Calix[4]arene-α-hydroxyphosphonic acids. Synthesis, stereochemistry, and inhibition of glutathione S -transferase

A series of dipropoxy-, tripropoxy-and tetrapropoxycalix[4]arenes bearing one or two fragments of α-hydroxymethylphosphonic acid at the upper rim of the macrocycle was prepared by the reaction of the corresponding mono-and di-formylcalixarenes with sodium salts of dialkyl phosphites or with trialkyl (tristrimethylsilyl)phosphites followed by dealkylation (desilylation) of the ester derivatives. The conformations of the macrocyclic skeleton and the stereoisomeric forms of the compounds obtained were investigated by 1 H NMR. The resulting α-hydroxymethylphosphonic acids were found to be able to inhibit the activity of glutathione S - transferase in vitro .


Introduction
][8][9][10] Also, the calix [4]arene derivatives were fixed on proteins. 11,12Calixarenes are considered as promising objects of biochemical research due to their ability to simulate the substrate-receptor interaction with biomolecules. 13,14reviously we have shown that calix [4]arenes functionalized with phosphonic, aminophosphonic and methylenebisphosphonic acid residues are effective inhibitors of alkaline phosphatases [15][16][17] and some of protein tyrosine phosphatases. 18,19Such compounds are more active than model acyclic inhibitors due to effects of the calixarene platform.6][17][18] The aim of this work was to design calixarene based α-hydroxymethylphosphonic acids and characterize their properties as inhibitors of glutathione S-transferase (GST).This enzyme is involved in a variety of biological processes including detoxication of xenobiotics in cells [20][21][22] and thereby can decrease effects of drugs causing multidrug resistance. 22,23Several heterocyclic compounds, 24,25 etacrinic acid derivatives 26 and some structural analogues of glutathione 27,28 have been described as reversible or irreversible inhibitors of GST.Some endogenous compounds such as bilirubin and bile salts are able to bind to GST and inhibit their activity. 23,29As we have already noted, 30 glutathione S-transferase can be a potential target for calixarene based phosphonic acids in vitro.It this paper we describe the synthesis, stereochemistry and inhibition ability of a series of lowerrim substituted calix [4]arenes bearing one or two α-hydroxymethylphosphonic acid fragments at the upper rim of the macrocycle.
Compounds 4-6 are colorless crystalline substances.Doublets of PCH proton (δ 4.5-4.7 ppm, JHP 13 Hz) and two doublets (or two multiplets) of AB spin system of axial and equatorial methylene protons ArCH2Ar groups are the most characteristic signals in the 1 H NMR spectra of the phosphonates 4-6.The difference in chemical shifts between the axial and equatorial protons in the dipropoxycalixarene 4 (∆δ 0.88 ppm) confirms the flattened cone conformation of the macrocyclic skeleton.In this conformation (symmetry С2v) two opposite phenol rings possess almost coplanar orientation and two O-propylated phenol rings possess almost perpendicular orientation to the main plane of the macrocycle formed by the methylene links.Intramolecular hydrogen bonds OH … OPr at the lower rim of the macrocycle stabilize this conformation. 34The value ∆δ 1.22 ppm shows the regular cone conformation (symmetry С2v) of tetrapropoxycalixarene 6. 35 Tripropoxycalixarene 5 with ∆δ 1.05 ppm occupies an intermediate position between these two conformations.
Two chiral carbon atoms in calixarene-bis-α-hydroxymethylphosphonates determine the formation of racemic form (RR+SS), meso form (RS) or their mixtures in the reaction products.2][33] For synthesis of an individual diastereoisomers we introduced diformylcalixarene 7 in the Pudovik reaction.The interaction of 7 with an excess of sodium salt of dialkyl phosphites in dioxane solution at 10-15 o C for 8 h proceeds diastereoselectively and leads to the meso form of calixarene-di-αhydroxymethylphosphonates 8-10 with d.e.almost 100% (Scheme 2).

Scheme 2. Synthesis of dipropoxycalixarene-bis-α-hydroxymethylphosphonates 8-13.
The reaction of diformylcalixarene 7 with trialkyl phosphites in the presence of anhydrous HCl is less diastereoselective (d.e.4-86%) than the Pudovik reaction.The ratio of racemic and meso forms of calixarene-bis-α-hydroxymethylphosphonates 11-13 is depended on the trialkyl phosphite alkyl length as well as the temperature of the reaction as shown in Table 1.The highest diastereomeric excess (86 %) was obtained by the reaction of calixarene 7 with tributyl phosphite at 10 o C. Decreasing the temperature significantly increases the reaction time but does not improve the diastereoselectivity. Pure racemic forms of calixarene-bis-α-hydroxymethylphosphonates 11-13 were obtained by crystallization.
The configuration of the stereogenic carbon atoms of СH(OH)P fragments in the racemic 11-13 and meso 8-10 forms of the calixarene-bis-α-hydroxymethylphosphonates were determined from signal patterns of aromatic protons in para-unsubstituted benzene rings (see 36 for a comparison).In the racemic form 11 possessing C2 symmetry axis the protons appear as three doublets of doublets.In the meso form 8 possessing Cs symmetry plane two doublets and two triplets of the protons are observed in the 1 H NMR spectra (Figure 1).The meso and racemic forms 8, 11 have also different chemical shifts of PCH protons as well as PH spin-spin coupling constants (δ 4.8 ppm, JHP 13 Hz and δ 4.9 ppm, JHP 10 Hz, respectively).Signals of phosphorus atoms of the racemic form (δ 19.0 ppm) are observed at higher field compared with those of the meso form (δ 22.1 ppm) in the 31 P NMR spectra.Diastereoselectivity of the diformylcalixarene phosphorylation depends significantly on the substitution pattern of the macrocycle lower rim.In contrast to dipropoxydiformylcalixarene 7, the reactions of tetrapropoxydiformylcalixarenes 14, 15 with the sodium salts of dialkyl phosphites or with trialkyl phosphites in the presence of anhydrous hydrogen chloride lead to equivalent (1:1) mixtures of the racemic and meso forms of calixarene-bis-α-hydroxymethylphosphonates 16-19 (Scheme 3).The diastereomeric ratio is not dependent on the size of the alkyl groups of the phosphorylating agents and conditions of the reaction.The two sets of signals for both forms are observed in their 1 H and 31 P NMR spectra.
It should be noted that the reaction of tetrapropoxydiformylcalixarenes 14, 15 with an excess of the sodium salts of dialkyl phosphites does not only form calixarene-bis-αhydroxymethylphosphonates 16-19.Phosphonate-phosphate rearrangement of one hydroxymethylphosphonate group occurs in the alkaline conditions at 20 o C for 48 h and leads to formation of phosphono-phosphates 20-22 (Scheme 4).Increasing the reaction time and temperature up to 60 o C does not cause rearrangement of the second phosphonate group and formation of the corresponding diphosphates.The desired calixarene-α-hydroxymethylphosphonic acids 23-29 were obtained in 90-95% yields by the reaction of the corresponding esters 4-6, 8, 11, 16, 18 with trimethylbromosilane and methanol (Scheme 5).
The reaction of carbonyl compounds with trimethylsilyl phosphites is a convenient method of synthesis of α-hydroxyphosphonic acids derivatives. 37,38Calixarene-α-hydroxyphosphonic acids 25-28 were obtained in near quantitative yields by the one pot reaction of formylcalixarenes 3, 7, 14, 15 with tristrimethylsilyl phosphite (THF, room temperature, 20-30 minutes) followed by cleavage of P-O-Si bonds of the silyl esters formed with methanol (Scheme 6).One equivalent of of tristrimethylsilyl phosphite is needed for phosphorylation of tetrapropoxymonoformylcalixarene 3, four equivalents are needed for dipropoxydiformylcalixarene 7 (phenolic OH groups destroy trimethylsilyl phosphite) and two equivalents for tetrapropoxydiformylcalixarenes 14, 15.
Acids 23-29 are colorless or slightly colored solids soluble in polar solvents such as dimethylsulfoxide, alcohols or aqueous alkaline solutions.They have no clearly defined melting points, which suggests that there is formation of dimeric and polymeric associates through intermolecular hydrogen bonds between hydroxymethyldihydroxyphosphonic fragments.This is confirmed by peaks of molecular ions of double mass in FAB mass spectra.

Inhibition of glutathione S-transferase
The research in vitro was undertaken in order to evaluate the properties of calix [4]arene-αhydroxymethylphosphonic acids as inhibitors of glutathione S-transferases from equine liver and human placenta.According to previous experiments, 30 the macrocyclic α-hydroxyphosphonates showed the highest affinity for these enzymes in comparison with corresponding αaminophosphonates or methylene bisphosphonates.The present study shows that some of calix [4]arene-α-hydroxymethylphosphonic acids have an affinity for the GST from equine liver with IC50 values in the micromolar range.Among dipropoxycalix [4]arenes tested towards GST from equine liver, the calix [4]arene 23 bearing one substituent at the upper rim was more effective in comparison with bis-substituted macrocycle 26.Introduction of four propyl groups to the lower rim of calix [4]arene leads to the analogue 25 being less effective in comparison with 23.At the same time, among dipropoxycalix [4]arenes, inhibiting properties of bisphosphonate derivatives 28 and 29 are higher than of monophosphonate 25.Compound 29, which bears two bromine atoms at the upper rim, has approximately the same activity as calix [4]arene 28.As is seen from Table 2, the IC50 values for tetrapropoxy substituted bisphosphorylated compound 28 are lower than these for bispropoxy substituted analogue 26.
It should be noted that impact of the calix [4]arenes on GST from equine liver is more pronounced than on placental GST.In the latter case the observed activities of bisphosphorylated compounds 26, 28 and 29 were higher than the activities of the monophosphorylated macrocycles 23 and 25 (Table 2).To elucidate the possible stereoselective effects of calix [4]arene-α-hydroxymethylphosphonates on the activity of glutathione S-transferase, the stereoisomeric forms of compounds 26 and 28 were docked computationally to the active site of GST P1-1 from human placenta.The docking calculations were performed using QXP/FLO+ program. 39A binding model was constructed automatically on the basis of known X-ray crystal structure of the enzyme 40 (PDB code: 20GS.pdb).The ligand was preliminary removed from the binding site of the enzyme.The phosphoryl residues of the inhibitors were in form of monoanions.Fulldock+, being the most exact in QXP/FLO+, was used as a docking method.According to the data obtained, stereoisomeric α-hydroxyphosphonates bind to the enzyme surface near to specific glutathione binding G-site.The binding free energies derived from docking (∆Edoc) indicate that all stereoisomeric forms of dipropoxycalix [4]arene 26 and and tetrapropoxycalix [4]arenes 28 can exhibit almost identical affinities for the enzyme (Table 3).The possible binding modes of the inhibitors mainly include hydrophobic interactions and hydrogen bonds in only some cases.For example, calixarene-bis-α-hydroxymethylphosphonic acids 26 which was identified and used in assay as meso form occupies the active site of GST P1-1 with narrow rim of the macrocyclic skeleton being oriented toward Arg13 and (R)-α-hydroxymethylphosphonate residue being positioned near Gln51.One of the aromatic rings of calix [4]arene platform is involved in interaction with Phe8.The hydrophobic interactions were found between macrocyclic platform of this compound and residues of Tyr108, Ile 104 belonging to H-site of the enzyme (Fig. 2).Although the glutathione S-transferases are classified into several groups, they all have highly specific glutathione binding G-site and xenobiotic binding H-site.Nonspecific H-site is capable of binding of very large substrates and shows structural plasticity among classes of these enzymes. 41Different affinities of calix [4]arene-α-hydroxymethylphosphonates to glutathione Stransferases from human placenta and equine liver (Table 2) may reflect some differences in the active sites of these enzymes and can be a result of more tight binding of the macrocyclic inhibitor to the enzyme from equine liver.It is known that inhibitors of glutathione S-transferases may interact with amino acid residues of both G-site and H-site.For example, kinetic studies of the enzyme from equine liver showed that S-(p-bromobenzyl)glutathione is a potent inhibitor with competitive kinetics with respect to glutathione (inhibition constant of approximately 1 μM) and mixed type of inhibition with respect to 1-chloro-2,4-dinitrobenzene (CDNB). 42In complex with GST P1-1, the glutathione part of the inhibitor occupies the G-site and p-bromobenzyl substituent is anchored to Phe8 and Tyr108 of the H-site. 43As seen in Figure 2, the residues of Phe8, Arg13 and Gln51, which are involved in glutathione binding at the G-site, 44 contact the bulky macrocyclic inhibitor 26, and this may influence the binding of glutathione to the enzyme.
The mode of binding of macrocycle 26 may also restrict the position of small molecule of CDNB which was located at the H-site.Taking into account the results of the docking study and variability of the family of calix[n]arenes (n = 4, 5, 6, 8 and more) possessing different size and conformation of their macrocyclic skeleton one can say that functionalization of the compounds with hydroxyphosphonic acid moieties is promising way of designing new potential inhibitors of glutathione S-transferase.

Conclusions
In summary, we have synthesized a novel series of calix [4]arenes bearing one or two fragments of α-hydroxymethylphosphonic acid at the upper rim.Dipropoxy-and tetrapropoxy-derivatives of calix [4]aren α-hydroxymethylphosphonic acids were designed to explore their inhibiting effects on activity of glutathione S-transferase from equine liver and human placenta in vitro.Some of compounds were found to have affinity for the enzyme from equine liver in the micromolar range.The number of α-hydroxymetylphosphonic acid and propoxy groups at the calixarene platform can influence the potency of compounds towards glutathione S-transferase.

Experimental Section
General. 1 H and 31 P NMR spectra were recorded on a Varian VXR spectrometer operating at 300 MHz and 121.5 MHz respectively.The chemical shifts are reported using internal tetramethylsilane and external 85% H3PO4 as references.The melting points were determined on a Boetius apparatus and are uncorrected.m-Nitrobenzyl alcohol was used as a matrix for registration of FAB-mass spectra.Analytical thin layer chromatography was carried out on Silufol plates.Column chromatography was carried out on silica gel Silufol L 40/100.Chloroform was distilled over phosphorus pentoxide.All reactions were carried out under dry argon.Formylcalix [4]arenes 1-3,7,14,15 were synthesized by the procedures described in the literature. 45

Synthesis of phosphonate phosphate substituted calixarenes 20-22. General procedure
To solution of diethyl phosphite (4 mmol) in dioxane (5 ml) sodium metal (0.4 mmol) was cautiously added in small portions.Formylcalixarene 14, 15 (0.2 mmol) was added to the resulting solution.The reaction mixture was stirred at 20 o C for 48 h and was quenched with water (100 ml).The product was extracted with chloroform.Organic phase was dried over Na2SO4.The solvent was evaporated and the product formed was purified by the column chromatography (eluent chloroformacetone, 10:1).

Synthesis of calixarene-α-hydroxymethylphosphonic acids 23-29 by dealkylation of corresponding esters. General procedure
To a solution of calixarene-α-hydroxymethylphosphonate 4-6, 8, 11, 16, 18 (0.1 mmol) in dry chloroform (5 ml) bromotrimethylsilane (ten-fold molar excess per each dialkoxyphosphonate group) was added.The reaction mixture was stirred at room temperature for 48 h and then was concentrated under reduced pressure.The residue was dissolved in absolute methanol (15 ml), the resulting mixture was stirred at 50 °C for 2 h, and then concentrated and dried in vacuo (0.05 mm Hg) at room temperature for 10 h.

One pot synthesis of calixarene-α-hydroxymethylphosphonic acids 25,26,28,29
To a solution of formylcalixarene 3, 7, 14, 15 (0.1 mmol) in dry THF (5 ml) tris-trimethylsilyl phosphite (0.13 mmol for 3, 0.4 mmol for 7 and 0.25 mmol for 14, 15) was added.The reaction mixture was stirred at room temperature for 8 h and then was concentrated under reduced pressure and dried in vacuo (0.05 mm Hg) at room temperature for 8 h.The residue was dissolved in absolute methanol (15 ml).The resulting mixture was stirred at 50 °C for 2 h, and then concentrated and dried in vacuo (0.05 mm Hg) at room temperature for 10 h.

Effect of inhibitors on the activity of glutathione S-transferase
Glutathione S-transferases from equine liver and human placenta were purchased from Sigma.According to the known method, 1-chloro-2,4-dinitrobenzene and L-glutathione were used as substrates of enzyme reaction. 25The activity of the enzymes were determined spectrophotometrically at 340 nm (where the product of reaction absorbs) using a molar extinction coefficient of 9600 M -1 cm -1 .The influence of the calix [4]arene-αhydroxymethylphosphonic acids on the rate of S-conjugate formation catalyzed by glutathione Stransferase was determined in 0.1 M Na-phosphate buffer at pH 6.5.The standard assay mixture contained 1 mM L-glutathione, 1mM CDNB and 0.1 mM EDTA and several concentrations of inhibitor.Compounds 23, 25, 26, 28 and 29 were preliminary dissolved in DMSO (the concentration of DMSO in reaction mixture was 2.5 vol.%).The mixture with substrates and inhibitor was incubated for 5 min at 25 °C, and the reaction was initiated by adding 5 μl or 10 μl of the enzyme (1 mg/ml).The IC50 values (inhibitor concentration that inhibits 50% of the enzyme activity) were determined in the presence of various amounts of calix [4]arene derivatives.The results given in Table 2 are presented as mean  S.D.

Table 1 .
Diastereoselectivity of the reaction of diformylcalixarene 7 with trialkyl phosphites