Reaction Pathway Sampling and Free-Energy Analyses for Multimeric Protein Complex Disassembly by Employing Hybrid Configuration Bias Monte Carlo/Molecular Dynamics Simulation

Abstract

Physicochemical characterization of multimeric biomacromolecule assembly and disassembly processes is a milestone to understand the mechanisms for biological phenomena at the molecular level. Mass spectroscopy (MS) and structural bioinformatics (SB) approaches have become feasible to identify subcomplexes involved in assembly and disassembly, while they cannot provide atomic information sufficient for free-energy calculation to characterize transition mechanism between two different sets of subcomplexes. To combine observations derived from MS and SB approaches with conventional free-energy calculation protocols, we here designed a new reaction pathway sampling method by employing hybrid configuration bias Monte Carlo/molecular dynamics (hcbMC/MD) scheme and applied it to simulate the disassembly process of serum amyloid P component (SAP) pentamer. The results we obtained are consistent with those of the earlier MS and SB studies with respect to SAP subcomplex species and the initial stage of SAP disassembly processes. Furthermore, we observed a novel dissociation event, ring-opening reaction of SAP pentamer. Employing free-energy calculation combined with the hcbMC/MD reaction pathway trajectories, we moreover obtained experimentally testable observations on (1) reaction time of the ring-opening reaction and (2) importance of Asp42 and Lys117 for stable formation of SAP oligomer.

Figure 1

Molecular structures of serum amyloid P (SAP) component homo pentamer. (A) SAP pentamer with annotation for each subunit. (B) Two individual inter-subunit interaction surfaces of SAP subunit, highlighted by blue and red colors; S₁ and S₂ are depicted as representative of subunit pair. (C) Initial stages of SAP pentamer disassembly processes predicted by structural bioinformatics approach. Reactions shown in (C) are supposed for any possible monomer and dimer.

Figure 2

Configurations of SAP subunits at the last (500th) hcbMC/MD cycle, where each subunit pair has no native contacts (NC), thus being supposed to be dissociated. Two individual inter-subunit interaction surfaces of each subunit are highlighted by blue and red colors as in the case of Figure 1. The five individual simulations discussed in (A)–(E) are indexed by lowercase alphabets (a–e) hereafter.

Figure 3

Disassembly of serum amyloid P (SAP) component homo pentamer through hcbMC/MD cycles. The five individual simulations discussed in (A)–(E) correspond to those indexed by lowercase alphabets (a–e) in Figure 2. A distribution of SAP (sub)complex species is illustrated by color. The green and red arrows indicate the initial stage of SAP disassembly (generation of tetramer or trimer) and the first cycle where SAP pentamer undergoes a ring-opening event.

Figure 4

Averaged number of native contacts between SAP subunit pair. Average and statistical error values are calculated using 2000 snapshot structures obtained from each of the five unbiased 200 ns MD simulations (an error bar is buried in the square, then not being apparent in this panel). Each inter-subunit native contact is distinguished by different colors. Annotation of subunit pair (S₁–S₂, etc.) and indexes of individual simulation (a–e) are given as in the case of Figure 1 and that of Figure 2, respectively. Statistical errors were estimated from standard error by considering 95% confidence interval.

Figure 5

Ring-opened SAP pentamer configuration obtained from hcbMC/MD simulation, indexed by “a” in Figure 2. The digits in (A) and (B) denote values of distance between centers of mass (COMs) of S₃ and S₄. The orange dotted line is depicted in the vicinity of broken subunit interaction interface. SAP subunit is distinguished by combination of alphabetic letter and digit as in the case of Figure 1.

Figure 6

Potential of mean force (PMF) of SAP pentamer ring-opening reaction (A)–(C), and reaction time evaluated with Eyring’s transition-state theory (D). Each hcbMC/MD simulation is indexed by lowercase alphabetical letter as in the case of Figure 2, and the corresponding PMF is distinguished from the remaining on panel by different colors (red: a; blue: b; green: c; purple: d; orange: e). In (A)–(C), PMFs calculated from the atomic coordinates obtained from the unbiased 10 ns NPT MD simulation, i.e., the initial structure for each of the hcbMC/MD simulations, are shown by gray lines. An error bar was calculated from standard deviation and denoted as CI95. SAP subunit pair is distinguished by combination of alphabetical letter and digit as in the case of Figure 1. In (A), the green line is found on the gray one. In (D), points above the red dotted borderline, netted with the orange square area, have the estimated reaction time falling within the observation period with AFM experiments.

Figure 7

Atomic contacts between subunit pair for snapshot structure at hcbMC/MD cycle for the initial ring-opening reaction. Annotation of SAP subunit pair (S₁–S₂, etc.) and indexes of individual simulation (a–e) are given as in the case of Figure 1 and that of Figure 2, respectively.

Figure 8

Positional relation between Asp42 and Lys117 at the subunit interface. (A) The initial structure of hcbMC/MD simulations. (B) Structure at the timing of initial ring opening in the simulation indexed with (B). (C) Distances between Cγ in Asp42 and Nζ in Lys117 for each interface. SAP subunit is distinguished by combination of alphabetic letter and digit as in the case of Figure 1. In (A), the blue dotted lines denote hydrogen bonding between Lys117 and Asp42. In panel (C), the red dotted line shows the level of 4 Å. Each hcbMC/MD simulation is indexed by lowercase alphabet letters (a–e) as in the case of Figure 2, and the index “i” denotes the atomic coordinates obtained from the unbiased 10 ns NPT MD simulation, i.e., the initial structure for each of the hcbMC/MD simulations.