Introduction

The number of deaths attributable to antimicrobial resistance (AMR), a condition where bacteria evolve and become resistant to antimicrobial agents and antibiotics, is increasing and has become a significant global challenge1. It is estimated that more than 35,000 deaths occur annually in the United States, and 33,000 deaths occur annually in Europe. It is estimated that the number of deaths related to AMR worldwide could reach 10 million per year by 20502. Methicillin-resistant Staphylococcus aureus (MRSA) is the most prevalent drug-resistant organism responsible for nosocomial and community-acquired infections3,4.

Staphylococcus aureus is a facultative anaerobic gram-positive bacillus5. It is in high numbers on humans’ and animals’ skin, nasal passages, and other body surfaces6. The yellowish nasal discharge commonly observed during a cold is a consequence of the pigmentation of this bacterium. It is widely distributed in nature and can be isolated ex vivo from air, soil, and dairy products. It is most prevalent in the crusts of wounds. S. aureus produces a multitude of toxin proteins that can result in a wide range of severe infections, including soft tissue infections such as pyogenesis, impetigo, folliculitis, and abscesses, as well as pneumonia, peritonitis, sepsis, and meningitis7,8. Moreover, it is a causative agent of food poisoning, toxic shock syndrome, and enteritis8. The organism is known to cause pneumonia, septicemia, and other severe infections in individuals with underlying diseases, such as the elderly, the sick, and young children with weakened immune systems9.

It is well documented that Staphylococcus aureus plays a role in enhancing antibiotic resistance through the formation of biofilms and persisters10,11. Consequently, this bacterium is attracting increasing attention worldwide as a significant pathogen responsible for opportunistic human infections. Nevertheless, a shortage of efficacious antimicrobial agents is currently available for clinical practice. In this study, we concentrated on SAL (S. aureus lipase). This enzyme has attracted considerable attention as a potential target for research due to its involvement in the growth of S. aureus12,13.

In light of the established correlation between SAL and the proliferation of S. aureus, it is postulated that the inhibition of this enzyme may impede the growth of the bacteria13. SAL is a protein with a molecular weight of 45.7 kilodaltons secreted as an enzyme consisting of 394 amino acid residues12. Glycerol ester hydrolase (Geh) is an enzyme that catalyzes the hydrolysis of triacylglycerol, a common lipid component. Its enzymatic function is based on the typical catalytic triad, comprising serine, aspartate, histidine residues, and oxyanion holes, which are analogous to those observed in serine proteases12.

The three-dimensional structure of SAL has already been determined, and the detailed structural mode of the inhibitor orlistat has been elucidated14. Orlistat is an anti-obesity pharmaceutical agent that inhibits the activity of gastric and pancreatic lipases15. Moreover, it has been demonstrated to exert inhibitory activity against SAL14. The detailed structural information was utilized in an in silico screening process to identify potential inhibitors of existing drugs.

In silico screening represents an effective methodology for identifying efficacious inhibitors within compound databases16. When crystal structures of enzyme-inhibitor complexes are available, docking studies employing programs such as Glide can rapidly identify potential compound candidates17,18. Nevertheless, the probability of successful drug design remains low, with a 2–3% correct response rate to identify compounds of the μmol order for which IC50 can be calculated.

The drug repositioning process, which encompasses in silico screening and repurposing of existing pharmaceuticals, has been attempted on numerous occasions in search of therapeutic agents that can effectively combat COVID-19 coronaviruses19. Given the substantial financial investment and extended approval timeline associated with the discovery and development of new drugs, repurposing existing, previously approved pharmaceuticals is becoming an increasingly appealing strategy20,21. This is because it entails reduced overall development costs, shorter development times, and the utilization of less risky compounds, which collectively contribute to a more efficient and cost-effective drug development process. The potential of drug repurposing (also known as drug repositioning) has been demonstrated; anti-obesity drugs, for instance, have been shown to exhibit inhibitory activity in SAL14.

In silico techniques have been employed in conjunction with databases such as KEGG to identify potential drug candidates for repurposing22,23. We were fortunate to identify an inhibitor among the 13 compounds obtained through in silico methods, which exhibited potent inhibitory activity (Supplemental Fig. S1). The compound in question is penfluridol, which has a long history of use as an antipsychotic (Fig. 1a)24.

Fig. 1
figure 1

Chemical structures of Penfluridol and Pimozide, with IC50 values. Relative activities were measured at different inhibitor concentrations. The two compounds exhibited a striking resemblance in their skeletal structures, with the sole distinction lying in the leftmost group depicted in the figure, which ultimately manifested in disparate solubility characteristics.

Full size image

Penfluridol is a highly potent first-generation diphenylbutylpiperidine antipsychotic that has been demonstrated to suppress the imbalance in the inflammatory response in LPS-induced macrophages by inhibiting the activation of the NLRP3 inflammasome25. It is a drug that is widely utilized in clinical practice.

Penfluridol is a relatively large compound comprising a biphenyl moiety with two benzene fluorides, a central heterocyclic ring system (comprising piperazine and piperidine groups) containing at least one nitrogen atom, which forms the backbone of the compound, and a benzene ring with a trifluoromethyl group and a chloro group attached to it, with three carbon atoms (Fig. 1a). As illustrated in Fig. 1a, three carbon chains link the two units.

One hundred compounds were selected for in silico screening through molecular docking calculations using the Glide program against 49,000 existing drug compounds in the KEGG database (Fig. S1)17,18,22,23.The results demonstrated that penfluridol exhibited potent inhibitory activity. This study had two objectives: first, to calculate the IC50 of penfluridol, which shows inhibitory activity against SAL, and second, to obtain a detailed structure with higher resolution and precision by attempting to crystallize the complex of SAL and penfluridol on Earth and under weightless conditions with the assistance of JAXA. For penfluridol, which demonstrated robust inhibitory activity, the half-inhibition constant (IC50) was calculated, and the complex of SAL and penfluridol was co-crystallized.

X-ray diffraction data for crystals of the SAL complexed with inhibitors were collected at the SPring-8 beamlines BL41XU and BL44XU. The structures were determined by the molecular replacement method, and molecular models were constructed from the obtained electron density maps to determine the steric structures of SAL and the existing drug Penfluridol.

Results

In silico screening and IC50 measurement by activity assay

In silico screening using Glide selected 100 compounds from the existing KEGG drug database22,23. The interaction between SAL and orlistat was analyzed using the coordinates of the structure of the complex of SAL and orlistat (pdbid: 6KSM). Consequently, a range of active site regions of SAL involved in binding to the inhibitor was selected, and the distribution of amino acids within these regions was verified. Furthermore, the electron donor, electron acceptor, and hydrophobic portions were determined on Glide, and compounds from the KEGG database that fit into the active site of SAL in terms of both inhibitor shape and electron distribution were selected. The selection of these 100 compounds was based on the Glidescore docking score value, which was utilized as an indicator. Compounds with Glidescore scores ranging from − 10 to − 8.6 were included in the list. Subsequently, the inhibitory activity of these compounds was validated sequentially, commencing with the most prioritized compounds (Fig. S1). As previously reported, the activity of the compounds was quantified by measuring the absorption wavelength at 405 nm of para-nitrophenol produced by the degradation of p-nitrophenyl butyrate (pNPB)14. A series of inhibitor solutions, with concentrations ranging from 1 to 100 mM, were added to cuvettes containing SAL, buffer, and water solutions. Subsequently, the alterations in absorbance were quantified by introducing a pNPB substrate solution dissolved in DMSO into the cuvettes. The initial absorbance was then measured over time and compared to the uninhibited solution’s slope to calculate the inhibition percentage.

Among the selected inhibitors, penfluridol, a dopamine receptor inhibitor, was identified as a SAL inhibitor, with an IC50 value of 7.3 μM (Figs. 1a, 2a). Penfluridol is a first-generation diphenylbutylpiperidine antipsychotic with potent antipsychotic activity. The IC50 values were found to be as potent as those of previously identified inhibitory compounds, including orlistat (IC50 = 2.4 μM) Petroselinic acid (IC50 = 3.4 μM) and Farnesol (IC50 = 570 μM) (Fig. S2)13,14,26.

Fig. 2
figure 2

Inhibition of SAL by Penfluridol and Pimozide. The IC50 value was determined using the IC50 calculator (https://www.aatbio.com/tools/ic50-calculator). SAL activity towards p-nitrophenylbutyrate. Enzymatic activity was measured by monitoring p-nitrophenol production. Each data point is presented as mean ± SEM with error bars representing SEM (n = 3). (a) The IC50 of Penfluridol versus SAL. (b) The IC50 of Pimozide versus SAL. (c) Comparison of increased SAL activity of the same concentrations of Penfluridol and Pimozide. (20 μM each). Values without inhibitors are taken as 100%. The statistical software used was Prism9 (GraphPad), with one-way ANOVA tests and Dunnet for multiple comparisons.

Full size image

X-ray structure analysis of SAL-penfluridol complex

Furthermore, co-crystals of the SAL-penfluridol complex were prepared. The solubility of penfluridol was found to be high, with the highest solubility observed in DMSO at 100 mg/mL. It was determined to be a suitable inhibitor of the co-crystallization process (see Fig. S1). The crystallization conditions were 2.8–3.01 M sodium formate and 0.1 M acetate buffer at pH 6.5, maintained at 295 K. The crystallization conditions were similar to those previously reported and yielded reproducible, diamond-shaped crystals in approximately one to two weeks14,26.

Penfluridol was the third Glidescore among the compounds that were available, and the higher scoring compounds Indacaterol (adrenergic β2 receptor agonist)27 and Olaparib (PARP inhibitor; for the treatment of breast and ovarian cancer)28 had lower solubility and were co-crystallized for data measurement (Fig. S1). The 2.41 Å resolution for Indacaterol (accompanied by six additional data sets) and the 2.19 Å resolution (along with 35 other data sets) were measured for Olaparib. The analysis revealed no binding at all, and since the crystal structure did not show binding of the inhibitor, the exact IC50 was not determined.

Moreover, the co-crystallization of the SAL-penfluridol complex was conducted at 293 K in a microgravity environment. In Projects MT8 and MT9, crystallization in microgravity conditions was successfully achieved in the thermostatic chamber of the Kibo Pressurized Module (PM) using a counter-diffusion method, resulting in the formation of highly pure hexagonal crystals (Fig. S3)29.

The data measurements of these composite crystals were conducted at BL41XU and BL44XU at SPring-8, and the resulting data exhibited a resolution of approximately 2.23 Å (on the ground, PDBid: 9L3C) and 2.57 Å (in space, PDBid: 9L3S), respectively. The SAL- penfluridol crystal (on the ground) was found to have space group P4122, with lattice constants a = b = 127.5 and c = 252.5 Å. It was nearly isomorphic to the native crystal, containing two molecules in the asymmetric unit (see Table 1)14. The structures were determined by the molecular replacement method using Phaser in CCP4 for structural analysis, and molecular modeling was performed in Coot30,31,32. The results of the molecular replacement analysis indicated the presence of an electron density near the active site in both the ground and space crystals, which appeared to be penfluridol (Fig. 3a, b). The active site’s electron density was highly reproducible (Fig. 3a, b).

Table 1 Data collection, processing, and refinement.
Full size table
Fig. 3
figure 3

(a) The primary interactions of penfluridol within the active site of the A chain of SAL are depicted using the LIGPLOT+ diagram on the right42 and the electron density diagram on the left (The 2Fo-Fc electron density map calculated at 1.0 sigma). (b) A. The primary interactions of penfluridol within the active site of the B chain of SAL are depicted in the LIGPLOT+ diagram on the right42 and the electron density diagram on the left. Fmt indicates formic acid.

Full size image

Coot's Ligand Builder was employed to fit a model molecule of penfluridol to the electron density32, and the structure of SAL and penfluridol was successfully determined through a process of repeated structure refinement analysis conducted using the Refmac software33.

Structural refinement resulted in the determination of a stereochemically correct structure with an Rfactor value of 20.2% and an Rfree value of 23.1%, indicating the certainty of the structure, using data at 2.23 Å resolution (Table 1).

Structure of the SAL-penfluridol complex

The monomeric structure of SAL exhibited similarity to that reported in previous studies and retained the typical α/β-hydrolase structure, consisting of 13 β-strands, 13 α-helices, and six 310-helices (Fig. 4)14,26. The SAL exhibited the characteristic of two large hydrophobic pockets extending from the active center, Ser11614, as observed in previous studies (Figs. 4, 5). The nucleophilic atom Ser116 was identified within the so-called "β-elbow" loop, situated between the β-chain and α-helix of the α/β hydrolase center34. Like other serine hydrolases, it contains three catalytic triad residues, namely His349, Asp307, and Ser116, which are essential for the catalytic mechanism. The nitrogen atoms in the leading chains of Phe17 and Met117 form oxyanion holes, which can serve to stabilize intermediate transition states during the reaction14,26.

Fig. 4
figure 4

The figure depicts the crystal structure of the monomer of the SAL-perfurylidol complex, with a ribbon model of the SAL structure and a fine decapod model. A sphere represents the penfluridol molecule. The residues involved in the interaction are indicated by yellow sticks. The fluorine atoms are illustrated as cyanide spheres. All molecular structure images were obtained from PyMOL (http://www.pymol.org/).

Full size image
Fig. 5
figure 5

The figure depicts the SAL dimer and penfluridol complexes. The monomers are illustrated in blue (chain A) and pink (chain B), and penfluridol bound to the A and B molecules is depicted as a green sphere. The packing between monomers exhibited the same symmetry as the complex structure that had been previously analyzed14,26.

Full size image

The SAL-penfluridol complex was also a dimeric molecule with one molecule of penfluridol bound to each of the A and B chains, and the binding mode of the two penfluridol molecules to SAL was identical (Fig. 5).

The electron density was observed to be reproducible at the active site (Fig. 3). The success of this complex analysis may be attributed, at least in part, to the high solubility of penfluridol. In comparison with the native crystal resolution of 2.0 Å, the penfluoridol complex crystal reflection exhibited relative weakness, yielding only approximately 2.5–2.7 Å. In this study, the highest resolution data of 2.23 Å was obtained. The acquisition of high-resolution data was facilitated by circumventing the conventional approach of making 180-degree measurements on single crystals, making 360-degree measurements, and integrating only those data deemed to be of sufficient quality. Structural analysis was also performed for all other 12 compounds, but due to the solubility of the inhibitors, precipitation often occurred, and complex structures could not be obtained (Fig. S1). Further activity assays showed that 12 species, other than penfluridol, had little or no inhibitory activity (Supplementary Fig. S1).

Interaction of SAL-penfluridol complex

The figure shows an enlarged view of the SAL-penfluridol complex in the vicinity of the active site (Fig. 3a, b). The structure of penfluridol was Y-shaped (Figs. 3, 4, 5, 6 and 7). Two fluorophenyls spread their hands, and the remaining residue extended from the junction. The IUPAC nomenclature for penfluridol is 1-[4,4-bis(4-fluorophenyl)butyl]-4-[4-chloro-3-(trifluoromethyl)phenyl]piperidin-4-ol, which is composed of four distinct parts (Fig. 1a). In other words, the molecule consists of a trifluoromethyl phenyl chloride group, a piperidine ring, and three alkyl chains that serve as linkers, followed by two fluorinated phenyl groups (Fig. 1a).

Fig. 6
figure 6

The interaction of penfluridol with amino acid residues at the catalytic site of SAL is illustrated. Blue sticks indicate the residues involved in binding. For penfluridol, carbon is drawn in green, oxygen in red, nitrogen in blue, and fluorine and chlorine in pale blue. The two formate ions are written in yellow. For further details, please refer to the accompanying text.

Full size image
Fig. 7
figure 7

The superimposition of structures around the active site was conducted with wild-type SAL and SAL-Penfluridol complex structures as monomers. The representation of penfluoridol bound to the A and B molecules is illustrated by sticks, with oxygen depicted in red, nitrogen in blue, chlorine in green, and fluorine in light blue. In a, wild-type SAL and complex A molecules are shown in green and magenta, respectively. The amino acid residues in the vicinity of the biphenyl moiety, which has undergone numerous structural changes, are indicated by stick notation. In b, wild-type SAL and B molecules of the complex are also shown in green and magenta, respectively.

Full size image

The binding mode of penfluridol differs from previous results, as evidenced by the interaction of the two phenyl fluoride rings across the hydrophobic pocket, which slightly widens the pocket (Fig. 7a, b)14,26. No direct covalent binding to the active residue, Ser116, was observed. The electron density map showed that Ser116 was involved in a binding interaction with the formic acid molecule in the crystallization reagent. The formic acid molecule was then placed in the so-called oxyanion hole, consisting of nitrogen atoms in the main chains of Phe17 and Met117, stabilizing the binding state of SAL and penfluridol. In addition, another molecule of formic acid B was present close to formic acid A (Figs. 3a, b, 6). This formic acid B interacted with the piperazine ring of penfluridol and also interacted with formic acid A, thereby mediating the bond between Ser116 and penfluridol (Figs. 3a, b, 6). The mode of bonding exhibited a high degree of similarity between the terrestrial and extraterrestrial environments. A similar phenomenon was observed with the alkyl chains that followed the biphenyl chain, where the electron density was thinner.

Indeed, SALs with excess penfluridol were removed by the Econo-Pac 10DG column, and their activity was found to be minimal, as shown in the Supplementary Data (Fig. S4ab). This suggests that the inhibitors may be covalently bound to the SAL. The data suggest that the inhibition is covalently bound in a manner distinct from the binding of the active site to Ser116.

The inhibition mechanism of this compound was found to be novel in that it does not involve direct covalent binding to the active residue Ser116. Alternatively, the active residue Ser116 interacts with the formic acid used for crystallization and binds to trifluoromethyl phenyl chloride via formic acid (Figs. 3a, b, 6). It is conceivable that water molecules, rather than formic acid, facilitate binding in solution.

The trifluoromethyl group was observed to interact with Ala174, Ala175, and Phe17 and Phe178. The chloro group was close to Leu242 (Figs. 3a, b, 6). The phenyl ring was located between Phe17 and Val309. The piperidine ring connected to the phenyl ring was located between Leu18 and Met288, where an electron density was observed that could be attributed to either water or formic acid (Fig. 3a, b). The alkyl chain connecting the piperidine ring to the two fluorophenyl rings exhibited a slightly lower density, with a clear electron density observed in the A chain and a slightly disconnected disorder in the B chain (Fig. 3a, b).

One of the biphenyl fluorides was observed to form hydrophobic interactions with a phenylalanine cluster consisting of three phenylalanine residues (Phe285, Phe286, Phe357) and Leu287 at the active site (Fig. 6). The second phenyl fluoride was captured by a hydrophobic scissors cluster consisting of Tyr29 and Pro30 and further interacted with a hydrophobic cluster of Tyr32, Val350, Ile353, and Val355 (Fig. 6).

The results of this analysis were comparable to those of the in silico analysis, although there were some discrepancies in the structural data (Fig. 8a, b). Notably, the predicted positions of the trifluoromethyl group and the chlorine moiety were reversed from left to right (Fig. 8a, b). In the prediction, the trifluoromethyl group was observed to be close to Ser116 (Fig. 8b). In addition, minor discrepancies were observed in the positions of the two fluorophenyl rings, with the hydrophobic residues surrounding these rings being shifted to expand the hydrophobic pocket (Fig. 7a, b).

Fig. 8
figure 8

The position of the inhibitor penfluridol within the complex structure is illustrated. (a) The blue sticks represent the results of the X-ray analysis. (b) The green molecule represents the predicted binding mode of penfluridol, as calculated by the Glide program. The chlorobenzotrifluoride group on the left has been relocated from its original position to the right in accordance with the findings of the X-ray analysis. The molecule is depicted as a green stick, and the hydrogen molecule is represented in white. (c) A similar approach was employed to predict the bonding diagram of the pimozide molecule. Similarly, the yellow sticks illustrate the predicted bonding diagram of the pimozide molecule. (d) The two molecules, resulting from the structure prediction and the X-ray analysis, are superimposed. One of the biphenyls exhibits a slight variation in the distribution of the phenyl group, while the other displays a comparable variation in the distribution of the biphenyl group.

Full size image

Fluorescence changes due to the interaction of penfluridol and SAL

The relationship between SAL and penfluridol was further investigated using data from previous studies and fluorescence measurements. A recent study conducted by a research group in China showed that penfluridol inhibited S. aureus growth in a concentration-dependent manner35. However, it is important to note that this study did not investigate the involvement of SAL35.

The supplementary Fig. S5a shows the fluorescence spectrum of penfluridol, which shows a concentration-dependent change in the fluorescence spectrum of SAL (Fig. S5a). Specifically, an increase in the concentration of SAL resulted in a shift to a lower wavelength (435 nm) in the fluorescence spectrum (Supplementary Fig. S5a). The observed relationship between penfluridol and growth inhibition suggests that this compound may be a potential diagnostic agent for detecting S. aureus. This would facilitate the rapid detection of S. aureus.

After studying the shift in the fluorescence wavelength peak, it was proposed that the effect of fluorescence resulting from the rotation of the benzene fluoride ring may have been modified by binding to the active site of SAL (Fig. S5b). It was postulated that the intramolecular motion of penfluridol, capable of rotation, would be inhibited by binding to the active site of SAL (Fig. S5b). This would result in suppressing the rotation of the benzene difluoride of penfluridol, which would lead to a change in fluorescence (Fig. S5ab). Indeed, the electron density was also slightly reduced in the region of the carbon chain, leading to the two benzene difluorides (Fig. 3a, b), which we interpreted as an indication of the magnitude of the intramolecular fluctuations.

Discussion

Penfluridol was identified as a promising candidate compound in the in silico screening process and was ranked 15th (Fig. S1). Pimozide, a compound with structural similarity to penfluridol, was the 84th compound to undergo evaluation in the in silico screening process (Fig. S1). Pimozide is a neuroleptic drug indicated primarily for the treatment of schizophrenia36. Furthermore, it has been utilized in treating autism in pediatric patients. The IC50 value of this drug, which has a similar structure to that of penfluridol, was determined. Moreover, an attempt was made to determine the structure of the complex. The IC50 of pimozide was determined to be approximately 48.5 µM (Fig. 2b). The value for penfluridol was 7.3 µM, indicating that penfluridol is approximately six times more potent in inhibitory activity than pimozide (Fig. 2c). A comparison of the structural formulas for these compounds is presented in Fig. 1. The IUPAC names for penfluridol and pimozide are 1-[4,4-bis(4-fluorophenyl)butyl]-4-[4-chloro-3-(trifluoromethyl)phenyl]piperidin-4-ol and 3-[1-[4,4-bis(4-fluorophenyl)butyl]piperidin-4-yl]-1H-benzimidazol-2-one, respectively. The structural similarity between the biphenyl fluoride and piperazine groups and the propyl chain moieties linking them, with a terminal methyl trifluoride group, is responsible for the fivefold difference in activity between pimozide and benzimidazole (Figs. 1a, b, 2a–c)36. The compound with the structure of diphenylbutylpiperidine is a prototypical antipsychotic drug (Fig. 1a, b).

In all 11 diffraction data sets of SAL-pimozide co-crystals measured at SPring-8, no evidence of pimozide binding to the pimozide-SAL complex crystal was observed. This was attributed to the low solubility and low inhibitory activity of pimozide (Fig. S1). In stark contrast to pimozide, which displayed markedly low solubility (Solubility of 18 mg/mL in DMSO)36, penfluridol exhibited markedly high aqueous solubility (100 mg/mL in DMSO) (Fig. S1). Due to its high solubility and pronounced inhibitory effect, the electron density of penfluridol was distinctly discernible (Fig. 3a, b), whereas no map was discernible for pimozide.

The results demonstrate that both the fluorinated biphenyl moiety and the terminal substituents substantially influence the binding of SAL to penfluridol. The predicted binding mode of pimozide was arranged such that the oxygen molecule of the carbonyl group of the terminal benzimidazole was observed to bind to the active site, Ser116 (Fig. 8c). Nevertheless, the structure could not be undetermined. Similarly, the prediction for this moiety differed from the actual experimental data for penfluridol (Fig. 8a, b). In the prediction, the methyl trifluoride group was oriented toward Ser116 (Fig. 8). However, in the actual experimental data, this orientation was reversed (Fig. 8a, b). The electron density map indicated that the orientations were, in fact, opposite one another (Fig. 3a, b).

The authors identified penfluridol as a potential inhibitor of SAL among existing pharmaceutical agents. Concurrently, a Chinese research team employed a comparable drug repositioning strategy and identified a compound that demonstrated inhibitory activity against the growth of Staphylococcus aureus among existing pharmaceuticals35. Notably, this compound was penfluridol35. Penfluridol demonstrated potent bactericidal efficacy against S. aureus, with reported minimum inhibitory concentrations (MIC) ranging between 4 and 8 μg/ml and minimum bactericidal concentrations (MBC) from 16 to 32 μg/ml35. Furthermore, the study revealed that penfluridol significantly impeded the formation of Staphylococcus aureus biofilms, with dose-dependent efficacy and the capacity to eradicate biofilms formed before 24 h35.

The findings indicated that penfluridol was efficacious in eradicating biofilms formed within 24 h35. Moreover, penfluridol effectively eradicates methicillin-resistant S. aureus (MRSA) persistent cells in both in vitro and in vivo35. In particular, it demonstrated significant efficacy in mouse models of subcutaneous abscess, skin wound infection, and acute peritonitis caused by MRSA35.

Prior research has indicated that farnesol, a terpenoid organic compound belonging to the class of linear sesquiterpenes (Fig. S2c), displays inhibitory activity against the growth of S. aureus13. The introduction of increasing quantities of farnesol was observed to result in a corresponding inhibition of bacterial growth13. Compared to farnesol, penfluridol demonstrated a more pronounced inhibitory effect against SAL (Fig. S2d), and a more pronounced inhibitory effect on bacterial growth13. This indicates that penfluridol may impede the proliferation of S. aureus by inhibiting SAL. These findings suggest that SAL inhibitors may have therapeutic potential as agents to prevent the growth of S. aureus.

Finally, the potential for SAL inhibitors with increased efficacy and reduced side effects is discussed. With regard to penfluridol, its pronounced toxicity in mouse models was considered to be a significant impediment to its practical application. Therefore, it is imperative to modify this compound’s structure to ensure its safety. In this regard, we have explored the potential of a compound that exhibits enhanced selectivity in binding to the SAL. The results of the analysis of the penfluridol complex were compared with those of the orlistat and petroselinic acid complexes that were previously analyzed and determined (Fig. 9). As illustrated by the structural formula, both orlistat and petroselinic acid consist exclusively of extended alkyl chains and are covalently bound to the active catalytic residue Ser116 (Fig. S2ab)14, 26. In addition, the IC50 values exhibited slightly greater potency in comparison to penfluridol (Fig. S2d).

Fig. 9
figure 9

A comparison was made between the SAL and the inhibitor that has been previously analyzed, using a structure overlay technique. (a) The structure of SAL and penfluridol is shown here. The catalytic residue Ser116 is marked with a white stick, and it is evident that it is not directly bound to penfluridol. The representation of penfluridol is indicated by magenta, oxygen by red, nitrogen by blue, chlorine by green, and fluorine by pale blue. (b) The structure of the SAL-Orlistat complex is illustrated, with Orlistat represented by sticks and colored cyan. Ser116 has been demonstrated to bind with Orlistat. (c) SAL and petroselinic acid complex structure. The green stick indicates petroselinic acid binding to Ser116. (d) The figure below presents a schematic representation of all inhibitors, superimposed on the binding sites of SAL.

Full size image

In contrast, the binding mode of penfluridol was entirely distinct, exhibiting no covalent bond with Ser116 and no alkyl chain (Fig. 9a). The long alkyl chains of orlistat exhibited significant overlap with the central skeletal portion of penfluridol (Fig. 9b, d). Conversely, there was minimal overlap with the alkyl chain portion of petroselinic acid (Fig. 9c, d). In the case of orlistat, the lactone ring moiety exhibited a reaction with and binding to Ser116, while in petroselinic acid, the double bond moiety and Ser116 were covalently linked (Fig. 9b, c)14, 26.

A novel drug design concept would be to create compounds that extend the alkyl chain from the chloro portion of the chlorobenzotrifluoride portion of penfluridol (Fig. 1a) or react with a lactone ring (Fig. 9, S2a). The creation of such compounds, if feasible, is anticipated to result in inhibitors of SAL that exhibit high selectivity and non-toxicity. The findings of this study suggest the feasibility of producing inhibitors of SAL that are both highly selective and non-toxic.

Methods

In silico analysis

The docking method was employed with the Glide program, based on the structural information of the SAL and orlistat complexes from a previous study14,17,18. The compounds were fitted against a scaffold that quantified the shape of the enzyme’s active site and whether it donates or accepts charge, thereby identifying the compounds with the best binding states. From a database comprising tens of thousands of compounds, binding prediction scores were calculated, and candidate inhibitors were evaluated through enzyme-compound docking studies. The top 100 inhibitor candidates were obtained from a 49,000 existing drug compounds database in the Kyoto Encyclopedia of Genes and Genomes (KEGG)22,23.

Expression and purification of pColdII-SAL

The glycerol stock containing the recombinant SAL gene in the E. coli pColdII vector harbored by BL21(DE3) was precultured in 50 mL LB medium containing 50 μg/mL ampicillin at 37 °C in 200 mL Meyer flasks, as previously described14,26,37. Ten milliliters of the preculture solution were transferred to one liter of LB medium containing ampicillin and incubated at 37 °C for approximately two hours. Subsequently, protein expression was induced by adding 0.1 mM IPTG at a final concentration. The mixture was incubated at 15 °C for 24 h. Subsequently, the cells were collected by centrifugation at 277 K for 30 min at 4000 rpm, as previously reported. The supernatant fluid was washed twice with TE buffer and suspended in lysis buffer (50 mM phosphate buffer at pH 8.0 containing 0.3 M NaCl and 10 mM imidazole). The suspension was sonicated at 273 K using an ultrasonic disruptor (UD-211, Takara Tomy) and centrifugation. The supernatant was then subjected to a series of filtration processes at 0.45 µm and 0.22 µm, respectively. The resulting supernatant was purified by immobilized metal affinity chromatography using Nuvia IMAC resin (Bio-Rad). The proteins were eluted with an imidazole concentration gradient (10–400 mM), and the SAL fractions were collected and dialyzed in buffer (0.2 M NaCl, 10 mM MES, pH 6.5) to remove imidazole. Subsequently, the sample was further purified through SP (Toyopearl) column chromatography. The purified protein was loaded onto the column and eluted with a NaCl concentration gradient (0.2–1.0 M). The SAL protein fractions were combined and dialyzed in 0.2 M NaCl, 10 mM MES, pH 6.0. Subsequently, the dialyzed SAL solution was subjected to MonoQ column chromatography (GE Healthcare) to remove extraneous material, and the flow-through fraction was collected as the final purified protein14,26,37. The purity of the protein fractions was evaluated using SDS-PAGE electrophoresis.

Measurement of IC50 half-inhibition values

Enzyme activity was quantified by measuring the absorbance at 405 nm, which indicates degradation of the p-nitrophenylbutyrate (pNPB) ester substrate, as previously reported14,26,37. The concentrations of SAL and pNPB were adjusted to 0.002 mM and 0.08 mM, respectively. Each assay was conducted in a 1-mL cuvette containing 50 mM HEPES (pH 7.5) and 10% DMSO at 25 °C. To determine the IC50 value, a series of inhibitor concentrations (0–1 mM) was prepared. Before initiating the SAL/pNPB reaction, SAL was incubated with the inhibitor at 25 °C for 10 min.

Subsequently, the substrate was added, and the increase in the absorbance of free p-nitrophenol was monitored at 405 nm. The relative specific activity of SAL was calculated from the slope of the graph for each condition. Inhibition experiments were conducted by adjusting the solution to concentrations that inhibited SAL activity by more or less than 50%. At least three measurements were taken for each experiment, and the mean data for each concentration were used to calculate the IC50 value.

Crystallization on the ground and in space

The protein solution was subjected to ultrafiltration using a Vivaspin15 device, resulting in a 30 mg/mL concentration. This solution was then employed in crystallization experiments.

The protein concentration values were determined using the Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA), with γ-globulin as the standard protein.

The inhibitors were dissolved in DMSO or buffer, added to concentrated SAL, and incubated at 4 °C overnight. The concentrations of the various inhibitors were adjusted to achieve molar concentrations that were 5 to 10 times greater than that of SAL.

The sitting drop vapor diffusion method was conducted at 295 K using a Cryschem plate. A total of 500 µL of the crystallization solution (reservoir) was added to the bottom of the plate, and 2 µL of the mixture of crystallization solution and protein solution was dispensed into the reservoir from the top. Subsequently, the reservoir was sealed.

Experiments to investigate the formation of protein crystals of the highest quality (PCG) have commenced at 293 K. Protein samples and other crystallization apparatus were transported to the International Space Station (ISS) aboard the U.S. SpaceX Dragon resupply ships (SpX-26 and SpX-28). There, astronauts conducted crystallization experiments in the thermostatic chambers of the Kibo laboratory from November to December 2022 and June to July 2023.

The counter-diffusion method represents a specific instance of the broader category of liquid–liquid diffusion methods29. The current JAXA PCG employs a silicon tube filled with agarose gel connected to a glass capillary filled with protein solution as the fundamental unit. The crystallization solution diffuses into the capillary through the gel tube, while the protein and other solutes in the capillary diffuse out of the capillary through the gel tube. This bidirectional diffusion allows for the simultaneous exploration of a wide range of crystallization conditions. The ratio of components was modified to (protein, penfluridol, crystallization solution) = 20:1:20. A volume of 12.5 µL per milliliter of crystallization solution and gel immersion solution was administered with the penfluridol and transported to the site.

Data collection, structure determination, and refinement

X-ray diffraction data were collected at SPring-8 beamlines BL41XU and BL44XU at 100 K in a nitrogen stream using a mother solution containing 30% glycerol as a cryoprotectant. The X-ray wavelength was adjusted to 0.9 Å, and the oscillation angle were set to 0.1°, resulting in 180°–360° images. The distance between the crystal and the detector were 240–300 mm. The diffraction images were integrated and scaled using the XDS program package39 with the KAMO automated programming system40 at SPring-8. The data set was obtained at the limit of resolution where the CC 1/2 value was approximately 0.541.

The molecular replacement method, as utilized by the Phaser program in the CCP4 package, was employed to determine the structure30,31. The previously reported native structure of SAL (PDB ID: 6KSI) was used as the search model14. The COOT program was used to construct the inhibitor molecule and fit the molecular model, and REFMAC5 was used to refine the structure32,33.

To fit the model to the electron density map, the inhibitor molecule was constructed with Coot's Ligand builder32. Real-space refinement was then employed and continued until convergence using Refmac software33.