ИСТИНА

Carotenoproteins are protein complexes with a stoichiometrically bound carotenoid component, therebyacquiring important biological functions. Despite the non-covalent nature of the bond, a strict specificityof interaction between carotenoids and specific protein regions is observed. The mechanisms of thisspecificity remain insufficiently studied, which is relevant in light of the development of newbiotechnological tools based on carotenoproteins and the demand for their targeted modification.The aim of this work was to identify the features of protein-carotenoid interaction through a detailedanalysis of the molecular structures of carotenoprotein holo-forms from the PDB database. Particularattention was paid to structures obtained using cryo-electron microscopy (cryo-EM), which providesdata on the native distribution of conformations in an ensemble of specific structures.The enrichment or depletion of the protein-pigment interface with specific types of residues relative totheir background occurrence in proteins deposited in the Swiss-Prot database was analyzed. Acomparative analysis was performed both for the environment of the entire polyisoprenoid chain of thecarotenoid and separately for the environment of the ionone rings. Special attention was given to thegeometric parameters of interactions between the ionone rings of carotenoids and the aromatic aminoacids of the protein. To identify ligand-specific patterns in the amino acid environment and visualizestructural differences, the UMAP dimensionality reduction method was applied.The analysis showed that the carotenoid binding interface is significantly enriched in hydrophobicamino acids and, conversely, strongly depleted in polar and charged residues, which is a key propertyfor specific interaction. Tryptophan and phenylalanine achieve the highest frequency of occurrencerelative to the background (4 times more frequent), while lysine achieves the lowest (128 times lessfrequent).Geometric analysis of interactions between ionone rings and aromatic amino acids showed apredominance of perpendicular (T-shaped) orientations, especially for phenylalanine. Tryptophandemonstrated the ability to form both T-shaped and parallel (stacking) interactions, while tyrosineshowed geometric non-specificity. Furthermore, ligand-specific differences in the amino acidcomposition of binding sites were found, which is confirmed by UMAP data clustering.Thus, the analysis succeeded in identifying specific organizational features of carotenoid-binding sites.