Simulations suggest double sodium binding induces unexpected conformational changes in thrombin

Abstract

Thrombin is a Na[Formula: see text]-activated serine protease existing in two forms targeted to procoagulant and anticoagulant activities, respectively. There is one Na[Formula: see text]-binding site that has been the focus of the study of the thrombin. However, molecular dynamics (MD) simulations suggest that there might be actually two Na[Formula: see text]-binding sites in thrombin and that Na[Formula: see text] ions can even bind to two sites simultaneously. In this study, we performed 12 independent 2-µs all-atom MD simulations for the wild-type (WT) thrombin and we studied the effects of the different Na[Formula: see text] binding modes on thrombin. From the root-mean-square fluctuations (RMSF) for the [Formula: see text]-carbons, we see that the atomic fluctuations mainly change in the 60s, 170s, and 220s loops, and the connection (residue 167 to 170). The correlation matrices for different binding modes suggest regions that may play an important role in thrombin's allosteric response and provide us a possible allosteric pathway for the sodium binding. Amorim-Hennig (AH) clustering tells us how the structure of the regions of interest changes on sodium binding. Principal component analysis (PCA) shows us how the different regions of thrombin change conformation together with sodium binding. Solvent-accessible surface area (SASA) exposes the conformational change in exosite I and catalytic triad. Finally, we argue that the double binding mode might be an inactive mode and that the kinetic scheme for the Na[Formula: see text] binding to thrombin might be a multiple-step mechanism rather than a 2-step mechanism.

Keywords: Allostery; Free energy; Molecular dynamics; Sodium binding; Thrombin.

Fig. 1

Thrombin is shown in transparent blue. Residues ASP189, ALA190, ARG221, and LYS224 are displayed via Licorice. The bound Na⁺ ions are shown as the yellow bead. Water molecules within 3.4 Å of the bound Na⁺ ions are also displayed

Fig. 2

Root-mean-square fluctuations (RMSF) for the α-carbons of unbound, double binding, outer binding, and inner binding. Residue number follows the sequential residue numbering scheme that all residues are numbered from 1 to 295 (MD notation)

Fig. 3

Thrombin-Na⁺ complex structure. The heavy chain is shown in transparent green. The light chain is shown in purple. The 60s (pink), 170s (violate), 220s (tan), and γ (black) loops, exosite I (orange) and II (yellow), and the connection (cyan) are the regions of interest. Sidechains of the catalytic triad are displayed in red. The bound Na⁺ is shown as the blue bead

Fig. 4

(a)–(d) are the correlation matrices for the thrombin in the double binding mode, unbound mode, inner binding mode, and outer binding mode. (e)–(g) are the visualizations of the correlation matrix for unbound-double, outer-double, and inner-double, respectively. The representation method is the same as what we used in our previous study. In graphs (e)–(g), light-red lines and light-blue lines represent pairs of residues with positive and negative subtractions (unbound/outer/inner-double) of correlation matrices whose absolute value is larger than 0.5. The values are shown in Tables S1–S3

Fig. 5

Amorim-Hennig Clustering Distribution for the different binding modes of thrombin. (a)–(b) and (e)–(h) represents the clustering results of the heavy atoms for the 60s loop, the connection, the 170s loop, the 220s loop, the γ loop, and the catalytic triad. The percentage for each cluster is on top of each column. (c)–(d) are the representative structures for the 60s loop and the connection, respectively. They are shown in NewCartoon representation with transparent green. Gray shadows represent the variances in the 60s loop or the connection

Fig. 6

(a) Conformational free energy surfaces of the regulatory loops. (b) Conformational free energy surfaces of the regulatory loops in different binding modes of thrombin. (c) Representative structures for the regulatory loops in well 0 to 8. They are shown in NewCartoon representation with transparent green. Sidechains in the regulatory loops are displayed by the Licorice in red. Gray shadows represent the variances in the regulatory loops. The catalytic triad is shown in blue

Fig. 7

(a) Conformational free energy surfaces of the catalytic pocket. (b) Conformational free energy surfaces of the catalytic pocket in different binding modes of thrombin. (c) Representative structures for the catalytic pocket in well 0 to 8. They are shown in NewCartoon representation with transparent green. Sidechains in the catalytic pocket are displayed by the Licorice in red. Gray shadows represent the variances in the catalytic pocket

Fig. 8

The probability distribution of the solvent-accessible surface area (SASA) for different binding modes. (a)–(d) plot the 220s loop, the γ loop, the catalytic triad, and the exosite I, respectively

Fig. 9

(a) Conformational free energy surfaces of the exosite I. (b) Conformational free energy surfaces of the exosite I in different binding modes of thrombin. (c) Representative structures for the exosite I in well 0 to 3. They are shown in NewCartoon representation with transparent green. Sidechains in the exosite I are displayed by the Licorice in red. Gray shadows represent the variances in the exosite I

Fig. 10

Four-state scheme for thrombin