The LDL receptor clustering motif interacts with the clathrin terminal domain in a reverse turn conformation

Abstract

Previously the hexapeptide motif FXNPXY807 in the cytoplasmic tail of the LDL receptor was shown to be essential for clustering in clathrin-coated pits. We used nuclear magnetic resonance line-broadening and transferred nuclear Overhauser effect measurements to identify the molecule in the clathrin lattice that interacts with this hexapeptide, and determined the structure of the bound motif. The wild-type peptide bound in a single conformation with a reverse turn at residues NPVY. Tyr807Ser, a peptide that harbors a mutation that disrupts receptor clustering, displayed markedly reduced interactions. Clustering motif peptides interacted with clathrin cages assembled in the presence or absence of AP2, with recombinant clathrin terminal domains, but not with clathrin hubs. The identification of terminal domains as the primary site of interaction for FXNPXY807 suggests that adaptor molecules are not required for receptor-mediated endocytosis of LDL, and that at least two different tyrosine-based internalization motifs exist for clustering receptors in coated pits.

Figure 1

Differential line-broadening of internalization peptides in the presence and absence of clathrin-AP cages. Corresponding regions of 1-dimensional ¹H NMR spectra are shown at absolute intensity for 2.5 mM peptide with or without 20 mg/ml assembled clathrin/AP cages. (A–F) Aromatic region of the spectrum from wild-type peptides FDNPVYQKTT alone (A), FDNPVYQKTT plus cages (B), FDNPVY plus cages (C), and Y₈₀₇S mutant peptides FDNPVSQKTT alone (D), FDNPVSQKTT plus cages (E), and FDNPVS plus cages (F). (G and H) Upfield region of wild-type peptide, FDNPVYQKTT, alone (G) and with cages (H). The resonances of the aromatic protons are at 7.29, 7.33, and 7.38 ppm for Phe and 6.81 and 7.12 ppm for Tyr and for the methyl protons at 0.83 and 0.89 ppm for Val(γ), 1.25 ppm for Thr₉(γ), and 1.18 ppm for Thr₁₀(γ).

Figure 4

Differential broadening of the wild-type and mutant hexapeptides in the presence of assembled clathrin cages reconstituted with and without AP-2. Corresponding regions of 1-dimensional ¹H NMR from the aromatic region of the spectra are shown at absolute intensity. The indicated peptide (final concentration 1.5 mM) was added to buffer alone or buffer plus 9.7 mg/ml protein: FDNPVY alone (A), plus clathrin cages (B), plus clathrin and AP-2 cages (C), FDNPVS alone (D), plus clathrin cages (E), or plus clathrin and AP-2 cages. NMR samples were prepared as described for Fig. 1. Because of differing concentrations of peptide and protein, the absolute signal intensity can not be directly compared with those in Fig. 1.

Figure 5

Differential line-broadening of internalization peptides in the presence and absence of recombinant clathrin terminal domains and hubs. Corresponding regions of 1-dimensional ¹H NMR from the aromatic region of the spectra are shown at absolute intensity. Samples contained 1.0 mM peptide with or without 175 μM clathrin terminal domain TD(1–579) at pH 6.2 as follows: FDNPVY alone (A), plus TD(1–579) (B) or FDNPVS alone (C), plus TD(1–579) (D). The interactions of the wild-type hexapeptide (1 mM) with 200 μM of either TD(1–579) or clathrin Hub(1074–1483) were compared at pH 7.2 in 150 mM KCl. Samples contained: FDNPVY alone (E), plus TD(1–579) (F), or plus Hub(1074–1483) (G).

Figure 2

NOESY spectra of wild-type and mutant internalization motif peptides in the presence of clathrin-AP cages. Samples were the same as those in the spectra of Fig. 1. Corresponding expansions from the aliphatic/amide + aromatic region of NOESY spectra are presented as contour plots at the same level from data acquired with 60 ms of mixing time. All samples contained 20 mg/ml cages and 2.5 mM peptide: FDNPVY (A), FDNPVS (B), FDNPVYQKTT (C), and FDNPVSQKTT (D).

Figure 3

NOE buildup rates for fixed internuclear distances. NOE buildup curves from the sample containing the wild-type decapeptide with cages are shown for the intraresidue Hβ/Hβ NOEs of Phe (squares), Asn (diamonds), Pro (circles), and Tyr (triangles; A), and for the intraresidue Hβ/Hγ NOEs of Val(Hγ₁) (squares), Val(Hγ₂) (diamonds), Thr₉ (circles), and Thr₁₀ (triangles; B). Crosspeak intensities were calculated in arbitrary units of volume using the program Felix 2.30 from spectra acquired at 60, 100, 140, and 220 ms mixing times.

Figure 6

Cage-bound structure of FDNPVY calculated from transferred NOE- derived restraints. (A) Contour plot of the transferred NOE between Tyr(HN) (7.71 ppm) and Val(HN) (8.07 ppm) from the hexapeptide with a 60-ms mixing time. (B) The superimposed backbones of 21 calculated structures obtained by simulated annealing. (C) A single representative structure of the AcFDNPVY peptide bound to clathrin-AP cages.

Figure 6

Cage-bound structure of FDNPVY calculated from transferred NOE- derived restraints. (A) Contour plot of the transferred NOE between Tyr(HN) (7.71 ppm) and Val(HN) (8.07 ppm) from the hexapeptide with a 60-ms mixing time. (B) The superimposed backbones of 21 calculated structures obtained by simulated annealing. (C) A single representative structure of the AcFDNPVY peptide bound to clathrin-AP cages.

Figure 6

Cage-bound structure of FDNPVY calculated from transferred NOE- derived restraints. (A) Contour plot of the transferred NOE between Tyr(HN) (7.71 ppm) and Val(HN) (8.07 ppm) from the hexapeptide with a 60-ms mixing time. (B) The superimposed backbones of 21 calculated structures obtained by simulated annealing. (C) A single representative structure of the AcFDNPVY peptide bound to clathrin-AP cages.