A Study on the Chemistry and Biological Activity of 26-Sulfur Analogs of Diosgenin: Synthesis of 26-Thiodiosgenin S-Mono- and Dioxides, and Their Alkyl Derivatives

A chemoselective procedure for MCPBA oxidation of 26-thiodiosgenin to corresponding sulfoxides and sulfone was elaborated. An unusual equilibration of sulfoxides in solution was observed. Moreover, α-alkylation of sulfoxide and sulfone was investigated. Finally, the biological activity of obtained compounds was examined.


Introduction
Saponins are a class of chemical compounds abundant in various plant species [1]. More specifically, they are amphiphilic glycosides producing the soap-like foam when shaken in aqueous solutions. The lipophilic triterpene or steroid aglycones (sapogenins) are combined in these compounds with one or more hydrophilic sugar moieties. The steroidal sapogenins usually contain 27 carbon atoms and the oxidized side chain, which forms a spiroacetal system characteristic for steroidal spirostanes, e.g., diosgenin (1) (Figure 1). These compounds have received considerable attention as precursors for synthesizing sex hormones and various steroidal drugs.
There is also a group of naturally occurring compounds based on a C 27 cholestane skeleton, which are essentially nitrogen analogs of spirostane sapogenins, e.g., solasodine (2) (Figure 1). These steroidal alkaloids (spirosolanes) are a class of secondary metabolites isolated from plants (mostly of the family Solanaceae), amphibians, and marine invertebrates [2][3][4][5]. Evidence accumulated in the last two decades demonstrates that steroidal alkaloids of the spirosolane group show a wide range of bioactivities, including anticancer, antimicrobial, anti-inflammatory, antinociceptive, etc., suggesting their great potential for pharmaceutical application. Several comprehensive review articles on the alkaloid bioactivity, especially anticancer activity, and the mechanism of their biological action have recently appeared [6][7][8].
The replacement of the F-ring oxygen atom in spirostane sapogenins with a different heteroatom severely affects the chemical properties of a steroid and may result in useful alterations to its biological activity. The potential of heterosteroids as novel drugs encourages organic chemists to undertake studies in this field. The sulfur analogs of steroidal sapogenins do not occur in nature but 26-thiodiosgenin (3), the diosgenin F-ring sapogenins do not occur in nature but 26-thiodiosgenin (3), the diosgenin F-ring thiacounterpart, was described in the literature many years ago [9]. However, efficient syntheses of this compound have been reported only recently. In particular, the one-pot Wang synthesis [10], which involves the treatment of a solution of diosgenin in dichloromethane with hydrogen sulfide gas in the presence of BF3·Et2O as a catalyst, was the most advantageous for us. It has been shown that replacing the F-ring oxygen atom of diosgenin with sulfur increases compound cytotoxicity against different cancer cell lines, especially in the case of glycosyl derivatives. For example, IC50 of a natural saponin, dioscine, against the lung cancer cell line A549 IC50 was found 4.02 μM versus 3.72 μM for 26-thiodioscine [11,12].   (3) is now readily available from diosgenin, its chemistry has not been explored yet. The Se-analog of diosgenin (4), in which selenium atom replaces the F-ring spiroketal oxygen, has also been described, but we have found the literature procedure difficult to reproduce [12,13], i.e., the product we obtained proved to be a mixture of stereoisomers. The development in partial and total syntheses of thiasteroids has been reviewed, but steroid analogs containing sulfur in the side chain were not included in this article [14]. We have recently described a simple synthesis of carbaanalogs of steroidal sapogenins along with their biological activity [15]. Unfortunately, the carbaanalog (5) was obtained as an inseparable mixture of cis/trans isomers (note that C22 and C25 are not stereogenic centers in this compound).

Chemistry
The presence of a soft sulfur atom in 26-thiodiosgenin (3) alters its chemical, physical, and biological properties, making it, among others, very susceptible to oxidation. In fact, diosgenin (2) can also be oxidized with different reagents; its oxidation sites are the C5-C6 double bond, the 3β-hydroxyl group, and the carbon atoms C20 or C23. However, the introduction of sulfur in place of oxygen changes the reactivity of compound and directs the oxidation of 3 to the sulfur atom. It is well known that the oxidation of sulfides is a two-step process. In the first step, only one oxygen atom is transferred from the oxidizing agent to sulfur, forming sulfoxide. Since a new stereogenic center is generated at sulfur during this step (provided that the starting sulfide is not symmetrical), two diastereomeric sulfoxides can be formed. Further oxidation of both sulfoxides leads to a single sulfone. We have recently described a simple synthesis of carbaanalogs of steroidal sapogenins along with their biological activity [15]. Unfortunately, the carbaanalog (5) was obtained as an inseparable mixture of cis/trans isomers (note that C22 and C25 are not stereogenic centers in this compound).

Chemistry
The presence of a soft sulfur atom in 26-thiodiosgenin (3) alters its chemical, physical, and biological properties, making it, among others, very susceptible to oxidation. In fact, diosgenin (2) can also be oxidized with different reagents; its oxidation sites are the C5-C6 double bond, the 3β-hydroxyl group, and the carbon atoms C20 or C23. However, the introduction of sulfur in place of oxygen changes the reactivity of compound and directs the oxidation of 3 to the sulfur atom. It is well known that the oxidation of sulfides is a two-step process. In the first step, only one oxygen atom is transferred from the oxidizing agent to sulfur, forming sulfoxide. Since a new stereogenic center is generated at sulfur during this step (provided that the starting sulfide is not symmetrical), two diastereomeric sulfoxides can be formed. Further oxidation of both sulfoxides leads to a single sulfone. The oxidation of sulfoxides to sulfones is relatively easy, and sometimes it is difficult to stop the oxidation of sulfides at an intermediate stage. The simplest method Molecules 2023, 28, 189 3 of 18 of sulfoxide and sulfone synthesis is the sulfide oxidation with halogen derivatives and metal-mediated oxidative systems [16]. However, due to growing concerns about chemical pollution and environmental protection, there is a tendency to use hydrogen peroxide or peroxy acids as atom-efficient and environmentally benign oxidizing agents [16,17]. The oxidation of organic sulfides with peroxycarboxylic acids or with hydrogen peroxide, which is usually activated by conventional acidic catalysts, often leads to various side reactions such as over-oxidation (sulfoxides to sulfones), epoxidation (if a double bond is present) and Baeyer-Villiger oxidation (if an oxo group is present). After testing several methods of the chemoselective sulfide oxidation, we turned our attention to the method employing hydrogen peroxide and N-hydroxysuccinimide (NHS) [18]. According to the literature, sulfoxides can be obtained with this reagent in acetone under reflux conditions, without over-oxidation to sulfones, and the method is compatible with the presence of sensitive groups, including alkenes and hydroxyl groups. Nonetheless, a complex mixture of products was formed when 26-thiodiosgenin (3) was subjected to the described reaction conditions. In turn, when compound 3 was treated with this oxidizing system under milder conditions (50 • C, 10 h), the corresponding sulfone 8 was obtained as the only product in 70% yield (Scheme 1). The formation of intermediate sulfoxides 6 and 7 was not observed, even after lowering the reaction temperature to 30 • C and reducing the reaction time. The yield of sulfone 8 then dropped to 68%. The oxidation of sulfoxides to sulfones is relatively easy, and sometimes it is difficult to stop the oxidation of sulfides at an intermediate stage. The simplest method of sulfoxide and sulfone synthesis is the sulfide oxidation with halogen derivatives and metal-mediated oxidative systems [16]. However, due to growing concerns about chemical pollution and environmental protection, there is a tendency to use hydrogen peroxide or peroxy acids as atom-efficient and environmentally benign oxidizing agents [16,17]. The oxidation of organic sulfides with peroxycarboxylic acids or with hydrogen peroxide, which is usually activated by conventional acidic catalysts, often leads to various side reactions such as over-oxidation (sulfoxides to sulfones), epoxidation (if a double bond is present) and Baeyer-Villiger oxidation (if an oxo group is present). After testing several methods of the chemoselective sulfide oxidation, we turned our attention to the method employing hydrogen peroxide and N-hydroxysuccinimide (NHS) [18]. According to the literature, sulfoxides can be obtained with this reagent in acetone under reflux conditions, without over-oxidation to sulfones, and the method is compatible with the presence of sensitive groups, including alkenes and hydroxyl groups. Nonetheless, a complex mixture of products was formed when 26-thiodiosgenin (3) was subjected to the described reaction conditions. In turn, when compound 3 was treated with this oxidizing system under milder conditions (50 °C, 10 h), the corresponding sulfone 8 was obtained as the only product in 70% yield (Scheme 1). The formation of intermediate sulfoxides 6 and 7 was not observed, even after lowering the reaction temperature to 30 °C and reducing the reaction time. The yield of sulfone 8 then dropped to 68%. Another reagent recommended for oxidation of sulfides to sulfoxides is m-chloroperoxybenzoic acid (MCPBA) [19]. The reaction was carried out with 1.1 equiv. of this reagent at −78 °C (dry ice/acetone bath). The progress of the reaction was monitored by TLC, which showed a single spot of a very polar product (more polar than sulfone 8). After two hours, the reaction came to completion and was quenched with dimethylsulfide.
Although the product appeared to be single by TLC, a detailed analysis of its spectra ( Figure 2) revealed that it was a mixture of sulfoxides, one of which was by far predominant. In the 1 H NMR spectrum, a tiny signal at about 3 ppm derived from protons of the C26 methylene group of the minor sulfoxide (the C26 methylene protons of the major one appeared as a doublet at δ 2.78 and triplet at δ 2.57 ppm). Furthermore, the LC-MS analysis unambiguously confirmed the presence of trace amounts of the second sulfoxide. Unfortunately, the separation of these isomers was practically impossible, even by crystallization of crude product. Another reagent recommended for oxidation of sulfides to sulfoxides is m-chloroperoxybenzoic acid (MCPBA) [19]. The reaction was carried out with 1.1 equiv. of this reagent at −78 • C (dry ice/acetone bath). The progress of the reaction was monitored by TLC, which showed a single spot of a very polar product (more polar than sulfone 8). After two hours, the reaction came to completion and was quenched with dimethylsulfide.
Although the product appeared to be single by TLC, a detailed analysis of its spectra ( Figure 2) revealed that it was a mixture of sulfoxides, one of which was by far predominant. In the 1 H NMR spectrum, a tiny signal at about 3 ppm derived from protons of the C26 methylene group of the minor sulfoxide (the C26 methylene protons of the major one appeared as a doublet at δ 2.78 and triplet at δ 2.57 ppm). Furthermore, the LC-MS analysis unambiguously confirmed the presence of trace amounts of the second sulfoxide. Unfortunately, the separation of these isomers was practically impossible, even by crystallization of crude product. The reaction of 26-thiodiosgenin (3) with 2.2 equiv. of MCPBA at −40 °C afforded sulfone 8 as the major product, in addition to 10% yield of sulfoxides 6/7 and 2% yield of the over-oxidized product 5,6-epoxysulfone (Scheme 2). The low temperature (−78 °C) MCPBA oxidation of 26-thiodiosgenin 4-nitrobenzoate (3a) was also carried out with the expectation of obtaining the well-crystallizing and easily separable products. Indeed, it turned out that the separation of isomeric sulfoxides was possible not only by HPLC but also by a silica gel (230-400 mesh) gravity flow column chromatography. The less polar minor product was obtained in 9% yield, while the major one in 71% yield (Scheme 3). The reaction of 26-thiodiosgenin (3) with 2.2 equiv. of MCPBA at −40 °C afforded sulfone 8 as the major product, in addition to 10% yield of sulfoxides 6/7 and 2% yield of the over-oxidized product 5,6-epoxysulfone (Scheme 2). The low temperature (−78 °C) MCPBA oxidation of 26-thiodiosgenin 4-nitrobenzoate (3a) was also carried out with the expectation of obtaining the well-crystallizing and easily separable products. Indeed, it turned out that the separation of isomeric sulfoxides was possible not only by HPLC but also by a silica gel (230-400 mesh) gravity flow column chromatography. The less polar minor product was obtained in 9% yield, while the major one in 71% yield (Scheme 3). The low temperature (−78 • C) MCPBA oxidation of 26-thiodiosgenin 4-nitrobenzoate (3a) was also carried out with the expectation of obtaining the well-crystallizing and easily separable products. Indeed, it turned out that the separation of isomeric sulfoxides was possible not only by HPLC but also by a silica gel (230-400 mesh) gravity flow column chromatography. The less polar minor product was obtained in 9% yield, while the major one in 71% yield (Scheme 3). A detailed analysis of 1 H NMR spectra of sulfoxides ( Figure 3) can unequivocally ascribe the configuration at the sulfur atom. The diagnostic for the configuration assignment proved to be the H-16α signal. Cone shielding anisotropy generated by the sulfinyl group is similar to that of the carbonyl group. The inspection of Dreiding models, as well as molecular modeling employing the MM+ calculations (Figure 4), show that the α proton at C16 is in close proximity to the sulfinyl group (deshielding zone) in the equatorial sulfoxide 7 (configuration R at sulfur), while the effect of the axial S-oxide 6 (configuration S at sulfur) on this proton is negligible. As can be seen from Table 1, the H-16α signal appears at δ 4.64 in 26-thiodiosgenin (3), at δ 4.55-4.58 for axial sulfoxides 6 (or 6a), and it is strongly deshielded (to δ 5.37-5.38) for equatorial sulfoxides 7 (or 7a).  A detailed analysis of 1 H NMR spectra of sulfoxides ( Figure 3) can unequivocally ascribe the configuration at the sulfur atom. The diagnostic for the configuration assignment proved to be the H-16α signal. Cone shielding anisotropy generated by the sulfinyl group is similar to that of the carbonyl group. The inspection of Dreiding models, as well as molecular modeling employing the MM+ calculations (Figure 4), show that the α proton at C16 is in close proximity to the sulfinyl group (deshielding zone) in the equatorial sulfoxide 7 (configuration R at sulfur), while the effect of the axial S-oxide 6 (configuration S at sulfur) on this proton is negligible. As can be seen from Table 1, the H-16α signal appears at δ 4.64 in 26-thiodiosgenin (3), at δ 4.55-4.58 for axial sulfoxides 6 (or 6a), and it is strongly deshielded (to δ 5.37-5.38) for equatorial sulfoxides 7 (or 7a). A detailed analysis of 1 H NMR spectra of sulfoxides ( Figure 3) can unequivocally ascribe the configuration at the sulfur atom. The diagnostic for the configuration assignment proved to be the H-16α signal. Cone shielding anisotropy generated by the sulfinyl group is similar to that of the carbonyl group. The inspection of Dreiding models, as well as molecular modeling employing the MM+ calculations (Figure 4), show that the α proton at C16 is in close proximity to the sulfinyl group (deshielding zone) in the equatorial sulfoxide 7 (configuration R at sulfur), while the effect of the axial S-oxide 6 (configuration S at sulfur) on this proton is negligible. As can be seen from Table 1, the H-16α signal appears at δ 4.64 in 26-thiodiosgenin (3), at δ 4.55-4.58 for axial sulfoxides 6 (or 6a), and it is strongly deshielded (to δ 5.37-5.38) for equatorial sulfoxides 7 (or 7a).      The sulfoxides are usually configurationally stable. However, in our case, the sulfoxides are stable in solid state only, but in solution (dichloromethane) a slow equilibration between the isomeric sulfoxides 6 and 7 occurs. The solution of 6a was allowed to stand under argon at room temperature and the progress of isomerization was monitored by TLC. The equilibrium was reached within 14 days, and then the ratio of 6a:7a was 3:2. The same isomeric ratio was obtained when the equatorial sulfoxide 7a was subjected to equilibration. This result is consistent with molecular mechanics calculations which showed a slightly lower steric energy for the axial sulfoxide ( Figure 4). The equilibration between sulfoxides can be reached much faster (10 min) if a catalytic amount of p-TsOH is added. However, when the reaction mixture was allowed to stand for a longer time, further isomerization processes occurred (presumably at C20 and C22); as a result, up to 8 isomers can be formed. The tentative mechanism of the sulfoxide equilibration is shown in Scheme 4.  The sulfoxides are usually configurationally stable. However, in our case, the sulfoxides are stable in solid state only, but in solution (dichloromethane) a slow equilibration between the isomeric sulfoxides 6 and 7 occurs. The solution of 6a was allowed to stand under argon at room temperature and the progress of isomerization was monitored by TLC. The equilibrium was reached within 14 days, and then the ratio of 6a:7a was 3:2. The same isomeric ratio was obtained when the equatorial sulfoxide 7a was subjected to equilibration. This result is consistent with molecular mechanics calculations which showed a slightly lower steric energy for the axial sulfoxide ( Figure 4). The equilibration between sulfoxides can be reached much faster (10 min) if a catalytic amount of p-TsOH is added. However, when the reaction mixture was allowed to stand for a longer time, further isomerization processes occurred (presumably at C20 and C22); as a result, up to 8 isomers can be formed. The tentative mechanism of the sulfoxide equilibration is shown in Scheme 4. This unusual behavior of sulfoxides 6 and 7 is caused by an easy cleavage of the C22-S bond under acidic conditions. The cleavage leads to the sulfenic acid and the relatively stable oxocarbenium ion. The reverse reaction recovers the sulfoxide, but the configuration at the sulfur atom may be inversed. It should be noted that if the sulfoxide is treated with acid for a longer time, the C20 proton may be abstracted from the oxocarbenium ion to give the C20-C22 double bond, and then an isomerization at these carbon atoms may also occur.
On the other hand, the C22-S bond in sulfoxides 6/7 and sulfone 8 should be resistant to basic conditions. If so, an α-alkylation should be possible. Consequently, the starting 26-thiodiosgenin (3) was converted to tert-butyldimethylsilyl ether 3b with TBDMS-Cl and imidazole. Then 3b was oxidized with 1.1 equiv. of MCPBA at −78 • C affording sulfoxide 6b. After increasing the reaction temperature to −40 • C and raising the amount of oxidant to 2.2 equiv., the corresponding sulfone 8b was also prepared. The obtained sulfoxide 6b and sulfone 8b were deprotonated with n-butyllithium and then treated with 2.5 equiv. of methyl iodide (Scheme 5). The reaction of sulfoxide 6b led to the formation of the dimethylated product 10b in 91% yield. The α-methylation of sulfone 8b also proceeded smoothly, but only the mono substituted product 11b was obtained. The difference in the reaction course is probably due to a larger steric hindrance at C26 in the sulfone than in the corresponding sulfoxide. The α-methylation of sulfone 8b proved to be highly stereoselective, providing only one product 11b in 82% yield. The S configuration at the newly formed stereogenic center at C26 was concluded on the basis of the 1 H NMR signal of H-26, which appeared at δ 3.14 as a doublet of quartets with coupling constants J = 6.9 and 11.1 Hz. The latter comes from the coupling of two axial protons at C26 (alpha) and C25 (beta). It implies that the new methyl group at C26 assumed the equatorial position. The reaction of sulfone 8b with ethyl iodide also provided the α-substituted product, albeit in slightly lower yield (76%). Attempts of α-alkylation of 26-thiodiosgenin 3-TBDMS-ether (3b) failed, probably due to a low acidity of the α-proton in this compound. This result could be expected given the lower acidity of dimethylsulfide (pK a = 45.0) compared with dimethylsulfoxide (pK a = 35.1) and dimethylsulfone (pK a = 31.1).
On the other hand, the C22-S bond in sulfoxides 6/7 and sulfone 8 should be resistant to basic conditions. If so, an -alkylation should be possible. Consequently, the starting 26-thiodiosgenin (3) was converted to tert-butyldimethylsilyl ether 3b with TBDMS-Cl and imidazole. Then 3b was oxidized with 1.1 equiv. of MCPBA at −78 °C affording sulfoxide 6b. After increasing the reaction temperature to −40 °C and raising the amount of oxidant to 2.2 equiv., the corresponding sulfone 8b was also prepared. The obtained sulfoxide 6b and sulfone 8b were deprotonated with n-butyllithium and then treated with 2.5 equiv. of methyl iodide (Scheme 5). The reaction of sulfoxide 6b led to the formation of the dimethylated product 10b in 91% yield. The α-methylation of sulfone 8b also proceeded smoothly, but only the mono substituted product 11b was obtained. The difference in the reaction course is probably due to a larger steric hindrance at C26 in the sulfone than in the corresponding sulfoxide. The α-methylation of sulfone 8b proved to be highly stereoselective, providing only one product 11b in 82% yield. The S configuration at the newly formed stereogenic center at C26 was concluded on the basis of the 1 H NMR signal of H-26, which appeared at δ 3.14 as a doublet of quartets with coupling constants J = 6.9 and 11.1 Hz. The latter comes from the coupling of two axial protons at C26 (alpha) and C25 (beta). It implies that the new methyl group at C26 assumed the equatorial position. The reaction of sulfone 8b with ethyl iodide also provided the α-substituted product, albeit in slightly lower yield (76%). Attempts of α-alkylation of 26-thiodiosgenin 3-TBDMSether (3b) failed, probably due to a low acidity of the α-proton in this compound. This result could be expected given the lower acidity of dimethylsulfide (pKa = 45.0) compared with dimethylsulfoxide (pKa = 35.1) and dimethylsulfone (pKa = 31.1).

Antimicrobial Activity of Tested Compounds
Diosgenin (1) belongs to the group of steroidal compounds known as sapogenins, which are obtained from their glycoside forms, i.e., saponins [20]. It demonstrates various biological activities, i.a. anticancer, antioxidant, and anti-inflammatory activity [21]. In order to verify if newly synthesized sulfur analogs of diosgenin (1) possess antimicrobial activity, the MIC (minimal inhibition concentration) and MBC (minimal bactericidal concentration) values were estimated, using two bacterial strains: Gram-positive Staphylococcus aureus 8325-4 and Gram-negative Escherichia coli 35218. The obtained MIC and MBC values are shown in Table 2. Diosgenin (1) has a strong antimicrobial activity against S. aureus (Gram-positive bacteria), and its growth is inhibited at the concentration of 3.9 µg/mL, whereas four-times higher concentration (15.615 µg/mL) was estimated as MBC for these bacteria. Much weaker activity was detected for diosgenin (1) against Gram-negative E. coli where MIC and MBC were 250 µg/mL and 500 µg/mL, respectively. This effect may be due to the characteristic structure of the outer membrane of Gram-negative bacteria, which contains unique component-lipopolysaccharide [22]. Exchanging the oxygen atom with a sulfur atom in 26-thiodiosgenin (3) performed via chemical modification of the diosgenin F-ring results in an increase of antibacterial activity for both the S. aureus and E. coli. The obtained MIC and MBC values were: 1.95 µg/mL and 3.9 µg/mL for S. aureus, and 62.5 µg/mL and 125 µg/mL for E. coli. This may be a consequence of the presence of a sulfur atom in the structure of these compounds. It is well know that sulfur functional groups are found in many pharmaceuticals, including penicillin, prevacid (lansoprazole), seroquel (quetiapine), dapsone, or sulfamethoxazole [23]. From the chemical point of view, they belong to different classes of sulfur compounds, e.g., cyclic sulfides, sulfoxides, sulfones, or sulfonamides. The oxidation of 26-thiodiosgenin (3) to sulfoxide or sulfone as well as their α-methylation decreased antibacterial activity of compounds 6, 8, 11 compared with 3. However, their antimicrobial activity was still stronger than that of diosgenin (1) (see Table 2). The weakest activity has been detected for carbaanalog 5 with the MIC values 7.8 µg/mL and 500 µg/mL for S. aureus and E. coli, respectively. This is in line with our previous findings for not methylated 5-and 6-membered F-ring carbaanalogs recently described [15]. The latter compound showed antimicrobial activity against S. aureus 8325-4 with the MIC = 4 µg/mL and against E. coli with the MIC = 512 µg/mL. The carbaanalog 5 demonstrated slightly lower antimicrobial potential against S. aureus 8325-4 (MIC = 7.8 µg/mL), probably due to the presence of an additional 27-methyl group. In the case of E. coli the activity was similar (MIC = 500 µg/mL for 5 and MIC = 512 µg/mL for its non-methylated analog). The determined MIC and MBC values allowed the conclusion that 26-thiodiosgenin (3) demonstrated the strongest antibacterial activity against S. aureus 8325-4 and E. coli ATCC 35218 among all tested compounds. Both its S-oxidation and α-methylation of the oxidized derivatives decreased their antimicrobial potential.

The Interaction of Studied Compounds with Proteins of Bacterial Cell Membranes
The results described above clearly showed that studied compounds exhibit antibacterial activity against S. aureus and E. coli. This feature may be due to the interaction between the molecule and specific component of the cell membrane, i.e., a protein or a phospholipid. In order to check if prepared thia-steroids possess the ability to interact with bacterial membranes proteins, we have analyzed fluorescence changes of tryptophan residues of membrane proteins as the marker of such interactions.
Based on fluorescence analysis, the Stern-Volmer plots were drawn ( Figure 5) using equation given below (Equation (1)) [24]. Referring to Figure 5 it can be concluded that all compounds have an affinity for Trp 214 residues in a hydrophobic pocket of Gram-positive (S. aureus) and Gram-negative (E. coli) bacterial membrane proteins, but the observed effect was visibly stronger for S. aureus. Based on the Equation (1), the Stern-Volmer constants have been calculated and are presented below (Table 3). Referring to Figure 5 it can be concluded that all compounds have an affinity for Trp 214 residues in a hydrophobic pocket of Gram-positive (S. aureus) and Gram-negative (E. coli) bacterial membrane proteins, but the observed effect was visibly stronger for S. aureus. Based on the Equation (1), the Stern-Volmer constants have been calculated and are presented below (Table 3). According to calculated K SV constants, it can be inferred that diosgenin (1) (3) and further α-methylation of the corresponding sulfone decreased the affinity of these compounds to tryptophan residues localized in the membrane proteins of S. aureus and E. coli to the values lower than that of diosgenin (1) and thiodiosgenin (3). Despite different affinity of studied compounds to Trp 214 residue of the bacterial membrane proteins, it can be concluded that their antibacterial activity is closely related to it. The lowest K SV in the interaction with S. aureus was obtained for carbaanalog 5 (K SV = (1.258 ± 0.189) × 10 4 M −1 ). It indicated the weakest affinity of this compound to the hydrophobic pockets of tryptophan in staphylococcal membrane proteins.
In order to check whether studied compounds form complexes with bacterial cells proteins, the quenching constants (k q ) have been calculated using equation below (Equation (2)) and are presented in Table 3.
where: k q -quenching constant, K SV -Stern-Volmer constant, τ 0 -fluorescence lifetime of fluorophore molecules. The calculated k q values for all tested compounds were greater than the one for the maximum scatter collision (2 × 10 10 M −1 s −1 ), thus it can be deduced that investigated molecules formed complexes with S. aureus and E. coli membrane proteins.

Cytotoxicity Study
The obtained sulfur analogs of diosgenin (3, 6-12) were briefly tested for cytotoxicity against three cancer lines (MCF7, K562, and HeLa) and normal human retina cells RPE-1. All tested compounds have not been shown to be toxic to normal RPE-1 cells, as well as towards MCF7 and K562 (Table 4). However, sulfoxides 7 and 10, as well as alkylated sulfones 11 and 12, exhibited a moderate toxicity against human cervical carcinoma cells (HeLa) with IC 50 34.3, 32.2, 92.3 and 44.1 µM, respectively. Interestingly, the carbaanalog 5 proved also slightly cytotoxic against HeLa cancer cell line (IC 50 38.2 µM). The reagents were purchased from Merck, Alfa Aesar, or Acros. All solvents were freshly distilled prior to use. The dry solvents were prepared by distillation over the following drying agents: DMF (4 Å molecular sieves), THF (Na/benzophenone), CH 2 Cl 2 (CaH 2 ).

Synthesis of 26-Thiodiosgenin 3β-4-Nitrobenzoate (3a)
26-Thiodiosgenin (3) (200 mg, 0.47 mmol) was dissolved in a mixture of dry pyridine (25 mL) and dichloromethane (10 mL), and then 4-nitrobenzoyl chloride (106 mg, 0.56 mmol) was added. The reaction mixture was stirred at room temperature. After completion of the reaction (7 days), it was poured into water, and extracted with dichloromethane (3 × 100 mL). The combined organic extracts were dried over anhydrous Na 2 SO 4 , and the solvent was evaporated in vacuo. The crude product was purified by dry flash chromatography with a hexane/ethyl acetate (99:1) mixture to afford ester 3a (245 mg, 91%).  The mixture of compounds 6 and 7 proved inseparable. These compounds were obtained in their pure forms by removing p-nitrobenzoyl groups from separated compounds 6a and 7a, what is described below.

General Procedure for Removing the 3β-4-Nitrobenzoyl Group
To a solution of steroidal sulfoxide (6a or 7a) (50 mg) (0.08 mmol) in dry methanol (10 mL) NaOH (80 mg, 2 mmol) was added. The reaction was stirred for 24 h. Then, the solvent was evaporated in vacuo, the crude product was dissolved in dichloromethane and washed by water. The organic extract was dried over anhydrous Na 2 SO 4 , and concentrated under reduced pressure. The crude product (6 or 7) was purified by column chromatography on silica gel with hexane/ethyl acetate (3:7) mixture elution. 30% hydrogen peroxide (45 µL, 0.4 mmol) and NHS (N-hydroxysuccinimide) (23 mg, 0.2 mmol) were added to a solution of 26-thiodiosgenin (3) (40 mg, 0.1 mmol) in acetone (5 mL). The mixture was heated at 50 • C for 10 h. Then, the reaction mixture was poured into aqueous solution of NaHSO 3 , and extracted with ethyl acetate (3 × 50 mL). The combined organic extracts were dried over anhydrous Na 2 SO 4 , and evaporated to dryness in vacuo. The crude product was purified by column chromatography on silica gel with hexane/ethyl acetate (3:1) elution to afford sulfone (8) (230 mg, 70%).