Reactions of Malononitrile Dimer with Isothiocyanates

The reaction of 2-amino-1,1,3-tricyanopropene (malononitrile dimer) with isothiocyanates leads to 1-substituted 4,6-diamino-2-thioxo-1,2-dihydropyridine-3,5-dicarbonitriles or 4,6-diamino-2-(phenylimino)-2H-thiopyran-3,5-dicarbonitrile, depending on the conditions. Quantum-chemical modeling of the IR spectra and reaction routes for the synthesized compounds was carried out. In silico predictive analysis of potential protein targets, compliance with bioavailability criteria, and ADMET parameters was performed.

DOI: 10.1134/S1070363221060013 2-Аminо-1,1,3-tricyanopropene 1, which is readily prepared by the dimerization of malononitrile [1], is a polyfunctional reagent widely used in synthetic organic chemistry. Heterocyclizations involving malononitrile dimer 1 are known since the middle of XX century and are fairly well studied (for reviews, see [2][3][4]). Nevertheless, there are contradictory data in the literature concerning the regioselectivity of the reactions of malononitrile dimer 1 with isothiocyanates. Thus, Eldin [5] described the reaction of dimer 1 with phenyl isothiocyanate in the presence of a base in DMF, leading to substituted pyridine 2. The structure of the latter is also confi rmed by its S-alkylation, for example, to thienopyridines 3 [6][7][8] (Scheme 1). At the same time, a similar reaction of dimer 1 with PhNCS in pyridine under refl ux leads, according to [9], to substituted pyrimidine 4. At the same time, it should be noted that the spectral data in [5,9] do not allow one to unambiguously assign structures 2 or 4 to the products of these reactions.
We reproduced the procedure in [5] to obtain, in a yield of 74%, a product that actually has (according to the IR and NMR spectra) the structure of 4,6-diamino-2thioxo-1-phenyl-1,2-dihydropyridine-3,5-dicarbonitrile 2 ( Table 1, entry 1; Scheme 2). Further studies showed that the reaction of PhNCS with dimer 1, in principle, does not require harsh conditions similar to those in [5]. For example, pyridine 2 was obtained in 84-92% yields by the reaction of dimer 1 with phenyl isothiocyanate at room temperature in EtOH in the presence of Et 3 N ( Table 1, entry 2). A similar reaction with allyl isothiocyanate gives 1-allyl-4,6-diamino-2-thioxo-1,2dihydropyridine-3,5-dicarbonitrile 5 in 58-73% yields ( Table 1, entry 6; Scheme 2). The use of stronger bases did not show obvious advantages. For example, the reaction of dimer 1 with allyl isothiocyanate in a superbasic medium (KOH in DMSO), too, leads to the isolation of compound 5 in a yield of 71% (Table 1, entry 7), but the resulting product requires additional purifi cation. Compound 2 is a yellow powder insoluble in water and EtOH, moderately soluble in acetone or in an AcOH-DMF mixture under heating. Compound 5 forms colorless crystals soluble in acetone and hot EtOH. The results of all experiments are summarized in Table 1.
Reproducing the method of synthesis of pyrimidine 4, described by Abdel-Latif et al. [9], gave ambiguous results. One way or another, we failed to obtain both compound 4 and, in general, a product having physicochemical or spectral characteristics reported in [9]. It was found that the direction of the reaction is strongly dependent on its conditions. For example, after the mixture of reagents in pyridine was fi rst allowed to stand at room temperature and then refl uxed, we obtained, in a low yield (8%), a beige substance melting above 300 °C and insoluble in boiling EtOH.
The reaction of dimer 1 with allyl isothiocyanate in boiling pyridine formed an orange product (yield 37-47%, based on the expected 1 : 1 adduct). The IR spectra of the product showed absorption bands corresponding to stretching vibrations of the N-H group (3323 and 3207 cm -1 ) and conjugated cyano group (2197 cm -1 ). However, the synthesized compound could not be characterized by 1 H and 13 C NMR spectroscopy because of its low solubility in organic solvents, including DMSO.   [5] [9] REACTIONS OF MALONONITRILE DIMER WITH ISOTHIOCYANATES Malononitrile dimer potassium salt 8 reacts with isothiocyanates differently from malononitrile dimer 1. For example, the reaction of salt 8 with allyl isothiocyanate give the same 1-allyl-4,6-diamino-2thioxo-1,2-dihydropyridine-3,5-dicarbonitrile 5 but in an appreciably lower yield (27%) (Scheme 2; Table 1, entry 8). At the same time, the reaction of salt 8 with PhN=C=S under analogous conditions leads to thiopyran 6 in a yield of 82% (Table 1, entry 9). It should be noted that we failed to recyclize thiopyran 6 into 2-thioxopyridine 2 even upon prolonged boiling in alcoholic KOH. Thiopyran 6 has never been described in the literature, but the formation of structurally similar 2H-thiopyrans was previously observed in the reactions of cyclohexylidenemalononitrile with PhNCS [22], ammonolysis of thiopyran salts [23], condensation of malononitrile with enamines β-ketothioanilides [24], and reactions of malononitrile or its dimer with carbon disulfi de in the presence of alkalis [25] (Scheme 3).
The structures of 2-thioxopyridines 2 and 5 and 2-iminothiopyran 6 were studied in detail by IR and NMR spectroscopy ( 1 H, 13 C DEPTQ, 1 H-13 C HSQC, 1 H-13 C HMBC). The spectra are presented in Supplementary Information. The 1 Н NMR spectra of compounds 2, 5, and 6 show signals of the substituent R and broadened signals of two amino groups. The 13 С (DEPTQ) NMR spectra of all the compounds show signals of two C≡N carbons and 5 characteristic signals of the heteroring carbons. Evidence for the structures is provided by the results of heteronuclear correlation experiments. Thus, the 1 H-13 C HMBC spectra of compounds 2, 5, and 6 contain cross peaks between the protons of one of the amino groups and two signals of the heteroring carbons in the high-fi eld region (C 3 and C 5 ), which unambiguously confi rms the presence of the N≡C-C 3 =C 4 (NH 2 )-C 5 -C≡N fragment (see Supplementary Information). At the same time, the spectra of isomeric compounds 2 and 6 differ from each other in the low-fi eld region: in the spectrum of compound 2, the pyridine C 4 and C 6 signals are observed at 155.6 and 155.9 ppm and the signal assignable to the thiocarbonyl C 2 =S carbon, at 181.0 ppm, whereas the respective signals in the spectrum of thiopyran 6 are present at 156.7 (C 6 ), 159.3 (C 4 ), and 160.4 (C 2 ) ppm. The IR spectra of compounds 2, 5, and 6 contain stretching vibration bands of the N-H bonds and conjugated С≡N groups (2204-2212 см -1 ). At the same time, the spectra of thiones 2 and 5 show vibration bands of the С=S bond (strong bands at 1344-1348 cm -1 and medium intensity bands at 1180-1190 cm -1 ).
Quantum-chemical study. Further evidence for the structure of compounds 2, 5, and 6 was obtained by the quantum-chemical calculations of spectral modes. For example, the molecular geometry ( Fig. 1) and IR spectra, calculated using the ORCA 4.2 software package [26,27], at the B3LYP-D3BJ level of theory [28][29][30][31] with the 6-311+G(2d,p) split-valence basis set, showed a good fi t to the experimental spectra (Tables 2 and  3). The calculated and experimental frequencies were compared using scale factors [0.9679 for high-frequency (>1000 cm −1 ) and 1.0100 for low-frequency modes (<1000 cm −1 )] [32]. The input files were generated using the Gabedit 2.5 program [33]. The ChemCraft 1.8 program was used to visualize the molecular geometries and calculated IR spectra.
Quantum-chemical calculations were also applied to interpret the experimental results, in particular, the different regioselectivities of the heterocyclizations of phenyl isothiocyanate with malononitrile dimer. The mechanisms of intramolecular cyclization were calculated  [26,27]. The search for transition states and reaction paths, as well as calculation of the vibration modes and Gibbs free energies were performed using the DFT method with the Grimme B97-3c scheme [34,35], which combines the B97 GGA functional and the def2-mTZVP basis set with the D3BJ dispersion correction [36].  Since the reaction was carried out with an excess of a base (triethylamine) in ethanol or pyridine (as a solvent), as the initial structure we considered anion 9 associated with two base molecules, one of which was protonated and thus formed an ion pair with anion 9 and the second was hydrogen-bonded to the H 22 proton (Fig. 2a). To determine the optimal geometry of the initial state, a search was for the most stable conformation of anion 9 was fi rst performed.
The cyclization leading to 2-thioxopyridine 2 occurs as a result of the nucleophilic attack of N 13 on C 6 , accompanied by H 45 transfer from the base molecule to N 8 and deprotonation of N 13 by the second base (triethylamine or pyridine) molecule (Figs. 2b and 2d). The reaction leading to thiopyran 6 involves the nucleophilic attack of S 12 on C 6 , accompanied by proton transfer from the base molecule to N 8 (Figs. 2c and 2e). The subsequent deprotonation of N 13 by the second base molecule has, according to the calculations, very low activation energy, and, therefore, it is not a rate-limiting stage.
The molecular structures of starting anion 9, transition states, and cyclization products are presented in Fig. 2 (reaction in pyridine). The calculations showed that the cyclization leading to thiopyran 6 is a kinetically controlled reaction. The activation energy is 70.2 kJ/mol. At the same time, the heat effect of this reaction is low (-9.1 kJ/mol). The activation energy of the cyclization, leading to 2-thioxopyridine 2, is 91.5 kJ/mol, but its heat effect is higher (-36.6 kJ/mol). Thus, under the thermodynamic control, 2-thioxopyridine 2 should be the main reaction product.
The simulation of analogous process in the ethanol medium (the resulting molecular structures are presented in Fig. 3) showed that the activation energy of the cyclization reaction in the presence of triethylamine as a base increases to 138.2 kJ/mol, and the heat effect of this reaction also increases (-48.2 kJ/mol). Presumably, the reason why this process has a higher activation energy compared to the same reaction in pyridine is the mutual repulsion of the two bulky triethylаmine molecules, which are getting closer together during reaction, and the resulting distortion of the molecular structure of the reacting substrate. Thus, the benzene ring on the N 13 atom takes a position almost perpendicular to the forming pyridine ring. The cyclization involving the sulfur atom, while having a slightly lower activation energy (65.2 kJ/mol), has a negative heat effect (4.5 kJ/mol), and, therefore, it is unlikely to occur. The energy profi les of the cyclization reactions in study are shown in Fig. 4.
Thus, having considered the simulated cyclization mechanisms allow us to conclude that 2-iminоthiopyrans are formed under kinetic control, whereas 2-thioxopyridines are formed under thermodynamic control. The observed formation of 2-iminоthiopyran in pyridine (25°С to refl uxing) can be explained by the fact that initially, still at 25°С (before heating is applied), the reaction occurs under kinetically controlled conditions to form 2-iminоthiopyran 6 which further is unable to undergo recyclization into 2-thioxopyridine 2 (the diffi culty of the latter process was confi rmed experimentally). The     different regioselectivities of the reactions of PhNCS with malononitrile dimer (Et 3 N, EtOH, 25°C) and with its potassium salt (EtOH, 25°C) can be explained by a specifi c role of the base catalyst. Successful formation of 2-thioxopyridine 2 requires involvement of two base molecules: to form an unstable multi-charged transition state 10 (Scheme 5) requires simultaneous deprotonation of the endocyclic nitrogen atom with a free base and protonation of the negatively charged exocyclic nitrogen atom with the pyridinium (triethylammonium) cation. However, in the case of malononitrile dimer salt 8, such base and its protonated form are absent from the reaction medium. Water and ethanol, that are present in the reaction medium, are weaker proton donor/acceptors than pyridine or triethylamine and their conjugate acids; therefore, in this case, the reaction proceeds exclusively by the way of formation of kinetic product 6.
In our opinion, the different results obtained with allyl isothiocyanate can be associated with more facile recyclization of the 2-iminothiopyran product. However, this issue calls for a more detailed study, which will be the subject of our further research.
As we can see from the data in Table 4, the сLogP values for compounds 2, 5, and 6 vary from -0.95 to -0.13, which suggests good absorption and permeation [52][53][54]. At the same time, the log S value of lower than -4.0 for compound 2 points to a lower solubility (<1×10 -4 M). The TPSA parameters for all the compounds have borderline values, even though they all formally fi t the oral bioavailability criteria. No toxicity risks predicted for compounds 2, 5, and 6. The drug likeness indices are low, but the drug scores, due to the formal compliance with the bioavailability criteria and predicted low toxicity, are fairly high (0.37-0.48).   To predict biological activity, we also used the open source software PASS Online [56,57] and AntiBac-Pred [58]. The resulting data predict an inhibitory effect on kinases for compound 2 with a probability of 0.765, an anti-ischemic effect for compound 5 with a probability of 0.800, as well as activity against the β-amyloid precursor protein for thiopyran 6 with a probability of 0.800. The highest antibacterial effect with respect to Campylobacter jejuni bacteria is predicted for compound 6 (confi dence 0.3158; confi dence > 0, if the probability of activity is greater than the probability of inactivity P a > P i ).
The SwissADME [59,60] and GUSAR [61,62] software packages were used to predict the ADMET parameters (absorption, distribution, metabolism, excretion, toxicity). According to the US EPA criteria, in terms of acute oral toxicity, compounds 2, 5, and 6 can be assigned to class III (for low toxicity compounds, 500 mg/kg < LD 50 < 5000 mg/kg). Gastroenteric absorption, lack of ability to penetrate the blood-brain barrier (BBB), as well as lack of inhibition of most P450 cytochromes are predicted for all the compounds ( Table 5).
The potential protein targets for the synthesized compounds were predicted using the new GalaxySagittarius protein-ligand docking protocol [63] based on the GalaxyWeb web server [64,65]. The geometry and minimum energy optimized 3D structures of compounds 2, 5, and 6 were generated using the ORCA 4.2 software package [26,27]. Docking using the GalaxySagittarius protocol was performed in the Binding compatability prediction and Re-ranking using docking modes. Table  6 shows the docking results compounds 2, 5, and 6 for 10 target-ligand complexes with the minimal free binding energies ΔG bind and the best estimates for the protein-ligand interaction. The predicted protein targets are specifi ed using the ID identifi ers in the Protein Data Bank (PDB) and the UniProt database.
As we can see from Table 6, compounds 2, 5, and 6 show affi nity to a wide range of proteins. In particular, the compounds are predicted to be the most affi ne urokinase plasminogen activator (uPA, PDB ID 1c5z, UniProt ID P00749) (Fig. 5), BRPF1 protein (PDB ID 5eww, UniProt ID P55201), and bromodomain-containing protein 9 BRD9 (PDB ID 5ji8, UniProt ID Q9H8M2), which regulates the apoptosis signal to kinase 1 ASK1/MAP3K (PDB ID 5up3, UniProt ID Q99683) and a number of other targets. Considering the important role of uPA [66] and BRD9 [67] in oncogenic processes and of ASK1/MAP3K protein kinases in the immune response [68], the synthesized compounds can be considered as inhibitors potentially useful for treating various forms of cancer, neurodegenerative processes, rheumatoid arthritis, etc. In general, the results of the study allow us to conclude the following. The reaction of malononitrile dimer with isothiocyanates can take two different paths to form 4,6-diamino-2-thioxo-1,2-dihydropyridine-3,5dicarbonitrile or 4,6-diamino-2-(phenylamino)-2Hthiopyran-3,5-dicarbonitrile. It should be noted that the formation of pyrimidine derivatives, reported in [9], has not been confi rmed. Quantum-chemical calculations established that 2-iminothiopyrans and 2-thioxopyridines are formed under kinetically and thermodynamically controlled conditions, respectively. In the case of phenyl isothiocyanate, the reaction is regiocontrolled, and can be directed, by changing conditions, to the exclusive formation of a derivative of 2-(phenylimino)thiopyran or 2-thioxo-1-phenylpyridine. The predictive analysis of biological activity provide evidence showing that the synthesized compounds hold promise as potential pharmaceuticals.

EXPERIMENTAL
The IR spectra were obtained on a Bruker Vertex 70 spectrophotometer equipped with an ATR unit with a diamond crystal, accuracy ±4 cm -1 . The NMR spectra were recorded on a Bruker Avance III HD 400 MHz instrument (400.17 MHz for 1 H and 100.63 MHz for 13 C) in DMSO-d 6 ; the residual solvent signals were used reference. The purity of the synthesized compounds was controlled by TLC on Sorbfi l-A plates (Imid, Krasnodar), eluent acetone-hexane (1: 1), visualization by exposure to iodine vapor and UV light. Malononitrile dimer 1 was prepared according to Mittelbach [1]. Malononitrile dimer potassium salt 8 was prepared by the same procedure (yield 75-90%) but without fi nal acidifi cation, and the product was reacted without additional purifi cation. Pyridine was dried over KОН.
1-Allyl-4,6-diamino-2-thioxo-1,2-dihydropyridine-3,5-dicarbonitrile (5  b. Synthesis from malononitrile dimer potassium salt 8. Allyl isothiocyanate (0.5 mL, 5.09 mmol) and then 2 mL of water were added to a stirred suspension of 0.87 g (5.09 mmol) of salt 8 in 10 mL EtOH. A slight exotherm developed, and the mixture homogenized to form a tea-colored solution. A day after, a crystalline precipitate began to slowly form. In 3 days, the pale pink crystals were fi ltered off and washed with EtOH and petroleum ether. The product was spectrally identical to compound 5 prepared as described above. Yield 329 mg (28%). Attempted isolation of an additional amount of the product by addition water or acid to the mother liquor failed. 4,6-Diamino-2-(phenyliminо)-2H-thiopyran-3,5dicarbonitrile (6). а. Synthesis from malononitrile dimer in pyridine. A mixture of PhNCS (1.2 mL, 0.01 mol) and malononitrile dimer 1 (1.32 g, 0.01 mol) in 30 mL of anhydrous pyridine was allowed to stand for 24 h at 25°С and then refl uxed for 2 h and cooled down. The resulting dark solution was poured onto ice, acidifi ed to pH 2 with HCl, and allowed to stand in a refrigerator for 24 h at 4°С. The precipitate was fi ltered off, washed with EtOH, and refl uxed for 3-5 min in EtOH for purifi cation. The beige precipitate was fi ltered off to obtain 0.20 g (8%) of compound 6, mp > 300°С{mp 225°C (EtОН) [9]}. Therewith, salt 8 quickly dissolved, and a yellow solution was formed. After ~2 h, a precipitate started to form. The mixture was stirred without heating for another 24 h. White powder of 2-iminothiopyran 6 was fi ltered off, washed with EtOH. Yield 11.57 g (82%). The product is spectrally identical to compound 6 obtained as described above.

OPEN ACCESS
This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

SUPPLEMENTARY INFORMATION
The online version contains supplementary material available at https://doi.org/10.1134/S1070363221060013