Molecular Docking and Screening Studies of New Natural Sortase A Inhibitors

To date, multi-drug resistant bacteria represent an increasing health threat, with a high impact on mortality, morbidity, and health costs on a global scale. The ability of bacteria to rapidly and permanently acquire new virulence factors and drug-resistance elements requires the development of new antimicrobial agents and selection of new proper targets, such as sortase A. This specific bacterial target plays an important role in the virulence of many Gram-positive pathogens, and its inhibition should produce a mild evolutionary pressure which will not favor the development of resistance. A primary screening using a fluorescence resonance energy transfer assay was used to experimentally evaluate the inhibitory activity of several compounds on sortase A. Using molecular docking and structure-activity relationship analyses, several lead inhibitors were identified, which were further tested for antimicrobial activity using the well diffusion test and minimum inhibitory concentration. The toxicity was assessed using the Daphnia magna test and used as a future screening filter. Three natural compounds were identified in this study as promising candidates for further development into therapeutically useful anti-infective agents that could be used to treat infections caused by multi-drug resistant bacterial pathogens which include sortase A in their enzymatic set.


Introduction
The multitude of infections caused by multi-drug resistant (MDR) bacteria has determined the need to discover or develop new ways to combat these microorganisms. Many bacterial species are able to adapt to the environment in which antibacterial agents are present using genetic mechanisms, mutations, or selection. Among the Gram-positive organisms, the most common are Staphylococcus sp. [1], Enterococcus sp. [2], and Streptococcus sp. [3], and from the Gram-negative class, we mention Acinetobacter, Pseudomonas, and various Enterobacteriaceae (including Klebsiella, E. coli, Serratia, and Proteus) [4].
Some strains of Enterococcus sp. and Staphylococcus aureus (S. aureus) have evolved to become highly aggressive nosocomial pathogens and are responsible for a wide range of infections such as endocarditis, urinary and biliary tract infections, bacteremia, meningitis, and wound infections [5,6]. The main issue surrounding these bacterial pathogens is that they often develop resistance to antibiotic drugs that target bacterial growth or viability. New strategies for developing the next-generation of antimicrobial therapeutics are being elaborated, one of which specifically targets virulence pathways that are non-essential for growth [7]. Unlike antibacterials which inhibit cell growth, anti-virulence therapies present a number of advantages since they produce a mild evolutionary pressure which will not favor the development of resistance, provide an increased repertoire of pharmacological targets, and generate agents with new mechanisms of action [8]. The possibility of numerous anti-virulence approaches is offered by the many virulence factors of S. aureus [9].
Among the factors regulating several central processes of bacterial biology (such as adhesion, colonization, and biofilm), are the transpeptidase enzymes known as sortases [10]. These enzymes catalyze a ligation reaction between a five amino-acid substrate motif (LPXTG) and oligoglycine nucleophiles [11] and play a significant role in the covalent attachment of surface proteins to the cell wall of bacteria [10].
The most studied sortase is sortase A (SrtA) from S. aureus due to the fact that this bacterium has increased pathogenicity and multi-drug resistance [12]. Its role in the virulence of many Gram-positive pathogens, including staphylococci, streptococci, enterococci, and Listeria monocytogenes, is proved by the fact that mutants lacking genes for SrtA display attenuated virulence, without affecting the growth of the bacteria [13]. The development and also the search for new inhibitors of SrtA use the S. aureus enzyme as a prototype.
An important source of bioactive molecules is represented by plants, and thus, they are being explored for the discovery and development of novel anti-virulence agents [14]. Of the many natural compounds that have been tested, flavonoids [15] and coumarins [16] are the most used for their antibacterial properties, but there are also anthraquinones [17] and alkaloids [18] with such properties. Due to their potential and accessibility, these compounds continue to be important sources of structural scaffolds for lead generation.
Even if the antibacterial activity is known for many of these compounds, their implications in anti-virulence is still uncertain. In this research, we continue our primary study of SrtA inhibitors [19] focusing on finding new inhibitors from natural sources. Molecular docking and structure-activity relationship analyses were performed to better understand the mechanism of inhibition and to reveal essential structural features. Antimicrobial activity was tested using the well diffusion test and the minimum inhibitory concentration (MIC) method, in order to find candidates with a minimum impact on bacterial cell growth. The toxicity was assessed using the Daphnia magna test as a prescreening method, thus considerably reducing the number of mammals required for future toxicity testing.

Sortase A Activity Assay
A total of 56 natural compounds were chosen for a primary screening on SrtA activity based on their molecular diversity and on a set of rules derived from our previous research [19]. The candidates were small molecules with a low molecular flexibility and a minimum of two hydrogen bond acceptors. Chemically, 19 compounds belonged to the flavonoid family, six were coumarins, five were derivatives of cinnamic acid, five were alkaloids, three were anthraquinones, six belonged to the terpene family, and the rest had various scaffolds. Of the 56 natural compounds assayed, 22 compounds produced a significant inhibitory effect, of which seven compounds demonstrated good inhibitory effects on SrtA activity. The SrtA inhibitory effect was not correlated with the chemical family.
Based on the preliminary screening, a series of 11 compounds were selected to be tested at various concentrations in order to evaluate the inhibitory activity. The inhibitory activity was defined as the concentration of compound causing a 50% decrease in SrtA activity (IC50), relative to the negative control. The results are presented in Table 1. All the other compounds had no significant effect on SrtA at 10 µM. The IC50 value was only calculated for three compounds because the 50% inhibition threshold was not obtained for the others; the limited solubility in water hindered the use of the fluorescence resonance energy transfer (FRET) method. Myricetin showed the most potent activity on SrtA, its IC50 being eight-fold lower than esculetin, and eleven-fold lower than palmatine. For these three compounds, a good correlation between the concentrations and the inhibitory effect ( Figure 1) on all moments of determination was observed (r 2 > 0.9). Piperine, quercetin, and rhein demonstrated promising inhibitory effects at 10 µM, but their low solubility in water limited the testing at higher concentrations. The IC50 value was only calculated for three compounds because the 50% inhibition threshold was not obtained for the others; the limited solubility in water hindered the use of the fluorescence resonance energy transfer (FRET) method. Myricetin showed the most potent activity on SrtA, its IC50 being eight-fold lower than esculetin, and eleven-fold lower than palmatine. For these three compounds, a good correlation between the concentrations and the inhibitory effect ( Figure 1) on all moments of determination was observed (r 2 > 0.9). Piperine, quercetin, and rhein demonstrated promising inhibitory effects at 10 μM, but their low solubility in water limited the testing at higher concentrations.

Protein-Ligand Interactions
A computational docking algorithm was used to predict the relative binding affinities for a series of 16 inhibitors and to allow us to observe the structure-inhibitory action relationships. Among the tested compounds, genistein showed the lowest estimated free binding energy, and should thus have theoretically exhibited the highest transpeptidase inhibition activity. However, the SrtA inhibition assay demonstrated that genistein is inactive. The lack of correlation between in silico and in vitro results can be explained by analyzing genistein's interaction with the enzyme. It was previously stated that the ionizable residues GLU105, SER116, THR180, THR183, ASP185, TRY187, LYS196, and ARG197 are situated proximal to CYS184 in the active site and thus facilitate transpeptidase enzymatic activity [20]. Genistein doesn't participate in any hydrogen-bond interactions with the aforementioned residues, Even though genistein manifests a high binding affinity for SrtA and is involved in van der Waals and π-σ interactions with ARG197 and THR180, respectively (Figure 2), the interactions formed in the binding site do not seem to meet the requirements for inhibiting the transpeptidase activity.

Protein-Ligand Interactions
A computational docking algorithm was used to predict the relative binding affinities for a series of 16 inhibitors and to allow us to observe the structure-inhibitory action relationships. Among the tested compounds, genistein showed the lowest estimated free binding energy, and should thus have theoretically exhibited the highest transpeptidase inhibition activity. However, the SrtA inhibition assay demonstrated that genistein is inactive. The lack of correlation between in silico and in vitro results can be explained by analyzing genistein's interaction with the enzyme. It was previously stated that the ionizable residues GLU105, SER116, THR180, THR183, ASP185, TRY187, LYS196, and ARG197 are situated proximal to CYS184 in the active site and thus facilitate transpeptidase enzymatic activity [20]. Genistein doesn't participate in any hydrogen-bond interactions with the aforementioned residues, Even though genistein manifests a high binding affinity for SrtA and is involved in van der Waals and π-σ interactions with ARG197 and THR180, respectively (Figure 2), the interactions formed in the binding site do not seem to meet the requirements for inhibiting the transpeptidase activity. On the other hand, myricetin exhibited both high inhibition activity and high binding affinity for SrtA, most likely due to its ability to form hydrogen bonds with two of the polar residues, GLU105 and ARG197. Moreover, myricetin also binds to GLY167, a residue which has been reported to play a major role in the SrtA inhibition activity of quercitrin [21]. Myricetin interacts through van der Waals forces with SER116, THR180, and other residues. The protein-ligand complex is stabilized by eight hydrophobic interacting residues and four types of π interactions ( Figure 3). Furthermore, the docked conformation of myricetin is likely to hinder the interaction between the enzyme and its substrate, thus inhibiting SrtA activity in an orthosteric fashion. Myricetin and genistein have similar chemical structures, but there are two significant differences, the position of the phenyl substitution on the chromen-4-one scaffold and the fact that myricetin contains two additional hydroxyl groups, enabling its orientation in a more efficient manner in the SrtA binding site. On the other hand, myricetin exhibited both high inhibition activity and high binding affinity for SrtA, most likely due to its ability to form hydrogen bonds with two of the polar residues, GLU105 and ARG197. Moreover, myricetin also binds to GLY167, a residue which has been reported to play a major role in the SrtA inhibition activity of quercitrin [21]. Myricetin interacts through van der Waals forces with SER116, THR180, and other residues. The protein-ligand complex is stabilized by eight hydrophobic interacting residues and four types of π interactions ( Figure 3). Furthermore, the docked conformation of myricetin is likely to hinder the interaction between the enzyme and its substrate, thus inhibiting SrtA activity in an orthosteric fashion. Myricetin and genistein have similar chemical structures, but there are two significant differences, the position of the phenyl substitution on the chromen-4-one scaffold and the fact that myricetin contains two additional hydroxyl groups, enabling its orientation in a more efficient manner in the SrtA binding site. Esculetin and palmatine chloride showed an inhibitory activity of SrtA six-fold and 10-fold lower than myricetin. Their binding affinities were also lower than the value predicted for myricetin. Like myricetin, esculetin can form conventional hydrogen bonds with ARG197 and a π-donor hydrogen bond with VAL168, which contribute to its ability to inhibit SrtA activity. Other interacting forces that can occur for both myricetin and esculetin are van der Waals interactions with THR180 and π-σ interactions with ILE199 ( Figure 4).  Esculetin and palmatine chloride showed an inhibitory activity of SrtA six-fold and 10-fold lower than myricetin. Their binding affinities were also lower than the value predicted for myricetin. Like myricetin, esculetin can form conventional hydrogen bonds with ARG197 and a π-donor hydrogen bond with VAL168, which contribute to its ability to inhibit SrtA activity. Other interacting forces that can occur for both myricetin and esculetin are van der Waals interactions with THR180 and π-σ interactions with ILE199 ( Figure 4). Esculetin and palmatine chloride showed an inhibitory activity of SrtA six-fold and 10-fold lower than myricetin. Their binding affinities were also lower than the value predicted for myricetin. Like myricetin, esculetin can form conventional hydrogen bonds with ARG197 and a π-donor hydrogen bond with VAL168, which contribute to its ability to inhibit SrtA activity. Other interacting forces that can occur for both myricetin and esculetin are van der Waals interactions with THR180 and π-σ interactions with ILE199 ( Figure 4).  Palmatine chloride has the highest binding energy amongst the active compounds, most likely due to its lack of hydrogen bond donor groups. Nonetheless, it can form a hydrogen bond with the ionizable residue TYR187 and participates in van der Waals interactions with GLU105, ARG197, and the catalytic residue, CYS184. The docked conformation is stabilized by alkyl and π-alkyl interactions and nine hydrophobic residues ( Figure 5). Therefore, palmatine chloride can prevent the substrate from binding to the SrtA catalytic site. Palmatine chloride has the highest binding energy amongst the active compounds, most likely due to its lack of hydrogen bond donor groups. Nonetheless, it can form a hydrogen bond with the ionizable residue TYR187 and participates in van der Waals interactions with GLU105, ARG197, and the catalytic residue, CYS184. The docked conformation is stabilized by alkyl and π-alkyl interactions and nine hydrophobic residues ( Figure 5). Therefore, palmatine chloride can prevent the substrate from binding to the SrtA catalytic site. Although both rutin and troxerutin had binding affinities comparable to myricetin, their in vitro inhibition activity of SrtA was rather mediocre. Unlike myricetin, rutin and troxerutin are glycosylated and thus have a larger contact surface with the protein. Rutin can participate in hydrogen bonding with ARG197 and troxerutin can form the same bond type with ARG197 and GLU105. Both ligands interact with the enzyme through a large number of van der Waals and hydrophobic interactions and are likely to block the active site of the enzyme (Figures 6 and 7). Possible reasons why these glycosylated flavonoids showed lower SrtA inhibition activities are that the protein-ligand complex might be prone to destabilization or the inability to enter the enzyme's active site pocket due to their large molecular volume or high electrostatic repulsion. Nevertheless, rutin can be transformed in vivo by enzymatic deglycosylation into quercetin, which has been proven to inhibit SrtA enzymatic activity [22].
Even though esculetin and umbelliferone have almost identical chemical structures and similar binding affinities, umbelliferone didn't exhibit significant SrtA inhibition activity. Also, umbelliferone doesn't participate in any hydrogen-bond interactions with the relevant ionizable residues, yet THR180 and SER116 are involved in π-σ and van der Waals interactions, respectively ( Figure 8). Most likely, umbelliferone was unable to inhibit SrtA activity due to its lack of a second hydroxyl group on the aromatic ring. Although both rutin and troxerutin had binding affinities comparable to myricetin, their in vitro inhibition activity of SrtA was rather mediocre. Unlike myricetin, rutin and troxerutin are glycosylated and thus have a larger contact surface with the protein. Rutin can participate in hydrogen bonding with ARG197 and troxerutin can form the same bond type with ARG197 and GLU105. Both ligands interact with the enzyme through a large number of van der Waals and hydrophobic interactions and are likely to block the active site of the enzyme (Figures 6 and 7). Possible reasons why these glycosylated flavonoids showed lower SrtA inhibition activities are that the protein-ligand complex might be prone to destabilization or the inability to enter the enzyme's active site pocket due to their large molecular volume or high electrostatic repulsion. Nevertheless, rutin can be transformed in vivo by enzymatic deglycosylation into quercetin, which has been proven to inhibit SrtA enzymatic activity [22].
Even though esculetin and umbelliferone have almost identical chemical structures and similar binding affinities, umbelliferone didn't exhibit significant SrtA inhibition activity. Also, umbelliferone doesn't participate in any hydrogen-bond interactions with the relevant ionizable residues, yet THR180 and SER116 are involved in π-σ and van der Waals interactions, respectively (Figure 8). Most likely, umbelliferone was unable to inhibit SrtA activity due to its lack of a second hydroxyl group on the aromatic ring.

Antibacterial Activity
Our molecular docking studies and the FRET-based assay evaluating the inhibitory effect on SrtA, indicated that the lead compounds should be further tested for antimicrobial activity. The antimicrobial activity of compounds was firstly screened by the well diffusion test and then subjected to the MIC method. The compounds were selected for the preliminary antibacterial screening if they produced a statistically significant SrtA inhibition, and are as following: artemisinin, chrysin, emodin, esculetin, genistein, myricetin, naringin, palmatine chloride, piperine, podophyllotoxin, quercetin, quinic acid, rhein, rutin trihydrate, troxerutin, and umbelliferone. Only emodin and rhein, two 1,8-dihydroxyanthraquinone derivatives, produced an observable antimicrobial effect on the tested bacteria strains. Table 2 shows the obtained values of the diameter inhibition zone for emodin and rhein. Of the tested compounds, emodin and rhein exerted a potent inhibitory effect on Gram-positive S. aureus ATCC 25923, S. epidermidis ATCC 12228, and B. cereus ATCC 11778. None of the compounds exhibited antimicrobial activity against Gram-negative bacteria. S. epidermidis was found to be the most sensitive to the antimicrobial activity of rhein, while emodin had a significant effect on S. aureus ATCC 25923. Both compounds exerted moderate antimicrobial activity against B. cereus ATCC 11778.

Antibacterial Activity
Our molecular docking studies and the FRET-based assay evaluating the inhibitory effect on SrtA, indicated that the lead compounds should be further tested for antimicrobial activity. The antimicrobial activity of compounds was firstly screened by the well diffusion test and then subjected to the MIC method. The compounds were selected for the preliminary antibacterial screening if they produced a statistically significant SrtA inhibition, and are as following: artemisinin, chrysin, emodin, esculetin, genistein, myricetin, naringin, palmatine chloride, piperine, podophyllotoxin, quercetin, quinic acid, rhein, rutin trihydrate, troxerutin, and umbelliferone. Only emodin and rhein, two 1,8-dihydroxyanthraquinone derivatives, produced an observable antimicrobial effect on the tested bacteria strains. Table 2 shows the obtained values of the diameter inhibition zone for emodin and rhein. Of the tested compounds, emodin and rhein exerted a potent inhibitory effect on Gram-positive S. aureus ATCC 25923, S. epidermidis ATCC 12228, and B. cereus ATCC 11778. None of the compounds exhibited antimicrobial activity against Gram-negative bacteria. S. epidermidis was found to be the most sensitive to the antimicrobial activity of rhein, while emodin had a significant effect on S. aureus ATCC 25923. Both compounds exerted moderate antimicrobial activity against B. cereus ATCC 11778.
Based on the preliminary assay, only emodin and rhein produced a significant effect to be tested for MIC determination. MIC values of emodin and rhein against the strains selected were obtained by well diffusion test screening (Table 3). It was noted that emodin had more inhibitory action against all bacterial strains than rhein. Even if the study on the bacterial growth influence was performed as a filter for a selective anti-virulence effect, the results of the antibacterial properties of the 1,8-dihydroxyanthraquinone derivatives could be developed.

Acute Toxicity Assessment Using Daphnia Magna
The majority of the compounds did not induce significant toxicity on Daphnia magna. Myricetin, palmatine chloride, naringin, rutin, esculetin, quercetin, and troxerutin induced a maximum of 30% lethality after 48 h. Rhein, arthemisin, and genistein induced a maximum toxicity between 40% and 60% at concentrations of 50 and 100 µM, respectively, but the results weren't dose dependent and therefore median lethal concentration (LC50) couldn't be calculated. A concentration-lethality dependence was obtained for piperine, chrysin, emodin, podophyllotoxin, and umbelliferone. For these compounds, the calculated LC50 values are presented in Table 4 and the lethality curves in Figure 9.
The highest toxicity on the crustacean model was produced by piperine, followed by podophyllotoxin, chrysin, emodin, and umbelliferone. The positive control, potassium dichromate, induced 100% toxicity after 48 h of exposure at concentrations in the 1.3-34 µM range. The calculated LC50 values after 24 h and 48 h demonstrated that all compounds produced a lower toxicity than the positive control. Based on the preliminary assay, only emodin and rhein produced a significant effect to be tested for MIC determination. MIC values of emodin and rhein against the strains selected were obtained by well diffusion test screening (Table 3). It was noted that emodin had more inhibitory action against all bacterial strains than rhein. Even if the study on the bacterial growth influence was performed as a filter for a selective anti-virulence effect, the results of the antibacterial properties of the 1,8dihydroxyanthraquinone derivatives could be developed. Table 3. MIC values of the tested compounds (μg/mL).

Acute Toxicity Assessment Using Daphnia Magna
The majority of the compounds did not induce significant toxicity on Daphnia magna. Myricetin, palmatine chloride, naringin, rutin, esculetin, quercetin, and troxerutin induced a maximum of 30% lethality after 48 h. Rhein, arthemisin, and genistein induced a maximum toxicity between 40% and 60% at concentrations of 50 and 100 μM, respectively, but the results weren't dose dependent and therefore median lethal concentration (LC50) couldn't be calculated. A concentration-lethality dependence was obtained for piperine, chrysin, emodin, podophyllotoxin, and umbelliferone. For these compounds, the calculated LC50 values are presented in Table 4 and the lethality curves in Figure 9.
The highest toxicity on the crustacean model was produced by piperine, followed by podophyllotoxin, chrysin, emodin, and umbelliferone. The positive control, potassium dichromate, induced 100% toxicity after 48 h of exposure at concentrations in the 1.3-34 μM range. The calculated LC50 values after 24 h and 48 h demonstrated that all compounds produced a lower toxicity than the positive control.

Discussion
A chemically diverse set of 56 natural compounds were tested for their ability to inhibit SrtA activity, as potential anti-virulence agents. Myricetin, palmatine, and esculetin were identified as SrtA inhibitors, with promising IC50 values. Comparing these values with various known SrtA inhibitors, the compounds are in the first quartile [19]. Only a series of 2-phenylpyridazin-3-one derivatives are described in the literature, with IC50 values lower than myricetin [23].
The small number of hits hindered the obtaining of clear structure-activity relationships (SAR), but some structural patterns emerged as leads for future screening.
Myricetin, like quercetin and its 3-O-glycosides, hyperoside and rutin, belongs to the class of flavonols, all of which share a 3-hydroxyflavone backbone. Based on the in vitro study and on the protein-ligand interactions, the number and position of hydroxyl groups on the flavonol scaffold proved to be essential for SrtA inhibition. The IC50 value for quercetin was not calculated due to its very low solubility in water. The SAR study of the flavonoid group was therefore limited. The results of both the enzymatic assay and molecular docking indicates that the flavonoid aglycones are more potent towards SrtA than the corresponding glycosides. This observation is in concordance with our previous findings on SrtA inhibitor structural features, showing that the number of rotatable bonds should be less than or equal to 4 [19].
Esculetin is the lactone derivative formed by intramolecular cyclization of the caffeic acid, a derivative of cinnamic acid. It demonstrated good SrtA inhibition, while the caffeic acid and its methoxy-derivative, trans-ferulic acid, proved to be inactive. The results of the docking studies revealed the cause of this difference, the cis/trans conformation, and the importance of the hydrogen bonds between the carbonyl group and ARG197. The significance of the 6,7-dihydroxy-substitution on the chromen-2-one scaffold was demonstrated by the small inhibitory effect produced by both umbelliferone and coumarin. Esculin is esculetin 6-O-glucoside and also produced low SrtA inhibition because it lacks one essential hydrogen bonding donor, the 6-hydroxy group.
Palmatine chloride is a protoberberine alkaloid and has a low chemical similarity with all the compounds tested in this study. The results of the docking indicate an interaction with the catalytic center of the enzyme, but further research is needed to better understand this mechanism.

Discussion
A chemically diverse set of 56 natural compounds were tested for their ability to inhibit SrtA activity, as potential anti-virulence agents. Myricetin, palmatine, and esculetin were identified as SrtA inhibitors, with promising IC50 values. Comparing these values with various known SrtA inhibitors, the compounds are in the first quartile [19]. Only a series of 2-phenylpyridazin-3-one derivatives are described in the literature, with IC50 values lower than myricetin [23].
The small number of hits hindered the obtaining of clear structure-activity relationships (SAR), but some structural patterns emerged as leads for future screening.
Myricetin, like quercetin and its 3-O-glycosides, hyperoside and rutin, belongs to the class of flavonols, all of which share a 3-hydroxyflavone backbone. Based on the in vitro study and on the protein-ligand interactions, the number and position of hydroxyl groups on the flavonol scaffold proved to be essential for SrtA inhibition. The IC50 value for quercetin was not calculated due to its very low solubility in water. The SAR study of the flavonoid group was therefore limited. The results of both the enzymatic assay and molecular docking indicates that the flavonoid aglycones are more potent towards SrtA than the corresponding glycosides. This observation is in concordance with our previous findings on SrtA inhibitor structural features, showing that the number of rotatable bonds should be less than or equal to 4 [19].
Esculetin is the lactone derivative formed by intramolecular cyclization of the caffeic acid, a derivative of cinnamic acid. It demonstrated good SrtA inhibition, while the caffeic acid and its methoxy-derivative, trans-ferulic acid, proved to be inactive. The results of the docking studies revealed the cause of this difference, the cis/trans conformation, and the importance of the hydrogen bonds between the carbonyl group and ARG197. The significance of the 6,7-dihydroxy-substitution on the chromen-2-one scaffold was demonstrated by the small inhibitory effect produced by both umbelliferone and coumarin. Esculin is esculetin 6-O-glucoside and also produced low SrtA inhibition because it lacks one essential hydrogen bonding donor, the 6-hydroxy group.
Palmatine chloride is a protoberberine alkaloid and has a low chemical similarity with all the compounds tested in this study. The results of the docking indicate an interaction with the catalytic center of the enzyme, but further research is needed to better understand this mechanism.
Myricetin, palmatine chloride, esculetin, rutin, and troxerutin, presented the best inhibitory effect on SrtA and showed a low toxicity on Daphnia magna. The highest lethality induced by these compounds was 30% after 48 h of exposure at 100 µM, a value significantly higher than the IC50 values calculated for SrtA inhibition. The selected compounds also have the advantage of not influencing the Gram-positive bacteria growth, and therefore exhibit a low risk of inducing bacterial resistance.
Piperine and podophyllotoxin demonstrated promising SrtA inhibitions, but their low aqueous solubility and the high level of toxicity produced in the Daphnia test impedes their development as anti-virulence agents. Still, future molecular docking studies and SARs could reveal interesting new compounds.

Protein and Ligand Preparation
The three-dimensional structure file of the S. aureus SrtA was downloaded from the RCSB PDB database and was prepared by adding missing polar hydrogen atoms using the Protoss tool provided by the Proteins Plus server [24]. All water molecules were removed and Gasteiger charges were assigned to the protein structure.

Protein-Ligand Interaction Prediction
Molecular docking was performed using the AutoDock Tools 1.5.6 graphical user interface, which runs AutoDock Vina 1.1.2 software (Scripps Research Institute, San Diego, CA, USA). Autodock Vina is a docking platform which is faster and more accurate than the previous docking software, AutoDock 4 [25]. This newer version generates nine conformations (poses) of the protein-ligand complex which are displayed from lowest to highest binding free energy (∆G).
Discovery Studio ® Visualizer 2016 (Accelrys Software Inc., San Diego, CA, USA) was used for the three-dimensional visualization and generation of two-dimensional plots of the protein-ligand interactions. Also, the number of residues which participated in hydrophobic interactions was predicted using the LigPlot Plus v.1.4 (European Bioinformatics Institute, Cambridge, UK) program.
Test microorganisms were stored in freezing conditions and activated by cultivation in Nutrient broth at 37 • C for 24 h. Overnight cultures were used to obtain the inoculum by transferring an amount of cells from the broth to a test tube containing 5 mL of demineralized water until a turbidity of 1-2 × 10 8 colony-forming unit (CFU)/mL was achieved, equivalent to 0.5 McFarland standard [29]. Dilutions of the inoculum were cultured on nutrient agar supplemented with 5% sheep blood to verify the absence of contamination and the validity of the inoculum.

Well Diffusion Test
Compounds were investigated by the agar well diffusion method. The plates containing Mueller Hinton agar medium were spread with 0.1 mL of the bacterial inoculum. The wells (6 mm diameter) were filled with 50 µL of the sample. The plates were incubated at 37 • C for 24 h. DMSO was used as a negative control and one standard antibiotic (gentamicin, 10 µg) as a positive control. Microbial inhibition was visually appreciated as the diameter of the inhibition zones surrounding the well and recorded in mm. All tests were run in triplicate for each combination of extract and microbial strain [26].

Microdilution Test
Bacterial strains susceptible to compounds in the well diffusion assay were studied for their MIC values using 96-well microtiter plates [27,28]. The experiments were conducted using the reference method for testing the in vitro activity of the antimicrobial agents where MIC is the lowest concentration of an antimicrobial agent that, under defined in vitro conditions, prevents the appearance of visible growth of the microorganism after its incubation overnight. A total of 250 µL of the bacterial suspension in sterile demineralized water was added to 11 mL Mueller-Hinton broth (MBH) to a concentration of 5 × 10 7 cells/mL. Compounds to be investigated were dissolved in MBH containing bacterial inoculum (5 × 10 4 CFU/well) to achieve the two-fold serial concentrations ranging between 4-250 µL/mL [30].
The microplates were covered with an adhesive seal and incubated at 37 • C for 24 h. DMSO combined with MHB was used as the negative control for each bacterial strain. The plates were visually read under normal laboratory lighting using a manual mirror viewer. Two replicates were conducted for each compound and each bacterial strain.

Acute Toxicity Assessment Using Daphnia Magna
The Daphnia magna test was performed in 4 mL 12-tissue culture wells, using 10 daphnids/well, with each sample being tested in duplicate [31]. Lethality was recorded after 24 h and 48 h, considering the organisms that did not move their appendages for 30 s as dead. All experiments were conducted in the dark, in a plant growth chamber (Sanyo MLR-351 H, San Diego, CA, USA) at 25 ± 1 • C [32,33]. The bioassay was performed at concentrations ranging from 0.5 to 100 µM (0.5, 1, 5, 10, 50 and 100 µM) for each compound and, based on preliminary data, the positive control, potassium dichromate, at concentrations ranging from 0.2 to 10 µg/mL (0.2, 0.4, 0.6, 2.0, 4.0 and 10 µg/mL corresponding to 0.6, 1.3, 2.0, 6.8, 13.6 and 34.0 µM). The reference test with potassium dichromate was conducted to check the sensitivity of Daphnia to meet the validity criterion according to the OECD (The Organisation for Economic Co-operation and Development) guideline 202, if the LC50 of potassium dichromate at 24 h is ranging from 0.6 to 2.1 µg/mL [34].

Statistical Analysis
Statistical analysis of the inhibitory effect on SrtA and the Daphnia magna bioassay was performed using GraphPad Prism version 5.01 software (GraphPad Software, Inc., La Jolla, CA, USA). A student's t-test (p < 0.05) was applied in order to analyze the results obtained from replicates. No statistical differences were found for three (inhibition on SrtA) or two (Daphnia magna bioassay) replicates. The inhibition (I %) and lethality percentage (L %) values were calculated and plotted against the logarithm of concentrations and the corresponding curves were calculated using the least squares fit method. The lethal concentration (LC50), which produces an L% value equal to 50, was determined by interpolation on the curves. Whenever the obtained results permitted, the upper and lower limits of the 95% confidence interval (95% CI) and the correlation coefficient (r 2 ) were calculated.