Molecular Interplay between the Dimer Interface and the Substrate-Binding Site of Human Peptidylarginine Deiminase 4

Abstract

Our previous studies suggest that the fully active form of Peptidylarginine deiminase 4 (PAD4) should be a dimer and not a monomer. This paper provides a plausible mechanism for the control of PAD4 catalysis by molecular interplay between its dimer-interface loop (I-loop) and its substrate-binding loop (S-loop). Mutagenesis studies revealed that two hydrophobic residues, W347 and V469, are critical for substrate binding at the active site; mutating these two residues led to a severe reduction in the catalytic activity. We also identified several hydrophobic amino acid residues (L6, L279 and V283) at the dimer interface. Ultracentrifugation analysis revealed that interruption of the hydrophobicity of this region decreases dimer formation and, consequently, enzyme activity. Molecular dynamic simulations and mutagenesis studies suggested that the dimer interface and the substrate-binding site of PAD4, which consist of the I-loop and the S-loop, respectively, are responsible for substrate binding and dimer stabilization. We identified five residues with crucial roles in PAD4 catalysis and dimerization: Y435 and R441 in the I-loop, D465 and V469 in the S-loop, and W548, which stabilizes the I-loop via van der Waals interactions with C434 and Y435. The molecular interplay between the S-loop and the I-loop is crucial for PAD4 catalysis.

Figure 1. Structure of homodimeric PAD4 and the hydrophobic residues located at the substrate-binding site and the dimer interface.

(A) Homodimer of human PAD4 (PDB ID: 1WDA). The bound calcium ions are indicated as green balls. The substrate analogue, benzoyl‐L‐arginine amide (BA), is shown as a sphere model with the carbon atoms in pink. The individual subunits are shown in gray and orange. (B) Detailed structure of the interface loop (I‐loop, cyan) and the substrate‐binding loop (S‐loop, magenta). The important residues on the loops are shown as stick models. (C) Active site of PAD4. The substrate‐binding residues are represented as yellow sticks, and the substrate BA is shown as pink sticks. The black dashed lines represent the polar contacts between the residues and the substrate. (D) Hydrophobic amino acid residues in the dimer interface of PAD4. The green and orange sticks indicate the residues associated with the different subunits.

Figure 2. Continuous sedimentation coefficient distributions of PAD4 WT and the hydrophobic substrate-binding-site mutants.

(A) WT. (B) W347A. (C) V469A. Three protein concentrations (0.1, 0.3, and 0.9 mg/ml) were analyzed as indicated in the figure. The sedimentation velocity data were globally fit using the SEDPHAT program to obtain the K_d (dissociation constant) of the PAD4 dimer (Table 1). M, monomer; D, dimer.

Figure 3. Continuous sedimentation coefficient distributions of PAD4 WT and the W548 series of dimer-interface mutants.

(A) WT. (B) W548A. (C) W548F. (D) W548K. Three protein concentrations (0.1, 0.3, and 0.9 mg/ml) were analyzed as indicated in the figure. The sedimentation velocity data were globally fit using the SEDPHAT program to obtain the K_d of the PAD4 dimer (Table 2). M, monomer; D, dimer.

Figure 4. Continuous sedimentation coefficient distributions of the PAD4 L6 and L279 series of dimer-interface mutants.

(A) L6A. (B) L6I. (C) L6D. (D) L279A. (E) L279I. (F) L279D. The K_d values of these mutants are shown in Table 2. M, monomer; D, dimer.

Figure 5. Continuous sedimentation coefficient distributions of the PAD4 V283 and V284 series of dimer-interface mutants.

(A) V283A; (B) V283I; (C) V283D; (D) V284A; (E) V284I; (F) V284D. The K_d values of these mutants are shown in Table 2. M, monomer; D, dimer.

Figure 6. MD simulations of PAD4 WT and the W548 mutants.

(A) RMSF for the Cα atoms in the C-terminal domain of the WT enzyme and the W548 mutants. The locations of the I-loop and the S-loop are labeled. (B) Structures of the I-loop (cyan) and the S-loop (magenta) in the PAD4 C-terminal domain. The PAD4 monomer after MD simulation is shown as a gray ribbon. To demonstrate the location of the loops in the dimer interface, we reconstituted the structure into a dimer in this panel; the other subunit is depicted as an orange ribbon. These important residues are highlighted as stick models and labeled. The bound calcium ion is depicted as a green ball. The substrate analogue BA is shown as a sphere model with the carbon atoms in pink. The possible polar contacts are represented by black dashed lines. (C) Frames of interaction between the residues in the I-loop and the S-loop during a 10-ns MD simulation. The black dashed lines represent the polar contacts between the two loops.

Figure 7. Continuous sedimentation coefficient distributions of the PAD4 I-loop and S-loop mutants.

(A) WT. (B) C434A. (C) Y435A. (D) R441. (E) D465A. The K_d values of these mutants are shown in Table 3. M, monomer; D, dimer.