4-Arylthiosemicarbazide derivatives as a new class of tyrosinase inhibitors and anti-Toxoplasma gondii agents

Abstract We report herein anti-proliferation effects of 4-arylthiosemicarbazides, with a cyclopentane substitution at N1 position, on highly virulent RH strain of Toxoplasma gondii. Among them, the highest in vitro anti-Toxoplasma activity was found with the meta-iodo derivative. Further experiments demonstrated inhibitory effects of thiosemicarbazides on tyrosinase (Tyr) activity, and good correlation was found between percentage of Tyr inhibition and IC50Tg. To confirm the concept that thiosemicarbazides are able to disrupt tyrosine metabolism in Toxoplasma tachyzoites, the most potent Tyr inhibitors were tested for their efficacy of T. gondii growth inhibition. All of them significantly reduced the number of tachyzoites in the parasitophorous vacuoles (PVs) compared to untreated cells, as well as inhibited tachyzoites growth by impeding cell division. Collectively, these results indicate that compounds with the thiosemicarbazide scaffold are able to disrupt tyrosine metabolism in Toxoplasma tachyzoites by deregulation of their crucial enzyme tyrosine hydroxylase (TyrH).


Introduction
Toxoplasma gondii (T. gondii) is an obligate intracellular protozoan parasite that infects birds and all mammals, including humans. Due to its unusually wide range of intermediate hosts, T. gondii is one of the most successful pathogens among eukaryotes and toxoplasmosis is one of the world's most common latent infection 1,2 . Although largely asymptomatic, chronic toxoplasmosis can be especially harmful to newborns infected congenitally, immunocompromised persons, or patients receiving immunosuppressive therapy for organ transplantation [3][4][5][6][7][8][9] .
T. gondii exists in three infectious stages, such as oocysts (with sporozoites), tissue cysts (with bradyzoites), and rapidly proliferated tachyzoites, which form the parasitophorous vacuole (PV), a structure created by apicomplexan parasites in the host cells. Oocysts are only produced in the definitive hostsmembers of the Felidae family (domestic cats and their relatives). Unsporulated oocysts are excreted in feline faeces of infected individuals. Each sporulated oocyst contains two sporocysts which hold four sporozoites. Even if a single sporulated oocyst is ingested or inhaled by the intermediate host, including all classes of warm-blooded vertebrates, sporozoites transform to tachyzoites, which localised in neural and muscle tissues including brain, eye, skeletal muscle, kidney, spleen, blood, liver, and heart tissue, and convert to bradyzoites enclosed in cysts that persist chronically 10 . In the case of infected pregnant women, tachyzoites can penetrate into the foetus via the bloodstream or placenta during the acute phase or reactivation of the infection. Ingestion of the tissue cysts in contaminated meat is also a source of infection, as bradyzoites transform back into tachyzoites upon entering a new host 11 .
The genome of T. gondii contains two genes, TgAaaH1 constitutively expressed in both tachyzoites and bradyzoites and TgAaaH2 whose expression is uniquely upregulated in bradyzoite cysts 12 . These genes encode tyrosine hydroxylase (TyrH) 12 that promote catalysis of L-tyrosine conversion into L-3,4-dihydroxyphenylalanine (L-DOPA) 13 , the precursor of dopamine. Reports have shown that T. gondii TyrH is related to the dopamine levels in the host's brain 14 . To address the role of TgAaaH genes in oocyst development, Sibley et al. and McFarland et al., 15,16 and others 17 investigated the effect of TgAaaH1/TgAaaH2 gene deletion and determined that the TgAaaH2 gene is not connected with the dopamine-dependency of Toxoplasma infection. An alternatively proposed role of TgAaaH genes involves the promotion of dityrosine crosslinking in oocyst wall proteins 18 . Although the role for TgAaaH1/TgAaaH2 genes in the tachyzoite or bradyzoite stages has not been reported thus far, close examination of L-tyrosine metabolism in Toxoplasma at these stages has revealed that they are both dependent on the addition of exogenous L-tyrosine for efficient growth. Additionally, the efficient production of PVs formation is limited in the absence of L-tyrosine 19 . Moreover, Parker et al. 20 described a family of plasma membrane-localised amino acid transporters, named apicomplexan amino acid transporters (ApiATs), that are ubiquitous in apicomplexan parasites, including T. gondii. Functional characterisation of these transporters in T. gondii shows that several are important for intracellular growth at the tachyzoite stage of the parasite, which is related with acute infections. One such protein is the exchanger for aromatic and large neutral amino acids, with exhibit particular importance for L-tyrosine uptake pathway and amino acid homeostasis, which is critical for parasite virulence. The collective reports [12][13][14][15][16][17][18][19][20] prove the concept that compounds that inhibit exogenous L-tyrosine uptake into T. gondii and/or participate in the disruption of the parasite L-tyrosine metabolism, have the potential to progress into vital medicines for the effective treatment of toxoplasmosis. Unfortunately, to date, there are no effective treatments against bradyzoites, and those medicines that target the tachyzoites (pyrimethamine [PYR] and sulfadiazine are the most effective) are associated with toxicity and hypersensitivity [21][22][23][24][25][26][27] . Other serious problems, such as drug resistance and incidence of relapses after discontinuation of therapy have been also reported 28,29 . Hence, novel, efficacious anti-Toxoplasma agents are indeed greatly needed.
To address these issues, we explored the therapeutic potentials of low molecular weight heterocyclic compounds with thiosemicarbazide scaffold. In Liesen et al. 30 on the example of small series of imidazole-thiosemicarbazides, reported for the first time anti-T. gondii potential for thiosemicarbazide-based compounds. The tested compounds showed better LD 50 values for both infected cells and intracellular parasites than the standard drugs sulfadiazine and hydroxyurea. Inspired by those promising results, as well as those reported for closely related thiosemicarbazones 31,32 efforts have been made by our research group to expand these initial findings with further details on anti-T. gondii activity of thiosemicarbazides-based compounds. Based on the systematic search, [33][34][35] several promising candidates (Scheme 1(A)) were identified to inhibit RH-strain tachyzoites proliferation in vitro more effectively than standard drugs sulfadiazine or trimethoprim (TRI) and at the same time did not inhibit the growth of human cell lines, suggesting that the anti-Toxoplasma strategy can be nontoxic towards mammals. However, crucial information regarding the molecular mechanisms involved in their anti-Toxoplasma effect has yet to be reported. One of the key steps in determining the cause of therapeutic intervention is through the identification of the essential pathways that can be targeted by small molecules.
Recent reports have shown the inhibitory activity of thiosemicarbazide and thiosemicarbazones against tyrosinase (Tyr) (Scheme 1(B)) which, similarly to TyrH, catalyses two initial sequential oxidations of L-tyrosine to L-DOPA [36][37][38][39] . Based on these results, we hoped that thiosemicarbazide-based compounds could show an inhibitory effect against T. gondii TyrH. Results of such studies have not been reported in the literature thus far. Using combined experimental and theoretical approaches, we hypothesised that the inhibitory effect of studied thiosemicarbazides on invasion and division of tachyzoites in host cells is related to their disruption of L-tyrosine metabolism in Toxoplasma tachyzoites and/or defects in exogenous L-tyrosine transport to parasitic cells.
These studies, described here, reflect our approach to support the hypothesis that the biological effect of the examined compounds results from disruption of T. gondii L-tyrosine metabolism by inhibiting TyrH. First, the problem of cytotoxicity of the studied compounds is addressed. Second, we measured the effect of the studied compounds on in vitro anti-T. gondii activity. Once the prepared compounds are determined as non-toxic and biologically active, we investigated if their effect on T. gondii is related to inhibitory activity against TyrH. However, this is a two-step process owing to the lack of access to enzyme preparations. Therefore, following literature precedents 40 , we used commercially available Tyr, which also catalyses the conversion of L-tyrosine to L-DOPA, as a proxy to the inhibition pattern of TyrH. Subsequently, a series of experiments screened for the inhibition of Tyr and its comparison with that of T. gondii proliferation in vitro. In order to evolve and strengthen the conclusion to our hypothesis, studies conducted on Tyr inhibition allowed for the interpretation of TyrH mechanism of action. Additionally, docking studies were performed on all compounds to both enzymes and showed the correspondence of the energetics of binding between them, as well as with T. gondii proliferation inhibition. In this case, we focussed on four compounds (2a, 2b, 6a, and 6b), which exhibit the highest inhibitory activity against Tyr. To further understand the mechanism of inhibition we comprehensively investigated the kinetics of inhibition followed by docking studies. These studies highlighted binding modes and further supported the assumed correspondence between inhibition of Tyr and TyrH. Hence, we examined the efficacy of compounds 2a, 2b, 6a, and 6b in inhibiting T. gondii growth in vitro into human foreskin fibroblast cells using confocal microscopy.

Results and discussion
Several promising candidates with thiosemicarbazide scaffolds have been shown to inhibit RH tachyzoites proliferation in vitro more effectively and selectively than antifolate drugs (sulfadiazine, TRI, or PYR) [33][34][35] . Unfortunately, no structural basis or molecular mechanisms of underlying anti-T. gondii activity of thiosemicarbazide has been reported thus far in literature. Hence, a series of previously reported 4-arylthiosemicarbazides bearing N1 imidazole ring was examined 34,35 (1a-9a, Table 1), which have proven nontoxic, as well as exhibited enhanced potency and selectivity compared to TRI, sulfadiazine, and combination. Additionally, to start the SAR exploration, we have also chosen to replace the N1 imidazole with more hydrophobic N1 cyclopentane to obtain a set of cyclopentane-thiosemicarbazides 1b-9b (Table 1). These compounds were prepared using reported methods 33-35 via a onestep reaction between cyclopentanecarbohydrazide and appropriate aryl isothiocyanate (Scheme 2). Detailed procedures for the synthesis and characterisation of compounds 1b-9b are presented in the Experimental Section. First, we evaluated the cytotoxic effects of 1b-9b, which is necessary prior to measuring anti-Toxoplasma and Tyr inhibition in order to exclude cytotoxic compounds. The cytotoxic effects of 1a-9a have been reported previously 34,35 . Cytotoxicity was estimated using mouse fibroblast (L929) cells line via MTT assay, according to the international standards ISO 10993-5:2009(E). T. gondii, an intracellular parasite, only multiplies inside the PV in the infected host cells, hence, host cells should not be destroyed by the cytotoxic concentration of the studied compounds. Therefore, the obtained results were expressed as CC 30 , which was defined as the highest concentration of the compound that causes 30% damage of cells and was considered according to ISO as non-toxic. CC 30 values of cyclopentane-thiosemicarbazides 1b-9b are presented in Table 1. CC 30 values of imidazole-thiosemicarbazides 1a-9a are also included for comparison.
As shown in Table 1, cyclopentane-thiosemicarbazides 1b-9b did not show significant cytotoxic effect on L929 cells after 24 h of incubation and are considered good candidates for biological assays. The analysis of MTT results (e.g. 1a vs. 1b) also reveals that, except for compounds 2b and 6b, cyclopentane-thiosemicarbazides 1b-9b show less cytotoxic effect than their N1 imidazole analogues 1a-9a. It is worth noting that CC 30 values for compounds 1b-9b are at least 1.92-14.47 higher than the cytotoxic concentration of TRI and, except for compound 6b, that at least 1.83 to 4.87 higher than the cytotoxic concentration of PYR.
At this point, we questioned if replacing N1 imidazole in the basic structure of 4-arylthiosemicarbazide by more hydrophobic N1 cyclopentane influences the anti-Toxoplasma effect. This strategy is based on observation that incorporation of more lipophilic substituent into thiosemicarbazide core structure improves anti-T. gondii activity 32 . To address this, we submitted 1b-9b to T. gondii in vitro proliferation assay. PYR commonly used for the treatment of human toxoplasmosis, and TRI, and used to treat animals in Europe 44 , were employed as positive controls, whereas 0.1% concentration of dimethyl sulfoxide (compounds and drugs solvent) was used as the negative control (data not shown). In addition, anti-T. gondii activity of the reference Tyr/TyrH inhibitor, kojic acid (KA), was established. IC 50Tg and selectivity ratio (SR) values of cyclopentane-thiosemicarbazides 1b-9b and KA are presented in Table 1. IC 50Tg values of imidazole-thiosemicarbazides 1a-9a reported previously 34,35 are also listed for comparison.
Although cyclopentane-thiosemicarbazides 1b-9b show lower potency and selectivity in inhibiting T. gondii proliferation in vitro compared with more hydrophilic precursors 1a-9a, they possess enhanced selectivity compared to TRI. Furthermore, consistent with data obtained for compounds 1a-9a, meta-iodo compound 6b (54.49 mM) is the most effective; it showed efficacy close to that of TRI (50.94 mM) and, at the same time, better selectivity (2.16 vs. 1.2). The presence of other substituents, both electron withdrawing and electron donating, also reduced potency most notably ortho-chloro 3b, ortho-methoxyl 7b, and 2,5-fluoro 9b.
Within the series of compounds 1b-9b the activity follows the trend m- , whereas activity of compounds 1a-9a follows a slightly different trend m- Once the cytotoxic profile and in vitro biological activity of compounds 1a-9a and 1b-9b on T. gondii were determined, we subjected the studied compounds to anti-Tyr activity assay to establish the role of thiosemicarbazides, towards the activity of Toxoplasma TyrH. Since there is currently no commercial Toxoplasma TyrH, following literature precedents 40 we used in our experiments mushroom Tyr that, similarly to TyrH, catalyses two initial sequential oxidations of L-tyrosine to L-DOPA. The reference inhibitor, KA, was used for the enzyme assay at the concentration of 25 mM at which 50% inhibition of Tyr was obtained in the preliminary tests (data not shown). A few general conclusions arise from this inhibitory study, as graphically summarised in Figure 1. Importantly, all studied compounds display efficacy below their cytotoxic concentration. Imidazole-thiosemicarbazides 1a-9a universally produce higher inhibitory effect than their N1 cyclopentane substituted 1b-9b analogues (e.g. 1a vs. 1b; Figure 1), which is in agreement with T. gondii proliferation measurements data. The observed differences may arise from the electronic effect of N1 imidazole on the whole thiosemicarbazide skeleton. Additionally, it may also be possible that the imidazole present in compounds 1a-9a promote structural benefits either by providing increased hydrophilicity to adopt favourable conformations in the active site or favourable interactions between the imidazole and enzyme. Furthermore, among the thiosemicarbazides tested, meta-iodo 6a (48.9 ± 3.0%) and 6 b (43.9 ± 3.4%), and para-nitro 2a (46.4 ± 3.6%) and 2b (45.6 ± 2.8%) display the highest inhibitory effects, similar to that of KA (50.4 ± 4.9%), thus suggesting that the presence of the electron-withdrawing substitution at the meta or para position of N4 phenyl ring and (or) steric effect of the substituents can induce improved anti-Tyr activity. In addition, among ortho-halogen compounds, the steric and/or electronic change from chloro (3a, 3b) to bromo (4a, 4b) to iodo (5a, 5b) increase inhibitory potency. Although the precise role of the electronegativity of ortho-halogens in anti-Tyr activity remains unclear, changes in pKa or enhancement in binding created through decreased positive charge of N4 phenyl ring may be a possibility. Finally, the trend of inhibitory potency; meta-iodo 6a (48.9 ± 3.0%) > meta-iodo 6b (47.9 ± 3.4%) > para-nitro 2a (46.4 ± 3.6%) > para-nitro 2b (45.6 ± 2.8%), can be related to the substitution pattern and (or) steric effect of the substituents.
According to the obtained results from the T. gondii in vitro proliferation assay, good correlation exists (Figure 2, Pearson correlation coefficient r ¼ 0.84; p ¼ .0053) between percentage of Tyr inhibition and IC 50Tg of both sets of compounds. This further supports the assumption that anti-Toxoplasma effect of thiosemicarbazides might be related to inhibition of Tyr, and thus, TyrH activity.
In order to test if Tyr inhibitors also act as T. gondii TyrH inhibitors, complementary docking studies were carried out. Currently, there is no reported 3 D structure of T. gondii TyrH, hence, the 3D structure of rat TyrH (PDB ID: 1TOH) 45 is employed since its catalytic domains from rat and T. gondii are highly comparable 12 . As shown in Figure 3, the sequence of TgAaaH1 (UniProtKB code B2L7T1) in the catalytic domain is highly homological with rat enzyme and their 3D structure similarity is illustrated by Figure 4. The catalytic domain contains conserved VPWFPR motif that begins at V240 of the T. gondii sequence and residue V164 of the rat TyrH. This catalytic domain also contains conserved iron-binding residues at positions His407, His412, and Glu452, which correspond to His331, His336, and Glu376 in rat TyrH, respectively 46 . In addition, validity of structural modelling of T. gondii catalytic domains using the crystal structure of rat TyrH proxy is corroborated by the sequence alignment and placement of key residues ( Figure 3 shows the alignment generated using MUSCLE software 47 ). Therefore, our computational studies of the crystallographic structure of rat TyrH (PDB ID: 1TOH) were used.
First, in order to successfully predict the binding affinities to the active site of T. gondii TyrH, compounds 1a-9a and 1b-9b were docked to the active site of TyrH (PDB ID: 1TOH) and its proxy (for which experiments were performed) mushroom Tyr (PDB ID: 5M6B) 48 . As shown in Figure 5(A), predicable good correlation is observed (Pearson correlation coefficient r ¼ 0.98; p < .0001) between the binding energies to 5M6B plotted and those predicted for 1TOH. Additionally, good correlation is observed between the binding energies of 5M6B or 1TOH plotted against the percentage of Tyr inhibition ( Figure 5(B,C), respectively). The strong correlation between binding energies of compounds 1a-9a and 1b-9b to Tyr and TyrH indicate a similar mode of binding in both enzymes which further validates the implementation of the experimental results obtained for the Tyr in the interpretation of TyrH mechanism.
The screening studies of Tyr inhibition highlighted four distinctive thiosemicarbazides (2a, 2b, 6a, and 6b) with significant  inhibitory activity. Therefore, these compounds were subsequently subjected to detailed kinetic studies to evaluate inhibition parameters. The Lineweaver-Burk plots (Figure 6), which are the linearised transformations of the Michaelis-Menten curves, clearly demonstrate mixed character of inhibition of studied compounds. Hence, compounds 2a, 2b, 6a, and 6b cannot only bind to the enzyme at its active site and prevent the substrate from binding, but also at the allosteric site. This possibility has been addressed by docking studies (see below). From the line corresponding to the uncatalysed reaction, the values of kinetic parameters; pseudo-first-order reaction rate, V max /K M , equal to 0.0583 min À1 and rate of saturated enzyme, V max , 7.8 mM/min has been evaluated. Dixon plots (Figure 7) were used in evaluating the K I values 49 , while the plots of [L-tyrosine]/V vs.
[I] represents the concentration of the inhibitor. Rates were monitored spectrophotometrically by detecting changes in absorbance at 492 nm. The results are collectively summarised in Table 2 and highlight the trend in inhibitory potency; para-nitro 2b (IC 50Tyr 31.9 mM) > meta-iodo 6a (IC 50Tyr 37.4 mM) > para-nitro 2a (IC 50Tyr 38.3 mM) > meta-iodo 6b (IC 50Tyr 53.6 mM), that is fortunately not linked directly with cytotoxic activity. Indeed, the most cytotoxic 6b (CC 30 ¼ 117.68 mM) is the weakest inhibitor, whereas compounds 2a and 6a with comparable inhibitory potency differ substantially in their cytotoxicity, CC 30 ¼ 682.98 and 213.27 mM, respectively.
Two scenarios of the molecular mechanism of Tyr inhibition can be proposed from our kinetic studies. Since mixed type of inhibition has been determined quite frequently for this enzyme 51 , it is apparent that the inhibitor can compete with the substrate for binding space within the enzyme active site, i.e. a pocket around the iron atom, ligated by six histidines. However, two possible structural options can be considered for the other component of the inhibition, originating from the EÁSÁI ternary complex. The first involves the simultaneous binding of the reactant and inhibitor to the active site of the enzyme. The second consists of the inhibitor binding to an allosteric site. Detailed docking simulations excluded the first possibility on the basis of failure of binding the substrate to the enzyme-inhibitor and of biding inhibitor to enzyme-substrate complexes within the active site. This result is not surprising since the active site is quite tight and buried. Thus, only the second option remains, which is allosteric inhibition Figure 3. Alignment of aromatic amino acid hydroxylases generated using MUSCLE alignment software 44 . Identical residues between 1TOH and TgAaaH1 are highlighted in black. The positions that have a single and fully conserved residue are marked with an asterisk ( Ã ). The conservation between residues of strongly similar properties are marked with a colon (:), while conservation between residues of weakly similar properties are marked with a period (.). The iron-binding residues are highlighted with blue asterisk ( Ã ). The residues present in the active site that interact with the aromatic amino acid substrate and biopterin cofactor are highlighted with red asterisk ( Ã ) 12 . at a remote site with unknown mechanism of action or near the active site most likely with steric hindrance preventing substrate from entering/exiting while the inhibitor is bound to the protein.
Assuming that the mechanisms of inhibition of T. gondii TyrH (the enzyme in which catalytic mechanism we are mainly interested) and Tyr are similar, we have used the latter as a proxy. This assumption is supported by the similarity of action and the fact that mixed inhibition of allosteric origin has been reported for this class of enzymes despite of structural differences 52,53 . To explore the binding to allosteric sites, docking of the most potent inhibitor compound 2b to AaaH1 surface rather than the active site was conducted. Docking poses clustered around three possible allosteric binding sites that are illustrated in Figure 9. The first (leftmost in the figure) is close to the entrance of the active site, which for TyrH is quite accessible. The second and third are neighbouring sites on the other side of the active site and in enzyme part analogous to that of the recently documented allosteric site in tryptophan hydroxylase 54 .
To confirm the concept that thiosemicarbazides may act as inhibitors of T. gondii TyrH involved in L-tyrosine metabolism, we examined the influence of the most potent anti-Tyr compounds (2a, 2b, 6a, and 6b) during T. gondii growth into human foreskin fibroblast cells. In these studies, the reference Tyr inhibitor, KA, was also included. In the first variant of the experiment (assay A), the effect of the compounds towards penetration of tachyzoites into the host cells was determined. Therefore, Hs27 cells with extracellular T. gondii tachyzoites were incubated for 3 h in the milieu of tested compounds or KA and then washed out. Next, infected cells were incubated further 24 h in normal medium. In this assay, the compounds were present only during penetration of T. gondii into the cells (Scheme 3). The confocal scanning laser microscope analyses of the host cells after assay A demonstrates that PV formation can still occur, but morphologically are smaller in comparison to untreated cells (Figure 10(A)).
In the next step, the obtained data from microscopic analysis (Figure 10(B)) of assay A with selected compounds 2a, 2b, 6a, 6b, and KA, were analysed by the two-way ANOVA test and the results are summarised in Figure 11. These studied compounds effectively inhibited tachyzoites penetration into host cells in comparison to untreated cells. This effect is observed by the significant reduction in the number of infected cells (Figure 11(A)) and tachyzoites per PV (KA > 2a ¼ 2b > 6b > 6a; Figure 11(B)) in comparison to untreated cells.
Additionally, we performed an alternative version to that described above (assay B), in which tachyzoites had 3 h to penetrate the host cells in the absence of compounds 2a, 2b, 6a, 6b, or KA. Then, the tested compounds were added and further incubated for 24 h, thus, were present only during the growth of intercellular T. gondii (Scheme 3). All tested Tyr inhibitors (2a, 2b, 6a, and 6b) as well as reference KA, similar to assay A, had an effect on the size of the formed PVs compared to untreated cells ( Figure 12(A)).
The results from microscopic analysis of assay B (Figure 12(B)) are summarised in Figure 13, and were analysed by the two-way ANOVA test. As expected, the difference in percentage of infected cells is not observable, in comparison to untreated cells, due to in assay B T. gondii has time (3 h without compounds) to enter the cell and for initial PV formation (Figure 13(A)). However, the significant reduction in the number of tachyzoites per PV is visible (Figure 13(B)). Compounds 2a, 2b, 6a, and 6b significantly inhibit intracellular parasite proliferation by impeding tachyzoites division (KA ¼ 6b > 2a > 6a ¼ 2b) in comparison to untreated cells.
Recently, Marino and Boothroyd 19 observed a decrease in the number of PV per cell in L-tyrosine-free medium as well as the average number of doublings tachyzoites post-infection, which is related to Toxoplasma tachyzoites requiring exogenous L-tyrosine for growth. The presence of the studied compounds can disrupt L-tyrosine metabolism in tachyzoites. In both situations, a significant decrease in the number of tachyzoites per PV is observed (Scheme 3).
Conclusively, the results obtained for in silico predictions of molecular basis of anti-T. gondii activity of thiosemicarbazidebased compounds combined with experimental influence of compounds during Toxoplasma growth (assay A and B) support conclusion that the studied compounds are able to disrupt L-tyrosine metabolism in tachyzoites by deregulation of TyrH. In order to further provide evidence for this proposition, enzymatic experiments should be conducted. An alternative explanation may be that thiosemicarbazide-based compounds target the plasma membranelocalised amino acid transporter(s), which is important for L-tyrosine uptake into parasite. This hypothesis originates from the studies of Marino and Boothroyd 19 that recently confirmed that Toxoplasma tachyzoites are auxotrophic for exogenous L-tyrosine, which must be salvaged from their host cells via direct uptake or degradation of host proteins. In addition, Wang et al. 55 reported that although recombinant AaaH1 and AaaH2 are capable of converting phenylalanine to L-tyrosine, in which their expression levels during tachyzoite growth are exceedingly low. These reports highlight that L-tyrosine transporter localised to the plasma membrane of tachyzoites might be an alternative key target for anti-Toxoplasma activity of thiosemicarbazide-based compounds. This hypothesis is beyond the scope of this study and will be considered in our future work.

Chemistry
The prepared compounds were characterised using 1 H and 13 C NMR spectroscopy. Each tested compound had !96% purity, as determined using elemental analysis. All chemicals used were obtained from Sigma-Aldrich (Saint Louis, MO) and Alfa Aesar (Karlsruhe, Germany) and used without further purification. TLC separations were performed on Merck aluminium plates with silica gel 60 F 254 (Darmstadt, Germany). Melting points were determined using a Fischer-Johns block and are uncorrected. Elemental analyses were determined by an AMZ-CHX elemental analyser (PG, Gdansk, Poland), and all determined elements were within ±0.4% of the theoretical values. 1 H and 13 C NMR spectra were recorded using a Bruker Avance spectrometer (Bruker BioSpin GmbH, Rheinstetten, Germany) using DMSO-d 6 as solvent. Chemical shifts are reported as d values in parts per million (ppm) and indirectly referenced to tetramethylsilane via the solvent signal. Physicochemical characterisations of compounds 1a-9a were described previously 34,35 . The structures of compounds 1b-3b, 7b, and 8b are known (CAS numbers: 891079-50-8, 891079-58-6, 891075-55-1, 891069-84-4, 891062-80-9, respectively), however, there are no references reporting their use, preparation or physicochemical characterisation. Therefore, this data is included. General method for the synthesis of thiosemicarbazides 1a-9a and 1b-9b The corresponding aryl isothiocyanate (0.01 mol) was added to a solution of 4-methylimidazole-5-carbohydrazide (0.01 mol) or cyclopentanecarbohydrazide (0.01 mol) in anhydrous ethanol (25 mL). The reaction mixture was heated under reflux until the isothiocyanate agent was consumed as determined by TLC (mobile phase: CHCl 3 / EtOH, 3:1, 2:1, or 1:1). The final product was filtered from the reaction mixture and purified by crystallisation from EtOH.   182.07, 174.93,  159.19, 128.72, 117.22, 117.09, 116.92, 116.79, 42.46, 30.18, 26.14. -1 ] represents the catalytic efficiency as a ratio of the maximum rate of reaction (V max ) to the Michaelis constant (K m ). b K M [mM] represents Michaelis-Menten constant, the substrate concentration at which the reaction rate was 50% of the maximum rate of reaction when all enzyme active sites were saturated with substrate (V max ). c K I [mM] represents the inhibitory constant, the compound concentration required to produce half maximum inhibition. d IC 50Tyr [mM] represents the concentration of the compound required for 50% inhibition of Tyr. e CC 30 [mM] represents the non-cytotoxic concentration of the compound required for 30% of cells proliferation inhibition in vitro.

Biological evaluation
All experiments were carried out in BSL-2 laboratory with biological safety cabinet. Compounds 1a-9a and 1 b-9b were dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich, St. Louis, MO) to 100 mM. The final concentration of DMSO in compound dilutions was below 1.00%. TRI and PYR (92131 and 46706, Sigma-Aldrich) was dissolved in DMSO to 25 mM. The final concentration of DMSO in TRI dilution was below 1.00%. Tyr inhibitor, KA (K3125, Sigma-Aldrich), was dissolved in DMSO to 100 mM. The final concentration of DMSO in KA dilution was below 1.00%. All dilutions were freshly prepared before experiments.
Cell viability assay Cell viability assay was performed according to international standards (ISO 10993-5:2009(E)), using tetrazolium salt (MTT, Sigma-Aldrich) and mouse fibroblasts L929 cells. Culture medium RPMI 1640, without phenol red (Biowest, Nuaille, France) supplemented with 10% FBS, 2 mM L-glutamine (Sigma-Aldrich), 100 I.U./ mL penicillin, 100 mg/mL streptomycin were used. Briefly, 1 Â 10 4 / well of L929 cells were placed in 96-well plates and incubated for 24 h at 37 C and 10% CO 2 . Afterwards, the old culture medium was replaced with 100 mL of the compound, drug or inhibitor, diluted in culture medium, and the cells treated for 24 h. Additionally, cells were treated with 0.03-4.0% concentration of DMSO as the solvent (data not shown). Then, 50 mL of 1 mg/mL of MTT solution in RPMI 1640 without phenol red was added to each well and incubated for 2 h at 37 C and 10% CO 2 . Next, cell culture medium was aspirated carefully and 150 mL of DMSO was added to each well, and the plates were gently mixed. Then, 25 mL 0.1 M glycine buffer (pH 10.5) (Sigma-Aldrich) was added. The optical density at 570 nm using a multi-mode microplate reader SpectraMax V R i3 (Syngen, Taipei City, Taiwan) was recorded. The results were expressed as a percentage of viability compared to untreated cells. All experiments were performed in triplicate. Cytotoxicity assay using human fibroblast Hs27 cells was also performed as described above (included in Supplementary Material).      and RH-GFP (ATCC V R 50940 TM , expression of green fluorescent protein) strains were maintained as tachyzoites according to the ATCC product sheet, in parasite culture medium, which contains DMEM medium with 3% HIFBS (Heat-Inactivated FBS; 1 h in 56 C). Infected tissue culture cells were incubated (37 C, 5% CO 2 ).

Antiparasitic assay
The influence of thiosemicarbazides 1a-9a, 1b-9b, TRI, PYR, and KA on T. gondii RH proliferation was examined using the following method: 1 Â 10 4 /well of Hs27 cells were seeded on 96-well plates in cell culture medium. After 72 h of incubation, the medium was removed and then 1 Â 10 5 /well tachyzoites of the RH strain were added to the cell monolayers, in parasite culture medium. After 1 h the thiosemicarbazides, drugs, or KA dilutions in the parasite culture medium were added to the Hs27 cells with T. gondii (concentration range of the compounds 0-500 mM, TRI 0-125 mM, and KA 0-1500 mM). Assay Aafter 48 h, the medium was removed and then 5 Â 10 5 /500 mL/well tachyzoites of the RH-GFP strain were added to the cell monolayers, in parasite culture medium with the selected compounds at IC 50Tg concentration: 2a, 2b, 6a, 6b, and KA, for 3 h. Then, the cells were washed to remove extracellular parasites and parasite culture medium (without any compounds) was added for 24 h.
Assay Bafter 48 h, the medium was removed and then 5 Â 10 5 /500 mL/well tachyzoites of the RH-GFP strain were added to the cell monolayers, in parasite culture medium for 3 h. Then, the cells were washed to remove extracellular parasites, and the cells were treated with the selected compounds at IC 50Tg concentration: 2a, 2b, 6a, 6b, and KA, for 24 h.
In both assays as the controls, Hs27 cells were infected, but not treated. After subsequent 24 h of incubation (assay A and B), the slides were washed with sterile PBS, fixed with formaldehyde solution (252549, Sigma-Aldrich), 3.7% in PBS for 20 min, and stained with DAPI for 10 min (1 mg/mL, ThermoFisher Scientific, Waltham, MA) and Texas Red TM -X Phalloidin for 40 min (1.65 mM; ThermoFisher Scientific). Microscopic analysis of 100 cells in turn, comprised counting the infected cell ratio and tachyzoites per each PVs of infected cell, was performed using a fluorescent microscope (Olympus IX50, Tokyo, Japan) at a magnification Â1000. Additionally, cells with PVs were visualised using a confocal scanning laser microscope, Leica TCS SP8, equipped with Leica LAS AF Lite Software for image processing. Three independent experiments on triplicate chamber slides using the same conditions for both assay were performed.

Docking studies
The docking calculations were performed using two protocols and crystal structures of TyrH (ID: 1TOH) 45 and mushroom Tyr (PDB ID: 5M6B) 48 deposited in PDB 56 . Docking to the active sites the FlexX docking algorithm 57 of the LeadIT environment as implemented in the LeadIT 2.1.9 program (BioSolveIT GmbH, Augustin, Germany) was employed. In the case of all analysed compounds, the first 100 top-ranked docking poses were saved for each docking run. In the search for allosteric binding sites the EADock DSS algorithm 58 , as implemented in the SwissDock 59 program. In the cluster analysis, the radius of 2 Å and the maximum cluster size of 8 entries were used. Graphical representation of the docking results was conducted using the Chimaera software 60 . Values with statistically significant differences are labelled by brackets and asterisks as follows: Ã p < .05; ÃÃ p < .01; ÃÃÃ p < .001; ÃÃÃÃ p < .0001. Data were compared using a two-way ANOVA.

Sequence alignments
Sequence alignments were created using MUSCLE multiple alignment software 47 . The catalytic domain of three enzymes from biopterin-dependent aromatic amino acid hydroxylase family (EC 1.14.16.) with tyrosine 3-monooxygenase activity were used: two aromatic amino acid hydroxylase of T. gondii (GenBank: TgAaaH1 -ACB99413.1) and TyrH from rat (GenBank: 1TOH -AAA42257.1). The conserved VPWFPR motif was identified as the beginning of the catalytic domain of enzymes 12,61 .

Graphs and statistical analyses
Statistical analyses and graphs were performed using GraphPad Prism version 9.0.0 on macOS (GraphPad Software, San Diego, CA). Compounds with CC 30 or IC 50 values greater than the highest concentration tested, they were calculated based on extrapolation of the curves using the GraphPad Prism. Additionally, to establish a relationship between cytotoxicity and antiparasitic activity, the SR values were calculated as the ratio of 30% cytotoxic concentration (CC 30 ) to 50% antiparasitic concentration (IC 50Tg ).

Conclusion
Herein, we have described anti-proliferation effects of two series of 4-arylthiosemicarbazides with imidazole or cyclopentane substitution at N1 position on highly virulent T. gondii RH strain. Among these series, compounds with meta-iodo substitution 6a and 6b showed efficacy close to TRI while exhibiting enhanced selectivity. Thus, compounds 6a and 6b could be used as lead structures in further structural optimisation and structure-activity relationship studies. Further experiments with compounds of both series demonstrated their inhibitory effects on Tyr activity, with good correlation being observed (Pearson correlation coefficient r ¼ 0.84; p ¼ .0053) between percentage of Tyr inhibition and IC 50Tg of both sets of thiosemicarbazides. Next, the most potent Tyr inhibitors, compounds 2a, 2b, 6a, and 6b, were explored to determine their efficacy in inhibiting T. gondii growth in human foreskin fibroblast cells using confocal microscopy. All studied compounds significantly reduced the number of tachyzoites in the PVs compared to untreated cells, and inhibited tachyzoites growth by impeding their division. In addition, the compounds effectively inhibited tachyzoites penetration into host cells. Collectively, the obtained results support the conclusion that the examined compounds with thiosemicarbazide scaffold were able of disrupting L-tyrosine metabolism in Toxoplasma tachyzoites by deregulation of TyrHthe crucial enzyme in the conversion of L-tyrosine to L-DOPA. The mechanism of the inhibition of this step included the binding to an allosteric site, which was in agreement with the mechanism recently reported for tryptophan hydroxylase.