Flavonoid Derivatives as New Potent Inhibitors of Cysteine Proteases: An Important Step toward the Design of New Compounds for the Treatment of Leishmaniasis

Leishmaniasis is a neglected tropical disease, affecting more than 350 million people globally. However, there is currently no vaccine available against human leishmaniasis, and current treatment is hampered by high cost, side-effects, and painful administration routes. It has become a United Nations goal to end leishmaniasis epidemics by 2030, and multitarget drug strategy emerges as a promising alternative. Among the multitarget compounds, flavonoids are a renowned class of natural products, and a structurally diverse library can be prepared through organic synthesis, which can be tested for biological effectiveness. In this study, we synthesised 17 flavonoid analogues using a scalable, easy-to-reproduce, and inexpensive method. All synthesised compounds presented an impressive inhibition capacity against rCPB2.8, rCPB3, and rH84Y enzymes, which are highly expressed in the amastigote stage, the target form of the parasite. Compounds 3c, f12a, and f12b were found to be effective against all isoforms. Furthermore, their intermolecular interactions were also investigated through a molecular modelling study. These compounds were highly potent against the parasite and demonstrated low cytotoxic action against mammalian cells. These results are pioneering, representing an advance in the investigation of the mechanisms behind the antileishmanial action of flavonoid derivatives. Moreover, compounds have been shown to be promising leads for the design of other cysteine protease inhibitors for the treatment of leishmaniasis diseases.


Introduction
Leishmaniasis comprises a group of vector-borne infectious diseases with a broad clinical spectrum. Classified as a neglected disease by the World Health Organisation (WHO), leishmaniasis affects more than 350 million people worldwide [1,2]. Although it represents a serious public health problem, there is still no vaccine for humans. In addition, Leishmaniasis comprises a group of vector-borne infectious diseases with a broad clinical spectrum. Classified as a neglected disease by the World Health Organisation (WHO), leishmaniasis affects more than 350 million people worldwide [1,2]. Although it represents a serious public health problem, there is still no vaccine for humans. In addition, the drugs used in therapy are expensive and highly toxic, they cause numerous side-effects, and the routes of administration are painful. Consequently, adherence to treatment is impaired, further strengthening the disease cycle, especially in developing countries [3].
The serious situation involving leishmaniasis resulted in the United Nations (UN) setting a goal in 2020 to combat this group of diseases: sustainable development goal (SDG) 3.3 aims to end the epidemics of several diseases, including neglected tropical diseases (NTDs) by 2030 [4]. This objective has made the discovery and development of new antileishmanial drugs that are more effective, cheaper, easily obtainable, and capable of being administered via alternative routes an emergency. To achieve this goal, different drug discovery approaches need to be used. The multitarget drug strategy has emerged in the last few decades; this approach is based on the complexity of the pathologies and considers that single-target drugs are insufficient to achieve the desired therapeutic effects [5]. Recently, the multitarget drug strategy was reported as a tool to accelerate the discovery of safer, more active, and patient-compliant drugs for the treatment of leishmaniasis [6].
In countries rich in biodiversity, such as Brazil, the use of secondary metabolites from natural sources as new prototypes is a compelling alternative. Flavonoids stand out as they comprise one of the most diverse groups of secondary metabolites, marked by their wide distribution in plants and different therapeutic potentials [7]. These compounds are structurally versatile due to their chemical core and are considered important prototypes for the development of multitarget drugs (Figure 1). It is essential to note that flavonoids have shown in vitro and in vivo antileishmanial activity [8,9]. Flavonols such as quercetin and fisetin inhibit the arginase enzyme, as well as modulate the host's immune response against the parasite, resulting in low patient toxicity [8]. However, the process of isolating and purifying these compounds is expensive and time-consuming. These methodologies generally require high-cost equipment, result in low yields, and use toxic solvents, reducing the sustainability of the process. In line with our interest in discovering and developing new sustainable, efficient methodologies for biologically active compounds and a patient-compliant alternative for leishmaniasis treatment [10][11][12][13][14][15][16][17], our research group aimed to yield natural-based bioactive compounds with multitarget properties. To achieve this purpose, we designed a scalable, easy-to-replicate, and inexpensive synthetic route to obtain flavonol and chalcone analogues. The antileishmanial activity was elucidated using in vitro and in silico methods, and the cytotoxicity was measured to determine the selectivity index (SI).

Chemistry
Our retrosynthetic analysis was based on green chemistry principles. Thus, we used mild reaction conditions and inexpensive catalysts and reagents ( Figure 2). The chalcone In line with our interest in discovering and developing new sustainable, efficient methodologies for biologically active compounds and a patient-compliant alternative for leishmaniasis treatment [10][11][12][13][14][15][16][17], our research group aimed to yield natural-based bioactive compounds with multitarget properties. To achieve this purpose, we designed a scalable, easy-to-replicate, and inexpensive synthetic route to obtain flavonol and chalcone analogues. The antileishmanial activity was elucidated using in vitro and in silico methods, and the cytotoxicity was measured to determine the selectivity index (SI).

Chemistry
Our retrosynthetic analysis was based on green chemistry principles. Thus, we used mild reaction conditions and inexpensive catalysts and reagents ( Figure 2). The chalcone analogue synthesis consists of a Claisen-Schmidt reaction, which is a condensation of aldehydes and carbonyl compounds, leading to α,β-unsaturated ketones in the presence of a base or Lewis acid [18]. In particular, the use of a base as a catalyst provides higher yields of flavonol-like compounds. The intramolecular H-bond of o-hydroxyacetophenones leads to an increase in acidity of its α-hydrogen and, in the presence of a base, aids in the generation of a strongly attacking enol anion [19]. Chalcone analogues were used in the synthesis of flavonol-like molecules through the Algar Flynn-Oyamada reaction, in which a chalcone undergoes an oxidative cyclisation under alkaline conditions to form a flavonol [20]. of a base or Lewis acid [18]. In particular, the use of a base as a catalyst provides higher yields of flavonol-like compounds. The intramolecular H-bond of ohydroxyacetophenones leads to an increase in acidity of its α-hydrogen and, in the presence of a base, aids in the generation of a strongly attacking enol anion [19]. Chalcone analogues were used in the synthesis of flavonol-like molecules through the Algar Flynn-Oyamada reaction, in which a chalcone undergoes an oxidative cyclisation under alkaline conditions to form a flavonol [20]. To obtain inexpensive bioactive compounds, we selected low-cost benzaldehydes and acetophenones. The number of synthetic compounds and the structural variations were designed to be statistically relevant for the elucidation of the structure-activity relationship (SAR). Among acetophenones, we also used halogenated ohydroxyacetophenones, as these substituents are reported for their high and selective antimicrobial actions, including antileishmanial activity [21,22]. Previously described aspects of the synthetic process were reviewed to obtain moderate to high yields of the target compounds (Table 1). Our protocol can also be applied to significantly larger reaction mixtures, and this scale-up also gives good yields.  To obtain inexpensive bioactive compounds, we selected low-cost benzaldehydes and acetophenones. The number of synthetic compounds and the structural variations were designed to be statistically relevant for the elucidation of the structure-activity relationship (SAR). Among acetophenones, we also used halogenated o-hydroxyacetophenones, as these substituents are reported for their high and selective antimicrobial actions, including antileishmanial activity [21,22]. Previously described aspects of the synthetic process were reviewed to obtain moderate to high yields of the target compounds (Table 1). Our protocol can also be applied to significantly larger reaction mixtures, and this scale-up also gives good yields. hydroxyacetophenones leads to an increase in acidity of its α-hydrogen and, in the presence of a base, aids in the generation of a strongly attacking enol anion [19]. Chalcone analogues were used in the synthesis of flavonol-like molecules through the Algar Flynn-Oyamada reaction, in which a chalcone undergoes an oxidative cyclisation under alkaline conditions to form a flavonol [20]. To obtain inexpensive bioactive compounds, we selected low-cost benzaldehydes and acetophenones. The number of synthetic compounds and the structural variations were designed to be statistically relevant for the elucidation of the structure-activity relationship (SAR). Among acetophenones, we also used halogenated ohydroxyacetophenones, as these substituents are reported for their high and selective antimicrobial actions, including antileishmanial activity [21,22]. Previously described aspects of the synthetic process were reviewed to obtain moderate to high yields of the target compounds (Table 1). Our protocol can also be applied to significantly larger reaction mixtures, and this scale-up also gives good yields. Higher yields of chalcones were achieved when halogenated o-hydroxyacetophenones were used. Substituting a halogen atom in the phenolic ring results in an increase in the acidity of the hydroxyl group, which further favours the formation of the enol anion over the formation of unsubstituted o-hydroxyacetophenones [23]. Although chalcones exist as trans (E) or cis (Z) isomers, the E isomer is more stable from the thermodynamic perspective [24]. The configuration of the Z isomer is unstable as a result of the strong steric effects between the carbonyl group and the A-ring, making the E isomer the predominant configuration obtained in our study.
The compounds were unambiguously characterised by NMR spectra. Chalcones 4a-4c presented most of the 13 C-NMR signals as doublets due to 1 J, 2 J, and 3 J fluorine (spin 1 /2) coupling with the respective carbons in the aromatic ring. Compared to the starting materials, the chalcones presented CH group signals at 7.30-7.90 ppm, with coupling constants (J) of 15.3 Hz, confirming the formation of the E isomer. These signals were not observed in the flavonol spectra. Furthermore, compounds f12a-f13c presented an OH group signal at 9.60-9.30 ppm, indicating the loss of 2-hydroxy from the starting material (chalcone), classically observed with a larger chemical shift due to the intramolecular H bond with the carbonyl group. The NMR data were compared with the literature, and the structures of the synthesised compounds were confirmed [25][26][27][28][29][30][31][32][33][34][35][36][37]. The spectra are available in the Supplementary Material. The discovery and development of new antileishmanial drugs is usually directed through phenotypic or target-based approaches [2]. The target-based strategy is based on previous evidence of the action of compounds on specific pharmacological targets. Driven by advances in molecular biology and the urgency to discover new effective drugs, this approach has been the dominant tool in the last three decades [38]. However, most of the molecules designed by the target-based strategy only demonstrated antipromastigote effects, although amastigote is the target form of the parasite, since this stage occurs in mammalian host cells [39].
Leishmania proteases stand out in this context. In particular, cysteine protease B (CPB) expression is elevated in the amastigote stage and plays an important role in the interaction between the parasite and its mammalian host [40]. The results of the inhibition of these proteases indicate their influence on macrophage infection and amastigote survival in host cells, as well as modulating the host's immune response [41]. The isolated bioflavonoids and their semisynthetic derivatives have demonstrated satisfactory activity against the isoforms rCPB2.8 and rCPB3 [42]; thus, these enzymes represent promising therapeutic targets for our study.
In the screening of compounds with rCPB3, it was found that chalcone analogues 1b, 2a, and 2b reduced the enzymatic activity by 67.01%, 70.92%, and 62.55%, respectively, at a concentration of 5 µM. Flavonols were unable to inhibit the enzyme by more than 50% at concentrations of 1 µM or 5 µM ( Figure 3B). In the inhibition of rH84Y with compounds at 5 µM, it was observed that chalcones 1b, 1c, and 2c reduced activity by 73.61%, 82.29%, and 70.77%, respectively, while the flavonoid f12c reduced activity by 56%. Only compound 2c was able to inhibit more than 50% at 1 µM ( Figure 3C).
Since all chalcones and flavonol analogues inhibited the three enzymes to some degree at a concentration of 5 µM, all were subjected to assays to determine their inhibitory potential (IC 50 ) against rCPB2.8 and isoforms.
Since all chalcones and flavonol analogues inhibited the three enzymes to some degree at a concentration of 5 µM, all were subjected to assays to determine their inhibitory potential (IC50) against rCPB2.8 and isoforms.
The difference in amino-acid sequence may have affected the inhibitory potential of 4c. The enzyme rH84Y differs from rCPB3 by a single amino-acid residue [43]. Since 4c was 8.74 times more potent in inhibiting rH84Y compared to rCPB3, it is possible to infer that the substitution of histidine for tyrosine was the variation that most affected the inhibitory potential of 4c.
In the evaluation of the inhibitory capacity of flavonoids against rCPB2.8, the compounds f12a (IC 50 = 4.72 ± 0.38 µM), f12b (IC 50 = 5.23 ± 0.32 µM), and f13a (IC 50 = 1.88 ± 0.07 µM) showed the highest inhibitory potentials. Methoxy added to the C4 carbon in the f13a structure seems to potentiate the enzyme inhibition. On the other hand, the presence of dimethylamine on C4 in f13c disfavoured the inhibition of rCPB2.8 by 7.62-fold. Compounds f12a (IC 50 = 7.71 ± 0.78 µM) and f12b (IC 50 = 7.06 ± 0.63 µM) showed the best IC 50 values on rCPB3. Again, it was found that the presence of dimethylamine did not favour rCPB3 inhibition, as compounds f12c (IC 50 = 16.67 ± 2.53 µM) and f13c (IC 50 = 29.75 ± 2.08 µM) were the least effective. Regarding the inhibition of rH84Y, the flavonoids f12a (IC 50 = 3.85 ± 0.27 µM) and f12b (IC 50 = 8.85 ± 0.33 µM) presented the best inhibitory potential, and compounds f12c and f13c had the lowest inhibitory capacity (Table 3). On the basis of these data, it can be observed that the presence of dimethylamine on C4 carbon negatively impacts the inhibition of the three enzymes.  The high inhibition capacity of all compounds against the three enzyme isoforms confirms their multitarget properties. Studies have shown the formation of larger lesions in BALB/c mice that received amastigotes expressing only CPB2.8 compared to those with amastigotes deficient in all three isoforms [44]. This result confirms that inhibition against the three isoforms is essential when developing a new effective antileishmanial compound based on the inhibition of the CPB enzymes. Among the compounds under study, the chalcone 3c and flavonols f12a and f12b were the most effective simultaneously against all enzymes tested, making them good candidates for prototypes. Therefore, these compounds were selected to evaluate the inhibition mechanisms of rCPB2.8, rCPB3, and rH84Y.
By evaluating the inhibition mechanisms of compounds 3c, f12a, and f12b on rCPB2.8, we verified the slope vs.
[inhibitor] and the intercept vs.
[inhibitor] (Figures S1D, S2D, and S3D, Supplementary Materials) of the parabolic profile. This means that two molecules participate in the inhibition mechanism process. Initially, there is binding of the first molecule, which may favour or impair the binding of the second. This molecular behaviour is responsible for respectively establishing the cooperativity as positive or The high inhibition capacity of all compounds against the three enzyme isoforms confirms their multitarget properties. Studies have shown the formation of larger lesions in BALB/c mice that received amastigotes expressing only CPB2.8 compared to those with amastigotes deficient in all three isoforms [44]. This result confirms that inhibition against the three isoforms is essential when developing a new effective antileishmanial compound based on the inhibition of the CPB enzymes. Among the compounds under study, the chalcone 3c and flavonols f12a and f12b were the most effective simultaneously against all enzymes tested, making them good candidates for prototypes. Therefore, these compounds were selected to evaluate the inhibition mechanisms of rCPB2.8, rCPB3, and rH84Y.
By evaluating the inhibition mechanisms of compounds 3c, f12a, and f12b on rCPB2.8, we verified the slope vs.
[inhibitor] and the intercept vs.
[inhibitor] (Figures S1D, S2D, and S3D, Supplementary Materials) of the parabolic profile. This means that two molecules participate in the inhibition mechanism process. Initially, there is binding of the first molecule, which may favour or impair the binding of the second. This molecular behaviour is responsible for respectively establishing the cooperativity as positive or negative. To determine the affinity constants of the compounds, it was necessary to perform the linearisation of the parabolas, obtaining the replots 1/K Slope vs. According to the replots 1/K Slope vs.
The results of the inhibitory mechanism of rCPB2.8 by f12a demonstrated a cooperativity inhibition as a function of parabolic replots. Their respective linearisation ( Figure S2C-F, Supplementary Materials) allowed the determination of K i = 12.4 ± 1.4 µM and αK i = 10.7 ± 1.1 µM, with α being~1, considering the standard deviation. The first molecule of f12a had the same affinity for binding to the free enzyme E or the ES complex. The values of βK i = 3.65 ± 0.39 µM and γK i = 0.25 ± 0.03 µM defined β = 0.29 and γ = 0.02 (Table 4). Therefore, binding of the first molecule favoured the formation of the IEI and IESI complexes 3.4-and 50-fold, respectively.
Parabolic replots were observed for the inhibition of rCPB3 by compound 3c (Figure S4B stants were K i = 15.4 ± 3.0 µM, αK i = 141 ± 27 µM, βK i = 2.21 ± 0.16 µM (β = 0.14), and γK i = 0.77 ± 0.14 µM (γ = 0.05) ( Table 4). The α value (α = 9.15) shows that 3c preferentially binds to the free enzyme. With the binding of the compound to the ES complex that forms ESI defined by the γ factor, the formation of the quaternary IESI complex will be favoured 20-fold. On the other hand, the formation of the EI complex by binding of the first molecule 3c favours the formation of IEI sevenfold. Therefore, 3c presented a non-competitive inhibition mechanism with positive cooperativity, while f12a and f12b presented a simple linear non-competitive inhibition mechanism.

Molecular Modelling Study
The results of inhibitory capacity and the mechanism of action against these isoforms by the obtained chalcone and flavonol analogues are unprecedented. However, previous work in the literature may shed light on understanding the potential binding mode of the compounds at the active site of the CPB isoforms. Leishmania mexicana type B cysteine proteases are L-like cathepsins. The inhibition activity of these enzymes by chalcones is already known. Studies by Raghav and Kaur found that the catalytic CYS 29 thiolate of cathepsin L was able to attack the nucleophilic sites of chalcones [45]. In other recent work, chalcones demonstrated in vitro antileishmanial activity on the amastigote and promastigote forms of L. infantum. The suggested mechanism of action was the inhibition of pro-cathepsin L-like by the formation of hydrogen bonds between the amino acid TRP 151 of the active site and the carboxyl group of the chalcones [46]. Isolated flavonoids were also able to inhibit cathepsins L and B in previous studies [47].
Other classes of small molecules have also been described as inhibitors that target multiple cathepsin L-like cysteine proteases, some with overlapping antiparasitic activity [40]. Among them, vinyl sulphones have been shown to be highly potent and selective inhibitors of cathepsins L and B and are also considered antiparasitic prototypes [48,49]. Interestingly, a remarkable 3D similarity is demonstrated by the structural overlap of a crystallographic analogue of vinyl sulphone and compounds 3c, f12b, and f12c; as a result, there is a compatible binding mode at the active site of the enzyme ( Figure 4A,B, respectively). In particular, the overlap with the vinyl sulphone was useful in understanding non-competitive inhibition with a positive cooperativity mechanism. Therefore, the crystal structure of a papain-like cysteine protease bound with the vinyl sulphone derivative [50] was used as a template in the homology modelling study.
The isoforms rCPB2.8, rCPB3, and rH84Y have a small number of modifications in their active sites (Table S1, Supplementary Material). However, the few amino-acid variations between these isoenzymes are important in modifying substrate specificities [43]. This change may have modified either the catalytic or allosteric site, even at a distance. This type of event was observed in substitutions distant from the active site that affect the catalytic activity of CheZ and the binding of CheYp, with possible propagation of structural or dynamic disturbance [51].
inhibitors of cathepsins L and B and are also considered antiparasitic prototypes [48,49]. Interestingly, a remarkable 3D similarity is demonstrated by the structural overlap of a crystallographic analogue of vinyl sulphone and compounds 3c, f12b, and f12c; as a result, there is a compatible binding mode at the active site of the enzyme ( Figure 4A,B, respectively). In particular, the overlap with the vinyl sulphone was useful in understanding non-competitive inhibition with a positive cooperativity mechanism. Therefore, the crystal structure of a papain-like cysteine protease bound with the vinyl sulphone derivative [50] was used as a template in the homology modelling study. The isoforms rCPB2.8, rCPB3, and rH84Y have a small number of modifications in their active sites (Table S1, Supplementary Material). However, the few amino-acid Interestingly, even with the amino-acid residue variations, chalcone 3c showed the same mechanism of action on the three isoforms. To investigate its possible binding mode at the active site of the enzymes, the energy values of each pose pointed by 3D overlapping at rCPB2.8 was calculated after geometry optimisation (Table S2, Supplementary Materials). The position with the lower potential energy was also used to investigate the intermolecular interactions at the active site of rCPB3 and rH84Y. Despite having the same mechanism of action, the calculated binding free energy values of compound 3c at the catalytic site of the isoforms were remarkable different (Table 5). Particularly, the chalcone had the most promising binding free energy result on rCPB3 (−24.70 kcal·mol −1 ). These values reinforce the hypothesis of a more favourable interaction in a molecular scenario between compound 3c and this isoform and corroborates with the in vitro results. On the contrary, the variation of amino-acid residues resulted in a different mechanism of action for flavonols f12a and f12b between the three isoforms. This change is due to the difference in the negative charge distribution of these residues, which necessarily results in significant changes in the electrostatic potential on the surface of the isoenzymes, in addition to providing the parasite with a series of hydrolytic activities [43,52]. Therefore, the change in the electrostatic potential on the surface of the isoenzymes promoted changes in the inhibition mechanisms, as well as differences in the affinity constants.
Following the simple linear non-competitive inhibition mechanism, the potential energy of the different positions of f12a at the binding site of rCPB3 was promising (Table S3, Supplementary Materials). The binding pose with the lower potential energy was used to investigate the intermolecular interactions of the flavonol analogues at the active site of rCPB3 and rH84Y, since the mechanism of action was the same at these two isoforms. As also occurred with compound 3c, the calculated binding free energy values of f12a at the binding site of the isoforms were notably different, with the lowest value for rCPB3 (−5.65 kcal·mol −1 ). This result corroborates the in vitro assays that demonstrate a higher affinity constant for this isoform (Table 5). Despite the structural similarity, compound f12b had a lower binding free energy at the simulations of the active site of rH84Y ( Table 5). The calculated energy values also demonstrated a great difference among the three isoforms (Table 5), which corroborates the considerably lower affinity constant for rCPB2.8.
Analysing the output results of the simulations of compound 3c, the variations of amino-acid residues at the binding site of the isoforms resulted in a similar occupation of the binding pockets, but noticeable differences in the intermolecular interactions ( Figure 5A,C,E). At the binding site of rCPB2.8, 3c mainly made hydrophobic interactions ( Figure 5B). However, the substitution of ASP 186 to ASN 186 resulted in the formation of a hydrogen bond, which was also observed with GLY 144 ( Figure 5D). This strong intermolecular interaction was conserved at the active site of rH84Y; however, the hydrogen bond with GLY 144 was not observed at the catalytic site of this isoform ( Figure 5F). The difference in intermolecular interactions, added to the binding-free energy values, corroborates the affinity constants obtained by in vitro assays.
Microorganisms 2022, 10, x FOR PEER REVIEW 11 o calculated energy values also demonstrated a great difference among the three isofor (Table 5), which corroborates the considerably lower affinity constant for rCPB2.8.
Analysing the output results of the simulations of compound 3c, the variations amino-acid residues at the binding site of the isoforms resulted in a similar occupation the binding pockets, but noticeable differences in the intermolecular interactions ( Fig  5A,C,E). At the binding site of rCPB2.8, 3c mainly made hydrophobic interactions ( Fig  5B). However, the substitution of ASP 186 to ASN 186 resulted in the formation o hydrogen bond, which was also observed with GLY 144 ( Figure 5D). This stro intermolecular interaction was conserved at the active site of rH84Y; however, hydrogen bond with GLY 144 was not observed at the catalytic site of this isoform ( Fig  5F). The difference in intermolecular interactions, added to the binding-free ene values, corroborates the affinity constants obtained by in vitro assays. At the active site of rCPB2.8, the molecules of flavonol f12a had a great occupation the binding pockets and made hydrogen bonds with the amino-acid residue ASP ( Figure 6A,B). The number of hydrogen bonds increased at the binding site of rCPB3, si the compound made interactions with TRP 310 and GLY 191 ( Figure 6D). Previous stud have already discussed the interaction between flavonoid derivatives and the GLY ami At the active site of rCPB2.8, the molecules of flavonol f12a had a great occupation of the binding pockets and made hydrogen bonds with the amino-acid residue ASP 189 ( Figure 6A,B). The number of hydrogen bonds increased at the binding site of rCPB3, since the compound made interactions with TRP 310 and GLY 191 ( Figure 6D). Previous studies have already discussed the interaction between flavonoid derivatives and the GLY amino-acid residue of the cathepsin L catalytic site, as well as its importance in stabilising the active compound at the binding site of the enzyme [45]. The occupation of f12a at the binding pockets of rCPB3 and rH84Y was very similar (Figure 6C,E). This resemblance was reflected by the interactions with the amino-acid residues of the two isoforms ( Figure 6F).
The binding poses of f12b with the rCPB isoforms were very similar to those found for f12a and highly resembled the binding poses of each isoform, even by the simple linear non-competitive inhibition mechanism ( Figure 7A,C,E). At the active site of rH84Y, the proximity with the amino-acid residue GLY 191 (3.953 Å) may be related to a better interaction with this isoform ( Figure 7F).
Microorganisms 2022, 10, x FOR PEER REVIEW 12 of 24 was reflected by the interactions with the amino-acid residues of the two isoforms ( Figure  6F). The binding poses of f12b with the rCPB isoforms were very similar to those found for f12a and highly resembled the binding poses of each isoform, even by the simple linear non-competitive inhibition mechanism ( Figure 7A,C,E). At the active site of rH84Y, the proximity with the amino-acid residue GLY 191 (3.953 Å) may be related to a better interaction with this isoform ( Figure 7F).

Antipromastigote Assay and Cytotoxicity Elucidation
Although flavonoid derivatives have demonstrated promising enzyme-inhibitory potential, the development of a drug candidate for leishmaniasis treatment depends on several pharmacological aspects. Among them, the cytotoxicity of compounds is crucial for the discovery of a new antileishmanial prototype, since high toxicity still represents a

Antipromastigote Assay and Cytotoxicity Elucidation
Although flavonoid derivatives have demonstrated promising enzyme-inhibitory potential, the development of a drug candidate for leishmaniasis treatment depends on several pharmacological aspects. Among them, the cytotoxicity of compounds is crucial for the discovery of a new antileishmanial prototype, since high toxicity still represents a serious limitation of the drugs used in current therapy [53]. The compounds need to be highly active against the Leishmania parasite and provide safety to host cells. This pharmacological characteristic is measured by SI, defined as the ratio of the 50% cytotoxic concentration of mammalian cells (GL 50 ) to the half-maximum inhibitory concentration on parasites (IC 50 ).
However, the determination of IC 50 by in vitro screening tests is a challenge with the Leishmania parasite itself. The Leishmania lifecycle requires the presence of a sand fly vector and a mammalian host that causes the existence of two distinct morphological forms (promastigote and amastigote). Although the amastigote form is found in host cells and is considered the target form of the parasite, determination of the IC 50 consists of a time-consuming and laborious procedure and is not suitable for a large-scale screening method [54]. In general, exploratory screening methods designed to accelerate the testing of many compounds are performed on the promastigote form [54,55]. Therefore, to determine the SI of all flavonoid derivatives, we used the half-maximum inhibitory concentration on promastigote forms. All 17 flavonoid derivatives had moderate to low solubility in water. This physical property represented an obstacle in determining the IC 50 of the compounds, as the test occurred in an aqueous medium. Consequently, the higher concentration tested (10 µg/mL) was lower than that generally described in the literature (50 µg/mL) [56]. Compounds with IC 50 greater than 10 µg/mL were considered nonactive. However, since even the higher concentration tested was lower than that reported for compounds considered active against the parasite, these molecules may not be excluded as potential antileishmanial prototypes.
Among the tested molecules, the chalcone analogues stood out. Compounds 2a, 3a, and 4c were more active than the standard drug pentamidine (IC 50 = 0.71, 0.60, and 0.50 µM, respectively). Meanwhile, f12c had the highest potency of the flavonol derivatives against the parasite (IC 50 = 0.73 µM). The results indicated that the substitution of a chloride atom in the phenolic ring of the chalcone and flavonol derivatives increased the activity against the Leishmania promastigote form (Table 6).
Compounds 3a and 4c, the most active against the promastigote form, were also noncytotoxic against mammal cells ( Figure S10, Supplementary Materials), with an optimal selective index (SI > 1752.48 and 1443.10, respectively) when compared to first-line drugs such as pentamidine and amphotericin B [56] (Table 6). It is important to emphasise that the compounds tested all demonstrated low cytotoxicity to mammal cells, resulting in high SI values.
The lipophilicity (logP) and water solubility (logS) properties of the flavonoid derivatives were also measured by in silico analysis. The importance of these chemical characteristics was first discussed by Lipinski et al. through the publication of the rule of five. In this study, among the physicochemical characteristics of a set of standard drugs, clogP ≤ 5 was postulated as being necessary for an ideal prototype [57]. Later, the development of an ADME in silico tool, based on the analysis of more than 2000 standard drugs, indicated that more than 80% of the drugs on the market have a (estimated) logS value greater than −4 [58]. All 17 flavonoid derivatives had logS values close to −4 and logP ≤ 5, following the characteristics postulated by the rule of five. Although further studies against amastigote form need to be carried out, the ADME results, added to the biological potential, indicate that all synthetic flavonoids are druglike and can be considered promising prototypes for the treatment of Leishmaniasis disease. LogP, octanol/water partition coefficient measured by SwissADME [59]; 2 LogS expressed as log (g/100 g water) measured by SwissADME [59], 3 IC 50 , half-maximum inhibitory concentration on promastigote forms; 4 n.a, not active (IC 50 > 10 µg/mL); 5 GL 50 , concentration that inhibited cell growth by 50%; 6 SI (selectivity index), IC 50 in mammalian cells/IC 50 in extracellular promastigotes; 7 n.d, not determined. The data are representative of three independent experiments.

Conclusions
Our synthetic protocols confirmed that the methods are versatile, scalable, easy to reproduce, and inexpensive for obtaining high yields of flavonoid derivatives. The compounds demonstrated the multitarget properties intended by our study to inhibit all tested rCPB isoforms of L. mexicana. All chalcones and flavonol analogues inhibited the three enzymes to some degree at a concentration of 5 µM and were subjected to assays to determine their inhibitory potential IC 50 against rCPB2.8 and isoforms. Regarding the activity of chalcones, the presence of chlorine attached on carbon C3 or C4 , fluoride on C3 , or hydroxyl on the C6 seems not to affect the inhibitory capacity of the compounds toward rCPB2.8. However, the presence of the 1,3-dioxolane group, in general, did not favour this enzyme inhibition. Interestingly, the chalcones bearing a chlorine atom at C3 (3a and 3c) or a methoxy group at C4 (3a and 4a) had the lowest IC 50 on rCPB3, on the rH84Y isoform and compound 4c had the lowest IC 50 on the rH84Y isoform. In the evaluation of the inhibitory capacity of flavonoids, it was observed that the presence of dimethylamine on the C4 carbon negatively impacted the inhibition of the three enzymes. Among the flavonoid analogues, compounds 3c, f12a, and f12b stood out for being effective against all isoforms simultaneously. Interestingly, the in vitro study of the mechanism of cysteine protease inhibition showed that small variations of amino-acid residues between the rCPB isoforms were able to change the mechanism of action and binding mode position of the compounds. These findings were confirmed by the in silico investigation that demonstrated the formation of strong intermolecular interactions between the compounds and the active site of each enzyme. The compounds were highly potent and demonstrated low cytotoxic action against mammalian cells, proving that the tested molecules are extremely selective. In addition, the calculated logP and logS values proved that the compounds have ADME properties compatible with those observed in standard drugs. The antileishmanial activity of all flavonoid analogues needs to be elucidated against amastigote form in future studies. However, our results show important progress in the investigation of the antileishmanial action of synthetic flavonoid derivatives and reinforce their potential as prototypes for the design of other cysteine protease inhibitors for the treatment of leishmaniasis.

Chemistry
All reagents were purchased from Sigma-Aldrich ® and were analytical grade, used without further purification. Reactions were monitored by TLC using Merck 60 F254 precoated silica plates, and spot visualisation was achieved with UV light (254-360 nm), molybdophosphoric acid (10% w/v), and a solution of sulphur vanillin (0.5 g vanillin in 100 mL of sulphuric acid/methanol (40:10)). All products were purified by recrystallisation from ethanol (EtOH). The solvents used in the reactions and recrystallisation were purified and dried according to procedures found in the literature [60]. A mixture of hexane and ethyl acetate was used in a 1:2 (v/v) proportion as the mobile phase to measure the retention factor (R f ) values of all purified compounds. All melting points were determined using a Quimis ® of Brazil model Q340S instrument. The 1 H-and 13 C-NMR spectra were recorded on a Bruker Avance DPX-300 or Bruker Ascend 500 spectrometer. Chemical shifts are reported as δ values (ppm) referenced to the residual solvent (CDCl 3 at δ 7.24 ppm, DMSOd 6 at δ 2.50 ppm). Peak multiplicities are abbreviated as follows: s (singlet); d (doublet); dd (doublet of doublets); tp (triplet of doublets); t (triplet); dt (doublet of triplets); m (multiplet). The coupling constants (J) are quoted in hertz and recorded at the nearest 0.1 Hz.

General Procedure for the Synthesis of Chalcone-Like Compounds by the Claisen-Schmidt Reaction (1a-4c)
The synthesis procedure followed the Claisen-Schmidt reaction methodology described in the literature, with modifications [33]. An aqueous solution of NaOH (3 M, 1.6 mL) was added to a solution of aromatic ketone (1 mmol) in EtOH. An ethanolic solution of substituted benzaldehyde was added dropwise to the reaction mixture. The mixture was stirred at room temperature for 24 h and then cooled. The reaction mixture was acidified with concentrated HCl (37%) to pH = 2 in an ice bath and under vigorous stirring. The precipitate formed was filtered, washed with cold water, and purified by recrystallisation from ethanol.

General Procedure for the Synthesis of Flavonol-Like Compounds by the Algar Flynn-Oyamada Reaction (f12a-f13c)
In a round-bottom flask, an aqueous solution of NaOH (1 M, 2 mL) was added to 1 mmol of chalcone in EtOH (5 mL). The solution was cooled until an ice-cold suspension was formed. An aqueous solution of H 2 O 2 (35%, 250 µL) was added to the ice-cold suspension; the mixture was allowed to warm to room temperature and stirred for 1-2 h. Distilled water (3 mL) was then added. The reaction mixture was acidified with concentrated HCl (37%) to pH = 2 in an ice bath and under vigorous stirring. The precipitate formed was filtered, washed with cold water, and purified by recrystallisation from ethanol.

Screening the Inhibitory Activity of Compounds
The screening assays for the inhibitory activity of the compounds in the rCPB2.8, rCPB3, and rH84Y enzymes were performed using 100 mM sodium acetate buffer containing 5 mM EDTA, 100 mM NaCl, 0.01% Triton X-100, and 20% glycerol, at pH 5.5. Enzyme aliquots were pre-incubated with 5 mM DTT for 5 min at 37 • C. After checking the initial rate of the reaction corresponding to the control, the enzymatic rate was measured at two concentrations of compounds, 1 µM and 5 µM. Enzyme activity was monitored by hydrolysis of the substrate Z-FR-MCA by measuring the fluorescence at λ Ex = 360 nm and λ Em = 480 nm on a Hitachi F2700 spectrofluorometer, obtaining the rate values in UAF/min (arbitrary fluorescence units by minute).

Determination of IC 50 Values for Inhibitors
Cysteine proteases rCPB2.8, rCPB3, and rH84Y were assayed in 100 mM sodium acetate buffer containing 5 mM EDTA, 100 mM NaCl, 0.01% Triton X-100, and 20% glycerol, at pH 5.5. The enzymes were pre-incubated in the presence of 5 mM DTT for 5 min at 37 • C in a 1 mL final volume with constant stirring. Enzyme activities were monitored using fluorogenic probe Z-FR-AMC (9.25 µM final concentration), and the fluorescence was monitored by spectrofluorometry using fluorometer F2700 (Hitachi, Tokyo, Japan) set to λ Ex = 360 nm and λ Em = 480 nm. The IC 50 evaluation was performed using a progressive increase in the concentration of the compounds; the IC 50 values were calculated using nonlinear regression, and the data were analysed by Grafit 5.0.13 software using Equation (1). (1)

Enzyme Kinetics and Determination of the Mechanism of Inhibition
Studies of cysteine proteases rCPB2.8, rCPB3, and rH84Y inhibition kinetics were performed in different concentrations of Z-FR-AMC in the presence and absence of compounds using 100 mM sodium acetate buffer containing 5 mM EDTA, 100 mM NaCl, 0.01% Triton X-100, and 20% glycerol, at pH 5.5. Aliquots of the enzymes were pre-incubated with 5 mM DTT for 5 min at 37 • C. For every kinetic measurement, the compounds were pre-incubated with each enzyme for 10 min before adding the substrate. All kinetic assays were performed in duplicate. Inhibition constants were determined using different equations, depending on the inhibition mechanism. The assumed K M values of rCPB2.8, rCPB3, and rH84Y for Z-FR-AMC were 3.23 µM, 2.99 µM, and 2.80 µM, respectively. The data of the activity rate and substrate concentration generated rectangular hyperbolic profiles that were linearised using the Lineweaver-Burk double-reciprocal plot. The replot profiles of the slope vs. inhibitor and intercept vs. inhibitor provided Ki and αKi parameters, respectively. If the replots present a parabolic profile, the system involves the participation of a second molecule of the compound, which can bind the enzyme complexed to the first compound (EI), forming IEI, and bind to the ESI complex, forming IESI. Linearisation is required, generating the 1/K slope vs. inhibitor and 1/K intercept vs. inhibitor replots, according to Equations (2) and (3).
where Ki is the inhibitory constant, [I] is the concentration of inhibitor, α is the factor of formation of the ESI complex (enzyme-substrate-inhibitor complex), β is the factor of formation of the IEI complex (inhibitor-enzyme-inhibitor complex), and γ is the factor of formation of the IESI complex (inhibitor-enzyme-substrate-inhibitor complex) according to the general mechanism ( Figure 8). Furthermore, β and γ measure the cooperativity between the binding of the first and second inhibitor molecules to form IEI and IESI, respectively.

Molecular Modelling
Compounds 3c, f12a, and f12b had their 3D structure drawn using the program MarvinSketch 16.9.5 (ChemAxon Ltd., Budapest, Hungary). The optimisation was carried out using the PM7 semiempirical method incorporated in the software MOPAC2016 [61]. A pH of 7.4 was considered for the definition of charges.
The three-dimensional structure of rCPB2.8 was obtained through the homology modelling methodology using the Swiss-Model program [62]. Therefore, we used the 3D structure of papain-like cysteine protease obtained from the Protein Data Bank (PDB ID: 1F2A) as a layout [50] and the primary structure of rCPB2.8 as the target sequence. The choice of the crystal was based on the similarity with rCPB2.8, as well as that between the tested compound and the crystallographic ligand. To determine the potential binding modes at the active site of rCPB2.8, different binding poses were obtained on the basis of the overlay between the tested molecules and the crystallographic ligand. For the compounds with non-competitive inhibition with a positive cooperativity mechanism, we manually added a second molecule in different positions. The 3D structures of rCPB3 and rH84Y were obtained according to the amino-acid residues differences at their active sites, as described in the literature, using the program UCSF Chimera [63]. All the binding poses were further optimised geometrically.
Geometry optimisations were made using the GROMACS 2018 package [64] and the CHARMM force field [65]. The ligand topology was obtained from the Swiss Param Server [66], and the properties of the solvent were mimetic based on the TIP3P water model. A cubic box was used to guarantee a space of 1.2 nm between the protein and the box walls, and ions were under physiological conditions (0.15 µM) in order to neutralise the system charges. The energy optimisation steps were performed using the steepest descent followed by the conjugated gradient algorithm. The convergent criterion was a

Molecular Modelling
Compounds 3c, f12a, and f12b had their 3D structure drawn using the program MarvinSketch 16.9.5 (ChemAxon Ltd., Budapest, Hungary). The optimisation was carried out using the PM7 semiempirical method incorporated in the software MOPAC2016 [61]. A pH of 7.4 was considered for the definition of charges.
The three-dimensional structure of rCPB2.8 was obtained through the homology modelling methodology using the Swiss-Model program [62]. Therefore, we used the 3D structure of papain-like cysteine protease obtained from the Protein Data Bank (PDB ID: 1F2A) as a layout [50] and the primary structure of rCPB2.8 as the target sequence. The choice of the crystal was based on the similarity with rCPB2.8, as well as that between the tested compound and the crystallographic ligand. To determine the potential binding modes at the active site of rCPB2.8, different binding poses were obtained on the basis of the overlay between the tested molecules and the crystallographic ligand. For the compounds with non-competitive inhibition with a positive cooperativity mechanism, we manually added a second molecule in different positions. The 3D structures of rCPB3 and rH84Y were obtained according to the amino-acid residues differences at their active sites, as described in the literature, using the program UCSF Chimera [63]. All the binding poses were further optimised geometrically.
Geometry optimisations were made using the GROMACS 2018 package [64] and the CHARMM force field [65]. The ligand topology was obtained from the Swiss Param Server [66], and the properties of the solvent were mimetic based on the TIP3P water model. A cubic box was used to guarantee a space of 1.2 nm between the protein and the box walls, and ions were under physiological conditions (0.15 µM) in order to neutralise the system charges. The energy optimisation steps were performed using the steepest descent followed by the conjugated gradient algorithm. The convergent criterion was a maximum force of 50 N on the atoms. The potential binding energy was measured, and the position of the binding mode with the best result was used to analyse the interactions with the binding site of all isoforms.

Cytotoxicity on Mammalian Cells
The cytotoxic effect of the test samples was evaluated on NIH/3T3 fibroblasts, obtained from the Rio de Janeiro Cell Bank (Brazil). Cells were seeded in 96-well plates (5 × 10 5 cells/well). After 24 h of fixation, cells were incubated for 48 h with test samples at 0.25-250 µg/mL, in triplicate. The tested compounds were dissolved in DMSO (Sigma-Aldrich ® SP/Brazil) while ensuring that the final concentration of the latter (0.25% at the highest sample concentration) did not interfere with cell viability. Doxorubicin (0.025-25 µg/mL was used as a positive control. Cells were fixed by adding 20% trichloroacetic acid and were subsequently stained with sulphorhodamine B (0.1%) diluted in acetic acid after 48 h of exposure [67]. Absorbance values were read on the PT-READER microplate instrument (Thermoplate ® ), and growth percentages were calculated according to procedures in the literature [68]. Cytotoxicity activity was expressed as the concentration of drug that inhibited cell growth by 50% (GL 50 ), and growth was determined by nonlinear regression using Origin 6.0 software (OriginLab). The statistical significance was analysed using an unpaired Student's t-test or a one-way analysis of variance. A p-value < 0.05 was considered statistically relevant. In addition, the SI was calculated using the ratio between cytotoxicity in NHI-3T3 cells (GL 50 ) and activity in the parasite forms (IC 50 ).

Parasites
A standard strain of Leishmania (Leishmania amazonensis) (IFLA/BR/1967/PH8) was used for the evaluation of in vitro antileishmanial activity. The promastigote forms were grown in Schneider's insect medium (Sigma-Aldrich ® , SP/Brazil) supplemented with 20% foetal bovine serum (Sigma-Aldrich ® , SP/Brazil), 10,000 U/mL penicillin, and 10 mg/mL streptomycin (Sigma-Aldrich ® , SP/Brazil). Skin lesions were routinely isolated from previously induced skin lesions in BALB/c mice and kept in axenic culture until the 20th serial passage.
Supplementary Materials: The following supporting information can be downloaded at www.mdpi. com/xxx/s1: Figures S1-S10; Tables S1-S3; Spectra 1-34. Figure S1: Determination of the affinity constants of the 3c compound in the inhibition of rCPB2.8; Figure S2: Determination of the affinity constants of the f12a compound in the inhibition of rCPB2.8; Figure S3: Determination of the affinity constants of the f12b compound in the inhibition of rCPB2.8; Figure S4: Determination of the affinity constants of the 3c compound in the inhibition of rCPB3; Figure S5: Determination of the affinity constants of the f12a compound in the inhibition of rCPB3; Figure S6: Determination of the affinity constants of the f12b compound in the inhibition of rCPB3; Figure S7: Determination of the affinity constants of the 3c compound in the inhibition of rH84Y; Figure S8: Determination of the affinity constants of the f12a compound in the inhibition of rH84Y; Figure S9: Determination of the affinity constants of the f12b compound in the inhibition of rH84Y; Table S1: Differences of amino acid