Antifungal Thiazolidines: Synthesis and Biological Evaluation of Mycosidine Congeners

Novel derivatives of Mycosidine (3,5-substituted thiazolidine-2,4-diones) are synthesized by Knoevenagel condensation and reactions of thiazolidines with chloroformates or halo-acetic acid esters. Furthermore, 5-Arylidene-2,4-thiazolidinediones and their 2-thioxo analogs containing halogen and hydroxy groups or di(benzyloxy) substituents in 5-benzylidene moiety are tested for antifungal activity in vitro. Some of the synthesized compounds exhibit high antifungal activity, both fungistatic and fungicidal, and lead to morphological changes in the Candida yeast cell wall. Based on the use of limited proteomic screening and toxicity analysis in mutants, we show that Mycosidine activity is associated with glucose transport. This suggests that this first-in-class antifungal drug has a novel mechanism of action that deserves further study.


Introduction
A limited number of chemical classes of compounds can be used against fungal pathogens at this time. In particular, the treatment of topical and systemic fungal diseases in humans and animals uses drugs belonging to the following classes: azoles, polyenes, nucleobase analogues, and echinocandins. Azole drugs inhibit the synthesis of ergosterol, which is a component of lipid membranes, via inactivation of the intracellular enzyme lanosterol demethylase. Polyenes bind to ergosterol on the cell surface, thereby disrupting the membrane structure. Flucytosine inhibits DNA and possibly RNA synthesis. Echinocandins disrupt the structure of the fungal cell wall by inhibiting 1,3-beta-glucan synthase [1][2][3]. Thus, these drugs target only a few molecular targets, which is why the problem of resistant forms of fungal pathogens arises more and more often in recent years. In addition, some of these drugs have significant side effects. Notably, during the past 30 years, only one new antifungal drug-Ibrexafungerp, selectively acting on the cell wall of pathogenic fungi, has entered clinical practice [4].
The search for new antifungal drugs and new targets is an extremely urgent task. One of the promising directions in the search for new antifungal drugs is the detection of chemical compounds acting on the cell wall of the fungus [5][6][7][8][9]. It was shown [12,13] that transposition of the benzyloxy group to the adjacent 3-position on the central aromatic ring had little effect on the PMT1 inhibitory activity. However, 3,4-bis(benzyloxy) substitution resulted in a significant increase in activity, both against the enzyme and C. albicans cells (IC50: 2.3 and 3.01 μM). Several compounds have been identified which inhibit C. albicans PMT1 with an IC50 in the range 0.2-0.5 μM (R=H, Me) [13].
The initial impulse for intensive study of the chemistry and biology of thiazolidines was initiated by the launch of hypoglycemic drugs-glitazones-containing various substituents in the 5-benzyl substituent of the thiazolidine core ( Figure 2) [37]. While these TZD derivatives are known to stimulate the PPAR-γ receptor, they also have multiple PPAR-γ independent effects and the specific role of PPAR-γ activation in the anticancer effects of TZDs is still under investigation [38,39].  It was shown [12,13] that transposition of the benzyloxy group to the adjacent 3-position on the central aromatic ring had little effect on the PMT1 inhibitory activity. However, 3,4-bis(benzyloxy) substitution resulted in a significant increase in activity, both against the enzyme and C. albicans cells (IC 50 : 2.3 and 3.01 µM). Several compounds have been identified which inhibit C. albicans PMT1 with an IC 50 in the range 0.2-0.5 µM (R=H, Me) [13].
The initial impulse for intensive study of the chemistry and biology of thiazolidines was initiated by the launch of hypoglycemic drugs-glitazones-containing various substituents in the 5-benzyl substituent of the thiazolidine core ( Figure 2) [37]. While these TZD derivatives are known to stimulate the PPAR-γ receptor, they also have multiple PPAR-γ independent effects and the specific role of PPAR-γ activation in the anticancer effects of TZDs is still under investigation [38,39].
The search for new antifungal drugs and new targets is an extremely urgent task. One of the promising directions in the search for new antifungal drugs is the detection of chemical compounds acting on the cell wall of the fungus [5][6][7][8][9].
Although the exact structure of the fungal cell wall is not fully understood, it consists of a complex mixture of proteins and polysaccharides, including glucan, mannans, and chitin. Most of the major cell wall components of fungal pathogens are not represented in humans. For these reasons, enzymes that assemble components of the fungal cell wall are excellent targets for antifungal chemotherapies and fungicides [10,11].
At the beginning of the 21st century, the first inhibitors of the mannosyl transferasederivatives of 2-(4-oxo-2-thioxothiazolidin-3-yl)acetic acid-were described (Figure 1, structure I) [12,13]. It was shown [12,13] that transposition of the benzyloxy group to the adjacent 3-position on the central aromatic ring had little effect on the PMT1 inhibitory activity. However, 3,4-bis(benzyloxy) substitution resulted in a significant increase in activity, both against the enzyme and C. albicans cells (IC50: 2.3 and 3.01 μM). Several compounds have been identified which inhibit C. albicans PMT1 with an IC50 in the range 0.2-0.5 μM (R=H, Me) [13].
The initial impulse for intensive study of the chemistry and biology of thiazolidines was initiated by the launch of hypoglycemic drugs-glitazones-containing various substituents in the 5-benzyl substituent of the thiazolidine core ( Figure 2) [37]. While these TZD derivatives are known to stimulate the PPAR-γ receptor, they also have multiple PPAR-γ independent effects and the specific role of PPAR-γ activation in the anticancer effects of TZDs is still under investigation [38,39].  TZDs have exhibited anti-tumor activity in a wide variety of experimental cancer models via effects on the cell cycle and induction of cell differentiation and apoptosis as well as by inhibiting tumor angiogenesis [40,41]. A study reported by Shah [40] showed that TZD derivative ciglitazone significantly decreased the production of VEGF in human granulosa cells in an in vitro model [42]. Mechanisms of anticancer activity of thiazolidin-4-ones and related heterocycles may be associated with their affinity to anticancer biological targets, such as non-membrane protein tyrosine phosphatase (SHP2), JNK-stimulating phosphatase-1 (JSP-1), tumor necrosis factor TNF-α, antiapoptotic biocomplex Bcl-XL-BH3, and integrin αvβ3 [43].
Epalrestat, a derivative of TZD, is also used in medical practice as a non-competitive and reversible aldose reductase inhibitor for the treatment of diabetic neuropathy [44] (Figure 3).
TZDs have exhibited anti-tumor activity in a wide variety of experimental cancer models via effects on the cell cycle and induction of cell differentiation and apoptosis as well as by inhibiting tumor angiogenesis [40,41]. A study reported by Shah [40] showed that TZD derivative ciglitazone significantly decreased the production of VEGF in human granulosa cells in an in vitro model [42]. Mechanisms of anticancer activity of thiazolidin-4-ones and related heterocycles may be associated with their affinity to anticancer biological targets, such as non-membrane protein tyrosine phosphatase (SHP2), JNK-stimulating phosphatase-1 (JSP-1), tumor necrosis factor TNF-α, antiapoptotic biocomplex Bcl-XL-BH3, and integrin αvβ3 [43].
Epalrestat, a derivative of TZD, is also used in medical practice as a non-competitive and reversible aldose reductase inhibitor for the treatment of diabetic neuropathy [44] ( Figure 3). Our group is interested in the antimicrobial activity of thiazolidine-2,4-diones, and in particular, their antifungal effects. A systematic study of a series of various 3-and 5substituted thiazolidines leads to the creation of a novel topical antifungal drug-Mycosidine ( Figure 3) [45][46][47][48].
Thus, this study aims to create novel, more efficient derivatives of TZD and to characterize the mechanisms responsible for their antifungal effects.

Ethyl 2,4-Dioxothiazolidine-3-Carboxylate and Other Thiazolidine-3-Carboxylates
The interactions of chloroformates with TZD derivatives containing hydroxyl groups in the arylidene fragment have recently been described in the literature [50,51]. The reactions were carried out in dry acetone in the presence of potassium carbonate. With an equimolecular ratio of reagents, the substitution on only the phenolic hydroxyl group is described [50], and with the double equivalent of the chloroformate and potassium carbonate, two reaction centers were involved at once with substitution at the ring nitrogen atom and the hydroxyl group of the arylidene radical [51]. The only N-ethoxycarbonyl derivative, obtained from the potassium salt of 5-nitrofurfurilidene-thiazolidine-2,4-dione with ethyl chloroformate in DMF, was described in a French patent [52]. However, other than the melting point, no data confirming the structure of the products was given.
We developed another method for obtaining new similar alkoxycarbonyl derivatives of TZD, on which the synthesis of Mycosidine is based: namely, reacting 5-arylidene TZD in a nonpolar solvent (e.g., toluene) with methyl chloroformate and triethylamine as a base at room temperature [45]. Using this technique from TZD (1) and ethyl chloroformate, ethyl 2,4-dioxothiazolidine-3-carboxylate (3) unsubstituted at position five was obtained. Our group is interested in the antimicrobial activity of thiazolidine-2,4-diones, and in particular, their antifungal effects. A systematic study of a series of various 3-and 5-substituted thiazolidines leads to the creation of a novel topical antifungal drug-Mycosidine ( Figure 3) [45][46][47][48].
Thus, this study aims to create novel, more efficient derivatives of TZD and to characterize the mechanisms responsible for their antifungal effects.

Ethyl 2,4-Dioxothiazolidine-3-Carboxylate and Other Thiazolidine-3-Carboxylates
The interactions of chloroformates with TZD derivatives containing hydroxyl groups in the arylidene fragment have recently been described in the literature [50,51]. The reactions were carried out in dry acetone in the presence of potassium carbonate. With an equimolecular ratio of reagents, the substitution on only the phenolic hydroxyl group is described [50], and with the double equivalent of the chloroformate and potassium carbonate, two reaction centers were involved at once with substitution at the ring nitrogen atom and the hydroxyl group of the arylidene radical [51]. The only N-ethoxycarbonyl derivative, obtained from the potassium salt of 5-nitrofurfurilidene-thiazolidine-2,4-dione with ethyl chloroformate in DMF, was described in a French patent [52]. However, other than the melting point, no data confirming the structure of the products was given.
We developed another method for obtaining new similar alkoxycarbonyl derivatives of TZD, on which the synthesis of Mycosidine is based: namely, reacting 5-arylidene TZD in a nonpolar solvent (e.g., toluene) with methyl chloroformate and triethylamine as a base at room temperature [45]. Using this technique from TZD (1) and ethyl chloroformate, ethyl 2,4-dioxothiazolidine-3-carboxylate (3) unsubstituted at position five was obtained.
The structure of compound 3 was confirmed using NMR spectroscopy and mass spectrometry. In the 1 H NMR spectrum of 3, there is a singlet of the methylene group at position five of the ring in the region of 4.00 ppm and signals of the ethyl group in the region of 4.38 ppm (2H, q, J = 7.16) and 1.34 ppm (3H, t, J = 6.82). In the 13 C NMR spectrum (in CDCl 3 ), signals of three carbonyl groups are observed at 147.49 (COOH), 167.55 (C 4 =O), and 167.65 (C 2 =O) ppm, the methylene group of the thiazolidine ring at 65.88 ppm, and the signals of the ethyl group of the ester CO-CH 2 CH 3 at δ 33.93 (CH 2 ) ppm and 13.81 (CH 3 ) ppm. However, it was not possible to carry out the Knoevenagel reaction of 3 with aromatic aldehydes by direct condensation under typical conditions. Under milder conditions (heating in ethanol with piperazine), decomposition products and unreacted compounds were isolated. In acetic acid with the catalysts (methylamine, piperidine), upon heating, the alkoxycarbonyl group was completely removed with the release of TZD as the main product. We managed to obtain the target 5-arylidene-3-carboxylates using an alternative method starting from 5-(4-chlorobenzylidene)thiazolidine-2,4-dione (4a) with various chloroformates (Scheme 1).
(in CDCl3), signals of three carbonyl groups are observed at 147.49 (COOH), 167.55 (C =O), and 167.65 (C 2 =O) ppm, the methylene group of the thiazolidine ring at 65.88 ppm, and the signals of the ethyl group of the ester CO-CH2CH3 at δ 33.93 (CH2) ppm and 13.81 (CH3) ppm. However, it was not possible to carry out the Knoevenagel reaction of 3 with aromatic aldehydes by direct condensation under typical conditions. Under milder conditions (heating in ethanol with piperazine), decomposition products and unreacted compounds were isolated. In acetic acid with the catalysts (methylamine, piperidine), upon heating, the alkoxycarbonyl group was completely removed with the release of TZD as the main product. We managed to obtain the target 5-arylidene-3-carboxylates using an alternative method starting from 5-(4-chlorobenzylidene)thiazolidine-2,4-dione (4a) with various chloroformates (Scheme 1). Scheme 1. General synthesis of 3-alkoxycarbonyl derivatives of TZD 5a-e. Reagents and conditions: (i) toluene, NEt3, rt; (ii) AcOH, MeNH2, reflux; (iii) EtOH, piperidine, reflux.
Compound 4a was obtained by the method described by us earlier [53] from compound 1 and 4-chlorobenzaldehyde in the acetic acid medium with 33% aqueous solution of methylamine as a catalyst. Condensation of 4a with various chloroformates under conditions similar to the preparation of 3 proceeded successfully with the production of target crystalline derivatives with good yields (68-85%). Structures were confirmed by 1D and 2D NMR and HRMS.
An example of the structure of 5a as a prototype of a series of compounds 5a-e is shown in Figure 4. The values of chemical shifts of 1 H and 13 C are shown in Tables 1 and 2, respectively. Scheme 1. General synthesis of 3-alkoxycarbonyl derivatives of TZD 5a-e. Reagents and conditions: (i) toluene, NEt 3 , rt; (ii) AcOH, MeNH 2 , reflux; (iii) EtOH, piperidine, reflux.
Compound 4a was obtained by the method described by us earlier [53] from compound 1 and 4-chlorobenzaldehyde in the acetic acid medium with 33% aqueous solution of methylamine as a catalyst. Condensation of 4a with various chloroformates under conditions similar to the preparation of 3 proceeded successfully with the production of target crystalline derivatives with good yields (68-85%). Structures were confirmed by 1D and 2D NMR and HRMS.
An example of the structure of 5a as a prototype of a series of compounds 5a-e is shown in Figure 4. The values of chemical shifts of 1 H and 13 C are shown in Tables 1 and 2, respectively.  Table 2. 13 C Chemical shifts of 5a in CDCl 3 (100 MHz).

C(3 ) C(5 ) C(4 )
δ (     The carbonyl group C(4) signal in the 13 C monoresonance spectrum is observed as a doublet with a coupling constant of 6.9 Hz. This information unambiguously confirms the attribution of this signal, because splitting on the proton H(6) can be observed only for C(4) but not for C (2). This measurement also clearly indicates the Z-configuration of the molecule, in which the nuclei C(4) and H(6), separated by three chemical bonds, are in a cis position to each other. For an alternative E-configuration in which these nuclei are in a trans position to each other, the spin-spin interaction constant between them should reach a value of 13-15 Hz. Thus, 5a, like the entire 5a-e series, exists exclusively in the form of a Z-conformer.
The description of NMR and mass spectra confirming the structure of the remaining alkoxycarbonyl derivatives 5a-e is given in the experimental section. In the IR spectra, there are three intense bands of valence vibrations of the CO group in the region: 1784-1789 (COOR), 1742-1747 (C 4 =O), and 1690-1700 cm −1 (C 2 =O). Valence vibrations are CH=C and the aromatic ring is in the region of 1606 and 1586 cm −1 .

5-Arylidene-3-Benzoylthiazolidine-2,4-Dione Derivatives
Several articles have described 3-Benzoyl derivatives of 5-arylidene-thiazolidine-2,4dione as potential herbicidal and antimicrobial agents. Their synthesis was based on 5arylidene-thiazolidine-2,4-dione and the corresponding benzoyl chloride and it had a good yield in anhydrous acetone in the presence of K2CO3 with heating [54] or in pyridine at 70 °C [55]. The carbonyl group C(4) signal in the 13 C monoresonance spectrum is observed as a doublet with a coupling constant of 6.9 Hz. This information unambiguously confirms the attribution of this signal, because splitting on the proton H(6) can be observed only for C(4) but not for C (2). This measurement also clearly indicates the Z-configuration of the molecule, in which the nuclei C(4) and H(6), separated by three chemical bonds, are in a cis position to each other. For an alternative E-configuration in which these nuclei are in a trans position to each other, the spin-spin interaction constant between them should reach a value of 13-15 Hz. Thus, 5a, like the entire 5a-e series, exists exclusively in the form of a Z-conformer.
The description of NMR and mass spectra confirming the structure of the remaining alkoxycarbonyl derivatives 5a-e is given in the experimental section. In the IR spectra, there are three intense bands of valence vibrations of the CO group in the region: 1784-1789 (COOR), 1742-1747 (C 4 =O), and 1690-1700 cm −1 (C 2 =O). Valence vibrations are CH=C and the aromatic ring is in the region of 1606 and 1586 cm −1 .
A more detailed study of the benzoylation reaction of TZD unsubstituted at position 5 was carried out by us earlier [56]. It was found that, depending on the synthesis conditions, there is a possibility of the formation of a 3-benzoyl derivative of TZD (6a) and thiazole-2,4-diyldibenzoate (6b) in addition to the main reaction product (Scheme 2). A more detailed study of the benzoylation reaction of TZD unsubstituted at position 5 was carried out by us earlier [56]. It was found that, depending on the synthesis conditions, there is a possibility of the formation of a 3-benzoyl derivative of TZD (6a) and thiazole-2,4-diyldibenzoate (6b) in addition to the main reaction product (Scheme 2).

Scheme 2. Benzoylation of TZD.
In this work, we tested the possibility of condensation of 3-benzoyl-TZD (6a) with aromatic aldehydes using the Knoevenagel reaction. The goal in this case was to obtain compounds with an unsubstituted hydroxyl group in the benzylidene fragment, because the acylation of the phenolic group of the arylidene residue may cause formation of unwanted O-acylated byproducts [57]. However, we were not able to obtain 5-arylidene derivatives 7a,b by direct condensation of 6a with aromatic aldehydes under any conditions. In this work, we tested the possibility of condensation of 3-benzoyl-TZD (6a) with aromatic aldehydes using the Knoevenagel reaction. The goal in this case was to obtain compounds with an unsubstituted hydroxyl group in the benzylidene fragment, because the acylation of the phenolic group of the arylidene residue may cause formation of unwanted O-acylated byproducts [57]. However, we were not able to obtain 5-arylidene derivatives 7a,b by direct condensation of 6a with aromatic aldehydes under any conditions. Then, a well-known alternative method for obtaining 7a,b by benzoylation of 5-arylidene TZD 4a,b was used (Scheme 3). In addition, some derivatives of 7a,b were obtained that did not contain free hydroxyl groups in the phenyl ring. The reaction was carried out according to a published procedure using acetone with K 2 CO 3 under heating [54].

Scheme 2. Benzoylation of TZD.
In this work, we tested the possibility of condensation of 3-benzoyl-TZD (6a) with aromatic aldehydes using the Knoevenagel reaction. The goal in this case was to obtain compounds with an unsubstituted hydroxyl group in the benzylidene fragment, because the acylation of the phenolic group of the arylidene residue may cause formation of unwanted O-acylated byproducts [57]. However, we were not able to obtain 5-arylidene derivatives 7a,b by direct condensation of 6a with aromatic aldehydes under any conditions. Then, a well-known alternative method for obtaining 7a,b by benzoylation of 5-arylidene TZD 4a,b was used (Scheme 3). In addition, some derivatives of 7a,b were obtained that did not contain free hydroxyl groups in the phenyl ring. The reaction was carried out according to a published procedure using acetone with K2CO3 under heating [54].
When analyzing the structure of 3-benzoyl derivatives by NMR (6a, 7a), it was noticed that upon storage in a DMSO-d6 solution, the benzoyl group undergoes decomposition with the formation of benzoic acid. This process occurs for both the unsubstituted TZD 6a (see Supplementary, p.98, Figure S1) and the 5-arylidene derivative 7a (see Supplementary, p.98, Figures S2 and S3). With no substituents in position 5 of the thiazolidinedione ring, the 3-benzoyl derivative TZD 6a hydrolyses in a solution of DMSO-d6 with a half-life of 5 h at 25 °C. The half-life of 5-substituted 3-benzoyl-TZD 7a in identical conditions was 11 h. A similar hydrolysis reaction also occurs in other solvents containing water. In the methanol solution, more complex reactions, involving the solvent, can occur (see Supplementary, p.99, Figure S3). Scheme 3. Synthesis of the target 3-benzoyl-5-arylidene-thiazolidine-2,4-dione 7a,b. Reagents and conditions: (iv) dry acetone, K 2 CO 3 , 60 • C.
When analyzing the structure of 3-benzoyl derivatives by NMR (6a, 7a), it was noticed that upon storage in a DMSO-d 6 solution, the benzoyl group undergoes decomposition with the formation of benzoic acid. This process occurs for both the unsubstituted TZD 6a (see Supplementary Figure S1) and the 5-arylidene derivative 7a (see Supplementary Figures S2 and S3). With no substituents in position 5 of the thiazolidinedione ring, the 3-benzoyl derivative TZD 6a hydrolyses in a solution of DMSO-d 6 with a half-life of 5 h at 25 • C. The half-life of 5-substituted 3-benzoyl-TZD 7a in identical conditions was 11 h. A similar hydrolysis reaction also occurs in other solvents containing water. In the methanol solution, more complex reactions, involving the solvent, can occur (see Supplementary Figure S3).
Stability is a highly important parameter for the active substance of a perspective drug. Thus, due to the instability of 3-benzoyl-substituted derivatives of TZD, we did not continue the study of this series of compounds containing a benzoyl group at the nitrogen atom of the thiazolidine ring.
To obtain acetic acid derivatives, we used well-known methods of synthesis, with slight changes and optimizations at some stages. We also prepared dioxothiazolidine derivatives and a new 2-thioxo-4-thiazolidinone analogue, previously described in the literature, for comparative activity studies. Both dioxothiazolidine and 2-thioxo-4-thiazolidinone derivatives were prepared and assayed to fully investigate this series.
Synthesis of the target 2-(2,4-dioxothiazolidin-3-yl)acetic acid derivatives 12a-e and their 2-thioxo analogs 17a,e is possible via two paths: either through 5-arylidene derivatives of TZD 4a-e or through 3-substituted TZD 9 or its 2-thioxo derivative 15. Condensation of commercially available 1, 14, and 15 with aromatic aldehydes was carried out according to a previously developed method in boiling acetic acid with a 33% methylamine as a catalyst (method A) [56]. Some of the 5-arylidene derivatives 4, 10, and 12 were obtained and described before: 4a [53,67-72]; 4c [69,70]; 4d [70]; 10a [57]; 12b, 17b [12]). The alkylation reaction for the compounds 4a-e and 15b with chloroacetic or bromoacetic acid esters was carried out in a DMF solution in the presence of K 2 CO 3 . Then, acidic hydrolysis of the ester group of the thioxo derivative 16 and dioxo derivatives 10-11a,b was performed. The action of a mixture of acetic and hydrochloric acids upon heating led to partial hydrolysis of benzyloxy groups and low yield of 12b. Under mild conditions of tert-butyl ester hydrolysis by trifluoroacetic acid in methylene chloride at room temperature, benzyloxy groups are not affected and thereby the yield of 12b increases. According to a published method [12], refluxing 2-(2,4-dioxo-1,3-thiazolidin-3-yl)acetic acid with 3,4-di(benzyloxy)benzaldehyde in an acetic acid with sodium acetate gave the product 12b in a very small yield.
Sulfur-containing counterparts 17b,e were obtained in a similar manner (Scheme 5). The compounds containing a hydroxyl group with two chlorine atoms on the phenyl ring (12d,e and 17e) were obtained by Knoevenagel condensation of (2,4-dioxothiazolidin-3yl)acetic acid or commercially available 13 and 14 with corresponding aromatic aldehyde under conditions previously described for other derivatives of TZD: acetic acid as a solvent with 33% methylamine as a catalyst. (ii) AcOH, MeNH 2 , reflux; (iii) EtOH, piperidine, reflux.
Condensation of 9 and 3,4-di(benzyloxy)benzaldehyde at different Knoevenagel conditions in ethanol with piperidine as a catalyst (method B), instead of acetic acid, with CH 3 COONa [12] increased the yield of 12b only to 14% after 28 h of reflux.
Sulfur-containing counterparts 17b,e were obtained in a similar manner (Scheme 5). The compounds containing a hydroxyl group with two chlorine atoms on the phenyl ring (12d,e and 17e) were obtained by Knoevenagel condensation of (2,4-dioxothiazolidin-3-yl)acetic acid or commercially available 13 and 14 with corresponding aromatic aldehyde under conditions previously described for other derivatives of TZD: acetic acid as a solvent with 33% methylamine as a catalyst.
CH3COONa [12] increased the yield of 12b only to 14% after 28 h of reflux.
Sulfur-containing counterparts 17b,e were obtained in a similar manner (Scheme 5). The compounds containing a hydroxyl group with two chlorine atoms on the phenyl ring (12d,e and 17e) were obtained by Knoevenagel condensation of (2,4-dioxothiazolidin-3yl)acetic acid or commercially available 13 and 14 with corresponding aromatic aldehyde under conditions previously described for other derivatives of TZD: acetic acid as a solvent with 33% methylamine as a catalyst. During the acquisition of the 1 H NMR spectra in DMSO-d6 and CDCl 3 solutions, a partial decomposition of compounds containing bis-benzyloxy substituents in the phenyl ring was observed, both for 12b and 4b. For the thioxo-containing analog 15b, the hydrolysis process begins in 10-15 min after dissolution in CDCl 3 (see Supplementary Material: the process of partial decomposition of 15b in CDCl 3 ).

Evaluation of the Biological Activity of Synthesized Compounds
In vitro minimum inhibitory concentration (MIC) of all target compounds was determined using the method recommended by the National Committee for Clinical Laboratory Standards (NCCLS) Clinical and Laboratory Standards Institute (CLSI) and the serial dilution method in 96-well plates [73,74]. All of the target compounds were evaluated for their antifungal activity against five important fungal pathogens. Fluconazole (FLC) was used as a reference drug. The in vitro antifungal activity results are summarized in Table 3.
As shown in Table 3, the substitution of the nitrogen atom of the thiazolidine ring with an alkoxycarbonyl or benzoyl radical led to a significant increase in antifungal activity compared to unsubstituted derivatives of TZD 4a-e, with respect to all types of yeast and filamentous fungi. The substitution of alkyl groups of alkoxycarbonyl derivatives with phenyl (5d) or benzyl (5e) radicals negatively affected antimicrobial activity. A similar result was observed with substituted acetic acid esters 10-11a,b. Table 3. Antifungal activity of synthesized compounds (MIC in mg/L). Experiments were performed at least 3 times.
In vitro minimum inhibitory concentration (MIC) of all target compounds was determined using the method recommended by the National Committee for Clinical Laboratory Standards (NCCLS) Clinical and Laboratory Standards Institute (CLSI) and the serial dilution method in 96-well plates [73,74]. All of the target compounds were evaluated for their antifungal activity against five important fungal pathogens. Fluconazole (FLC) was used as a reference drug. The in vitro antifungal activity results are summarized in Table  3. Dibenzyloxy derivatives of TZD 4b and its thioxo analog 15b did not show high antifungal activity. Furthermore, the 2,4-dichloro derivative 12c exhibited high activity against A. fumigatus ATCC 46645 (2 mg/L) and moderate activity against other types of filamentous fungi (16-32 mg/L).
Other acetic acid derivatives 12d-e containing chloro and hydroxy groups in the benzylidene radical showed moderate antifungal activity against filamentous fungi (MIC = 16 µg/mL) and the 2-thioxothiazolidin-4-one derivative The bis-benzyloxy derivative of TZD 12b and its thioxo analog 17b did not show good MIC values, despite the high inhibitory activity of 17b toward mannosyltransferase reported in some articles [12,13]. This fact indicates that the effect of thiazolidine derivatives on the fungal cell wall may not be associated with the inhibition of PMT1; it also extends to other targets that have yet to be clarified. When the 2-hydroxy group was combined with 3,5-dichloro substituents, the antifungal activity became more pronounced on C. albicans yeasts, so an additional experiment on clinical isolates of Candida spp. was conducted.
On clinical strains of Candida spp. in comparison with Fluconazole and Mycosidine, 3-substituted TZD derivatives 5a and 7a showed moderate activity similar to Mycosidine with MIC ranging from 8 to 64 mg/L (Table 4). High activity of 2-hydroxy-3,5-dichloro derivative 12e was demonstrated on yeasts C. albicans ATCC 24433 and C. parapsilosis ATCC 22019 (MIC 0.125-0.5 mg/L), whereas activity on other Candida spp. was much lower.

Cellular Response to Mycosidine
In order to gain a deeper understanding of the effects of TZD derivatives on fungal cells, we assayed the response of a model fungal organism-Saccharomyces cerevisiae-to Mycosidine, which was available in the highest quantities. To do this, we initially determined the MIC (15.6 mg/L) and assayed the number of cells with permeabilized membranes using propidium iodide (PI). Notably, at MIC and below it, the population of dead cells was~20% after 6 h of treatment, which suggests that the breakdown of the cell wall is not the primary mechanism that stops cell division. We also tested whether Mycosidine had a fungicidal effect on the cells by semi-quantitatively assaying the number of CFU using a spotting assay. This test showed that Mycosidine had a noticeable fungicidal effect, but only after 24 h, with no massive cell death observed after 2, 4, or 6 h of incubation, i.e., cell death in response to the drug is rather slow ( Figure 5A).
Then, we tested whether cells treated with this drug at sublethal concentrations experience some specific changes of protein levels. This was done by monitoring the levels of a select number of proteins tagged with GFP using flow cytometry. The proteins we tested were mainly selected from those considered as "sentinel-proteins" [75] as well as several additional proteins (65 proteins and autofluorescence control-see Supplementary Material). We used different concentrations of the drug-MIC and 0.5×MIC for screening all of the proteins-and then at 0.5-4×MIC for the confirmation tests. Notably, all of the experiments were performed with simultaneous staining for cells with membranes, and the presented data are for living cells only.
Our results show that Mycosidine caused specific increases in the levels of at least two proteins-Pdr5 and Hxt3 ( Figure 5B)-when compared to multiple other drugs that were tested in a similar manner (manuscript in preparation). Notably, this increase was only observed at 0.5×MIC and disappeared at higher concentrations. Pdr5 is an ABC-transporter, involved in the efflux of multiple compounds, including antifungal drugs such as azoles. Hxt3 is a low-affinity glucose transporter, which has not, to our knowledge, been previously implicated in responses to antifungal drugs. cells, we assayed the response of a model fungal organism-Saccharomyces cerevisiae-to Mycosidine, which was available in the highest quantities. To do this, we initially determined the MIC (15.6 mg/l) and assayed the number of cells with permeabilized membranes using propidium iodide (PI). Notably, at MIC and below it, the population of dead cells was ~20% after 6 h of treatment, which suggests that the breakdown of the cell wall is not the primary mechanism that stops cell division. We also tested whether Mycosidine had a fungicidal effect on the cells by semi-quantitatively assaying the number of CFU using a spotting assay. This test showed that Mycosidine had a noticeable fungicidal effect, but only after 24 h, with no massive cell death observed after 2, 4, or 6 h of incubation, i.e., cell death in response to the drug is rather slow ( Figure 5A). Then, we tested whether cells treated with this drug at sublethal concentrations experience some specific changes of protein levels. This was done by monitoring the levels of a select number of proteins tagged with GFP using flow cytometry. The proteins we tested were mainly selected from those considered as "sentinel-proteins" [75] as well as several additional proteins (65 proteins and autofluorescence control-see Supplementary Material). We used different concentrations of the drug-MIC and 0.5×MIC for screening all of the proteins-and then at 0.5-4×MIC for the confirmation tests. Notably, Having identified that these proteins are induced by drug treatment, we tested whether these proteins played a role in the sensitivity of cells to this drug by using strains containing deletions of the corresponding genes [76]. Deletion of Pdr5 had no effect ( Figure 5C), which shows that Mycosidine might not be a substrate of the Pdr5 transporter and that its induction might be a non-specific response of the cell to stress caused by the drug. A similar effect was recently shown for the protonophore pentachlorophenol [77].
Interestingly, the deletion of Hxt3 increased the sensitivity of the yeast to the drug ( Figure 5C), which suggests that the response of glucose transporters may be necessary for adaptation to the drug. Because yeast has numerous glucose transporters, it might be that one or several of them are targets for Mycosidine or that glucose transport is a resistance mechanism; this requires further study.
Lastly, monitoring of the amount of the GFP-tagged histone Htb2 allows us to monitor the dynamics of the cell cycle of drug treated cells ( Figure 5D). This analysis shows that Mycosidine at 0.5×MIC reduces the share of cells with a 2C complement of DNA (because Htb2 levels and DNA levels are usually well correlated) and increases the height of the area between the 1C and 2C peaks, suggesting that progression of the S-phase of the cell cycle is impaired. The MIC concentration of Mycosidine causes most cells to exhibit a 1C distribution, while also causing emergence of an additional population of cells with lower fluorescence, which might suggest apoptotic cell death. This suggests impaired cell cycle progression prior to the initiation of replication as well as, possibly, apoptosis-like cell death.

Effects on the Cell Wall
SICM is a non-contact type of scanning probe microscopy applicable for the investigation of live biological samples in liquid [78][79][80][81]. This method was used to study the influence of antifungal thiazolidines on the topography of living Candida cells. It is reasonable to use effective concentrations to see any changes in topography induced by the drugs in question and for this reason concentration equal to MIC and 10×MIC for both Mycosidine and 17b were chosen. The MIC for Mycosidine is two times lower than the same one for 17b, but 17b was still used as it is reported to be a highly active inhibitor of mannosyltransferase. Spikes that were 100-200 nm in diameter were observed after treatment of the Candida cells with Mycosidine and 17b compounds. The appearance of such spikes on the Candida cell surface is most probably related to the disruption of the cell wall. Only cells attached to the substrate can be investigated with SCIM and probably for this reason we can easily find some resistant cells without any spikes. The obtained results clearly demonstrate that the action of both Mycosidine and 17b causes disruption of the cell wall but Mycosidine achieves the same effect with much lower concentration ( Figure 6).

Materials and General Methods
All the reagents were obtained commercially and used without further purification. Thiazolidine-2,4-dione 1 and (4-oxo-2-thioxo-3-thiazolidinyl)acetic acid 3 were purchased from Sigma-Aldrich. Purity of the compounds was checked by thin layer chromatography using silica-gel 60 F254-coated Al plates (Merck) and spots were observed under UV light. 1 H NMR and 13 C NMR spectra were recorded on a Bruker Avance spectrometer (600 and 150 MHz, respectively) and Varian VXR-400 spectrometer at 400 MHz and 101 MHz, respectively, at 298 K in CDCl3 or DMSO-d6 at a concentration of samples of 5-15 mmol, with TMS as internal reference for 1 H and 13 C NMR spectra. The signal assignments of compound 11, 12, 15, and 17 were performed using 2D spectra (DQF-COSY, 13 C-1 H HSQC, and 13 C-1 H HMBC); the chemical shifts are expressed in ppm (δ scale) using DMSO and CDCl3 as an internal standard, and the coupling constants are expressed in Hz. The massspectral measurements were carried out by ESI method on micrOTOF-QII (Brucker Daltonics GmbH). Analytical HPLC was performed on a Shimadzu LC-20AD system using Kromasil-100-5-C18 (Akzo-Nobel) column, 4.6 × 250 mm, temperature 20 °C, UV detection, mobile phase A-0.2% HCOONH4), mobile phase B -MeCN, (pH 7.4), and fl-1ml/min.

Materials and General Methods
All the reagents were obtained commercially and used without further purification. Thiazolidine-2,4-dione 1 and (4-oxo-2-thioxo-3-thiazolidinyl)acetic acid 3 were purchased from Sigma-Aldrich. Purity of the compounds was checked by thin layer chromatography using silica-gel 60 F254-coated Al plates (Merck) and spots were observed under UV light. 1 H NMR and 13 C NMR spectra were recorded on a Bruker Avance spectrometer (600 and 150 MHz, respectively) and Varian VXR-400 spectrometer at 400 MHz and 101 MHz, respectively, at 298 K in CDCl 3 or DMSO-d 6 at a concentration of samples of 5-15 mmol, with TMS as internal reference for 1 H and 13 C NMR spectra. The signal assignments of compound 11, 12, 15, and 17 were performed using 2D spectra (DQF-COSY, 13 C-1 H HSQC, and 13 C-1 H HMBC); the chemical shifts are expressed in ppm (δ scale) using DMSO and CDCl 3 as an internal standard, and the coupling constants are expressed in Hz. The mass-spectral measurements were carried out by ESI method on micrOTOF-QII (Brucker Daltonics GmbH). Analytical HPLC was performed on a Shimadzu LC-20AD system using Kromasil-100-5-C18 (Akzo-Nobel) column, 4.6 × 250 mm, temperature 20 • C, UV detection, mobile phase A-0.2% HCOONH 4 ), mobile phase B-MeCN, (pH 7.4), and fl-1ml/min.

Synthesis of the Starting Materials
The synthetic routes used to synthesize starting materials are outlined in Schemes 1-3. The detailed description of the method and physico-chemical properties of 6a, 8, and 9 were described elsewhere [56,60].
Raw data were analyzed using CytExpert software (Beckman Coulter), and graphs were obtained by exporting data to MS Excel. Heatmaps were obtained using MS Excel and they depict the log 2 transformed ratio between the GFP fluorescence in the experimental sample divided by the control sample.

Sample Preparation for Cell Wall Investigation
C. parapsilosis 22,019 cells and C. albicans 10,231 cells were used for experiments. The topography of yeast cells was obtained in physiological saline at room temperature. A daily exposition of Candida suspension (10 3 CFU/mL) with drugs was performed in 5 mL petri dishes (GBO, Kremsmünster, Austria) modified with a layer of SYLGARD™ 527 A&B Silicone Dielectric Gel (Dow Chemical, Midland, MI, USA) to increase cell-surface adhesion and prevent sample drifting and reattachment. The surface of a petri dish was covered with the silicone gel with a ratio of components A and B in 1: 1 (by mass fraction) and kept in a laboratory drying cabinet Ulab UT-4686 (ULAB, Dhaka, Bangladesh) for an hour at 60 • C; then, the dishes were washed with distilled water. Concentration of cells in the suspension in physiological saline (PanEco, RF) was determined using hemocytometer (MiniMed, RF) and added to the modified petri dish. Dry drugs were dissolved in DMSO 99.7% (Sigma-Aldrich, St. Louis, MO, USA) and added to cells in physiological saline with final concentrations 16, 160 µg/mL of Mycosidine and 32, 320 µg/mL of 17. After incubation of cells for 24 h, the dishes were washed and filled with saline.

Scanning Ion-Conductance Microscopy (SICM)
SICM experiments were performed at room temperature by using SICM by ICAPPIC (ICAPPIC Ltd., London, UK) and borosilicate glass capillaries (Sutter Instruments, Novato, CA, USA). Nanopipettes were made from capillaries with an external diameter of 1.2 mm and an internal diameter of 0.69 mm on a P-2000 puller (Sutter Instruments, Novato, CA, USA). Topography was measured in saline [78] on cells immobilized on a layer of silicone dielectric gel. SICM measurements were performed using hopping mode protocol [79]. Setpoint parameter of ion current feedback control was 0.2%, providing a minimal influence of nanopipette probe on the sample. The nanopipette radius ranged from 20 to 35 nm to achieve a nanoscale resolution of topography imaging. The radius of the nanopipette and the mechanical characteristics were calculated using formulas reported in previous studies [80]. Cell images were recorded on areas of 10 × 10 µm at 74 nm resolution and on areas of 1.5 × 1.5 µm for 21 nm resolution. The image dimensions were 256 × 256 pixels. For each concentration point, three to five large images were obtained for a total of 5 to 10 cells. Image processing was performed using the "SICMImageViewer" software.

Conclusions
Novel derivatives of Mycosidine-3,5-substituted thiazolidine-2,4-diones were synthesized in excellent yield using Knoevenagel synthesis and alkylation or acylation of the imide nitrogen. Furthermore, 5-Arylidene-2,4-thiazolidinediones and 2-thioxo analogs containing halogen and hydroxy groups or a di(benzyloxy) group in the 5-benzylidene moiety were evaluated for their in vitro antifungal activity. Some of them exhibited strong activity against filamentous fungi (Trichophyton rubrum and Microsporum canis) and Candida spp. yeasts. The compounds were characterized by chromatographic and spectrometric methods. The compounds exhibited both fungistatic and fungicidal activity and resulted in the emergence of morphological changes in the cell wall of the Candida yeast. Using limited proteomic screening and the analysis of toxicity in mutants, we show that Mycosidine activity is dependent on glucose transport. This suggests that this first-in-class antifungal and its derivatives have a novel mechanism of action, which deserves further study.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ph15050563/s1, Figure S1: Fragments of 1 H NMR spectra of solution of 6a in DMSO-d6, recorded at 25 • C and 600 MHz at various time intervals after sample dissolution. Labeled are representative signals of initial compound and product of hydrolysis; Figure S2: Fragments of 1 H NMR spectra of a solution of 7a in DMSO-d6, recorded at 25 • C at 600 MHz at various time intervals after sample dissolution. Figure S3: Fragments of 1 H NMR spectra of a solution of 7a in CD 3 OH, recorded at 25 • C at 600 MHz at various time intervals after sample dissolution. Figure S4: Hydrolysis of 6a and 7a to 1,4a and benzoic acid in DMSO-d6 and CD 3 OD.