order of addition of the peptide ligands, confirming the independence of the two binding sites. The titration curves and corresponding binding affinities are also independent of the occupancy of the second site demonstrating that the formation of the ternary complex is noncooperative. (B) Pex5p site mutations of the Pex13p SH3 domain. The 3D structure of Pex13p SH3 mutants is not disrupted since ¹H, ¹⁵N correlation spectra of the ¹⁵N-labeled mutants are comparable to those of the wild-type Pex13p SH3 domain. Backbone amide resonances in the proximity of the point mutations show slightly different relative positions (e.g., A311, M334, V352). Saturation points used for Pex5p titrations are: wt, 1:2; A335Y, 1:4; L333A, 1:3; F310A, 1:10; (F310A, L333A), 1:4 (no effect); (R353E, K355E), 1:4 (no effect). Reference spectra in the absence of peptide and protein-ligand complexes at equimolar concentrations are shown in black and blue contours, respectively. The chemical shift change upon Pex5p addition is traced by a green line to saturation. Green contours correspond to protein-peptide molar ratios at saturation of binding. For the two double mutants, no binding of the Pex5p peptide is observed even at 4-fold excess peptide. To demonstrate that Pex14p binding is not affected in the Pex13p mutants, the trace of the chemical shift change of Y361 upon Pex14p ligand binding as well as the saturation point are shown in magenta for the wild-type SH3 domain and the two double mutant SH3 domains. Saturation of Pex14p binding is achieved at a molar 1:2 protein-peptide ratio in the double mutants, which is comparable to the wild-type protein. The concentration of the wild-type and mutant proteins is 0.2 mM.

Figure 6. In Vivo Analysis of Mutants in the Pex5p Binding Site of Pex13p SH3

(A) pex13Δ cells expressing wild-type Pex13p (open diamonds), Pex13p (E320K) (open squares), Pex13p (A335Y) (closed squares), Pex13p (L333A) (closed circles), Pex13p (F310A) (open triangles), Pex13p (R353E/K355E) (closed triangles), Pex13p (L333A/F310A) (crosses), or no insert (empty vector) (closed diamonds) were precultured in 0.3% minimal glucose medium and inocculated at OD₆₀₀ of 0.05 in minimal oleate medium, and growth was followed with time by measuring the optical density at 600 nm. (B) The strains as described in (A) were cotransformed with either GFP-PTS1 or PTS2-GFP, cultured on liquid oleate medium, and examined under the fluorescence microscope to visualize the distribution of the GFP fusion protein.