CoVAMPnet: Comparative Markov State Analysis for Studying Effects of Drug Candidates on Disordered Biomolecules

Abstract

Alzheimer's disease (AD) is characterized by the deposition of misfolded tau and amyloid-beta (Aβ). Tramiprosate (TMP) and its metabolite 3-sulfopropanoic acid (SPA) are phase 3 therapeutics believed to target Aβ oligomers. It is of paramount importance to understand how TMP/SPA modulate the conformations of Aβ. Here, we studied the Aβ42 alone and in the presence of TMP or SPA by adaptive sampling molecular dynamics. Next, to quantify the effects of drug candidates on Aβ42, we developed a novel Comparative Markov State Analysis (CoVAMPnet) approach: ensembles of learned Markov state models were aligned across different systems based on a solution to an optimal transport problem, and the directional importance of inter-residue distances for assignment to the Markov states was assessed by a discriminative analysis of aggregated neural network gradients. TMP/SPA shifted Aβ42 towards more structured conformations by interacting non-specifically with charged residues and destabilizing salt bridges involved in oligomerization. SPA impacted Aβ42 the most, preserving α-helices and suppressing aggregation-prone β-strands. Experimental biophysical analyses showed mild effects of TMP/SPA on Aβ42, and activity enhancement by the endogenous metabolization of TMP into SPA. The CoVAMPnet method is broadly applicable to study the effects of drug candidates on conformational behavior of intrinsically disordered biomolecules.

Machine Learning-Guided Protein Engineering

“…This approach was applied in conjunction with the end-to-end deep learning model VAMPnet, which was used to find an optimal projection for the system under study. 229 More recently, the CoVAMPnet framework 230 was created by expanding on the VAMPnet approach to add interpretability capabilities and a method for MSM alignment that facilitates comparisons of MSMs for two sets of simulations. If applied to different variants of a given protein, CoVAMPnet could potentially be used to evaluate the effects of discriminative mutations on protein dynamics.…”

Kinetic Ensemble of Tau Protein through the Markov State Model and Deep Learning Analysis

“…36,38,39 A recent study developed a novel Comparative Markov State Analysis (CoVAMPnet) approach to quantify the effects of drug candidates on Aβ 42 . 38 Furthermore, the application of State Predictive Information Bottleneck (SPIB) framework can efficiently and accurately learn a reaction coordinate from MD trajectories, which holds great potential for future IDPs research. 39 In this study, we combine the AlphaFold database, accelerated MD (aMD), extensive MD simulations, and MSM to characterize the conformational landscape of the K18 transition pathway.…”

“…Huang and colleagues utilized MSMs to identify the metastable conformational states of the intrinsically disordered monomer of the human islet amyloid polypeptide (hIAPP) and the transitions between them . In recent years, the deep learning method have been increasingly employed to unravel the biological importance of IDPs. ,, A recent study developed a novel Comparative Markov State Analysis (CoVAMPnet) approach to quantify the effects of drug candidates on Aβ 42 . Furthermore, the application of State Predictive Information Bottleneck (SPIB) framework can efficiently and accurately learn a reaction coordinate from MD trajectories, which holds great potential for future IDPs research …”