The regulation of conformational stability in multidomain proteins (MDP) is a central challenge in protein engineering with major implications for antigen optimization, enzyme activity modulation, and signal transduction. The function of these proteins depends on the properties of their constituent domains as well as on interdomain orientation and interface rearrangement during conformational transitions. Previous studies show that conformational transition is largely governed by changes in a small number of key residue pairs. However, such dynamic inter-residue contacts are typically sparse, transient, and coupled to large-scale domain motions, making them difficult to resolve directly by experiments or simulations. Existing deep-learning methods of dynamic contact prediction also lack specialized modeling for multidomain systems. Here, we present the Domain-Aware Residual Feature Modulation Network (DARMN), a deep learning framework for dynamic inter-residue contact prediction in MDP. Built on AlphaFold2 representations, DARMN uses coevolutionary information from multiple sequence alignments to capture sparse interdomain contact signals and applies Residual Feature-wise Linear Modulation to efficiently fuse single and pair representations. Furthermore, we design a domain-aware weighted focal loss function to distinguish between intradomain and interdomain contacts, thereby alleviating class imbalance and enhancing the learning of interdomain contacts. DARMN outperforms existing dynamic contact prediction and conformational ensemble prediction models, especially in long-distance contact identification, and generalizes well to unseen MDP. DARMN thus provides a useful computational framework for sequence design targeting conformational stabilization and mechanistic studies of MDP.
Jing Xiao, W. Wang, Yi Bo Ma et al.· Journal of Chemical Theory a...· 0 citations
Predicting the impact of single-point mutations on protein thermodynamic stability is crucial for protein engineering of therapeutic and industrial applications. By effectively capturing the three-dimensional structural information of proteins and the spatial physical environment of each residue, the inverse folding models (IFMs) upon fine-tuning, such as ThermoMPNN, achieved state-of-the-art performance in predicting thermostability changes in proteins caused by mutations. However, IFMs are limited in their capacity to capture protein deep evolutionary information, whereas protein language models (pLMs) excel. Here, we present SA-MPNN, a lightweight, end-to-end hybrid framework that dynamically integrates the protein sequence representations from a protein language model (ESM2) into the ThermoMPNN architecture to improve protein stability prediction by combining evolutionary representations with geometric structural embeddings. By evaluating various feature fusion strategies, we selected a self-attention-based integration mechanism to effectively combine the two modalities. Trained on the large-scale Megascale data set, SA-MPNN achieved modest but consistent gains over ThermoMPNN on various benchmark data sets, with particularly noticeable improvements in several correlation analysis and screening-oriented evaluations. Finally, wet-lab validation was performed on the top-ranking variants of Acetivibrio thermocellusβ-glucosidase (AtBgl1A) as a case study. The experimental results demonstrated that multiple designed mutants exhibited improved thermostability, and the optimal variant, GC20, achieved a melting temperature (Tm) of 76.98 °C, representing a 5.97 °C increase over the wild-type, thereby supporting the practical applicability of SA-MPNN in protein engineering.
Xin-Yue Zhang, Xiangshan Zheng, Ze-Yuan Dong et al.· Journal of Chemical Informat...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
