In silico targeting of red complex bacteria virulence factors of periodontitis with β-defensin 1

Background Periodontitis is a multi-factorial infection with red complex bacteria playing a crucial role in the pathogenesis. As bacteria are tending to develop resistance against conventional antibiotics, new treatment modalities need to be developed. Antimicrobial peptides (AMPs) are potential tools for drug development and are gaining widespread interest. β-defensin 1 is an important AMP and forms the first-line host defense mechanism. The present study analyzed the structure and molecular docking of β-defensin 1 with the virulence factors of red complex bacteria of periodontitis. The physico-chemical properties of β-defensin 1 were determined by various online tools such as ProtParam, ProteinPredict, ToxinPred, and BioPep web servers. The structure of β-defensin 1was predicted by the SWISS-MODEL web server and the structure was evaluated by different web tools. The structure of lipopolysaccharide of Porphyromonas gingivalis was drawn using Chem3D ultra 11.0 software. The structure of important protein virulence factors of red complex bacteria of periodontitis was determined by the SWISS-MODEL web server. The interaction study between β-defensin 1 and virulence factors was carried out by molecular docking using Auto dock version 4.0 software and pyDock WEB server. Results Using online tools, β-defensin 1 was predicted to be stable and non-toxic. SWISS-MODEL web server predicted Ramachandran score as 94.12% and clash score 0.0 for β-defensin 1. Auto dock version 4.0 software and pyDock WEB server analyzed the interaction to have low binding energies and hydrogen bonds were formed between the peptide and virulence factors. Conclusion β-defensin 1 was found to have good binding interaction with the disease-causing factors of red complex bacteria of periodontitis and in turn could play a role in reducing the severity of infection. β-defensin 1 could be a potential candidate for drug development for periodontitis.

bacteria generally are not found alone, but together in the periodontal pockets, implicating that these bacteria work in a cooperative manner leading to the destruction of periodontal tissues. Red complex bacteria produce various virulence factors that cause the destruction of periodontal tissues and also disseminate into the bloodstream modulating host immune responses which help bacteria evade host immune responses [3]. Important virulence factors from each bacterium have been chosen for the study and explained below. P. gingivalis produces disease-causing factors, such as lipopolysaccharide (LPS), gingipains, fimbriae/pili, collagenase, lectins, capsule, protease, and superoxide dismutase [4]. LPS lipid A endotoxin structure produced by P. gingivalis can undergo acylation by two different modes-tetra-acylated lipid A designated LPS 1435/1449 based on its molecular weights of 1435 and 1449 Da and the penta-acylated LPS designated LPS 1690 with the molecular weight of 1690 Da [5]. Gingipains are arginine-and lysine-specific cysteine proteinases that cleave peptide and protein substrates after arginine (gingipain R) and lysine residues (gingipain K). There are two types of gingipain R, namely RgpA and RgpB and one type of gingipain K, Kgp. Gingipains degrade extracellular matrix components, including fibronectin and collagen, deregulate inflammatory responses by cleaving cytokine, immunoglobulin, and complement factors [6]. Some of the virulence factors identified in T. forsythia include trypsin-like proteases, sialidases SiaH, leucine-rich repeat cell-surface-associated and secreted protein bacteroides surface protein A (BspA), hemagglutinin, karilysin, and mirolysin. BspA by activating the Toll-like receptor-2 (TLR-2)-dependent pathway triggers the release of bone-resorbing proinflammatory cytokines and chemokines from monocytes and gingival epithelial cells [7]. Mirolysin and karilysin show synergistic inhibition on the complement pathway of the host system [8]. Dentilisin is one of the major virulence factors of T. denticola and plays a role in disease progression by degrading host cell-matrix proteins and modulate immune responses by degrading cytokines such as interleukin-1β (IL-1β) and IL-6 [9]. Cystalysin from T.denticola is an enzyme that is responsible for hemolytic activities and damage to the gingival and periodontal tissues [10]. Studying the mechanism of action of virulence factors of red complex bacteria and their inhibition is an important research area and would open new avenues for the treatment of periodontitis. Antimicrobial peptides (AMPs) are important constituents of the innate immune system and have inhibitory effects against different microbes [11]. Due to the rise in resistance against conventional antibiotics, efforts are being made to bring AMPs to clinical use [12]. Defensins are small cationic AMPs that exhibit their antimicrobial activity mainly by disrupting membranes of pathogens via forming pores leading to osmotic imbalance [13]. Human β-defensin 1 is constitutively expressed in many epithelial tissues including kidney, gut, respiratory epithelium, and in the stratified epithelium of the oral cavity [14], and the present study analyzed the interaction between β-defensin 1 and virulence factors of red complex bacteria of periodontitis.

Computing physico-chemical properties of β-defensin 1
The FASTA sequence of β-defensin 1 was retrieved from the UniProt server (https:// www. unipr ot. org/ unipr ot/ P60022) and the accession number is P60022. The amino acid sequence of β-defensin 1 was submitted to the Prot-Param parameter analysis tool (https:// web. expasy. org/ protp aram/) for the computation of various physical and chemical properties such as molecular weight, estimated half-life, instability index, and grand average of hydropathicity (GRAVY).
Other important properties of β-defensin 1 such as solvent accessibility, cell penetration, bioactivity, and toxicity were calculated using different web tools as given below. Solvent accessibility of the amino acid residues in terms of three-dimensional structure and sub-cellular localization were calculated using the ProteinPredict web server (https:// predi ctpro tein. org/). Penetration through the blood-brain barrier was also predicted using the BBPpred online server (http:// bbppr ed. xialab. info/). The toxicity of β-defensin 1 was predicted using the ToxinPred web server (https:// webs. iiitd. edu. in/ ragha va/ toxin pred/ prote in. php). The antimicrobial activity was predicted by comparing the sequence of interest with the pre-existing antimicrobial peptide database AMPA (http:// tcoff ee. crg. cat/ apps/ ampa/ do: ampa). The bioactivity of the peptide was predicted using the BioPep web server (http:// www. uwm. edu. pl/ bioch emia/ index. php/ en/ biopep).

Homology modeling of β-defensin 1
Homology modeling of β-defensin 1 was carried using the SWISS-MODEL web server (https:// swiss model. expasy. org/ inter active). Sequence ID P60022 retrieved from the UniProt server was utilized for modeling. The quality of the model was assessed and evaluated using online servers such as ERRAT, Procheck, Verify3D, and PROVE (https:// saves. mbi. ucla. edu/).

Three-dimensional structure prediction of virulence factors of red complex bacteria
Virulence factors LPS (LPS 1435/1449 and LPS 1690 ), gingipain K (Kgp), gingipain R (RgpB) of P. gingivalis; BspA, karilysin, mirolysin of T. forsythia, and cystalysin, dentilisin of T. denticola were selected for the study. Since the LPS structure of P.gingivalis was not available in PDB, tetra (LPS 1435/1449 ), and penta-acylated (LPS 1690 ) LPS structures of P.gingivalis were drawn using Chem3D ultra 11.0 software and energy minimized and cleansed using AVOGADRO software. Protein sequences of other virulence factors of red complex bacteria were downloaded from the UNI-PROT webserver and subjected to protein blast. The UNIPROT accession numbers are Gingipain K (Kgp)-P72194, Gingipain R (RgpB)-P95493, BspA-O68831, Karilysin-A0A1D3UH21, Mirolysin-A0A0F7IPS1, Cystalysin-Q56257, Dentilisin-P96091. The sequences with high similarity index were then taken into consideration and used for protein modeling. Protein modeling of different conformers was carried out using the SWISS-MODEL web server and the same were analyzed using structural assessment tool in the Swiss bioinformatics database. Conformers showing high Ramachandran plot value and low clash score were selected for further process. The Qmean value of the model was also estimated to analyze the integrity of the developed homology protein structures.

β-Defensin 1 and virulence factors molecular docking study β-Defensin 1 and LPS interaction
Docking between β-defensin 1 and LPS was carried out using Auto dock version 4.0 software. The auto grid, the component of the Auto dock, was used to compute the grid map. The best conformations search was done by adopting a genetic algorithm with the local search (GA-LS) method. The docking parameters were set to default values with 100 independent docking runs using the software Auto dock Tool Kit. Molecular graphics and visualization were performed with the UCSF Chimera package.

β-Defensin 1 and protein interaction
The docking between β-defensin 1 and protein virulence factors was carried out using the pyDockWEB server (https:// life. bsc. es/ pid/ pydoc kweb) [15]. Initially, the protein complexes were prepared by carrying out protein homology modeling using SWISS-MODEL as mentioned above. Prepared models were evaluated using structural assessment by calculating the Qmean value. Models with better results were energy minimized and cleansed using protein preparation wizard (Schrodinger suite). Finally, both protein and peptide complexes were uploaded to the pyDockWEB server.

Physico-chemical properties of β-defensin 1
The amino acid sequence of β-defensin 1 was subjected to the ProtParam protein parameter analysis tool and the data obtained is given in Table 1.
Solvent accessibility of the amino acid residues was calculated using the ProteinPredict web server and the results are given below in Fig. 1. Out of 68 residues, 69.12% was found to have been exposed to solvent accessibility while 30.88% was found to be buried inside the core of the protein.
Subcellular localization of β-defensin 1 peptide, i.e., distribution and sedimentation nature of the biomolecule was studied and the results are depicted in Fig. 2. In the archaea domain, the protein was able to sediment at the cell walls of the organism whereas in the bacterial domain the protein was able to penetrate through the membrane, and finally for higher eukaryotic cells the protein was not able to penetrate even the outer cell walls. Hence, β-defensin 1 could be considered non-toxic for eukaryotic cells. Similarly, penetration through the blood-brain barrier was also predicted using the BBPpred online server and results are tabulated in Table 2.
Toxicity prediction of β-defensin 1 was carried out using the ToxinPred web server which cuts the protein into small peptides of 10 residue lengths and the prediction was carried out. The tool confirmed the non-toxic nature of the peptide. Antimicrobial activity was predicted by comparing the sequence of interest with the pre-existing antimicrobial peptide database and the results are shown below in Fig. 3.
Bioactivity of β-defensin 1 was predicted using the BioPep web server and predictions suggested that the peptide has excellent bioactivity toward a wide array of activity including angiotensin-converting enzyme (ACE) inhibitor, glucose uptake stimulating peptide, neuropeptide, antioxidative peptide, and anti-inflammatory peptide.

Structure prediction of β-defensin 1by homology modeling
The structure of β-defensin 1 was predicted by the SWISS-MODEL web server and the quality of the model was assessed and evaluated using different web tools. The results are depicted in Fig. 4 and Table 3.    The quality of the model was assessed and evaluated using online servers such as ERRAT, Procheck, Verify3D, PROVE, and the results are tabulated in Table 4.

Three-dimensional structures of virulence factors of red complex bacteria
The structure of LPS of P. gingivalis is given in Fig. 5.
The 3D structures of virulence factors are given below with their respective homology modeling scores in Fig. 6 and Table 5. Q Mean score represents quality of the protein structure predicted by SWISS-MODEL web server with a score of one being good.

Molecular docking of virulence factors with β-defensin 1 Docking of β-defensin 1 and LPS structure of P. gingivalis
The docking results of β-defensin 1 against LPS 1435/1449 and LPS 1690 of P. gingivalis are shown in Fig. 7 and Table 6. The binding energy (Kcal/mol), inhibition constant (Ki), and hydrogen bonds formed were used to evaluate the binding affinity of β-defensin 1 with LPS. The binding energies of LPS 1435 , LPS 1449 , and LPS 1690 with β-defensin 1 were found to be − 7.45 Kcal/mol, − 8.31 Kcal/mol, and − 8.67 Kcal/mol respectively. LPS 1690 had the best binding energy among the three LPS structures.

Docking of β-defensin 1 with gingipains of P. gingivalis
The docking results of β-defensin 1 against gingipains Kgp and RgpB of P. gingivalis are summarized below in Fig. 8 and Table 7. β-defensin 1 formed 3 hydrogen bonds with gingipain Kgp and 4 hydrogen bonds with gingipain RgpB. The binding energy (Kcal/mol), interacting residues, and hydrogen bonds formed were used to evaluate the binding affinity of β-defensin 1 with gingipains of P. gingivalis. The binding energies of gingipains Kgp and RgpB with β-defensin 1 were found to be − 40.814 Kcal/ mol and − 55.277 Kcal/mol. Gingipain RgpB had the best binding affinity with β-defensin 1.

Docking of β-defensin 1 with virulence factors of T. denticola
The docking results of β-defensin 1 against cystalysin and dentilisin of T. denticola are summarized in Fig. 10 and Table 9. β-defensin 1 interacted by forming 3 hydrogen bonds with cystalysin and dentilisin. The binding energies of cystalysin and dentilisin with β-defensin 1 were found to be − 40.846 Kcal/mol and − 41.028 Kcal/mol.     Both the virulence factors of T. denticola showed similar binding energies. The binding energies of β-defensin 1 from the docking study indicated that it had good interaction with different virulence factors of red complex bacteria of periodontitis.

Discussion
Oral cavity provides a favorable environment for colonization by various microorganisms so oral flora requires myriad defense mechanisms in order to prevent infection. Since red complex bacteria are important periodontal Table 7 Interaction results of β-defensin 1 against gingipains Kgp and RgpB of P. gingivalis  pathogens, numerous research works are being carried out to study the disease pathogenesis and strategies to inhibit their virulence mechanisms. AMPs show broadspectrum antimicrobial activity and are important constituents in the host defense against microbial challenge. They are expressed constitutively or induced in response to microorganisms. β-defensin 1 in addition to its antibacterial role performs other immune-related activities such as chemoattractants and promotes angiogenesis [16]. Studies have shown that AMPS such as lactoferrin and histatin have antibacterial efficacy against P. gingivalis [17,18]. Inhibition of virulence factors could prevent or slow down the progression of periodontitis and several inhibitors from natural sources are being developed [19]. Molecular modeling plays a pivotal role in computeraided drug design and is one of the most important virtual screening methods to study drug-receptor interaction [20]. Docking is a computational process for finding a suitable ligand that fits the protein's binding site energetically and geometrically [21]. The present study analyzed the interaction between the antimicrobial peptide β-defensin 1 and the virulence factors of red complex bacteria of periodontitis. Physico-chemical properties of β-defensin 1 were determined by the ProtParam web tool. The instability index computed by the tool predicted that the peptide was stable. The distribution nature of the peptide was studied by the ProteinPredict tool which predicted that the peptide was able to penetrate through the membrane of bacteria while it was not able to penetrate the higher eukaryotic membrane. β-defensin 1 can exert its antimicrobial effects without harming normal eukaryotic cells due to the positive charge on the peptide which can interact with negatively charged membranes of microbes, while the membrane composition of eukaryotic cells is mainly uncharged zwitterion (neutral) phospholipids, sphingomyelins, and cholesterol and hence cannot affect eukaryotic cells [22].
The structure of β-defensin 1 was predicted by the SWISS-MODEL web server which gave the Ramachandran score of 94.12% and clash score of 0.0. The 3D structures of virulence factors of red complex bacteria of periodontitis were predicted by the SWISS-MODEL server. The structure of LPS of P.gingivalis was drawn by Chem3D ultra 11.0 software.
Docking of β-defensin 1 with LPS of P. gingivalis was studied by Autodock version 4.0 software. Docking of β-defensin 1 with other virulence factors was carried out by pyDockWEB server. Docking studies with Auto-Dock software and pyDockWEB server revealed that β-defensin 1 was able to bind with low docking energies indicating their high affinity with the selected proteins of bacteria. β-defensin 1 showed significant interaction with the virulence factors and could play a role in the Fig. 10 Docking results of β-defensin 1 against a cystalysin and b dentilisin of T. denticola Table 9 Interaction results of β-defensin 1 against cystalysin and dentilisin of T. denticola inhibition of pathogenesis of periodontitis mediated by red complex bacteria.

Conclusion
β-defensin 1 is one of the important components of host's natural innate immunity and in this study, through molecular docking, it has been found that β-defensin 1 interacted with various virulence factors of periodontitis which suggests that it could be used as an adjunct/ alternative to antibiotics for the treatment of periodontitis. Through various online tools, β-defensin 1 was found to be non-toxic. It could be a promising candidate in the treatment of periodontitis and has a great potential application in the drug development process.