Exploring structural requirements of HDAC10 inhibitors through comparative machine learning approaches

Histone deacetylases (HDACs) are called the “REMOVER” enzymes as these eliminate the -acetate moiety from acetylated ϵ-amino groups of histone lysine and other non-histone proteins [[1], [2], [3], [4]]. They belong to the category of Zn2+ or nicotinamide adenine dinucleotide (NAD+)- dependent proteolytic enzymes, involved in transcriptional subduing as well as chromatin condensation. HDACs execute post-translational changes, such as ubiquitination and methylation, and can exert effects on gene transcription by increasing the interactivity between DNA and histone [[5], [6], [7]]. The deacetylation nature of the HDACs takes part as a modulator or regulator in several types of bio-signaling pathways [8]. Thus, they have crucial pathophysiological role in several deadliest maladies including neurodegenerative dysfunctions, inflammation, metabolic disorders, autoimmune disorders, and cancers etc. [[8], [9], [10], [11]]. The histone deacetylases family comprises 18 members [12,13] which are further divided into four classes depending on their homological sequences with yeast HDACs [14,15]. The four classes are class 1 HDACs (HDAC 1,2,3 and 8) [14]; class II HDACs, which are further subclassified into class IIa (HDAC 4,5,7, and 9) and class IIb (HDAC 6 and 10) [4]; class III which resembles yeast SIR2 (SIRT 1–7) protein structures and class IV that comprises only HDAC11 [16].

HDAC10 is an essential representative of class IIb HDACs and configurationally indistinguishable from the HDAC6 isotype. It is localized in cytoplasm and has poor lysine deacetylase activity. It can identify polyamines as substrate and thus, can be named as polyamine deacetylase [2,3,17,18]. The HDAC10 gene is confined to chromosome number 22 [18], includes 20 exons with two spliced transcripts [19,20] and contains an N-terminal catalytic domain and a leucine-rich domain with a C-terminal end. The N-terminal catalytic domain of HDAC10 is analogous to the deacetylase domain of other class II HDACs, but the C-terminal catalytic domain does not contain any residues that are imperative for enzymatic actions. The presence of both catalytic domains in the HDAC10 structure may confer resistance to trapoxin B and sodium butyrate [18,19]. Christianson and his co-workers have found that Zn2+ binding groups are present in HDAC10 and have additional interactions with inhibitors when accommodating with them [21]. They have also illustrated the structure of HDAC10 complexed with trifluoromethylketone (AAT) (PDB ID:1TD7), where canonical catalytic domain (PDAC domain) and smaller, non-catalytic ΨDAC domain assemble with unique butterfly-like architecture (Fig. 1) [22].

In recent years, HDAC10 inhibitors are emerging as a useful toolkit to combat numerous diseases. Therefore, the development of potent and selective inhibitors is much in demand. To develop better active HDAC10 inhibitors, it is necessary to focus on the structure-activity relationship (SAR) of the previously synthesized molecules as there is no crystal/NMR structure available for human HDAC10. In the present study, we have used different classification-based quantitative structure-activity relationship (QSAR) techniques, namely Bayesian classification, recursive partitioning and other machine learning (ML) based techniques like decision tree, random forest, gradient boosting, linear discriminant analysis (LDA) on a diverse set of molecules having HDAC10 inhibitory activities. The models were developed on the basis of prepared training set compounds and validated by the test set compounds. Moreover, molecular docking studies were applied to understand the interaction pattern of important structural fingerprints of the inhibitors with the HDAC10 enzyme. The current study has two main objectives: (i) the development of statistically validated ML models for prediction of effective HDAC10 inhibitors from different databases and (ii) an initiative to identify the favorable and unfavorable structural fingerprints that modulate HDAC10 inhibition.

留言 (0)

沒有登入
gif