Mutation of Glu-166 blocks the substrate-induced dimerization of SARS coronavirus main protease

Abstract

The maturation of SARS coronavirus involves the autocleavage of polyproteins 1a and 1ab by the main protease (Mpro) and a papain-like protease; these represent attractive targets for the development of anti-SARS drugs. The functional unit of Mpro is a homodimer, and each subunit has a His-41cdots, three dots, centeredCys-145 catalytic dyad. Current thinking in this area is that Mpro dimerization is essential for catalysis, although the influence of the substrate binding on the dimer formation has never been explored. Here, we delineate the contributions of the peptide substrate to Mpro dimerization. Enzyme kinetic assays indicate that the monomeric mutant R298A/L exhibits lower activity but in a cooperative manner. Analytical ultracentrifugation analyses indicate that in the presence of substrates, the major species of R298A/L shows a significant size shift toward the dimeric form and the monomer-dimer dissociation constant of R298A/L decreases by 12- to 17-fold, approaching that for wild-type. Furthermore, this substrate-induced dimerization was found to be reversible after substrates were removed. Based on the crystal structures, a key residue, Glu-166, which is responsible for recognizing the Gln-P1 of the substrate and binding to Ser-1 of another protomer, will interact with Asn-142 and block the S1 subsite entrance in the monomer. Our studies indicate that mutation of Glu-166 in the R298A mutant indeed blocks the substrate-induced dimerization. This demonstrates that Glu-166 plays a pivotal role in connecting the substrate binding site with the dimer interface. We conclude that protein-ligand and protein-protein interactions are closely correlated in Mpro.

Figure 1

Active center of the SARS-CoV Mpro. The interactions between the P1 substrate-binding subsite from chain A (cyan) and the N-finger and domain III from chain B (magenta) of SARS-CoV Mpro (PDB code 1UK4) are shown. The substrate analog and the side chain of the Gln-P1 residue are yellow. The R298A monomeric structure (green) (PDB code 2QCY) is superimposed on the wild-type chain A. The dashed lines indicate hydrogen bonds for 1UK4 (red) and 2QCY (black). This figure was produced using PyMOL (35).

Figure 2

Initial velocity patterns of Mpro (A–C) and dependence of the initial velocity of Mpro on enzyme concentration (D–F). (A–C) Plots of the velocity difference at various substrate concentrations for wild-type, R298A, and R298L mutants, respectively. The lines represent results fitted according to the Michaelis-Menten equation (Eq. 1) for wild-type and the Hill equation (Eq. 2) for mutants. The kinetic parameters are shown in Table 1. (D–F) Difference in velocity at various enzyme concentrations for wild-type, R298A, and R298L mutants, respectively. The concentration of substrate was 600 μM. The line represents the best fit to the nonlinear dependence equation (Eq. 3).

Figure 3

SV patterns of Mpro. (A) Typical trace of absorbance at 250 nm of the enzyme during the SV experiment. Symbols represent experimental data and lines the results fitted to the Lamm equation using the SEDFIT program (25,26). (B–E) Continuous (c(s)) distribution of wild-type, R298A, R298L, and R298A/Q299A mutants, respectively. The residual bitmaps are shown in the insets. Protein concentration was 1.0 mg/ml. The distributions in 10 mM phosphate buffer are shown by solid lines and those in 10 mM phosphate containing 600 μM TQ6-pNA substrate by dashed lines. The left vertical dotted line indicates the monomer, and the right the dimer position.

Figure 4

Effects of substrate concentration on the quaternary structure of the R298A mutant. (A) Continuous (c(s)) distribution of the R298A mutant at TQ6-pNA concentrations of 0 (solid circles), 20 (open circles), 120 (solid triangles), 250 (open triangles), 400 (solid squares), and 600 μM (open squares). (B) Sedimentation coefficient shift of R298A major species at different TQ6-pNA concentrations.

Figure 5

Size-exclusion chromatography of Mpro. The enzyme in phosphate-buffered saline buffer was applied to a preequilibrated Superose 12 10/300 GL column. The Mpro and its mutants, without or with preincubation in 600 μM substrate, are wild-type (solid circles), wild-type with substrate (open circles), R298A (solid triangles), R298A with substrate (open triangles), R298A/Q299A (solid squares), and R298A/Q299A with substrate (open squares).

Figure 6

Isothermal calorimetric titration for the substrate TQ6-pNA binding to Mpro and its mutants. (A–D) Binding of TQ6-pNA to wild-type, R298A, R298L, and R298A/Q299A mutants, respectively. Wild-type protein concentration was 5.7 μM and protein concentration of the mutants was 28.6 μM. The TQ6-pNA (1 mM) was titrated into 2.7 ml of protein solution using 25–30 injections at a rate of 10 μl/injection. Values for the N, K_d, and ΔH were determined by ligand binding analysis with the Digitam program (TA instruments). The solid circles show the observed values and the lines represent fitted results.

Figure 7

Sedimentation velocity experimental values for Glu-166 mutants E166A (A) and E166A/R298A (B) in D₂O. The protein concentration was 0.33 mg/ml. The distributions in D₂O are represented by solid lines and those in D₂O with 600 μM TQ6-pNA substrate by dashed lines. The left vertical dotted line indicates the monomer position and the right, the dimer position. (Insets) Residual bitmaps for mutants without (upper) and with (lower) substrate.

Figure 8

Schematic model for the catalytic process of SARS-CoV Mpro. Substrate binding triggers the monomer-dimer equilibrium to favor the active protomer (square). The active protomer now is prompt to associate, forming the catalytic-competent dimer. After the catalytic cycle, the associated dimer will dissociate again to form the inactive protomer (circle).