Naringin dihydrochalcone potentially binds to catalytic domain of matrix metalloproteinase-2: molecular docking, MM-GBSA, and molecular dynamics simulation approach

Abstract Matrix metalloproteinase-2 (MMP2), an extracellular matrix remodulating protein’s increased activity causes cancer-metastasis. Potential MMP2 inhibitors showed sever side-effects in clinical trials. Present study is focused on identification natural MMP2 inhibitor by applying molecular docking, MM-GBSA binding energy estimation and molecular dynamics (MD) simulations. Commercially available flavonoid compound library was used to screen the molecules potentially binding with catalytic domain of MMP2 protein compared to standard MMP2 inhibitor ARP100. Naringin dihydrochalcone (NDC) showed interaction with the important residues (His120, Leu82 and Val117) present at the MMP2 catalytic domain in comparison to known inhibitor ARP100 (dock score ≈ −13 and −8 kcal/mole respectively). Lower ligand-protein binding energy (-67.31 kcal/mole) obtained in MM-GBSA and the MD simulation trajectory analysis showed significant stable and energetically favourable binding of NDC at the catalytic site of MMP2. In conclusion, anti-metastatic potential of NDC should be validated in in vitro and in vivo experiments. Graphical Abstract


Introduction
Cancer (Ca) remains one of the major causes of death in modern world resulting in to approximately 10 million deaths in 2020. Among various Ca hallmarks, metastasis is the primary cause of death due to the disease (Singh et al. 2021). MMP2 belongs to the family of matrix metalloproteinases (MMPs) enzymes which play a central role in development, reproduction, tissue remodelling, and angiogenesis. MMP2 is a central player in facilitating the invasion and metastasis of cancer cells by degrading the major components of basement membrane and type IV collagen. Overexpression of MMP2 in various cancers has been positively correlated with the tumour size, shortened overall survival and metastasis. Inhibition of MMP2 has been reported to reduce the invasion and metastatic potential of cancer cells (Vandenbroucke and Libert 2014). MMP inhibitors such as batimastat, marimastat, tanomastat, prinomastat, rebimastat, and andecaliximab have been discontinued at the advance stages of clinical trial due to sever adverse effect and increased toxicity (Winer et al. 2018). Flavonoids and their derivatives have gained enormous attention in preclinical anti-cancer drug discovery due to their low toxicity, minimal side effects, low cost and easy availability (Kumar and Pandey 2013). In silico and in vitro report showed anti-metastatic potential in flavonoids. Previously, natural MMP2 inhibitors such as eckol, lemnalol, cimicifugic acid B, genipin, kaempferol, were identified by using computer based drug discovery approach and later validated in pre-clinical experimental models (Bae et al. 2017;Kumar et al. 2019). Furthermore, various studies have been conducted to identify the natural inhibitors of MMP2 protein (Abu Bakar et al., 2019;Albelwi et al., 2021;Das et al., 2021). In the present study, commercially available flavonoid compound's library was utilised to identify the compound with MMP2 active site binding potential using molecular docking, MM-GBSA and molecular dynamics simulation.

Results and discussion
In the present study we evaluated the binding potential of commercially available flavonoids with catalytic domain of MMP2 using molecular docking studies. The binding potential was further validated by using MM-GBSA and molecular dynamics simulation study in a comparative manner. Docking analysis revealed significantly higher binding potential of test flavonoids (Naringin dihydrochalcone, daidzin, and sophoricoside) with the catalytic site of MMP2 compared to standard inhibitor ARP 100 (Figure S1-S3). Docking score, amino acids involved in interaction, and type of interaction of the test compounds having higher dock score than ARP 100 are summarised in Table S1. Result showed that the standard inhibitor formed only two H-bonds with the amino acids at catalytic residue. The lead compound showed more number of hydrogen binding (n ¼ 07). It should be noted that NDC formed H-Bond with the similar two amino acids which were involved in the binding with the standard inhibitor (Table S1 and Figure  S2). NCD showed better docking score compared to test compounds and interacted with the His120, Leu82 and Val117 residues of MMP2 catalytic domain. Standard inhibitor ARP100 and NCD showed p-cation and Van der Waal interaction respectively with the zinc ion present at the catalytic site of MMP2. Proença et al. (2019) reported that His120, Leu82 and Val117 amino acids play important role in stabilising the ARP 100 molecule (competitive inhibitor of MMP2) at the catalytic active site of MMP2 protein. Recently, in vitro anti-metastatic inhibitor potential of ARP 100 has been studied in retinoblastoma (Rb) and triple negative breast cancer cells and proposed that inhibiting MMP2 with the inhibitor might act as new strategy to target metastasis in cancer cells (Webb et al. 2017). Binding energy of the top three (Naringin dihydrochalcone-NDC, Daidzin and Sophoricoside) docked compounds and ARP 100 in MM-GBSA analysis was À67.31, À61.31, À56.19, and À35.98 kcal/mole, respectively ( Figure S4). Contributors of different type of interaction in total binding energy are depicted in Figure S4.
Further to validate the stability and energetically favourable binding of lead molecules at the catalytic site of the target protein molecular dynamics (MD) simulation of different systems (Unbound MMP2, ARP 100 -MMP2 complex, NDC-MMP2 complex and daidzin-MMP2 complexes) was run for 100 ns time period. Temperature and pressure of all the four test complexes were analysed to assure the quality of the conducted MD simulations ( Figure S5). Average root mean square deviation (RMSD), root mean square fluctuations (RMSF), radius of gyration (Rg), solvent assessable surface area (SASA), protein-ligand/protein-solvent/intramolecular hydrogen bond (H-Bond) values of the test complexes are provided in the Table S2. The NDC-MMP2 complex has significantly lower RMSD, residual fluctuations, Rg, and SASA values in comparison to the unbound MMP2, ARP 100 -MMP2 complex, and daidzin-MMP2 complex (Figures S6 and  S7). These molecular dynamics parameters indicate that the binding of NDC results in less structural deviations and residue fluctuation in the target protein as well as stabilised the overall compactness of the protein structure. A total of 112, 111, 106 and 107 intramolecular hydrogen bonds were formed in ARP 100 -MMP2, NDC-MMP2, unbound MMP2, and daidzin-MMP2 complex respectively ( Figure S8A). Protein-solvent H-Bond pattern was highest in unbound MMP2 followed by diadzein-MMP2 complex, ARP 100 -MMP2 complex, and NDC-MMP2 complex ( Figure S8B). Analysis of hydrogen bond formation between protein and ligand clearly indicates the binding strength of a ligand with the target protein. Binding of NDC and daidzin with MMP catalytic domain involved more number of average H-Bonds compared to ARP 100 ( Figure S9). The residue involved in H-bond formation was calculated for the entire simulation period at the interval of 10 ns and the results are depicted in Figures S10-S12. Plots of proteinligand hydrogen bond analysis illustrated stable pattern of hydrogen bond formation between MMP2 and NDC which indicates the stable binding efficacy of the flavonoid with the MMP2 catalytic domain. Two principal components (PC1 and PC2) were selected to perform the principal component analysis (PCA) on the backbone atoms of the catalytic domain of MMP2 in unbound and compound (ARP 100 , NDC, and daidzin) bound state. Result showed that binding of ARP 100 , NDC, or daidzin significantly changed the conformational space compared to unbound protein ( Figure S13). PCA allows us to understand and distinguish the conformational behaviour and protein dynamics in unbound and ligand bound state. PCA results indicated that binding of lead compounds altered the atomic motions of the catalytic domain of MMP2. The two dimensional and three dimensional Gibbs free energy landscapes (FELs) of unbound MMP2 (a central global and few small energy minimum states), ARP 100 bound MMP2 (two central global and few small energy minimum states), NDC bound MMP2 (a central global and more small energy minimum states), Daidzin bound MMP2 (three central global and few small energy minimum states) systems are depicted in Figure S14. Result showed that binding of flavonoids (NDC and daidzin) induced a stable and different conformational state in the catalytic domain of MMP2.

Experimental
Supplementary material related to this article is available online, alongside Tables S1-S2 and Figure S1-S14.

Conclusion
Present study identifies NDC as a natural inhibitor of MMP2 protein which showed comparatively better binding potential at the catalytic site of the protein compared to known experimentally validated MMP2 inhibitor (ARP 100 ). Further, the MMPBSA and MD-simulation studies revealed that NDC formed stable and energetically favourable complex with the MMP2. Thus, NDC is a promising candidate as natural MMP2 inhibitor for its therapeutic implication in cancer treatment and should be further tested in in vitro and in vivo experiments.

Disclosure statement
No potential conflict of interest was reported by the authors.