Role of Stereochemistry on the Biological Activity of Nature-Inspired 3-Br-Acivicin Isomers and Derivatives

Chiral natural compounds are often biosynthesized in an enantiomerically pure fashion, and stereochemistry plays a pivotal role in biological activity. Herein, we investigated the significance of chirality for nature-inspired 3-Br-acivicin (3-BA) and its derivatives. The three unnatural isomers of 3-BA and its ester and amide derivatives were prepared and characterized for their antimalarial activity. Only the (5S, αS) isomers displayed significant antiplasmodial activity, revealing that their uptake might be mediated by the L-amino acid transport system, which is known to mediate the acivicin membrane’s permeability. In addition, we investigated the inhibitory activity towards Plasmodium falciparum glyceraldehyde 3-phosphate dehydrogenase (PfGAPDH) since it is involved in the multitarget mechanism of action of 3-BA. Molecular modeling has shed light on the structural and stereochemical requirements for an efficient interaction with PfGAPDH, leading to covalent irreversible binding and enzyme inactivation. While stereochemistry affects the target binding only for two subclasses (1a–d and 4a–d), it leads to significant differences in the antimalarial activity for all subclasses, suggesting that a stereoselective uptake might be responsible for the enhanced biological activity of the (5S, αS) isomers.


Introduction
Natural products (NPs) have historically played a major role in drug discovery [1] and have been pinpointed as privileged scaffolds for interacting with protein drug targets [2]. The unique chemical diversity and structural complexity of NPs allowed the expansion of the known chemical space explored by medicinal chemists [3]. The majority of NPs are chiral, and they are biosynthesized in an enantiomerically pure fashion [4]. Generally, stereochemistry has a crucial impact on drug action since it affects target binding, metabolism, and distribution. For different compounds classes, stereochemistry is the driver for potency and pharmacokinetics [5]. In addition, it has been shown to affect the protein transport systems, resulting in a stereospecific uptake of drugs, as described for β-lactam antibiotics [6].
The natural compound (5S, αS) acivicin (AT-125, Figure 1A), produced by Streptomyces sviceus, and its synthetic analogue 3-Br-acivicin (3-BA, 1a, Figure 1A) have been described as L-glutamine analogues capable of irreversibly inhibiting several glutamine-dependent amidotransferases, including CTP synthetase (CTPS), carbamoyl phosphate synthetase and XMP aminase [7][8][9], and γ-glutamyl transpeptidase [10]. The acivicin/3-BA covalent mechanism of action involves the nucleophilic attack of an activated catalytic cysteine residue of the target enzyme to the Cl/Br-substituted C-3 of the 4,5-dihydroisoxazole ring ( Figure 1C) [11,12]. Considering the inhibitory activity towards Trypanosoma brucei CTPS, 3-BA was three-fold more potent than acivicin, highlighting that the leaving group plays an important role in the irreversible mode of action. The differences in enzyme inhibition translated into enhanced in vitro and in vivo antitrypanosomal activity for 3-BA [13]. Considering different targets, we had previously reported that acivicin was inactive towards Plasmodium falciparum glyceraldehyde 3-phosphate dehydrogenase (Pf GAPDH), while 3-BA was able to irreversibly inhibit Pf GAPDH, a key enzyme of the Plasmodium glycolytic pathway [14]. In the erythrocytic stages, the malaria parasite relies heavily on anaerobic glycolysis for energy production [15], leading to the identification of glycolytic enzymes as promising targets for the development of antimalarial agents.
Molecules 2023, 28, x FOR PEER REVIEW 2 transport systems, resulting in a stereospecific uptake of drugs, as described for β-la antibiotics [6]. The natural compound (5S, αS) acivicin (AT-125, Figure 1A), produced by Strep ces sviceus, and its synthetic analogue 3-Br-acivicin (3-BA, 1a, Figure 1A) have bee scribed as L-glutamine analogues capable of irreversibly inhibiting several glutamin pendent amidotransferases, including CTP synthetase (CTPS), carbamoyl phosphate thetase and XMP aminase [7][8][9], and γ-glutamyl transpeptidase [10]. The acivicin/ covalent mechanism of action involves the nucleophilic attack of an activated cat cysteine residue of the target enzyme to the Cl/Br-substituted C-3 of the 4,5-dihydro azole ring ( Figure 1C) [11,12]. Considering the inhibitory activity towards Trypano brucei CTPS, 3-BA was three-fold more potent than acivicin, highlighting that the lea group plays an important role in the irreversible mode of action. The differences i zyme inhibition translated into enhanced in vitro and in vivo antitrypanosomal ac for 3-BA [13]. Considering different targets, we had previously reported that acivicin inactive towards Plasmodium falciparum glyceraldehyde 3-phosphate dehydroge (PfGAPDH), while 3-BA was able to irreversibly inhibit PfGAPDH, a key enzyme o Plasmodium glycolytic pathway [14]. In the erythrocytic stages, the malaria parasite heavily on anaerobic glycolysis for energy production [15], leading to the identificati glycolytic enzymes as promising targets for the development of antimalarial agents. We had previously modified the amino acidic portion of 1a and synthesized a s of analogues, keeping the stereochemistry of the natural product unchanged. The (2a, 3a) and amide (4a) derivatives of 3-Br-acivicin were investigated for their antima activity on P. falciparum cultures and for the ability to covalently inhibit PfGAPDH compounds (1a-4a) showed antiparasitic activity against D10 (chloroquine sensitive W2 (chloroquine resistant) strains of P. falciparum, with IC50 being lower than 1 μM (F 1B). The methyl and benzyl esters displayed the highest second-order rate constan act/Ki) used to characterize the covalent binding of irreversible inhibitors to PfGAPDH 10.7; 3a: 6.6 M −1 •s −1 ; Figure 1B) [14,16]. These outcomes clearly indicated that the simi with L-Gln was not essential for the inhibition of GAPDH.
Despite the fact that chemical modifications in the amino acidic portion of 1a been extensively explored [14,[16][17][18], the effect of stereochemistry on the biological r nition processes, protein binding, and antiparasitic activity has not yet been investig Herein, we disclosed the synthesis of the unnatural isomers of 1a-4a, and we explore We had previously modified the amino acidic portion of 1a and synthesized a series of analogues, keeping the stereochemistry of the natural product unchanged. The ester (2a, 3a) and amide (4a) derivatives of 3-Br-acivicin were investigated for their antimalarial activity on P. falciparum cultures and for the ability to covalently inhibit Pf GAPDH. All compounds (1a-4a) showed antiparasitic activity against D10 (chloroquine sensitive) and W2 (chloroquine resistant) strains of P. falciparum, with IC 50 being lower than 1 µM ( Figure 1B). The methyl and benzyl esters displayed the highest second-order rate constant (k inact /K i ) used to characterize the covalent binding of irreversible inhibitors to Pf GAPDH (2a: 10.7; 3a: 6.6 M −1 ·s −1 ; Figure 1B) [14,16]. These outcomes clearly indicated that the similarity with L-Gln was not essential for the inhibition of GAPDH.
Despite the fact that chemical modifications in the amino acidic portion of 1a have been extensively explored [14,[16][17][18], the effect of stereochemistry on the biological recognition processes, protein binding, and antiparasitic activity has not yet been investigated. Herein, we disclosed the synthesis of the unnatural isomers of 1a-4a, and we explored the importance of stereochemistry for the in vitro antimalarial activity and the inhibitory activity towards recombinant Pf GAPDH, which is known to be involved in the multi-target mechanism of 3-BA.

Synthesis of Enantiomerically Pure Unnatural Isomers
The synthesis of the enantiomerically pure target compounds started by using the chiral synthons (S)-12 or (R)-12. The diastereoisomeric mixtures 5 and 6 were obtained through the 1,3-dipolar cycloaddition of bromonitrile oxide to alkenes (S)-12 and (R)-12, respectively, as previously described [19]. Treatment of 5 with a 5:1 mixture of AcOH/H 2 O provided alcohols 7a and 7b (Scheme 1A) that were separated by flash chromatography. Analog conditions led to the synthesis of alcohols 7c and 7d, starting from 6 (Scheme 1B). Alcohols 7a-d were oxidized to the corresponding carboxylic acids 8a-d using Fe(NO 3 ) 3 ·9H 2 O, TEMPO, and KCl under an air flux as an O 2 source [20]. This is a cleaner method compared to the traditional oxidation procedures used to convert alcohols into carboxylic acids, which typically involve the use of chromium-based reagents, and it allowed us to obtain the desired products in high yields (80-88%). The final amino acids 1a-d were obtained after treatment of the N-Boc precursors (8a-d) with a 15% TFA solution in dichloromethane. The zwitterionic compounds were isolated after ion-exchange chromatography employing the DOWEX Marathon C (H + form) resin.
Molecules 2023, 28, x FOR PEER REVIEW importance of stereochemistry for the in vitro antimalarial activity and the inhibito tivity towards recombinant PfGAPDH, which is known to be involved in the multi mechanism of 3-BA.

Synthesis of Enantiomerically Pure Unnatural Isomers
The synthesis of the enantiomerically pure target compounds started by usi chiral synthons (S)-12 or (R)-12. The diastereoisomeric mixtures 5 and 6 were ob through the 1,3-dipolar cycloaddition of bromonitrile oxide to alkenes (S)-12 and respectively, as previously described [19]. Treatment of 5 with a 5:1 mixture of AcOH provided alcohols 7a and 7b (Scheme 1A) that were separated by flash chromatog Analog conditions led to the synthesis of alcohols 7c and 7d, starting from 6 (Schem Alcohols 7a-d were oxidized to the corresponding carboxylic acids 8a-d Fe(NO3)3•9H2O, TEMPO, and KCl under an air flux as an O2 source [20]. This is a c method compared to the traditional oxidation procedures used to convert alcoho carboxylic acids, which typically involve the use of chromium-based reagents, and lowed us to obtain the desired products in high yields (80-88%). The final amino aci d were obtained after treatment of the N-Boc precursors (8a-d) with a 15% TFA so in dichloromethane. The zwitterionic compounds were isolated after ion-exchange matography employing the DOWEX Marathon C (H + form) resin.  Compounds 2b-d, 3b-d, and 4b-d were synthesized following the procedure previously described for the (5S, αS) isomers 2a, 3a, and 4a [16]. Briefly, the N-Boc-protected Molecules 2023, 28, 3172 4 of 20 carboxylic acids 8b-d were treated with trimethylsilyldiazomethane and methanol at room temperature to obtain the methyl esters 9b-d (Scheme 2). The reaction of 8b-d with benzyl bromide under basic conditions at 50 • C afforded the benzyl esters 10b-d. The coupling reaction between the carboxylic acids 8b-d and benzylamine using EDC hydrochloride and HOBt provided the amides 11b-d. Intermediates 9b-d, 10b-d, and 11b-d were treated with a 15% TFA solution in dichloromethane to cleave the Boc-protecting group and provide the free amines 2b-d, 3b-d, and 4b-d, respectively (Scheme 2). r.t., 24 h; (iv) (1) 15% TFA, DCM, r.t., 1 h; (2) column purification with Dowex Marathon C (H + form) resin, 10%Py/H2O (eluent).
Compounds 2b-d, 3b-d, and 4b-d were synthesized following the procedure previously described for the (5S, αS) isomers 2a, 3a, and 4a [16]. Briefly, the N-Boc-protected carboxylic acids 8b-d were treated with trimethylsilyldiazomethane and methanol at room temperature to obtain the methyl esters 9b-d (Scheme 2). The reaction of 8b-d with benzyl bromide under basic conditions at 50 °C afforded the benzyl esters 10b-d. The coupling reaction between the carboxylic acids 8b-d and benzylamine using EDC hydrochloride and HOBt provided the amides 11b-d. Intermediates 9b-d, 10b-d, and 11b-d were treated with a 15% TFA solution in dichloromethane to cleave the Boc-protecting group and provide the free amines 2b-d, 3b-d, and 4b-d, respectively (Scheme 2).

Structure
Compound clogD a P. falciparum D10 IC 50 (µM) b P. falciparum W2 IC 50 (µM) b significant differences between the two strains (D10 and W2, Table 1). While the natural (5S, αS) isomers (1a-4a) were potent antimalarial agents, the isomers possessing the (5R, αR) absolute configuration displayed a moderate antiplasmodial activity (1 < IC50 < 10 μM). The (5S, αR) and (5R, αS) absolute configurations resulted in a huge drop in the antimalarial potency, leading to poorly active compounds (Table 1). For the methyl ester derivatives, the enantiomer of 2a [(5R, αR)-2d] was approximately 10-fold less potent than 2a towards P. falciparum D10 and W2 strains, whereas the diastereoisomers (2b and 2c) were inactive towards both strains (IC50 > 15 μM, Table 1). The same trend was observed for the other subclasses (1a-d, 3a-d, and 4a-d), with the exception of 1c, which presented an IC50 < 10 μM (Table 1). Thus, the stereochemistry led to significant differences in the antimalarial activity, with the natural isomers always being the most potent molecules. Since all the isomers in each subclass (1a-d, 2a-d, 3a-d, and 4a-d) have the same clogD, the differences in antiplasmodial potency could not be correlated to a different passive diffusion across the cell membranes. The natural compound (5S, αS) acivicin is known to compete with amino acids for cellular uptake via transporters. Indeed, acivicin uptake is mediated by the L-amino acid transport system that is responsible for the Na + -independent transport of neutral amino acids [21,22]. The microenvironment of P. falciparum-infected human erythrocytes is controlled by two membranes. Normal mature erythrocytes are relatively impermeable to L-Gln, and therefore also acivicin. The influx of L-Gln and other amino acids is controlled by different transporter systems, each varying their substrate specificity [23,24]. Our results suggested that the stereochemistry might be relevant for recognition by the putative transporters, not only for the amino acid 3-BA (1a), but also for its ester and amide derivatives (2a-4a). Indeed, only the (5S, αS) natural isomers (1a-4a) showed a significant potency in inhibit- significant differences between the two strains (D10 and W2, Table 1). While the natural (5S, αS) isomers (1a-4a) were potent antimalarial agents, the isomers possessing the (5R, αR) absolute configuration displayed a moderate antiplasmodial activity (1 < IC50 < 10 μM). The (5S, αR) and (5R, αS) absolute configurations resulted in a huge drop in the antimalarial potency, leading to poorly active compounds (Table 1). For the methyl ester derivatives, the enantiomer of 2a [(5R, αR)-2d] was approximately 10-fold less potent than 2a towards P. falciparum D10 and W2 strains, whereas the diastereoisomers (2b and 2c) were inactive towards both strains (IC50 > 15 μM, Table 1). The same trend was observed for the other subclasses (1a-d, 3a-d, and 4a-d), with the exception of 1c, which presented an IC50 < 10 μM (Table 1). Thus, the stereochemistry led to significant differences in the antimalarial activity, with the natural isomers always being the most potent molecules. Since all the isomers in each subclass (1a-d, 2a-d, 3a-d, and 4a-d) have the same clogD, the differences in antiplasmodial potency could not be correlated to a different passive diffusion across the cell membranes. The natural compound (5S, αS) acivicin is known to compete with amino acids for cellular uptake via transporters. Indeed, acivicin uptake is mediated by the L-amino acid transport system that is responsible for the Na + -independent transport of neutral amino acids [21,22]. The microenvironment of P. falciparum-infected human erythrocytes is controlled by two membranes. Normal mature erythrocytes are relatively impermeable to L-Gln, and therefore also acivicin. The influx of L-Gln and other amino acids is controlled by different transporter systems, each varying their substrate specificity [23,24]. Our results suggested that the stereochemistry might be relevant for recognition by the putative transporters, not only for the amino acid 3-BA (1a), but also for its ester and amide derivatives (2a-4a). Indeed, only the (5S, αS) natural isomers (1a-4a) showed a significant potency in inhibit- significant differences between the two strains (D10 and W2, Table 1). While the natural (5S, αS) isomers (1a-4a) were potent antimalarial agents, the isomers possessing the (5R, αR) absolute configuration displayed a moderate antiplasmodial activity (1 < IC50 < 10 μM). The (5S, αR) and (5R, αS) absolute configurations resulted in a huge drop in the antimalarial potency, leading to poorly active compounds (Table 1). For the methyl ester derivatives, the enantiomer of 2a [(5R, αR)-2d] was approximately 10-fold less potent than 2a towards P. falciparum D10 and W2 strains, whereas the diastereoisomers (2b and 2c) were inactive towards both strains (IC50 > 15 μM, Table 1). The same trend was observed for the other subclasses (1a-d, 3a-d, and 4a-d), with the exception of 1c, which presented an IC50 < 10 μM (Table 1). Thus, the stereochemistry led to significant differences in the antimalarial activity, with the natural isomers always being the most potent molecules. Since all the isomers in each subclass (1a-d, 2a-d, 3a-d, and 4a-d) have the same clogD, the differences in antiplasmodial potency could not be correlated to a different passive diffusion across the cell membranes. The natural compound (5S, αS) acivicin is known to compete with amino acids for cellular uptake via transporters. Indeed, acivicin uptake is mediated by the L-amino acid transport system that is responsible for the Na + -independent transport of neutral amino acids [21,22]. The microenvironment of P. falciparum-infected human erythrocytes is controlled by two membranes. Normal mature erythrocytes are relatively impermeable to L-Gln, and therefore also acivicin. The influx of L-Gln and other amino acids is controlled by different transporter systems, each varying their substrate specificity [23,24]. Our results suggested that the stereochemistry might be relevant for recognition by the putative transporters, not only for the amino acid 3-BA (1a), but also for its ester and amide derivatives (2a-4a). Indeed, only the (5S, αS) natural isomers (1a-4a) showed a significant potency in inhibit- significant differences between the two strains (D10 and W2, Table 1). While the natural (5S, αS) isomers (1a-4a) were potent antimalarial agents, the isomers possessing the (5R, αR) absolute configuration displayed a moderate antiplasmodial activity (1 < IC50 < 10 μM). The (5S, αR) and (5R, αS) absolute configurations resulted in a huge drop in the antimalarial potency, leading to poorly active compounds (Table 1). For the methyl ester derivatives, the enantiomer of 2a [(5R, αR)-2d] was approximately 10-fold less potent than 2a towards P. falciparum D10 and W2 strains, whereas the diastereoisomers (2b and 2c) were inactive towards both strains (IC50 > 15 μM, Table 1). The same trend was observed for the other subclasses (1a-d, 3a-d, and 4a-d), with the exception of 1c, which presented an IC50 < 10 μM (Table 1). Thus, the stereochemistry led to significant differences in the antimalarial activity, with the natural isomers always being the most potent molecules. Since all the isomers in each subclass (1a-d, 2a-d, 3a-d, and 4a-d) have the same clogD, the differences in antiplasmodial potency could not be correlated to a different passive diffusion across the cell membranes. The natural compound (5S, αS) acivicin is known to compete with amino acids for cellular uptake via transporters. Indeed, acivicin uptake is mediated by the L-amino acid transport system that is responsible for the Na + -independent transport of neutral amino acids [21,22]. The microenvironment of P. falciparum-infected human erythrocytes is controlled by two membranes. Normal mature erythrocytes are relatively impermeable to L-Gln, and therefore also acivicin. The influx of L-Gln and other amino acids is controlled by different transporter systems, each varying their substrate specificity [23,24]. Our results suggested that the stereochemistry might be relevant for recognition by the putative transporters, not only for the amino acid 3-BA (1a), but also for its ester and amide derivatives (2a-4a). Indeed, only the (5S, αS) natural isomers (1a-4a) showed a significant potency in inhibit- For the methyl ester derivatives, the enantiomer of 2a [(5R, αR)-2d] was approximately 10-fold less potent than 2a towards P. falciparum D10 and W2 strains, whereas the diastereoisomers (2b and 2c) were inactive towards both strains (IC 50 > 15 µM, Table 1). The same trend was observed for the other subclasses (1a-d, 3a-d, and 4a-d), with the exception of 1c, which presented an IC 50 < 10 µM (Table 1). Thus, the stereochemistry led to significant differences in the antimalarial activity, with the natural isomers always being the most potent molecules. Since all the isomers in each subclass (1a-d, 2a-d, 3a-d, and 4a-d) have the same clogD, the differences in antiplasmodial potency could not be correlated to a different passive diffusion across the cell membranes. The natural compound (5S, αS) acivicin is known to compete with amino acids for cellular uptake via transporters. Indeed, acivicin uptake is mediated by the L-amino acid transport system that is responsible for the Na + -independent transport of neutral amino acids [21,22]. The microenvironment of P. falciparum-infected human erythrocytes is controlled by two membranes. Normal mature erythrocytes are relatively impermeable to L-Gln, and therefore also acivicin. The influx of L-Gln and other amino acids is controlled by different transporter systems, each varying their substrate specificity [23,24]. Our results suggested that the stereochemistry might be relevant for recognition by the putative transporters, not only for the amino acid 3-BA (1a), but also for its ester and amide derivatives (2a-4a). Indeed, only the (5S, αS) natural isomers (1a-4a) showed a significant potency in inhibiting P. falciparum proliferation. To investigate whether the different antiplasmodial activities could be related also to a distinct interaction of the isomers with biological targets, we tested the inhibitory activity towards Pf GAPDH, which is a known target for 1a-4a [16].

In Vitro Inhibitory Activity towards PfGAPDH
3-Br-acivicin (1a) biological activity is correlated to its ability to inhibit multiple targets. Among these, Pf GAPDH had been previously suggested to be involved in the antimalarial effect of 1a and other structurally related compounds [16]. For this reason, the four diastereoisomers of each compound were investigated for their ability to inhibit the Pf GAPDH activity. All molecules were screened at a concentration of 100 µM after a 3 h incubation time. While 3-Br-acivicin (1a) produced a Pf GAPDH inhibition higher than 50%, its enantiomer 1d and diastereoisomers (1b and 1c) were inactive (Figure 2A). A similar trend was observed for the amide-bearing isomers 4a compared to 4b-d. In contrast, all four isomers of the methyl (2a-d) and benzyl (3a-d) ester derivatives were able to fully inhibit Pf GAPDH at a 100 µM concentration, leading to a residual Pf GAPDH activity lower than 1% (Figure 2A). Therefore, compounds 2a-d and 3a-d were selected for a further investigation and tested at 50 µM for 30 min. All diastereoisomers (2a-d and 3a-d) showed enzyme inhibition >50%, with the natural isomer 2a being the most active (37% residual activity at 50 µM; Figure 2A).

Molecular Modeling
Compounds 1a-1d, 2b-2d, 3b-3d, and 4b-4d were subjected to conformational analysis and flexible docking simulations in a complex with PfGAPDH using the same computational procedure previously applied to 2a, 3a, and 4a [16]. The present results (Tables S1-S38) were integrated with those previously obtained, analyzed, and related to the inhibitory activities of the compounds. The aim was to simulate the binding of the compounds to PfGAPDH just before undergoing nucleophilic attack by the catalytic cysteine residue (C153). According to the flip-flop model proposed for the catalytic reaction mechanism of GAPDH, the substrate rotates around C153, moving the C-3 phosphate from a first (PI) to a second (PII) interaction site [25,26]. In line with these findings, different starting complexes were considered and two possible binding approaches to C153, named BA1 and BA2, were obtained for each compound, which corresponded to the approach from the PI or PII site, respectively (Figures 3 and 4; Tables S15-S32). The covalent mechanism of action of 2a and 3a had been already characterized though mass spectrometry, using undigested and trypsin-digested Pf GAPDH after incubation with the tested compounds [14]. We also proved the selective modification of the catalytic Cys153 with no involvement of other Cys residues [14]. In addition, the k inact /K i ratio typically used to characterize the covalent binding of irreversible inhibitors to the target protein was measured for compounds 1a-4a [16]. Therefore, we can assume that the unnatural isomers 2b-4b, 2c-4c, and 2d-4d displayed the same irreversible mechanism of inhibition observed for their corresponding natural isomers (5S, αS) 1a-4a. Compounds 1a-1d, 2b-2d, 3b-3d, and 4b-4d were subjected to conformational analysis and flexible docking simulations in a complex with Pf GAPDH using the same computational procedure previously applied to 2a, 3a, and 4a [16]. The present results (Tables S1-S38) were integrated with those previously obtained, analyzed, and related to the inhibitory activities of the compounds. The aim was to simulate the binding of the compounds to Pf GAPDH just before undergoing nucleophilic attack by the catalytic cysteine residue (C153). According to the flip-flop model proposed for the catalytic reaction mechanism of GAPDH, the substrate rotates around C153, moving the C-3 phosphate from a first (P I ) to a second (P II ) interaction site [25,26]. In line with these findings, different starting complexes were considered and two possible binding approaches to C153, named BA1 and BA2, were obtained for each compound, which corresponded to the approach from the P I or P II site, respectively (Figures 3 and 4; Tables S15-S32).    The geometry and the conformational energy of the docked ligands were compared to those of the conformers obtained by the conformational analysis (Table S33). The solvent-accessible surface area (SASA) of the leaving group (bromine atom) was calculated (Table S34), and the ligand-protein interactions were analyzed (Tables S35-S38).

Molecular Modeling
It resulted that the active isomers 2a-2d could approach C153 assuming both binding modes. In BA1 (Figures 3A and S1), the protonated amine function and the ester function interacts with the PI site, while the bromine atom points towards NAD + .
In BA2 (Figures 3B and S2), the protonated amine and the ester function are located at the PII site, and the bromine atom is positioned at the PI site pointing towards the activesite segment. Considering the orientation of the leaving group, this binding approach is favored by a larger solvent-accessible surface area (SASA) of the bromine atom (Table S34). However, assuming BA1 ( Figures 3A and S1), the bromine atom could leave the active site together with NAD+, similarly to the hydride ion of the substrate. The putative binding conformations of 2a-2d resulted all within 5 kcal/mol from the global energy minimum (GM) ( Table S33). (green), and 1d (magenta) approaching from the P I site (BA1; A) or P II site (BA2; B); (ii) 4a (orange), 4b (yellow), 4c (green), and 4d (magenta) approaching from the P I site (BA1; C) or P II site (BA2; D). All structures are superimposed on the starting Pf GAPDH conformation by fitting Cα atoms. The starting protein conformation (gray) is displayed as solid ribbons and the Connolly surface of the active site is shown. The backbone of the docked complexes is in line ribbons, key interaction residues are displayed and labeled, and heteroatoms are colored by atom type (N, blue; O, red; S, yellow; Br, brown). The van der Waals volume of the sulfur atom of C153 is scaled by 50%. Hydrogen atoms are omitted for clarity, except those involved in ligand-protein hydrogen bond interactions (black dashed lines).
The geometry and the conformational energy of the docked ligands were compared to those of the conformers obtained by the conformational analysis (Table S33). The solvent-accessible surface area (SASA) of the leaving group (bromine atom) was calculated (Table S34), and the ligand-protein interactions were analyzed (Tables S35-S38).
It resulted that the active isomers 2a-2d could approach C153 assuming both binding modes. In BA1 (Figures 3A and S1), the protonated amine function and the ester function interacts with the P I site, while the bromine atom points towards NAD + .
In BA2 (Figures 3B and S2), the protonated amine and the ester function are located at the P II site, and the bromine atom is positioned at the P I site pointing towards the activesite segment. Considering the orientation of the leaving group, this binding approach is favored by a larger solvent-accessible surface area (SASA) of the bromine atom (Table S34). However, assuming BA1 ( Figures 3A and S1), the bromine atom could leave the active site together with NAD + , similarly to the hydride ion of the substrate. The putative binding conformations of 2a-2d resulted all within 5 kcal/mol from the global energy minimum (GM) (Table S33).
Similar results were obtained for the active diastereoisomers 3a-3d, although they were characterized by a more hindered benzyl moiety compared to the methyl group of 2a-2d ( Figures 3C,D, S3 and S4). In particular, assuming BA1, the phenyl ring is positioned in the large cleft between the S-loop (aa182-210) and the S7-S8 loop (aa122-130), being slightly differently oriented according to the stereochemistry of the C5 carbon (occupying the P II site in the 5S diastereoisomers 3a and 3c; Figure 3C). On the other hand, in the BA2, the phenyl ring is positioned according to the stereochemistry of the Cα carbon being oriented toward the P I site or the P II site in the αR (3c, 3d) and αS (3a, 3b) diastereoisomers, respectively ( Figure 3D).
On the other hand, the results obtained for 1a-1d suggest that a salt-bridge interaction between the negatively charged carboxylic group of the ligands and R237 of Pf GAPDH (involved in phosphate binding and translocation) could be responsible for the drastic reduction in the inhibitory activity of 1b-1d. Indeed, when the carboxylic group is positioned at the P I site (BA1; Figures 4A and S5), the salt-bridge with R237 is responsible for ensuring the electrophilic carbon of the 3-bromo-4,5-dihydroisoxazole ring (C3) stays away from the sulfur atom of C153 (>4 Å; 1b and 1c) or, in the case of 1d, placing the 4,5-dihydroisoxazole nitrogen atom between the C3 carbon and the C153 sulfur atom. In any case, such ionic interaction hampers the approach to C153 and the formation of the carbon sulfur bond (~1.80 Å).
When the carboxylic group is positioned in the P II site (BA2; Figures 4B and S6), the diastereoisomers 1b-1d still present the salt-bridge with R237 together with a chargeassisted hydrogen bond interaction with T214, drastically reducing the SASA of the bromine atom positioned at P I (Table S34). Importantly, the carboxylic group of 1a (the only active diastereoisomer) is not involved in any interaction with R237, and, accordingly, the SASA of the bromine atom is not reduced (Tables S34 and S35). In line with what was observed for the carboxylic group, for the amide derivatives 4b and 4d, the introduction of the NH group is responsible for additional interactions with the protein, which move the ligand away from C153 ( Figure 4C,D; Tables S27-S30; Figures S7 and S8). As previously reported, this is also true for 4a (the only active diastereoisomer), but only assuming the BA2 [16]. Finally, diastereoisomer 4c docked conformations are highly distorted in both binding approaches ( Figures S9 and S10) and show an energy difference from the global minimum conformer higher than 15 kcal/mol (Table S33). Therefore, the computational results are consistent with Pf GAPDH inhibitory activities and provide useful molecular models to drive future structural modifications.

General
All reagents, solvents, and starting materials were purchased from Sigma-Aldrich (Italy), Fluorochem, or TCI Europe. The diastereoisomeric mixtures 5 and 6 were prepared as previously described [19]. Compounds 1a, 2a, 3a, and 4a were prepared from intermediate 8a as previously described [13,14,16]. 1 H NMR and 13 C NMR spectra were recorded with a Varian Mercury 300 (300 MHz) spectrometer. NMR spectra were obtained in deuterated solvents, such as CDCl 3 or D 2 O. The chemical shift (δ values) is reported in ppm and corrected to the signal of the deuterated solvents. Peak multiplicities are reported as: s (singlet), d (doublet), dd (doublet of doublets), ddd (doublet of doublet of doublets), t (triplet), dt (doublet of triplets), q (quartet), m (multiplet), and br (broad). Chemical shifts (δ) are expressed in ppm and coupling constants (J) in Hertz (Hz). Thin layer chromatography (TLC) plates were purchased from Sigma-Aldrich Italy (silica gel 60 F254 aluminum sheets, with fluorescence indicator 254 nm); a dilute alkaline solution of KMnO 4 or a ninhydrin solution were used to visualize the compounds. Flash chromatography was performed using silica gel, with a pore size of 60 Å, and a mesh particle size of 230-400. Specific rotations ([α] D T ) were calculated based on the optical rotation measurements obtained with a Jasco P1010 polarimeter coupled with a Haake N3-B thermostat.

General Procedure B
Compounds 8b-d (1.0 eq) were treated with a 15% DCM solution of trifluoroacetic acid (10 eq) at 0 • C, and the solution was stirred at rt for 1 h. The volatiles were removed under vacuum and the crude was purified by ion-exchange chromatography using Dowex Marathon C (H + form) resin, as well as elution with a 10% solution of pyridine in water. The solvent was evaporated under a reduced pressure to obtain the compounds 1b-d. 3.1.4. General Procedure C TMSCHN 2 2.0 M in hexane (2 eq) was added dropwise to a cooled solution of 8b-d (1 eq) in toluene (10 mL/mmol) and MeOH (0.4 mL/mmol). After stirring at room temperature for 1 h, the mixture was concentrated under reduced pressure, and purification using silica gel column chromatography afforded the compounds 9b-d.

General Procedure D
Compounds 8b-d (1 eq) were dissolved in DMF (6 mL/mmol). Benzyl bromide (1.2 eq) and KHCO 3 (1.2 eq) were added to the solution. The reaction mixture was stirred at 50 • C for 1 h. The reaction mixture was diluted with EtOAc and the organic layer was washed with 1N HCl, 5% NaHCO 3 , and brine. The organic layer was dried over anhydrous Na 2 SO 4 , filtered, and evaporated to reach dryness. Purification using silica gel column chromatography afforded the compounds 10b-d.

General Procedure E
Benzylamine (1 eq), EDC hydrochloride (1 eq), and HOBt (0.5 eq) were added to a solution of 8b-d (1 eq) in dry THF (35 mL/mmol). The reaction mixture was stirred at room temperature for 2 h, then the solvent was removed under reduced pressure. Purification using silica gel column chromatography afforded the compounds 11b-d.

Conformational Analysis
The molecular models of the new compounds of 1a-1d, 2b-2d, 3b-3d, and 4b-4d were built (Small Molecule tool of Discovery Studio 2017; Dassault Systèmes BIOVIA, San Diego, CA, USA, 2017), atomic potentials and charges were assigned using the CFF forcefield [27]. The resulting structures were subjected to molecular mechanic (MM) energy minimization (ε = 80 * r) until the maximum RMS derivative was less than 0.001 kcal/Å, using the conjugate gradient as the minimization algorithm [28]. The obtained conformers were used as starting structures for the subsequent systematic conformational analysis (Search Small Molecule Conformations; Discovery Studio 2017). The conformational space was sampled by systematically varying the rotatable bonds sp3-sp3 and sp3-sp2 with an increment of 60 • . The RMSD cutoff for the structure selection was set to 0.01 (Å). Finally, to ensure a wide variance of the input structures to be successively fully minimized, an energy threshold value of 10 6 kcal/mol was used as the selection criteria. The generated structures were then subjected to MM energy minimization (CFF forcefield; ε = 80 * r) until the maximum RMS derivative was less than 0.001 kcal/Å, using the conjugate gradient as the minimization algorithm. Finally, the resulting conformers were ranked by their potential energy values (i.e., ∆E from the global energy minimum). The conformers within 5 kcal/mol from the global minimum were classified on the basis of dihedral angle values.

Docking Studies
Docking calculations were performed by using as a protein structure the previously developed atomic model of Pf GAPDH [16]. Although in the docking simulation all the systems were perturbed by a Monte Carlo/minimization procedure, the docking procedure formally requires a reasonable starting structure. In order to define the starting conformation of the new compounds, all the conformers within 5 kcal/mol from the global minimum were placed in the GAPDH catalytic site considering the two binding modes of the glyceraldehyde 3-phosphate analogue 2-(2-phosphono-ethyl)-acrylic acid 4-nitro-phenyl ester (PDB ID: 1ML3). The conformations with the lowest potential energy that did not show a significant steric overlap with catalytic-site amino acids were selected as the starting conformations for the docking calculations.
A docking procedure, which considers all the systems as flexible (i.e., ligand and protein), was applied. Flexible docking was achieved using the Affinity module in the Insight 2005 suite, setting the SA_Docking procedure [29] and using the cell multipole method for non-bonded interactions [30]. The docking procedure included a Monte-Carlo(MC)-based conformational search of the ligand within the active site of Pf GAPDH. During the first step, starting from the previously obtained roughly docked structures, the ligand was moved by a random combination of translation, rotation, and torsional changes to sample both the conformational space of the ligand and its orientation with respect to the protein (MxRChange = 3 Å; MxAngChange = 180 • ). During this step, the van der Waals (vdW) and Coulombic terms were scaled to a factor of 0.1 to avoid very severe divergences in the vdW and Coulombic energies. If the energy of a complex structure resulting from the random moves of the ligand was higher by the energy tolerance parameter than the energy of the last accepted structure, it was not accepted for minimization. To ensure a wide variance of the input structures to be successively minimized, an energy tolerance value of 10 6 kcal/mol from the previous structure was used. After the energy minimization step (conjugate gradient; 2500 iterations; ε = 1), the energy test, with an energy range of 50 kcal/mol, and a structure similarity check (rms tolerance = 0.3 kcal/Å) was applied to select the 20 acceptable structures. Each subsequent structure was generated from the last accepted structure.
All Pf GAPDH atoms were left free to move during the entire course of the docking calculations, whereas, in order to avoid unrealistic results, a tethering restraint was applied on the structurally conserved regions (SCRs) of the protein. To identify the SCRs, the Pf GAPDH sequence was analyzed using the structure prediction and sequence analysis server PredictProtein (http://www.predictprotein.org/ (accessed on 12 September 2022).
For each compound, the resulting complexes were superimposed on the starting structure by fitting all the Cα atoms, and the Cα RMSD of each residue and its average value were calculated. Then, the complexes were again superimposed on the starting structure by fitting the Cα atoms of the residues characterized by an average value of RMSD ≤ 0.2 Å. Considering this latter superimposition, the Cα RMSD of the catalytic residues and the RMSD of NAD + were calculated. The χ1 torsion angle of C153 and the geometric criteria of the hydrogen bond between C153 and H180 were also evaluated for each generated complex. In particular, the angle D-H-A and X-D-A of this hydrogen bond was calculated, assuming as D the sulfur atom of C153, as A the Nτ hydrogen atom of H180, and as X the Cβ of C153.
Finally, for each generated complex, the non-bonded interaction energy (vdW and electrostatic energy contribution; group-based method [33]; CUT_OFF = 100; ε = 2 * r; Discover_3 Module of Insight 2005) was calculated. In our previous publication, the docked complexes of 2a, 3a, and 4a were not analyzed using these criteria [16]. Accordingly, to properly compare the results obtained for new diastereoisomers with those obtained with 2a, 3a, and 4a, the latter were also included in this analysis.
For each docking calculation, the complex with the most favorable interaction energy was characterized by (i) the Cα RMSD of C152 and H179 with respect to the starting structure ≤ 3 Å; (ii) the gauche(-) conformation (from −30 • to −90 • ) of the torsion angle χ1 of C153 (i.e., the conformation needed to establish the hydrogen bond with H180); and (iii) the angles of the hydrogen bond between C153 and H180 > 90 • [34] were selected.
The selected docked complexes were then subjected to MM energy minimization, applying only the restraint on the hydrogen bond between the catalytic residues C153 and H180 (RMS derivative < 0.5 kcal/Å; Steepest Descent algorithm; ε = 80 * r; Module Discover; Insight 2005). The optimized complexes were again filtered by using, in addition to the above reported criteria, the distance between the electrophilic carbon C3 of the 3-bromo-4,5-dihydroisoxazole ring and the sulfur atom of C153 ≤ 4 Å.
The complex with the most favorable interaction energy meeting the filtering criteria was chosen as the structure representing the most probable calculated approach of the compounds to the catalytic cysteine of Pf GAPDH. The quality of the selected docked complexes was checked using Procheck structure evaluator software [35].

Expression and Purification of Pf GAPDH
Recombinant His-tagged Pf GAPDH was produced in Escherichia coli, as has already been described [14].

Enzyme Assays
The GAPDH activity was evaluated using a modified version of the Ferdinand assay in a buffered solution containing 10 mM of TEA, 10 mM of sodium arsenate, 5 mM of EDTA, 1.5 mM of NAD + , and 2.2 of mM DL-glyceraldehyde 3-phosphate, as has already been described [14]. GAPDH was added at a final concentration of 33 nM, and the NADH formation was monitored at 340 nm using a Cary4000 spectrophotometer (Agilent Technologies) with the cell holder maintained at 25 • C.

Parasite Growth and Drug Susceptibility Assay
The CQ-sensitive (D10) and CQ-resistant (W2) strains of P. falciparum were sustained in vitro, as described by Trager and Jensen [36,37]. All the strains were cultured at 5% hematocrit (human type A-positive red blood cells) in RPMI 1640 (EuroClone, Celbio) medium with the addition of 1% AlbuMax (Invitrogen, Milan, Italy), 0.01% hypoxanthine, 20 mM of Hepes Buffer, and 2 mM of glutamine. Parasites were maintained at 37 • C in a standard gas mixture consisting of 1% O 2 , 5% CO 2 , and 94% N 2 . For the drug sensitivity assay, the compounds were dissolved in DMSO and then diluted with medium to achieve the required concentrations (final DMSO concentration < 1%, which is nontoxic to the parasite). The drugs were placed in 96-well flat-bottom microplates (COSTAR), and serial dilutions were made. Asynchronous cultures with a parasitemia of 1-1.5% and 1% final hematocrit were added into the plates and incubated for 72 h at 37 • C. The parasite growth was determined spectrophotometrically (OD650) by measuring the activity of the parasite lactate dehydrogenase (pLDH), according to a modified version of Makler's method in control and drug-treated cultures [38]. The antiplasmodial activity was expressed as 50% inhibitory concentrations (IC 50 ). Each IC 50 value is the mean ± standard deviation of at least three separate experiments performed in duplicate.

Conclusions
Herein, we investigated the importance of stereochemistry for the biological activity of nature-inspired 3-Br-acivicin and its derivatives. For each molecule, the four isomers were prepared and evaluated in vitro towards P. falciparum. Only the (5S, αS) natural isomers showed a significant antimalarial activity, suggesting that their uptake might be mediated by the L-amino acid transport system. Considering the Pf GAPDH inhibition, the most potent compounds turned out to be the ester derivatives 2a-2d and 3a-3d, with no significant differences among the diastereoisomers. In contrast, stereochemistry affected the target binding for the other two subclasses (1a-1d and 4a-4d). A clear correlation between the Pf GAPDH inhibitory activity and the antimalarial potency was not observed; therefore, additional targets might be involved in the biological effects. The racemic form of the 3-bromo-4,5-dihydroisoxazole scaffold had been exploited in chemoproteomic profiling methods to engage reactive cysteine residues in the human proteome [39]. Our enantiomerically pure 3-bromo-4,5-dihydroisoxazole derivatives represent valuable tools for proteomic experiments on malaria parasite P. falciparum, aiming to reveal new ligandable sites in proteins and to guide a future structure-based drug design.

Funding:
The authors are grateful to the UNIMI GSA-IDEA project for partial financial support. C.B. was supported by "L'Oréal Italia for Women in Science" fellowship.
Data Availability Statement: Not applicable.

Conflicts of Interest:
The authors declare no conflict of interest.