MicroRNAs (miRNAs) are small non-coding single stranded RNAs (~18-22 bp) found in most eukaryotes, playing a crucial role in a wide range of biological processes, such as host-viral interaction, apoptosis, stem cell maintenance, and metabolism, etc. miRNAs are involved in gene silencing mechanisms by targeting un-translated regions of mRNA through sequence complementarity [1,2], and approximately 30% of protein-coding genes are regulated by miRNAs [3]. As many genes can be regulated by a single or multiple miRNAs, an immense interest has been created to understand miRNA-based gene regulation through the identification of miRNAs, as well as their target genes. The Argonaute (AGO) class of proteins plays a crucial functional role in the molecular recognition of miRNA-mRNA hybrids during the process of gene regulation [4,5], and mutations in either the sequence of miRNAs or the miRNA target sites have a significant contribution in the manifestation of cancer due to the loss of gene regulation [5,6].

Leukemia is a cancer of white blood cells that is commonly classified as acute myeloid leukemia (AML), acute lymphoblastic leukemia, and chronic lymphoblastic leukemia. AML is a well-characterized blood cancer in which the unnatural growth of immature white blood cells takes place, where several genes involved are either up-regulated or down regulated by miRNAs. Though the importance of miRNAs in gene regulation [2,3,4,5,7,8] and their regulation by RNA-induced silencing complexes [7] are well established, the mechanism of miRNA-induced gene silencing at the molecular level is still unclear [9,10]. Despite several detailed research studies, controlling abnormal cell proliferation at the molecular level through miRNA-mediated gene silencing is still a major challenge. Therefore, with the current knowledge, attempts have been made to study the AGO protein-assisted molecular interaction between miRNAs and their target genes, and a computational approach [11] in this regard will be of great aid and value. Here, in this study, the HOXA9 gene was selected from the literature as an important target in AML, and in silico approaches, like miRNA target site prediction, miRNA-mRNA duplex prediction, and molecular docking between AGO protein and four miRNAs, such as mir-145, mir-126, let7a, and mir-196b [8,12,13], were conducted to elucidate the AGO protein-assisted molecular interaction between miRNAs and the HOXA9 gene in association with AML [14,15] computationally.

The current study focused on in silico verification of miRNAs for targeting the HOXA9 gene, associated in AML, and the molecular interaction of catalytic AGO protein with these miRNAs to study the gene silencing mechanism computationally. As the literature has reported [12,13], four mature miRNAs-namely, miR-145, mir-126, let-7a, and mir-196b-are found to have target sites in the HOXA9 gene and were hence taken for further study. The binding sites in HOXA9 were predicted for these miRNAs using miRTarBase (Supplementary Tables 1 and 3), which was followed by the secondary structure prediction and miRNA-mRNA duplex formation. Two binding sites were predicted for mir-196b in the HOXA9 gene (Supplementary Table 1), denoted mir-196b.1 and mir-196b.2, respectively; so, two duplexes are formed between mir-196b and mRNA of the HOXA9 gene (Supplementary Table 3). The results suggested a strong and energetically favorable binding mode between all miRNA-mRNA duplexes (Supplementary Table 4), which is also supported by the reported interaction between them (Fig. 1). The good MFE score (-5.90 kcal/mol) also reflected that mir-145 had the highest binding affinity for the HOXA9 gene (Supplementary Table 4) among all miRNAs. Furthermore, to study the molecular interaction between AGO protein and miRNA-mRNA duplexes (miR-145, mir-126, let-7a, mir-196b.1, and mir-196b.2), molecular docking was carried out. The presence of strong hydrophobic amino acid residues (LEU 652, LEU 267, ALA 644, and ALA 648 in mir-145; LEU 267, LEU 277, LEU 279, LEU 281, ALA 278, ALA 648, and VAL 606 in mir-126; LEU 267, LEU 279, LEU 281, and ALA 278 in let-7a; ALA 111, VAL 58, and LEU 132 in mir-196b.1; and VAL 549, VAL 152, and ALA 170 in mir-196b.2) within a distance of 3.5 Å in the binding pocket of miRNAs implicates strong binding interactions at the molecular level between AGO protein and all chosen miRNAs (Table 1, Fig. 2A-2E). Also, amino acids with aromatic rings (TYR 642, PHE 649, and PHE 647 in mir-145; PHE 649 in mir-126; and TYR 171 and TYR 135 in mir-196b.2) were also observed within a distance of 3.5 Å, which supports the binding interaction (Table 1, Fig. 2A-2E) between AGO protein and miRNAs. The amino acid residues participated in hydrogen bonding (Table 2, Fig. 3A-3E) with miRNAs (miR-145, mir-126, let-7a, mir-196b.1, and mir-196b.2) within a distance of 3.0 Å, showing strong evidence of binding and a molecular interaction between AGO protein and all chosen miRNAs. The current investigation proved that mir-145, mir-126, let-7a, mir-196b.1, and mir-196b.2 have strong binding affinity for the HOXA9 gene and also throws light on AGO protein-assisted molecular interactions with miRNA-mRNA duplex structures. This report would be helpful in understanding miRNA-based gene silencing mechanisms for the HOXA9 gene in AML disease. Furthermore, a molecular dynamic simulation approach might be implemented to study the dynamic behavior of these interactions, which would help us better understand miRNA-based gene silencing mechanisms in AML computationally. Once validated experimentally, this methodology would further help us to design suitable miRNAs against respective genes rationally.