Structural Patterns in α Helices and β Sheets in Globular Proteins

Application of Machine Learning in the Quantitative Analysis of the Surface Characteristics of Highly Abundant Cytoplasmic Proteins: Toward AI-Based Biomimetics

Proteins in the crowded environment of human cells have often been studied regarding nonspecific interactions, misfolding, and aggregation, which may cause cellular malfunction and disease. Specifically, proteins with high abundance are more susceptible to these issues due to the law of mass action. Therefore, the surfaces of highly abundant cytoplasmic (HAC) proteins directly exposed to the environment can exhibit specific physicochemical, structural, and geometrical characteristics that reduce nonspecific interactions and adapt to the environment. However, the quantitative relationships between the overall surface descriptors still need clarification. Here, we used machine learning to identify HAC proteins using hydrophobicity, charge, roughness, secondary structures, and B-factor from the protein surfaces and quantified the contribution of each descriptor. First, several supervised learning algorithms were compared to solve binary classification problems for the surfaces of HAC and extracellular proteins. Then, logistic regression was used for the feature importance analysis of descriptors considering model performance (80.2% accuracy and 87.6% AUC) and interpretability. The HAC proteins showed positive correlations with negatively and positively charged areas but negative correlations with hydrophobicity, the B-factor, the proportion of beta structures, roughness, and the proportion of disordered regions. Finally, the details of each descriptor could be explained concerning adaptative surface strategies of HAC proteins to regulate nonspecific interactions, protein folding, flexibility, stability, and adsorption. This study presented a novel approach using various surface descriptors to identify HAC proteins and provided quantitative design rules for the surfaces well-suited to human cellular crowded environments.

Communities in the iron superoxide dismutase amino acid network

Amino acid networks (AANs) analysis is a new way to reveal the relationship between protein structure and function. We constructed six different types of AANs based on iron superoxide dismutase (Fe-SOD) three-dimensional structure information. These Fe-SOD AANs have clear community structures when they were modularized by different methods. Especially, detected communities are related to Fe-SOD secondary structures. Regular structures show better correlations with detected communities than irregular structures, and loops weaken these correlations, which suggest that secondary structure is the unit element in Fe-SOD folding process. In addition, a comparative analysis of mesophilic and thermophilic Fe-SOD AANs' communities revealed that thermostable Fe-SOD AANs had more highly associated community structures than mesophilic one. Thermophilic Fe-SOD AANs also had more high similarity between communities and secondary structures than mesophilic Fe-SOD AANs. The communities in Fe-SOD AANs show that dense interactions in modules can help to stabilize thermophilic Fe-SOD.

Conformational preferences of non-polar amino acid residues: an additional factor in amyloid formation

Amyloid consists of β-sheet polymers and is associated with disease and with functional assemblies. Amyloid-forming proteins differ widely in native structures and sequences. We describe here how conformational preferences of non-polar amino acid residues can affect amyloid formation. The most non-polar residues promote either β-strands (Val, Ile, Phe, and Cys, VIFC) or α-helices (Leu, Ala, and Met, LAM), while the most polar residues promote only α-helices. For 12 proteins associated with disease, the localizations of the amyloid core regions are known. Eleven of these contain segments that are biased for VIFC, but essentially lack segments that are biased for LAM. For the amyloid β-peptide associated with Alzheimer's disease and an amyloidogenic fragment of the prion protein, observed effects of mutations support that VIFC bias favors formation of β-sheet aggregates and amyloid, while LAM bias prevents it. VIFC and LAM profiles combine information on secondary structure propensities and polarity, and add a simple criterion to the prediction of amyloidogenic regions.