Searching for New Antibacterial Compounds Against Staphylococcus aureus: A Computational Study on the Binding Between FtsZ and FtsA

Abstract

Background: Staphylococcus aureus is a pathogen that has become resistant to different antibiotics, which makes it a threat to human health. Although the first penicillin-resistant strain appeared in 1945, nowadays, there are just a few alternatives to fight it. To circumvent this issue, novel approaches to develop drugs to target proteins of the bacteria cytoskeleton, essential for bacteria’s binary fission, are being developed. FtsZ and FtsA are two proteins that are key for the initial stages of binary fission. On one side, FtsZ forms a polymeric circular structure called the Z ring; meanwhile, FtsA binds to the cell membrane and then anchors to the Z ring. According to the literature, this interaction occurs within the C-terminus domain of FtsZ, which is mainly disordered. Objective: In this work, we studied the binding of FtsZ to FtsA using computational chemistry tools to identify the interactions between the two proteins to further use this information for the search of potential protein-protein binding inhibitors (PPBIs). Methods: We made a bioinformatic analysis to obtain a representative sequence of FtsZ and FtsA of Staphylococcus aureus. With this information, we built homology models of the FtsZ to carry out the molecular docking with the FtsA. Furthermore, alanine scanning was conducted to identify the key residues forming the FtsZ–FtsA complex. Finally, we used this information to generate a pharmacophore model to carry out a virtual screening approach. Results: We identified the key residues forming the FtsZ-FtsA complex as well as five molecules with high potential as PPBIs.

1. Introduction

Staphylococcus aureus (S. aureus) is a pathogen that has become multi-resistant to different antibiotics, which makes it a threat to human health, hence, the developing of novel therapies to face is compulsory [1,2,3,4,5,6].

Bacteria multiply through a mechanism known as binary fission, for which the bacteria’s cytoskeleton, formed by structural proteins, provides the cell with the structure needed to divide. The proteins that form the bacteria cytoskeleton are potential targets for developing novel antibiotics [7].

FtsA and FtsZ proteins are two main elements of the cytoskeleton, and they start the binary division process to end up with two cells that will continue to propagate the pathogen [5,6,8,9,10].

FtsZ is an enzyme that exerts GTPase activity, plays an essential role in the cell division processes, and is present in most bacteria [11,12]. Its structure (Figure S1A) and biological roles are known, which makes it an attractive biological target for developing novel drugs [13,14].

FtsZ protein dwells in the cytoplasm of many bacteria, and it forms a contractile ring by assembling itself in filamentous fragments that divide to yield two cells [15] (Figure S1B). It is formed by two globular subdomains (N- and C-termini) joined by an α-helix (H7) (Figure S1A). FtsZ is conserved across various bacteria, with a coverage ranging from 53% to 100% and protein-level identity between 57% and 99%, which highlights the importance of its function as part of the bacterial cytoskeleton, hence considered the prokaryotic equivalent of eukaryotic tubulin. The N-terminal domain folds in a Rossmann-type folding and comprises six β segments surrounded by α-helices joined by loops that form a cavity known as the nucleotide-binding site (Figure S1A). In contrast, the C-terminus domain is formed by four β segments joined by two α-helices on each side plus the catalytic loop T7 that displays the GTP hydrolysis and the polymerization of FtsZ [16,17,18] (Figure S1B).

At the end of the C-terminus domain, there is a disordered region of an approximated size of 80 residues, which is not resolvable by crystallography and works as the binding site that recognizes proteins, such as FtsA [9,19].

FtsA is a protein of approximately 480 residues that exerts ATPase activity and is the second most conserved protein among bacteria species. Together with FtsZ, it provides the structure for the Z ring in S. aureus (Figure S2B), which makes it an interesting target for designing novel drugs (Figure S2B) [10,20,21,22,23,24,25]. FtsA is the second most conserved protein in bacteria, with coverage ranging from 49% to 100% and protein-level identity between 47% and 100%. In 2014, the structure of S. aureus’ FtsA (FtsA) was resolved by crystallography showing four subunits in the asymmetric crystallographic cell: 1 A, 1C, 2 A, and 2B (Figure S2A) [26,27]. The binding of FtsA with FtsZ happens in many bacteria species, including S. aureus [20,28,29,30]. During the binding, one end of FtsA anchors to the cytoplasmatic membrane of bacteria, whereas the other interacts with the C-terminus disordered end of FtsZ [26]. This binding maintains the structural integrity of the Z ring, allowing the correct segregation of chromosomes [31,32].

Lately, in the literature there are reports of compounds known as anti-FtsZ obtained from natural sources, such as algae and plants [33,34]. In contrast, synthetic ones are benzamide inhibitors, namely, PC190723 [30], TXH9179, and TXH9179 [35], that mimic GTP and inhibit the polymerization of FtsZ.

In 2021, a docking study reported the binding of a xanthone (2-methylxantho-9-none) [36] that prevents the formation of the divisome and leads to a defective arrangement of the Z ring, which finally results in wrong bacteria division; nonetheless, such a compound has not reached clinical trials yet [37,38,39,40].

Our study focused on the interaction between FtsZ and FtsA, aiming to understand the nature of this protein–protein binding. Insights from this research could inform the design of compounds that disrupt this interaction, potentially offering a new strategy to combat multi-drug-resistant strains of S. aureus. To understand how FtsZ and FtsA interact with each other, we carried out molecular docking studies between them using two distinct protein–protein docking approximations (H. Dock and Molegro Virtual Docker) [41,42]. After studying the interactions between FtsZ and FtsA, we used FtsA as the receptor and the C-terminus sequence of FtsZ as the ligand and looked for molecules that may inhibit the formation of the FtsZ and FtsA complex.

2. Results and Discussion

2.1. Consensus Sequence Determination of FtsZ and FtsA of S. aureus

A total of 428 sequences of the FtsZ protein from Staphylococcus aureus were retrieved from the NCBI nr protein database using the sequence of the structurally resolved FtsZ (PDB ID: 3VOA) as the reference in a BLASTp search. The BLOSUM62 scoring matrix was applied, and the search was restricted to Staphylococcus aureus species. After filtering out partial and incomplete sequences—except for those derived from PDB-reported structures— 304 sequences were retained for further analysis.

The multiple sequence alignment (MSA) of the 304 FtsZ sequences was performed using Clustal X server (Dublin, Irland) with default parameters, confirming the high conservation of the protein. Estimates of evolutionary divergence between sequences, expressed as the number of amino acid substitutions per site, are provided in a Supplementary Table. Standard errors, calculated using a bootstrap procedure with 500 replicates, are shown above the diagonal table (Table S1). The analysis, conducted using the Poisson correction model [43], involved 304 amino acid sequences, with all the ambiguous positions removed for each sequence pair (pairwise deletion option). A total of 401 positions were included in the final dataset. The evolutionary analyses were performed using MEGA11 [44].

The overall average of the evolutionary divergence between the sequences was 0.0072, indicating a low level of variation and emphasizing the conserved nature of FtsZ. However, despite the overall conservation, variations were observed mainly in sequences linked to PDB-reported structures, where sequence incompleteness may contribute to the observed differences. Most of these changes were located in the C-terminus region, although sporadic substitutions appeared across the full length of the protein.

A consensus sequence was generated from the MSA to identify the most conserved regions. This consensus sequence was subsequently used to model residues 316 to 339 of the FtsZ protein in the 3VOA structure. Although this region exhibited the highest degree of variability, the consensus sequence maintained over 92% identity at each position, providing a robust template for structural modeling.

A total of 1283 sequences of the FtsA protein from Staphylococcus aureus were retrieved from the NCBI nr protein database using the sequence of the structurally resolved FtsA (PDB ID: 3WT0) as the reference in a BLASTp search. The BLOSUM62 scoring matrix was applied, and the search was restricted to Staphylococcus aureus species. After filtering out truncated sequences and those with 100% redundancy, 936 sequences were retained for further analysis.

2.2. C-Terminus Homology Modeling of FtsZ from T. maritima and S. aureus

As mentioned above, the C-terminus region of all bacterial FtsZs is disordered, and the structures deposited in the PDB are not complete. Therefore, we built two models of this entire protein, one corresponding to S. aureus (Figure 1A) (our system of interest) and the other corresponding to T. maritima (our reference system) (Figure 1B). FtsZ corresponding to S. aureus comprises a disordered segment of approximately 50 aa before the C-terminus helix begins, while FtsZ from T. maritima has its C-terminus helix positioned immediately without the disordered fragment. However, it has been shown that for other bacterial microorganisms, such as E. coli [23], this disordered segment is present in their C-terminus domains, like in S. aureus’s C-terminus.

The FtsZ crystallographic structures of S. aureus deposited in PDB have 308 aa in their principal chain, that is, the structured part of FtsZ. As mentioned above, the missing part of the protein corresponding to its C-terminus disordered region was modeled using SWISS-MODEL. The FtsA and FtsZ complex of the T. maritima has been used as a reference to try to understand the interaction of FtsA and FtsZ of S. aureus in previous works [25]. Therefore, we decided to model the complete structure of FtsZ of T. maritima and thus have a reference system to compare our findings and the results for the S. aureus proteins.

2.3. Molecular Docking of the Complete Structure of FtsA with FtsZ

According to the literature, Gly47, Val236, His240, Asp244, Arg301, Glu304, and Lys308 residues, located in the H7 and H8 helices of the 2B subdomain of FtsA from T. maritima (PDB: 4a2a) [20], are important for the interaction with FtsZ. Hence, we carried out pointed mutations over the mentioned residues. To use this information in our system of interest, we studied the similarities between the FtsA of T. maritima and S. aureus. Their alignment shows an RMSD value of 1.032 for the H7 helix and 1.871 Å for the H8 helix (Figure S3). Therefore, we focused on the FtsZ residues, Asp372, Ile373, Pro374, Phe376, and Ile377, and we considered them important for the binding of the C-terminus segment of FtsZ with FtsA for the reasons previously described [28].

For the molecular docking of these proteins, we downloaded the PDB file of FtsA of each bacterium. To obtain the structure and complete sequence of FtsZ, we searched for the complete sequence in Uniprot (2023 Release). The molecular docking was performed with HDOCK by directing the binding sites of the two complete proteins, yielding complexes in which the C-terminus helix seems to fit on the H7 and H8 helices of FtsA. However, many other aa of the protein also interact with the helices of FtsA. Figure 2 shows the poses in which the C-terminus helix of FtsZ approaches the H7 and H8 helix of FtsA.

The molecular dockings of FtsA and FtsZ from S. aureus carried out with the HDOCK web server (http://hdock.phys.hust.edu.cn/, accessed on 1 October 2024). The results of these dockings, which we will call from now on Models 1–10, show that there is a possibility that the C-terminus helix of FtsZ binds over the H7 and H8 helices belonging to region 2B of FtsA. However, when analyzing the resulting complexes carefully, we found that an error occurs with the docking, since the structure of FtsZ results in a cut in its disordered region. Nonetheless, we decided to briefly analyze how the C-terminus helix of FtsZ interacts with FtsA. In model 1 (Figure 3), the interacting residues of FtsA are 230, 234, and 238 of helix H7; the interacting residues of helix H8 are 296 and 299. The C-terminus helix of FtsZ interacts with the residues just mentioned but does not involve any of the key residues in this union. The FtsA–FtsZ interaction energy value is −173.94 Kcal for this pose.

In Model 10 (Figure 4), the main part of the C-terminus region of FtsZ with its residues Pro374 and Phe376 binds with the top of the H7 helix of FtsA, interacting with residues 230, 234, and 238. The final part of the C-terminus helix of FtsZ binds with residues 296, 299, and 303 present in the H8 helix of FtsA. This union is executed giving a score of −156.39 kcal/mol. A part of the C-terminus region of the protein (before starting with the unstructured region) interacts with the top of the H7 and H8 helices of FtsA. Like the previous complex, this one was also cut off in the disordered part.

Due to the problems with the complexes resulting from the previous experiment, we decided to cut the FtsZ protein, keeping the final region of the disordered part of the protein (C-terminus helix), which goes from residue 368 to 390. Then, we performed a new molecular docking approximation, maintaining the FtsA structure from the PDB:4A2A crystal, and used the C-terminus helix of FtsZ as a ligand. In the following figures, we show two of the best results regarding the position that the C-terminus helix takes with respect to FtsA, according to our reference system.

Model 6 of this docking shows how the C-terminus helix adheres to the H7 helix. Asp371, which is part of the CTT region of FtsZ, is positioned next to Asp372 (key residue), interacting with the key residues of FtsA, 299 and 303, belonging to H8 helix. Some of the key aa of H7 that interact with FtsZ are 230, 234, and 238.

2.4. Homology Modeling of the Disordered Segment of FtsZ from T. maritima and S. aureus

After modeling the complete structures of FtsZ from both species, we decided to model only the final part of this domain, with its disordered segments (Figure 5) and the pure C-terminus helix of S. aureus (Figure 5A) and T. maritima (Figure 5B), which interacts with the 2B region of FtsA. We did this to find the best approximation to study the interaction between FtsA and FtsZ from T. maritima and then extrapolate it to S. aureus. We also decided to cut the C-terminus peptide of Ftsz from S. aureus, leaving the part where the key aa for such union are positioned (Figure 5A right side).

S. aureus’s C-terminus tail comprises a long and disordered sequence of aa and two main regions known as the C-terminus tail (CTT) and the C-terminus variable (CTV) [28]. The CTT region comprises 11 residues (T, K, E, D, D, I, P, S, F, I, and R), highly conserved among different strains and species of bacteria, and interacts with proteins, such as FtsA. On the other hand, the CTV region, which are 12 residues in size and follows CTT, is also crucial, as its mutations can disrupt FtsZ interactions with other modulatory proteins, as well as affect binding towards other FtsZ monomers [19,28]. Therefore, the generation of reliable homology models for this part of the FtsZ was crucial. We obtained the best results from two different servers (Figure 6): SWISS-MODEL and PepFold 3 (INSERM, Paris, France).

To model the C-terminus of FtsZ from S. aureus, we first used the 72 residues that compose the helix and the disordered part of the protein. The SWISS-MODEL and PepFold sites generated the final part of the disordered structure as a short α-helix, in agreement with experimental data previously reported for T. maritima and E. coli. We also modeled the 23 residues that form the CTT and CTV regions of the C-terminus.

Based on the results above, we cut the C-terminus helix part of FtsZ to use it as a ligand and perform another docking experiment to try to reproduce the FtsZ–FtsA complex of T. maritima [20] (Figure 7A). This was successful because model 1, which resulted from the docking, completely reproduced the conformation observed in the crystal structure (Figure 7B).

We started from that point, using this system as a reference, to compare our in silico experiments with the FtsA and FtsZ of S. aureus. From these docking calculations, the results from Model 6 obtained with Hdock were the most reliable, according to our system of reference (see Figure 8). The CTV part of FtsZ binds to some residues of the H7 and H8 helices of FtsA.

Model 10 (Figure 9) is the closest to our control system (Figure 3) in terms of the position taken by the C-terminus helix when it rests on the FtsA helices; its CTV region binds to the H7 helix, and part of its CTT region binds to the H8 helix of FtsA.

FtsA involves six key aa of the seven considered important in this binding: 230, 238, and 234 of the H7 helix, and 296, 299, and 303 of the H8 helix participate.

FtsZ (C-terminus helix) has F376 interacting with residue 241 of the H7 helix of FtsA. As in Model 1, the CTV part of FtsZ binds to some residues of the H7 and H8 helices of FtsA.

To further refine this work, we decided to cut the C-terminus helix even further. Of the 23 residues that make up the CTV and CTT segments, we cut residues Thr368 to Arg378, which correspond to the CTV region, where the key residues for the interaction between FtsA and FtsZ are located. We also ran molecular docking in Hdock. In the resulting model 2 (Figure 10), the 10-aa C-terminus helix mostly sticks to helix H7, but some of its residues contact helix H8. Asp372 and Pro374 of the C-terminus peptide interact with residues of helix H7; among them, Asp238, which is one of the key residues in FtsA that we are interested in interacting with, Arg296, and Glu299 of helix H8 contact residues at the end of the CTV.

In model 5 of this docking (Figure 11), the residues of the H7 helix of FtsA, Asp238, interact with residues Asp372 and Ile373 of the C-terminus helix. In addition, PRO374 and ILE377 of this same small structure interact with the Arg296 present in the H8 helix of FtsA. Glu299 of the H7 helix of FtsA has three interactions with the key residues Asp372, Ile373, and Pro374 (key residues of the C-terminus helix of FtsZ).

As can be shown, we have a better reproducibility of the experimental results of the binding of the C-terminal of FtsZ with FtsA of S. aureus when we considered a small part of the CTT. Nevertheless, considering the importance of including the 23 residues of the C-terminal of FtsZ for the molecular docking, especially to see the effect on the homology model after the alanine scanning, we evaluated the importance of each aa from the CTT to the binding with FtsA. For this, we used the model obtained from PepFold that contains the 23 aa (PepFold23) to evaluate its binding over FtsA of S. aureus, using Molegro Virtual Docker to contrast the results with the ones obtained by HDock. The resultant complexes obtained from the docking of FtsA, and the peptide modeled with PepFold23 resulted in negative binding energies (as expected), and the ligands bound to the H7 and H8 helices of the protein (

Table 1. Binding energies for the complex formed with PepFold23.

Complex	Mol Dock Score (kcal/mol)	Hydrogen Bonds
1	−480.939	−12.02
2	−480.822	−10.37
3	−446.585	−7.08

) (Figure 12).

We used the previously described poses as seeds (PepFold23s1, PepFold23s2, and PepFold23s3) to run a new docking experiment, in which we allowed some degree of flexibility to helices H7 and H8 from FtsA, where the key residues to the interaction with FtsZ, namely, Gly47, Val236, His240, Asp244, Arg301, Glu304, and Lys308 (PepFold23), are located.

2.5. Flexible Molecular Docking Between FtsA and PepFold23s1/PepFold23s2/PepFold23s3 Models

Figure 13 displays the first result of the flexible docking run with FtsA–PepFold23s1, FtsA–PepFold23s2, and FtsA–PepFold23s2 showing a negative binding energy (

Table 2. Interaction energies (Kcal/mol) between FtsA and the seedC-terminal helixes.

Complex	Mol Dock Score	Hydrogen Bonds
PF23s1	−619.002	−15.07
PF23s2	−609.84	−17.68
PF23s3	−628.598	−9.88

) bound to H7 and H8 and interacting more with the key residues (

Table 3. Key residues in FtsA and ways of interaction with the seed helixes. The "X" is used to denote the type of interactions that are taking place in each complex.

Complex	Residues of FtsA	Asn 43	Met 230	Asp 234	Asp 238	Arg 296	Glu 299	Glu 303
	Interaction Type
FtsA-PF23s1	Hydrogen bond	X			X	X		X
	Electrostatic interaction				X	X	X	X
FtsA-PF23s2	Hydrogen bond	X		X		X	X	X
	Electrostatic interaction			X
FtsA-PF23s3	Hydrogen bond				X	X
	Electrostatic interaction				X	X	X	X

).

2.6. Flexible Molecular Docking Between PepFold23s3 (Mutated) and FtsA

We performed in silico mutations on the five different C-terminus from the wild-type peptides in those residues considered important for binding with FtsA by substituting them with alanine. The main objective of these mutations was to study potential anomalies that might affect the peptides’ positions on the helices of FtsA and their binding energies. Specifically, we mutated the following residues to alanine: Asp372, Ile373, Pro374, Phe376, and Ile377. The names of the mutated peptide-terminal helixes are PF23-D372A, PF23-I373A, PF23-P374A, PF23-F376A, and PF23-I377A, respectively. The favored poses are displayed in Figure 14.

The energy values of each complex are shown in (

Table 4. Binding energies (Kcal/mol) between FtsA and the mutated C-terminal helixes.

Complex	Mol Dock Score	Hydrogen Bonds
PF23-D372A	−483.63	−13.0545
PF23-I373A	−594.24	−11.562
PF23-I374A	−574.44	−10.2531

) and the interacting residues, as well as the type of interaction they make are shown in (

Table 5. Important residues of FtsA and the type of interaction with the mutated C-terminal helixes. The “X” is used to denote the type of interactions that are taking place in each complex.

FtsA’s Complex	FtsA’s Residues	Asn 43	Met 230	Asp 234	Asp 238	Arg 296	Glu 299	Glu 303
	Interaction Type
PF23-D372A	HB	X				X	X	X
	Electrostatic			X		X	X	X
PF23-I373A	HB	X			X	X
	Electrostatic			X	X
PF23-P374A	HB			X		X		X
	Electrostatic			X		X		X

).

2.7. Identification of Protein–Protein Binding Inhibitors with the Potential to Disrupt FtsA and FtsZ Binding

Using the Pharmit web server (URL: https://pharmit.csb.pitt.edu/search.html, accessed on 1 October 2024), we searched for drug candidates that could interrupt the formation of the FtsA–FtsZ complex in S. aureus. We did this by introducing, in the PDB format, FtsA and the CTV region (Thr368-Arg378) of the C-terminus helix (10 aa) of FtsZ, which served as a template for the model generation of pharmacophore models (Figure 15).

Once the PDB file of our ligand was loaded (Figure 15), we “cleaned” the molecule of structural and spatial features, such as aromaticity, hydrogen bond donors, hydrogen bond acceptors, positive ions, negative ions, and hydrophobic AA. We did this with information from the two complexes resulting from the molecular docking between FtsA and the C-terminus helix of FtsZ (CTV region). In the end, to obtain a suitable molecule, we left the following characteristics shown in Figure 16 and Figure 17 for each complex (models 2 and 5 described in Section 2.7); these characteristics are arranged in the AA of interest shown in Figure 16.

Once the characteristics for each model were selected, we proceeded to search for pharmacophoric candidates in the multiple databases of bioactive compounds (ZINC, PubChem, CHEMBL34) [45]. Among the multiple pharmacophoric models generated focusing on this molecular docking model 2, Figure 17 shows the best candidates for pharmacophoric models, chosen based on virtual screening for that position acquired in the docking.

Model 5 (Figure 18) of this same group of results of docking FtsA molecules with the C-terminal CTT segment of FtsZ (10 residues), also showed an interesting conformation, since the small helix placed some of its key aa (D372, I373, P374 and I377) on key residues of the H7 and H8 helices of FtsA (R296 and E299).

In Figure 19 we show the pharmacophoric candidates obtained after performing the search in the different databases based on the coordinates mentioned in the previous figure.

3. Discussion

The literature reports that the C-terminus of the protein FtsZ is made by a disordered sequence of amino acids. This region, which is not solvable by crystallography, has two fragments, known as the CTT and CTV, believed to form a short α-helix. CCT comprises 11 residues highly conserved among different bacteria species and the CTV fragment [19].

We modeled a 23-residue peptide formed by these regions and obtained three similar conformations all with a α-helix structure.

We ran rigid docking with these modeled peptides and found that only one of the models (PepFold23) bound to the key amino acids of FtsA. Then, we ran flexible docking using the best three poses that scored the more favorable results from the rigid docking, since the wild C-terminus peptides interacted with five out of seven FtsA residues, and the binding energies were lower than those of the rigid docking.

Nonetheless, in two out of the five experiments (PF23-F376A y PF23-I377A), the C-terminus peptide bound in a non-favorable fashion, and the mutations led to a change in the conformation and hence to the instability of the peptide. In general, the results show a correlation with the experimental data previously reported in the literature, which sustains the key role of the studied amino acids from which Arg296 is involved in the rigid and flexible dockings, hence, we can consider it a key residue to the binding of FtsA towards FtsZ.

4. Computational Methods

4.1. FtsZ and FtsA Structure Selection of Staphylococcus aureus

Among the available crystals of FtsA at the protein data bank (PDB), we chose the 3WT0 structure (2.00 Å) [26] because the crystallographic structure is complete; that is, unlike the others, it does not present cuts or mutations. Similarly, we chose the crystal structure of FtsZ of S. aureus (PDB:3VOA) to have the structure with the most information [14]. We carried out a structural analysis of the results with Chimera version 1.13.1 (NIH, Bethesda, MA, USA) [46].

4.2. Consensus Sequence Generation of FtsZ and FtsA in Staphylococcus aureus

To generate the consensus sequence of FtsZ from Staphylococcus aureus, the protein sequence of chain 1 of FtsZ from PDB file 3VOA was used as a reference. A BLASTp search was performed against the NCBI non-redundant (nr) protein sequence database, filtering out sequences exclusively from Staphylococcus aureus. This search yielded a total of 428 sequences [43]. Incomplete sequences were excluded from the dataset, except for those corresponding to proteins with structures reported in the PDB database. In addition, two sequences that were misannotated as S. aureus were removed. A total of 304 sequences were retained.

Similarly, for FtsA, the protein sequence of chain A from PDB file 3WT0 was used as a reference. A BLASTp search applying the same filtering criteria yielded 1283 sequences. After removing truncated sequences and those with 100% redundancy, 936 sequences were retained for further analysis.

Multiple sequence alignments were carried out for FtsZ and FtsA using ClustalX (version 2.1, European Bioinformatics Institute, Hinxton, UK) and MUSCLE (Wellcome Genome Campus, Cambridge, UK). Consensus sequences were generated from the respective alignments and used for further analysis. Finally, the amino acids were colored according to the Clustal color code to highlight conserved regions.

4.3. Homology Modeling of C-Terminus of FtsZ of S. aureus

For this study, we needed the complete structure of FtsZ, but this is not available experimentally because of its disordered C-terminus. Therefore, to build the homology model of the missing C-terminus inFtsZ (3VOA), we used the FASTA file of the protein “P0A030” and the one obtained from our consensus sequence. For the construction of these structures, we used the SWISSMODEL server [47]. To study the binding of this protein, we constructed three types of homology models: a complete one, only the disordered C-terminus (residues from 316 to 390), and the final sequence of 23 aa related directly to the binding (residues from 368 to 390, according to the experimental data), which give rise to a small helix at the end of the disordered segment. These aa make the following regions: C-terminus variable (CTV) and C-terminus tail (CTT). Some of the CTT residues are highly conserved among different bacterial species. For the two final models, we modeled the disordered fragment using SWISSMODEL and Pep-Fold.

4.4. FtsA and FtsZ Molecular Docking

The molecular docking experiments between FtsA and FtsZ were conducted using the HDOCK server (Huazhong University of Science and Technology, Wuhan, China) [41] and Molegro Virtual Docker (Molexus ApS, Odder, Denmark) [42]. The molecular dockings between FtsA and the three versions of FtsZ, namely, the full protein, only the complete C-terminus, and the isolated helix, were carried out in a directed manner on HDOCK, thus specifying the binding site residues in each protein. We ran the docking calculations of the small section of FtsZ with FtsA of S. aureus on Molegro Virtual Docker [42].

4.5. Mutation of C-Terminus Fragments of FtsZ’s Models

We ran docking experiments with Molegro Virtual Docker (MVD) and HDOCK to study the effects of mutations [20] on the binding between FtsZ and FtsA. The C-terminus fragment from FtsZ comprises a sequence of 23 residues (from 368 to 390), TKEDDIPSFIRNREERRSRRTRR, which are conserved among different bacteria species [19]. We carried out mutations on this fragment and exchanged the residues from 372 to 377 to alanine (Figure 20).

4.6. Molecular Docking of the Wild and Mutated Peptides FtsZ to FtsA

The literature reports that the FtsZ of E. coli, B. subtilis, and T. maritima has a short peptide in the C-terminus, which binds to the subdomain 2B of FtsA protein [20,44]. FtsA anchors to the cell membrane to form the Z ring formed by polymers of FtsZ. The C-terminus fragments modeled have few atoms, hence, we considered them the ligands and ran docking experiments with HDOCK and MVD to bind the C-terminus fragments and peptides to the FtsA (3WT0). According to what is reported in the literature, the helix H7 and H8 (Figure 21) of the SaFtA receptor protein were considered the region where the binding site is located. This was conducted following experimental reports where mutations in residues within this region of Thermotoga maritima FtsA (TmFtsA) affected its interaction with Thermotoga maritima FtsZ (TmFtsZ). These proteins share structural and sequence similarities with FtsZ and FtsA of S. aureus with an RMSD of 1.9A and 1.4A, respectively.

For the docking of FtsA with FtsZ, for the complete structure plus the disordered segment, the disordered segment plus the helix, and for the helix alone, the residues specified for the receptor protein binding site (FtsA) were Met230-Leu243 and Ser278-Gly312, and those of the FtsZ protein were Thr367-Arg390.

To run the docking calculations in MVD, we used the following settings: for the rigid docking, we used the MolDock SE, 10 runs, 1500 iterations, and a maximum population size of 50 individuals. We used the MolDock Score GRID as the scoring function, using a grid of 0.3 and a search radius of 20 Å. We used the best result—henceforth referred to as the pose—which was chosen based on the binding energy and position of the ligand (helix). This result was used as a seed pose to run the flexible docking approach. We ran the flexible docking approach using these settings: 2000 local iterations and 2000 global iterations using a 20 Å radius and the same settings described above for the search and the scoring function. However, we did conduct an extra step and aligned the structures of the H7 and H8 helices from TmFtsA and FtsA.

To visualize and analyze the results, we used MVD [42], and USCF Chimera 1.13.1 [46]. To validate our results, we decided to run molecular docking with the FtsZ and FtsA proteins of the microorganism Thermotoga maritima using the same procedure that we performed with FtsA and FtsZ of S. aureus. That is, we used the T. maritima system to define the parameters and procedures to reproduce the experimental data by means of computational approximations.

4.7. Analyses of the C-Terminus Structure as the Pharmacophore to Find Protein–Protein Binding Inhibitors

For the pharmacophore model generation, we used the best pose obtained from the molecular docking between the FtsZ C-terminus segment and the FtsA of S. aureus. This selection was based on the similarity upon the experimental data, like the key residues for the binding of FtsZ and the location of the binding site in FtsA. Additionally, for the pose selection, the shape and orientation of the C-terminus segment of FtsZ were considered based on those of T. maritima.

Different pharmacophore models were developed considering the different pharmacophore features of the key residues. Finally, the selection of the best model was based on the results gleaned from the virtual screening.

5. Conclusions

In this study, we carried out docking calculations with the FtsZ protein from Staphylococcus aureus to understand how it binds to itself and compared our results to the crystal structure reported in the literature [13].

Our results were similar to those reported in the experimental reference, in which two subunits of FtsZ interact through head–tail binding, where the T7 loop binds towards the DNA-binding domain of the lower subunit. As our results reproduced this binding mode, we consider that they are reliable, since they show a close correlation with the experimental data.

The modeling of the peptides resulted in structures in an α-helix form, which is an advantage for the design of novel inhibitors.

We identified four residues that could be key to the interaction between FtsA and the C-terminus peptide FtsZ: -Asp234, Arg296, Glu299, and Glu303, conserved among S. aureus and T. maritima. Two of these residues, Ar296 and Glu299, are involved in all the rigid and flexible docking, which we identified as the relevant residues to the interaction with the C-terminus tail of FtsZ.

With this study, we identified two potential PPBIs of the FtsZ and FtsA complex formation. Nevertheless, further studies are needed to validate these results.