A Study of the Bisphosphonic Derivatives from the Pudovik Reaction of Dialkyl α-Oxophosphonates and >P(O)H Reagents: X-ray Structure and Bioactivity

New hydroxy-methylenebisphosphonic derivatives were prepared with different P-functions. The outcome of the reaction of α-oxophosphonates (YC(O)P(O)(OR)2) and dialkyl phosphites or diarylphosphine oxides depended on the Y substituent of the oxo-compound, the nature of the P-reagent and the amount of the diethylamine catalyst. Starting from dimethyl α-oxoethylphosphonate, in the presence of 5% of diethylamine, the corresponding Pudovik adduct was the single product. While using 40% of the catalyst, the rearranged species with the >P(O)–O–CH–P(O)< skeleton was the exclusive component. A similar reaction of α-oxobenzylphosphonate followed the rearrangement protocol. X-ray crystallography revealed not only the spatial structures of the three products, but also an intricate pattern evolving from the interplay of slight chemical differences, solvent inclusion and disorder as well as H-bridge patterns, which invite further investigation. In vitro activity of the compounds was assessed on different tumor cell cultures using end-point-type cell tetrazolium-based measurements. These structure–activity studies revealed a cytostatic effect for four rearranged derivatives containing aromatic units. One of them had a pronounced effect on MDA-MB 231 and Ebc-1 cells, showing IC50 = 37.8 and 25.9 µM, respectively.


Introduction
Tetraalkyl methylenebisphosphonates and related derivatives are important intermediates, e.g., they may be modified by substitution on the central carbon atom. A number of methods were elaborated for alkylation [1][2][3][4][5][6][7][8] and acylation [9]. On the other hand, substituted hydroxy-methylenebisphosphonic derivatives form a prominent group called dronates that are used for bone diseases such as osteoporosis and cancer [10][11][12]. Dronic acid derivatives may be synthesized by the reaction of substituted acetic acids with phosphorus trichloride or phosphorous acid in solvents like methanesulfonic acid or sulfolane [13][14][15]. The senior author of this paper with colleagues was the one who elaborated the optimized synthesis of alendronate [16,17], ibandronate [16,17], risedronate [16,18] and Table 1. The reaction of dimethyl α-oxoethylphosphonate (1) with dialkyl phosphites or secondary phosphine oxides under different conditions. It was a noteworthy observation that ionic liquid additives promoted the efficiency of the reactions [19][20][21][22]. The other possibility for the preparation of dronic acid derivatives involved the addition of dialkyl phosphites to α-oxophosphonates [23][24][25][26][27][28][29], which is called the Pudovik reaction. In this article, we aimed at the synthesis of methylenebisphosphonic derivatives with mixed P-functions involving a phosphine oxide moiety. We also explored the rearrangement side-reaction, which afforded products with a >P(O)-O-CH-P(O)< moiety.

Synthesis
In the first series of experiments, dimethyl α-oxoethylphosphonate (1) was reacted with dimethyl phosphite at 0 °C in diethyl ether for 8 h. The outcome depended on the quantity of the diethylamine (DEA) catalyst applied: using 5%, the Pudovik reaction took place selectively to afford α-hydroxy-methylenebisphosphonate 2a (Table 1/Entry 1); however, in the presence of 40% of the catalyst, tetramethyl phosphonate-phosphate 3a, formed by a rearrangement of the primary hydroxy-methylenebisphosphonate 2a, was the exclusive product (Table 1/Entry 2). The addition of diethyl phosphite and dibutyl phosphite to oxophosphonate 1 applying 5% DEA also selectively provided the adducts 2b and 2c, respectively (Table 1/Entries 3 and 6). At the same time, after stirring the mixture in the presence of 40% of the catalyst at 0 °C for 8 h, the reaction mixtures comprised comparable portions of the adduct (2) and the rearranged product 3 (Table 1/Entries 4 and 7). In these cases, stirring at 26 °C for 3 days was necessary to achieve complete rearrangement (Table 1/Entries 5 and 8). It is noteworthy that the rearrangement of the phosphonate-phosphate compounds led to both possible isomers 3b-1/3b-2 and 3c-1/3c-2.
In the next round, the secondary phosphine oxides diphenylphosphine oxide, bis(4methylphenyl)phosphine oxide and bis (3,5-dimethylphenyl)phosphine oxide were added to the carbonyl group of dimethyl α-oxoethylphosphonate (1). After adding 40% of the catalyst, the reactions were completed after stirring at 0 °C for 8 h. In these cases, no rearranged products were formed (Table 1/Entries 9-11).
In summary, adducts 2a-c, 2d-f, along with rearranged products 3a, 3b and 3c were obtained in 62-87% yields after column chromatography or recrystallization. All compounds were characterized by 31 P, 13 C and 1 H NMR, as well as HRMS. Tetramethyl bisphosphonate 2a and the rearranged version 3a were described earlier [24,30]. [a] On the basis of relative 31 P NMR intensities. [b] The data set shaded in yellow refer to the best experiments.
In the next round, the secondary phosphine oxides diphenylphosphine oxide, bis(4-methylphenyl)phosphine oxide and bis (3,5- added to the carbonyl group of dimethyl α-oxoethylphosphonate (1). After adding 40% of the catalyst, the reactions were completed after stirring at 0 • C for 8 h. In these cases, no rearranged products were formed (Table 1/Entries 9-11).
In summary, adducts 2a-c, 2d-f, along with rearranged products 3a, 3b and 3c were obtained in 62-87% yields after column chromatography or recrystallization. All compounds were characterized by 31 P, 13 C and 1 H NMR, as well as HRMS. Tetramethyl bisphosphonate 2a and the rearranged version 3a were described earlier [24,30].
Finally, diethyl α-oxobenzylphosphonate (4) was reacted with the three diarylphosphine oxides also used above. The results are summarized in Table 2. It was not possible to stop at the adduct stage as there was an increased inclination for the rearrangement. Carrying out the reaction in diethyl ether in the presence of 40% DEA at 0 °C for 8 h, only isomers 5-1 and 5-2 of the rearranged products 5(d-f)-1 and 5(d-f)-2 were formed. Compound 5d, comprising isomers 5d-1 and 5d-2 in a comparable 6:4 proportion, was obtained as a mixture of isomers. However, the major isomers 5-1 of products 5e and 5f were prepared in a pure form by column chromatography. Compounds 5e-f were fully characterized new species. Previously, it was also found that the adducts derived from αoxobenzylphosphonates are less stable than those obtained from α-oxoethylphosphonates [29]. [a] On the basis of relative 31 P NMR intensities. [b] The data set shaded in yellow refer to the best experiments.

X-ray Structure of the Three Adducts
Two Pudovik adducts, dimethyl phosphonate-phosphine oxide derivatives 2d and 2e together the earlier described diethyl phosphonate-phosphine oxide 6 [29], were subjected to single crystal X-ray analysis. The results are presented in Figures 1-6. The stereostructures of 2d, 2e and 6 are shown in Figures 1, 3 and 5, respectively, while selected geometries were included in Table 3. Connection of the molecules in the crystal structure can be seen in Figures 2, 4  [a] On the basis of relative 31 P NMR intensities. [b] The data set shaded in yellow refer to the best experiments.

X-ray Structure of the Three Adducts
Two Pudovik adducts, dimethyl phosphonate-phosphine oxide derivatives 2d and 2e together the earlier described diethyl phosphonate-phosphine oxide 6 [29], were subjected to single crystal X-ray analysis. The results are presented in Figures 1-6. The stereostructures of 2d, 2e and 6 are shown in Figures 1, 3 and 5, respectively, while selected geometries were included in Table 3. Connection of the molecules in the crystal structure can be seen in Figures 2, 4  Finally, diethyl α-oxobenzylphosphonate (4) was reacted rylphosphine oxides also used above. The results are summarize possible to stop at the adduct stage as there was an increased inclin ment. Carrying out the reaction in diethyl ether in the presence of h, only isomers 5-1 and 5-2 of the rearranged products 5(d-f)-1 an Compound 5d, comprising isomers 5d-1 and 5d-2 in a comparab obtained as a mixture of isomers. However, the major isomers 5were prepared in a pure form by column chromatography. Com characterized new species. Previously, it was also found that the a oxobenzylphosphonates are less stable than those obtained from α [29]. [a] On the basis of relative 31 P NMR intensities. [b] The data set shaded i experiments.

X-ray Structure of the Three Adducts
Two Pudovik adducts, dimethyl phosphonate-phosphine ox 2e together the earlier described diethyl phosphonate-phosphine jected to single crystal X-ray analysis. The results are presented in ostructures of 2d, 2e and 6 are shown in Figures 1, 3 and 5, resp geometries were included in Table 3. Connection of the molecules can be seen in Figures 2, 4 and 6. It is clear, that 2d is present as an 2e and 6 are H-bonded dimers.
Bond distances in the P1 atom vicinity show nearly identical dimensions (Table 3). At first glance, this statement does not seem to apply to the P2 atom environment, inclining one to suppose that an eventual charge imbalance may be visible through slightly alternating bond lengths of phosphonates 2d, 2e and 6. Such an assumption may eventually lead to the disparity of the H-bonding pattern difference between 2d (continuous chain) and 2e (discrete dimers). Nevertheless, this supposition may not really be supportable as the disorder in the methoxy groups of species 2e obviates interpretations of bond differences in the immediate region of the P2 atom. One can only speculate that the disorder is a reason or a consequence of the dimer formation in the case of compound 2e. It seems probable that a dimer-type of H-bridge buildup may be the consequence of the effect of the p-methyl substitution in the benzene ring, the solvent inclusion, and the disorder of the methoxy groups at P2. Hydroxy-methylenebisphosphonate 6 also forms H-bonded dimers in its crystal. As a very simple tool of assessing packing tightness in respect of 2d, 2e.0.5 C 3 H 6 O and 6, it is instructive to compare their calculated densities in their crystals (See Experimental). The chain-forming 2d had the highest value in this series, while 2e.0.5 C 3 H 6 O had the lowest, even with the aid of a solvent molecule. It is also worth noting that the real assembly in the crystal had a perfect twofold symmetry rotor image with acetone solvent sitting on the symmetry axis. Thus, the real 2:1 stoichiometry is 2 × 2e.C 3 H 6 O. The density of derivative 6 took on an intermediate value between the two others, thus suggesting that dimer formation tends to yield to looser packing.
Bond distances in the P1 atom vicinity show nearly identical dimensions (Table 3). At first glance, this statement does not seem to apply to the P2 atom environment, inclining one to suppose that an eventual charge imbalance may be visible through slightly alternating bond lengths of phosphonates 2d, 2e and 6. Such an assumption may eventually lead to the disparity of the H-bonding pattern difference between 2d (continuous chain) and 2e (discrete dimers). Nevertheless, this supposition may not really be supportable as the disorder in the methoxy groups of species 2e obviates interpretations of bond differences in the immediate region of the P2 atom. One can only speculate that the disorder is a reason or a consequence of the dimer formation in the case of compound 2e. It seems probable that a dimer-type of H-bridge buildup may be the consequence of the effect of the p-methyl substitution in the benzene ring, the solvent inclusion, and the disorder of the methoxy groups at P2. Hydroxy-methylenebisphosphonate 6 also forms H-bonded dimers in its crystal. As a very simple tool of assessing packing tightness in respect of 2d, 2e . 0.5 C3H6O and 6, it is instructive to compare their calculated densities in their crystals (See Experimental). The chain-forming 2d had the highest value in this series, while 2e . 0.5 C3H6O had the lowest, even with the aid of a solvent molecule. It is also worth noting that the real assembly in the crystal had a perfect twofold symmetry rotor image with acetone solvent sitting on the symmetry axis. Thus, the real 2:1 stoichiometry is 2 × 2e . C3H6O. The density of derivative 6 took on an intermediate value between the two others, thus suggesting that dimer formation tends to yield to looser packing.     The crystal structure contains one acetone molecule for every two molecules of 2e (in a sp sition). The solvent molecule was omitted for clarity. DIAMOND [31] representation; therm soids are drawn at 50% probability level. The solvent molecule was omitted for clarity. DIAMOND [31] representation; thermal ellipsoids are drawn at 50% probability level.
Molecules 2023, 28, x FOR PEER REVIEW 7 of 16 Figure 6. View of the hydrogen-bonded dimers around a symmetry center of 6 in the crystal. DIA-MOND [31] representation; thermal ellipsoids are drawn at 50% probability level.
Results indicated that several compounds elicited a cytostatic effect on the human tumor cell lines. Considering the percentage of cytostasis values at c = 50 µM, we concluded that compounds 3b, 5f-1 and 7 had a moderate cytostatic effect on the cells. However, three compounds were more effective. Adduct 2b was efficient on A431 human epidermoid carcinoma cells and showed a cytostasis = 49.9%. Phosphine oxide-phosphate 5d induced a higher cytostasis on MDA-MB 231 human breast adenocarcinoma and Ebc-1 human lung carcinoma cells, showing a cytostasis of 48.9 and 45.3%, respectively, whereas In vitro cytostatic activity of the compounds was determined after an overnight treatment using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. (For the details see Experimental.) Cytostasis (%) and 50% inhibitory concentration (IC 50 ) were determined for each compound on MDA-MB 231 human breast adenocarcinoma, A431 human epidermoid carcinoma, PC-3 human prostate adenocarcinoma, and Ebc-1 human lung squamous cell carcinoma cell lines.
Results indicated that several compounds elicited a cytostatic effect on the human tumor cell lines. Considering the percentage of cytostasis values at c = 50 µM, we concluded that compounds 3b, 5f-1 and 7 had a moderate cytostatic effect on the cells. However, three compounds were more effective. Adduct 2b was efficient on A431 human epidermoid carcinoma cells and showed a cytostasis = 49.9%. Phosphine oxide-phosphate 5d induced a higher cytostasis on MDA-MB 231 human breast adenocarcinoma and Ebc-1 human lung carcinoma cells, showing a cytostasis of 48.9 and 45.3%, respectively, whereas 5e-1 elicited an outstanding effect on these cell lines: cytostasis = 69.9 and 72.4%, respectively). Overall, the MDA-MB 231 human breast adenocarcinoma cell line proved to be the most sensitive for all effective compounds. Cytostasis values at c = 50 µM are shown in Table 4. The above tendency was also confirmed when the effect was determined on the basis of the calculated IC 50 values. While the effect of compounds 2a-f and 3a-c were not very efficient, a higher cytostatic effect was observed for compounds 5d, 5e-1, 5f-1, and 7.  Table 5. Daunomycin, an often-used reference compound, was also tested on the cell lines we applied. Its efficiency was much better [32,33] than that of our compounds. For details, see footnote "a" in Table 5.

General
The 31 P, 13 C, 1 H-NMR spectra were taken on a Bruker DRX-500 or Bruker Avance-300 spectrometer operating at 202.4, 125.7, and 500 MHz or 121.5, 75.4 and 300 MHz respectively. The couplings were given in Hz. LC-MS measurements were performed with an Agilent 1200 liquid chromatography system, coupled with a 6130 quadrupole mass spectrometer equipped with an ESI ion source (Agilent Technologies, Palo Alto, CA, USA). High-resolution mass spectrometric measurements were performed using a Thermo Velos Pro Orbitrap Elite hybrid mass spectrometer in positive electrospray mode.

Single Crystal X-ray Diffraction Studies
Single crystals of compound 2d, 2e.0.5 C 3 H 6 O and 6 suitable for X-ray diffraction were obtained by slow evaporation of the respective acetone solution. The crystals were introduced into perfluorinated oil and a suitable single crystal was carefully mounted on the top of a thin glass wire. Data collection was performed with an Oxford Xcalibur 3 diffractometer equipped with a Spellman generator (50 kV, 40 mA) and a Kappa CCD detector, operating with Mo-K α radiation (λ = 0.71071Ǻ).
Data collection and reduction were performed using CrysAlisPro software [35]. Absorption correction using the multiscan method [35] was applied. The structures were solved with SHELXS-97 [36], refined with SHELXL-97 [37] and finally checked using PLATON [38]. Details of the data collection and structure refinement are summarized in Table 6.

In Vitro Cytotstasis Assays Cell Lines and Culture Conditions
The in vitro cytostatic effect of the compounds was studied on MDA-MB 231 human breast adenocarcinoma [39], A431 human epidermoid carcinoma [40], PC-3 human prostate adenocarcinoma [41], and Ebc-1 human lung squamous cell carcinoma [42] cell lines. Cells were cultured in a DMEM medium supplemented with 10% FBS, 2 mM L-glutamine, penicillin-streptomycin antibiotic mixture (50 IU/mL and 50 µg/mL, respectively), 1 mM sodium pyruvate and 1% non-essential amino acid mixture. The cell cultures were maintained at 37 • C in a humidified atmosphere with 5% CO 2 . The cells were grown to confluent state and then they were harvested by trypsinization and divided into 96-well tissue culture plates (initial cell number was of 5.0 × 10 3 cells/well). Cells were allowed to attach for 24 h at 37 • C when the culturing medium was removed and they were treated with the compounds in 2, 10, 50, and 250 µM concentration in a serum-free medium. (The treating solutions contained 1.0 v/v% DMSO). Control cells were treated only with serum-free medium or with DMSO (c = 1.0 v/v%) under the same conditions. After overnight incubation, cells were washed twice with a serum-free medium, and then cultured for another 72 h in 10% serum-containing medium at 37 • C. An MTT-solution (at c = 0.37 mg/mL final concentration) was added to each well and incubated for 3 h. The cells were centrifuged for 5 min at 900 g, and then the supernatant was removed. The obtained formazan crystals were dissolved in DMSO (100 µL) and the optical density (OD) of the samples was measured with an ELISA Reader (iEMS Reader, Labsystems, Vantaa, Finland) at detecting wavelength = 540 and reference wavelength = 620 nm. OD620 values were subtracted from the OD540 values, and then cytostasis % was calculated from this corrected OD value by the following equation: Cytostatic effect (%) = [1 − (OD treated /OD control )] × 100 where OD treated and OD control correspond to the optical densities of the treated and control wells, respectively. In each case, two independent experiments were carried out with 4 parallel measurements. Statistical data analysis was performed using Student's t-test at a 95% confidence level. A 50% inhibitory concentration (IC 50 , expressed in micromolar units) was determined from the dose-response curves: cytostasis was plotted as a function of concentration on which a sigmoidal curve was fitted using Microcal™ Origin 2018 software [43,44].

Conclusions
The outcome of the reaction of α-oxophosphonates (ZC(O)P(O)(OR) 2 ) and Y 2 P(O)H reagents depended on the nature of the Z substituent of the oxo-compound, the Y substituent of the P-reagent, and the amount of the diethylamine catalyst. In case of Z = Me, new hydroxy-methylenebisphosphonic derivatives with different P-functions were synthesized. Performing the reactions in the presence of an increased amount (40%) of the catalyst, or starting from an α-oxobenzylphosphnate, rearranged species comprising the >P(O)-O-CH-P(O)< motif were the products. The molecular dimensions mostly conformed to those expected for this kind of P-compound. The intermolecular connection pattern may be realized in centrosymmetric H-bridge dimers, but in one case a catameric chain structure was experienced. Solvent inclusion as well as the presence of disorder were also present in one of the crystals hampering deeper insight into the solid-state relations. Part of the compounds we synthesized showed significant in vitro cytotoxic activity on human tumor cell cultures of different tissue origin. The rearranged derivatives with aromatic units possessed considerable antiproliferative activity characterized by low IC 50 values.

Data Availability Statement:
The data presented in this study are available on request from the corresponding authors.