4,5-Dihydro-5-Oxo-Pyrazolo[1,5-a]Thieno[2,3-c]Pyrimidine: A Novel Scaffold Containing Thiophene Ring. Chemical Reactivity and In Silico Studies to Predict the Profile to GABAA Receptor Subtype

The isosteric replacement of the benzene with thiophene ring is a chemical modification widely applied in medicinal chemistry. Several drugs containing the thiophene ring are marketed for treating various pathologies (osteoporosis, peripheral artery disorder, psychosis, anxiety and convulsion). Taking into account this evidence and as a continuation of our study in the GABAA receptor modulators field, we designed and synthesized new compounds containing the thiophene ring with 4,5-dihydro-5-oxo-pyrazolo[1,5-a]thieno[2,3-c]pyrimidine and pyrazolo[1,5-a]thieno[2,3-c] pyrimidine scaffold. Moreover, these cores, never reported in the literature, are isosteres of pyrazolo[1,5-a]quinazolines (PQ), previously published by us as GABAAR subtype ligands. We introduced in the new scaffold those functions and groups (esters, ketones, alpha/beta-thiophene) that in our PQ derivatives were responsible for the activity, and at the same time, we have extensively investigated the reactivity of the new nucleus regarding the alkylation, reduction, halogenation and hydrolyses. On the six final designed compounds (12c–f, 22a,b) molecular docking and dynamic simulation studies have been performed. The analysis of dynamic simulation, applying our reported model ‘Proximity Frequencies’, collocates with high probability 12c, 22b, in the agonist class towards α1β2γ2-GABAAR.


Introduction
Thiophene is known to be an isostere of the benzene ring, thus representing a good tool in the field of medicinal chemistry to develop new drugs. The concept of isosterism applied to the thiophene ring was established in 1932 by Erlenmeyer, which proposed the equivalence between -CH = CH-and -S-(in benzene and thiophene, respectively) in terms of size, mass and capacity to provide an aromatic lone pair. Over time, compounds containing thiophene have been extensively investigated in medicinal chemistry, since they show a variety of pharmacological activities, such as anti-inflammatory, antioxidant, antimicrobial, antitumor, and antidepressant action [1][2][3][4]. Moreover, several thiopheneheterocycles-fused compounds have been approved by FDA as therapeutic agents for the treatment of osteoporosis (Raloxifene), peripheral artery disorder (Ticlopidine), psychosis (Olanzapine), anxiety and convulsion (Etizolam); see Figure 1. Starting from these evidences, we addressed our research towards the synthesis of heterocyclic compounds containing the thiophene ring. We applied this strategy to the synthesis of potential GABA A R subtypes ligands, which represent a field of research extensively investigated by us for many years, by obtaining some interesting derivatives with pyrazolobenzotriazine, pyrazolopyrimidine and pirazoloquinazoline scaffold [5][6][7][8]. This choice is also supported by the fact that in the literature are present some GABA A R ligands (Ro 19-4603, TB21007, Comp. 4, Comp. 16), in which the thiophene ring is fused or bonded to other cycles ( Figure 2) [9][10][11]. In particular, we report here the synthesis of new compounds with 4,5-dihydro-5-oxo-pyrazolo [1,5-a]thieno [2,3-c]pyrimidine and pyrazolo [1,5-a]thieno [2,3-c]pyrimidine scaffold as result of the isosteric replacement of the benzene with thiophene ring in our pyrazolo [1,5-a]quinazolines previously published as GABA A R subtype ligands [8,12] (Figure 3). Moreover, these new compounds can be considered as analogues of the abovementioned GABA A R ligand , in which we formally operated a contraction of the central diazepine ring. As a first approach, in this new scaffold, we tried to introduce at position 3 those functions and groups (esters, ketons, alpha/beta-thiophene) responsible for activity in our pyrazolo [1,5-a]quinazolines and in compounds of the literature (i.e.,  and Comp 16) (Figure 2).
The pyrazolo [1,5-a]thieno [2,3-c]pyrimidine core is not found in the literature (SciFinder, Reaxys), with the exception of the ethyl 4,5-dihydro-5-oxo-pyrazolo [1,5-a]thieno [2,3c]pyrimidine-3-carboxylate (RN 942034-93-7), of which neither the synthesis nor the characterization is reported, however. In addition, this single compound is not mentioned in any published work. Thus, it was very intriguing to investigate the feasibility and the reactivity of this nucleus toward the most common reactions, such as alkylation, reduction, halogenation, and hydrolysis. Finally, molecular docking studies and evaluation of the 'Proximity Frequencies' (exploiting our reported model) [8,13] were performed on all the final compounds to predict their profile on the α1β2γ2-GABAAR subtype.

Chemistry
The synthetic pathways for obtaining derivatives with pyrazolo[1,5-a]thieno [2,3e]pyrimidine scaffold are depicted in Schemes 1-6. In this synthetic section, we report not only the procedures for obtaining the final designed products mentioned in Figure 3, but also some reactivity studies on this new scaffold which, furthermore, have produced interesting results. NMR spectra, elemental analysis and other structural information are reported in Supplementary Materials, Table S1.
A particular reactivity was evidenced for not isolated intermediate 9e, the 3-(thien-3-yl)-5-chloropyrazolo[1,5-a]thieno [2,3-e]pyrimidine. In fact, when the reduction is performed with NaBH 4 , starting and final products mixture was anyway recovered, also largely changing the reaction conditions. On the other hand, performing a catalytic transfer hydrogenation (CTH) with HCOONH 4 and Pd/C in EtOH [17], it was possible to highlight in TLC a spot that could be the 4,5-dihydro derivate, which rapidly and spontaneously converts into the 3-(thien-3-yl)pyrazolo[1,5-a]thieno[2,3-e]pyrimidine 11e. Moving to the alkylation reactions, which afforded the final desired compounds 12c-f, the lactam derivatives (3)(4)(5)(6)(7)(8) were all treated following the classical method (DMF/K 2 CO 3 /CH 3 I), but depending on the substituent at the 3-position, different reactivity was evidenced, specifically: The regioselective O-alkylation observed for compound 5 could be due to a prevalence of the tautomeric form −N = C-OH with respect −NH-C = O; the predominance of the first one could indeed be associated with the amide function at position 3, since the CONH 2 group could form H-bonds with N-4, thus favoring the tricyclic heteroaromatic structure (see Figure 4). On the other hand, the rising temperature (reflux) can promote a free rotation of the C3-CO bond of the amide group, no longer involved in an H-bond, and thus, allowing alkylation at N-4 (compound 12b). Scheme 3 describes the different reactivity of the 3-ethyl carboxylate derivatives 4 and 12c towards the alkaline or acid hydrolysis. The starting ethyl 5-oxo-4,5-dihydropyrazolo [1,5-a]thieno[2,3-e]pyrimidine-3-carboxylate 4 behaves in the usual manner to alkaline or acid hydrolysis, giving the corresponding 3-carboxylic acid 14, which in turn undergoes decarboxylation in HCl 12M, at reflux temperature, yielding compound 15. Instead, starting from the ester 12c, also using different reaction conditions (NaOH 10% or LiOH in THF/water or AcOH/HCl or conc. HCl), it has never been possible to obtain the 3-carboxylic acid, but only the 3-decarboxylate derivative 16 is recovered. Thus, to get the 4-methyl-5-oxo-4,5-dihydropyrazolo[1,5-a]thieno[2,3-e] pyrimidine--3carboxylic acid, as a key intermediate for obtaining the final designed esters, we followed a reported procedure [18], which involves diazotization of the 3-carboxamide 12b (Scheme 4); the reaction did not afford the desired 3-carboxylic acid, but a mixture of two compounds, one identified as the 3-decarboxylate 16 and a possible mechanism of decarboxylation process is reported. The second compound, colored green, was assumed to be a 3-nitroso derivative 17; this hypothesis could be supported by the fact that, after decarboxylation, a high concentration of nitrosonium ion (NO + ) in the chemical environment could be able to make an electrophilic attach at position 3 of the pyrazolothienoquinazoline scaffold. The 1 H-NMR spectrum of the supposed compound 17, in addition to the absence of the proton at position 3, shows a shift of the methyl bound to N-4 (4.23 ppm) that is not consistent with the chemical shift values of the 4-methyl-5-oxo-4,5-dihydropyrazolo[1,5-a]thieno[2,3e]pyrimidine derivatives, whose N-methyl group constantly falls in the 3.3-3.7 ppm range. Moreover, the chemical shift of the proton in position 2 is lower than in other products with different substituents (CN, COOEt, CONH 2 , COOH), suggesting a different electronic/steric environment resulting from the substituent in position 3. The mass analysis confirmed the structure of compound 17. The same reaction performed on the 5-methoxypyrazolo [1,5-a]thieno[2,3-e]pyrimidine-3-carboxyamide 13b also gave a mixture of two products, but in this case, the green 3-nitroso derivative 19 was obtained together with the desired 5-methoxypyrazolo[1,5-a]thieno[2,3-e]pyrimidine-3-carboxylic acid (18). The mass analysis again confirmed the structure. A possible mechanism of decarboxylation is reported in Figure 5.  In order to obtain the desired 4-methyl-5-oxo-4,5-dihydropyrazolo [1,5-a]thieno[2,3e]pyrimidine-3-carboxylic acid 21, we explored a further synthetic strategy reported in Scheme 5. The 3-unsubstituted compound 16 was treated with HMTA obtaining the 4methyl-5-oxo-4,5-dihydropyrazolo[1,5-a]thieno[2,3-e]pyrimidin-3-carboxaldehyde derivative 20 and its further oxidation (KMnO 4 /water/acetone/sodium hydroxide) finally gave the desired 3-carboxylic acid 21. From acid 21, the final desired esters 22a,b were obtained by treatment with thionyl chloride and further addition of the suitable alcohol (t-BuOH and 2-thiophenemethanol respectively) in CH 2 Cl 2 .
The hydrolysis of the ester function to carboxylic acid also created problems on the ethyl pyrazolo[1,5-a]thieno[2,3-e]pyrimidine-3-carboxylate 11c (Scheme 6). In fact, the desired product 23 was recovered in a meagre yield together with a big amount of the 3-decarboxylate 24 only if the hydrolysis of 11c was performed in an alkaline medium; the reaction then evolved spontaneously towards the total formation of compound 24. On the contrary, in acid medium, compound 11c underwent a decarboxylation, directly giving the 3-unsubstituted derivative 24.
Therefore, we explored a different synthetic way, starting from the 5-oxo-4,5-dihydropyrazolo[1,5-a]thieno[2,3-e]pyrimidin-3-carboxylic acid 14, by using LiAlH 4 as reducing agent for the lactam function. Thus, after quenching the reaction and performing the standard workup, the residue was refluxed in toluene and Pd/C. The presence in 1 H-NMR spectrum of H5 at 9.06 ppm confirmed that the pyrazolothienopyrimidine core was indeed dehydrogenated but, at the same time, the carboxylic function was transformed in a methyl group, as evidenced by the peak at 2.33 ppm, compound 25, again preventing the desired 23.

Molecular Dynamic Studies
On the six final designed compounds (12c-f and 22a,b), a molecular docking study and an evaluation of the 'Proximity Frequencies' [8,13] were performed to predict their profile on the α1β2γ2-GABA A R subtype.
The value of Proximity Frequencies (PFs), used in a linear discriminant function (LDA), was able to correctly collocate 70.6% of agonists and 72.7% of antagonists by combining a double PF (αVal203-γThr142) with a triple PF (αHis102-αTyr160-γTyr58). The predictive capacity was evaluated on an appropriate training set of molecules with a cross-validation 'leave one out' (LOO) procedure. During a molecular dynamic simulation (60 ns), the agonist compounds were simultaneously close to the αVal203 and γThr142 amino acids, with a frequency of 37% compared to the frequency of 16% found by the antagonist compounds, while the antagonist compounds were simultaneously close to the αHis102, αTyr160, and γTyr58 amino acids, with a frequency of 35% against a frequency of 13% for agonist compounds. All the 3D structures of the molecules, as a training set and new final compounds, were designed [19] (DS ViewerPro 6.0 Accelrys Software Inc., San Diego, CA, USA) and placed in the binding site of the BDZs with the AUTODOCK 4.2 [20] docking program. The structure of the BDZ binding site was obtained from the recently solved GABA A R structure (PDB ID 6D6T) [21].
The docking program performed on the selected compounds (12c-f and 22a,b) gave a number of clusters of conformation(s) for each compound (rmsd 2.0). The evaluation of trajectories in the dynamic simulation was performed on the conformations that covered at least 90% of poses; the dynamic simulations were performed on an isolated portion of the protein between the α and γ chains comprising all amino acids within a radius of 2 nm from the center of the benzodiazepine binding site. Applying the PF model to the new selected compounds, it emerges that all six ones are collocated in the agonist class, 12d-f with low probability, while 12c and 22b, the 3-ethyl and the 3-(2-thienylmethyl)carboxylate, respectively, with a percentage of prediction of 74% and 78%, are more probable. The 3-tbuthylcarboxylate 22a shows a percentage of prediction slightly lower (69%), but always in the agonist class; see Table 1. Compounds 12c and 22b have in the position 3 an ester group which is able to engage a strong hydrogen bond interaction with γThr142 through the carbonyl moiety. Additionally, 22a has an ester group in the 3 position but the steric hindrance of the t-butyloxycarbonyl fragment makes less probable the hydrogen bond interaction of the carbonyl group with the γThr142 residue; see    These results are in accordance with our previously reported data [6,12], which evidenced the importance of the carbonyl group of the ester moiety to engage a strong hydrogen bond interaction with receptor protein. Compounds missing the ester group (12d-f) show a weak interaction with γThr142 in agreement with the low prediction percentage.
General procedure for the synthesis of compounds 9c,d. Compounds 4, 6 and 7 (0.6 mmol) were suspended in a mixture of POCl 3 (5.5 mL) and PCl 5 (0.91 mmol, 0.190 g) and refluxed for three hours. The evaporation to dryness to eliminate the excess of POCl 3 gave a residue recuperated with ice/water filtered and purified with a suitable solvent, obtaining 9c,d, starting from 4 and 6, respectively. From compound 7, the 5-chloro intermediate 9e was not isolated but used as such, see below.  General procedure for the synthesis of compounds 10c,d. Compounds 9c,d (0.4 mmol) was dissolved in a mixture of CH 2 Cl 2 /EtOH (7.5 mL/15 mL) and NaBH 4 (3.6 mmol, 0.136 g) was added in small portions. The reaction was maintained at room temperature for 40 min, and then the evaporation to dryness of the solvent gave a residue which was recovered with water, filtered and purified with a suitable solvent, obtaining 10c,d, respectively.    (C, H, N).

3-(Thiophen-3-yl)pyrazolo[1,5-a]thieno[2,3-e]pyrimidine (11e).
Compound 7, 3-(thiophene-3-yl)pyrazolo[1,5-a]thieno[2,3-e]pyrimidin-5(4H)-one (0.6 mmol), was suspended in a mixture of POCl 3 (5.5 mL) and PCl 5 (0.91 mmol, 0.190 g) and refluxed for three hours. The evaporation to dryness to eliminate the excess of POCl 3 gave the corresponding 5-chloro derivative (9e), not isolated but used as such for the next reduction step through a CTH (catalytic transfer hydrogenation). Thus, this intermediate suspended in EtOH (20 mL) was added of ammonium formate (4.08 mmol, 0.275 g) and 10% Pd/C as catalyst. The reaction was maintained at reflux temperature for several hours, during which it was possible to evidence, by TLC, the formation of the 4,5-dihydro derivative in a mixture with the final 4,5-dehydro compound 11e. When the reaction finished, the catalyst was filtered off, the solution evaporated to dryness, and the residue recovered with water. General procedure for the synthesis of compounds 12a,c-f and 13b,e,f. A solution of DMF abs. (5 mL), compounds 3-8 (0.40 mmol) and K 2 CO 3 anhydrous (0.80 mmol) was maintained for 15 min at room temperature. After this time, methyl iodide (0.80 mmol) was added and enhanced temperature to 80 • C. After one hour and monitoring the reaction by TLC, adding water gave a precipitate, filtered and purified by a suitable solvent. In the case of compounds 3, 4 and 6, only 4-N-CH 3 derivatives were formed (12a,c,d). From compound 4, only the 5-methoxyderivative 13b was recovered, while if the reaction is performed at reflux temperature, only the 4-methyl derivative 12b was obtained. From 7 and 8, a mixture of two products was recovered at the end of alkylation. The chromatographic separation permits isolating the 4-NCH 3 (12e,f) and the 5-OCH 3 derivatives (13e,f).   (C, H, N). The treatment of compound 12a with sulfuric acid at 60 • C and the subsequent addition of ice/water gave the precipitate, 12b recovered by filtration.   (C, H, N).  179.37, 156.18, 145.24, 144.91, 141.71, 138.40, 135.92, 135.62, 129.12 (C, H, N).  General procedure for the synthesis of compounds 17 and 19. A suspension of 12b or 13b (0.32 mmol) in H 2 SO 4 conc. (8 mL) was stirred until a solution was obtained and then cooled at 0 • C; to this solution, sodium nitrite (0.22g, 3.2 mmol/5 mL of water) was slowly added and the green suspension was maintained for 3 h at 0 • C. The suspension was made alkaline and extracted with ethyl acetate. After the standard work-up, the evaporation of the organic layer gave a green residue that was purified and characterized.   (21). The aldehyde 20 (150 mg, 0.73 mmol) was suspended in acetone and water (5 mL/5 mL) and a solution of potassium permanganate (1.1 mmol) in water was added after the suspension was made alkaline with sodium hydroxide 10%. The reaction was heated for 8 h, and after cooling and elimination of the manganese dioxide by filtration, the alkaline aqueous phase was extracted to eliminate the starting material not reacting. The next acidification of the aqueous phase gave the corresponding carboxylic acid that was recovered by extraction. Yield 60%, white crystals, mp 220-223 General procedure for the synthesis of compounds 22a,b. The carboxylic acid 21 (0.5 mmol) was transformed into the corresponding 3-carbonyl chloride by reaction with excess SOCl 2 in anhydrous conditions. After the standard work-up, the residue was suspended in dichloromethane (6 mL), and the suitable alcohol (excess 0.15 mL) was added; TLC monitored the reaction until the disappearance of the starting material. Then, the final solution was evaporated to dryness, and the residue recuperated with isopropyl ether and recrystallized.

Molecular Docking and Molecular Dynamic Simulation
The structure of the binding site was obtained from the Human α1β2γ2-GABAA receptor subtype in complex with GABA and flumazenil, conformation B (PDB ID 6D6T) [21], considering all the amino acids within a distance of about 2 nm from the structure of the Flumazenil. The ligands were placed at the binding site through AUTODOCK 4.2 [20]. The molecular dynamics simulations of ligand binding-site complexes were performed on a minimum number of conformations (maximum 2) such as to cover at least 90% of the poses found by AUTODOCK. A 60 ns MD simulation were performed for all complexes using GROMACS v5.1 program, and it was conducted in vacuum [22]. The DS ViewerPro 6.0 program [19] was used to build the initial conformations of ligands. The partial atomic charge of the ligand structures was calculated with CHIMERA [23] using AM1-BCC method, and the topology was created with ACPYPE [24] based on the routine Antechamber [25]. The OPLS-AA/L all-atom force field [26] parameters were applied to all the structures. To remove bad contacts, the energy minimization was performed using the steepest descent algorithm until convergence is achieved or for 50,000 maximum steps. The next equilibration of the system was conducted in two phases: (1) Canonical NVT ensemble, a 100 ps position-restrained of molecules at 300 K was carried out using a temperature coupling thermostat (velocity rescaling with a stochastic term) to ensure the proper stabilization of the temperature [27]. (2) Isothermal isobaric NPT ensemble, a 100 ps position-restrained of molecules at 300 K and 1 bar was carried out without using barostat pressure coupling to stabilize the system. These were then followed by a 60 ns MD run at 300 K with position restraints for all protein atoms. The Lincs algorithm [28] was used for bond constraints to maintain rigid bond lengths.
The initial velocity was randomly assigned taken from Maxwell-Boltzman distribution at 300 K and computed with a time step of 2 fs, and the coordinates were recorded every 0.6 ns for MD simulation of 60 ns. During the simulated trajectory, 100 conformations were collected. The 'Proximity Frequencies' (PFs) [1,2] with which the 100 conformations of each binding-site ligand complex intercepts two or more amino acid during the dynamic simulation have calculated. The 'Proximity Frequency' (PF) is the frequency with which the ligand was, during the molecular dynamic simulation, at a distance of less than 0.25 nm from an amino acid of the binding-site and also, simultaneously, from 2, 3 and 4 amino acids of the binding site.
The introduction in this new scaffold of those fragments responsible for the activity in our PQ [6,12] gave six final designed compounds (12c-f and 22a,b), which were, in turn, studied in the 'Proximity Frequency' model [13] to predict their potential profile on the α1β2γ2-GABA A R.
The results indicate for all six products an agonist profile, highlighting the suitability of the nucleus and confirming the importance of the carbonyl group of the ester moiety to engage strong hydrogen bond interaction with the receptor protein. In particular, the esters derivatives (12c and 22b) are able, through the carbonyl group, to interact with the amino acid residue γThr142 with a strong hydrogen bond (2.07 Å). Additionally, 22a has an ester group in position 3, but the steric hindrance of the t-butyloxycarbonyl fragment makes the hydrogen bond interaction of the carbonyl group with the γThr142 residue less probable. Compounds missing the ester group (12d-f) show a weak interaction with γThr142 in agreement with the low prediction percentage.
In conclusion, the synthesis of the 5-oxo-4,5-dihydropyrazolo [1,5-a]thieno [2,3-e]pyrimidine scaffold allowed us to identify a new chemical class of compounds potentially active on GABA A receptor subtype, and the in silico results should be completed and confirmed with biological assays.