Ammonium Formate-Pd/C as a New Reducing System for 1,2,4-Oxadiazoles. Synthesis of Guanidine Derivatives and Reductive Rearrangement to Quinazolin-4-Ones with Potential Anti-Diabetic Activity

1,2,4-Oxadiazole is a heterocycle with wide reactivity and many useful applications. The reactive O-N bond is usually reduced using molecular hydrogen to obtain amidine derivatives. NH4CO2H-Pd/C is here demonstrated as a new system for the O-N reduction, allowing us to obtain differently substituted acylamidine, acylguanidine and diacylguanidine derivatives. The proposed system is also effective for the achievement of a reductive rearrangement of 5-(2′-aminophenyl)-1,2,4-oxadiazoles into 1-alkylquinazolin-4(1H)-ones. The alkaloid glycosine was also obtained with this method. The obtained compounds were preliminarily tested for their biological activity in terms of their cytotoxicity, induced oxidative stress, α-glucosidase and DPP4 inhibition, showing potential application as anti-diabetics.


Introduction
1,2,4-Oxadiazoles are five-membered heterocycles that have been widely studied for their valuable applications ranging from the pharmaceutical field to material science [1]. The use of a 1,2,4-oxadiazole ring as an isosteric replacement of esters and amides increased their presence in medicinal chemistry projects [2] with biological applications ranging from the treatment of nonsense mutations [3,4] and Alzheimer's disease (AD) [5][6][7] to antibiotics [8][9][10] and antitumorals [11,12]. Applications in the field of materials chemistry, such as liquid crystals [13] metal sensors [14], OLED [15], energetic materials [16] and gas sorbing/releasing systems [17][18][19], have been reported. The high tendency to react and rearrange of this low-aromatic ring is one of the features that is mainly related to O-N bond breaking [20].

Reduction Reactions
We used 3,5-diphenyl-oxadiazole 1a as a model substrate in the optimization of the reaction conditions (Scheme 1). As shown in Table 1, the corresponding acylamidine 2a was obtained only in the presence of polar protic solvents (entries [8][9][10][11]; however, the desired product was obtained in good yield and in a shorter time by heating the reaction in methanol (entry 12). Interestingly, formic acid could also be used as a reducing agent (entry [10][11]. The use of zinc dust or the absence of catalyst did not produce the desired product (entry 6-7). The scope and limitations of these reduction reactions under optimized conditions using NH4CO2H and Pd/C (5%) in MeOH at 60 °C, were investigated with variously substituted oxadiazoles 1b-o (Scheme 2) and 3a-m (Scheme 3). Acylamidine, acylguanidine and diacylguanidine derivatives 2b-o were obtained in good to excellent yields with the only exception of chlorinated compound 2l, which, in addition to the reduction of the O-N bond, gave the corresponding de-halogenated product 2k (Scheme 2). Scheme 1. Reduction of 3,5-diphenyl-1,2,4-oxadiazole 1a.
As shown in Table 1, the corresponding acylamidine 2a was obtained only in the presence of polar protic solvents (entries 8-11); however, the desired product was obtained in good yield and in a shorter time by heating the reaction in methanol (entry 12). Interestingly, formic acid could also be used as a reducing agent (entry [10][11]. The use of zinc dust or the absence of catalyst did not produce the desired product (entry 6-7). The scope and limitations of these reduction reactions under optimized conditions using NH 4 CO 2 H and Pd/C (5%) in MeOH at 60 • C, were investigated with variously substituted oxadiazoles 1b-o (Scheme 2) and 3a-m (Scheme 3). Acylamidine, acylguanidine and diacylguanidine derivatives 2b-o were obtained in good to excellent yields with the only exception of chlorinated compound 2l, which, in addition to the reduction of the O-N bond, gave the corresponding de-halogenated product 2k (Scheme 2).
Furthermore, the reaction of ortho-aminophenyl-1,2,4-oxadiazoles 3 was considered as a potential reductive ring-rearrangement that allowed us to obtain the corresponding quinazolin-4-ones 4 in good to excellent yields (Scheme 3). The observed limitations of the proposed reaction are referred to the simultaneous dechlorination of compound 4k to compound 4a and by the reductive debenzylation process of 3m to compound 4a. The reaction performed on N-alkyl substituted oxadiazoles 3b,d-g,j provided the regioselective formation of N(1) substituted quinazolinones 4.
From a mechanistic point of view, the reductive rearrangement of oxadiazoles 3 into quinazolin-4-ones 4 could be rationalized with an initial reduction of the O-N bond, to amidine I accompanied by an intramolecular cyclization into intermediate II, followed by the elimination of ammonia, to obtain compound 4 (Scheme 4) [35].
The synthesized compounds were screened for their in vitro DPPIV inhibitor activity. The DPPIV enzyme is a therapeutic target in diabetic disease, and its inhibition increases the glucagon-like peptide levels with a consequent decrease of postprandial glycemia [36]. DPPIV activity was determined by cleaving the substrate to yield a fluorescent product (λ ex = 360/λ em = 460), proportional to the enzymatic activity. As shown in Figure 1, the tested compounds showed an important inhibitor activity at 100 µM, in particular, the alkaloid Glicosine 4d [37] and its derivatives.
The guanidine derivatives 2k, 2n and 2e inhibit the enzyme at the same concentration ( Table 2). The other tested compounds were not active or gave spectral interference with the assay. For those compounds with a good inhibitor activity (>83%), the IC 50 was determined. Table 2 shows the IC 50 values for the most active compounds, revealing a good potency in the sub-micromolar range for all compounds and an IC 50 value for Glicosine of the same order of that of reference drug Sitagliptin.
Furthermore, the reaction of ortho-aminophenyl-1,2,4-oxadiazoles 3 was considered as a potential reductive ring-rearrangement that allowed us to obtain the corresponding quinazolin-4-ones 4 in good to excellent yields (scheme 3). The observed limitations of the proposed reaction are referred to the simultaneous dechlorination of compound 4k to compound 4a and by the reductive debenzylation process of 3m to compound 4a. The reaction performed on N-alkyl substituted oxadiazoles 3b,d-g,j provided the From a mechanistic point of view, the reductive rearrangement of oxadiazoles 3 into quinazolin-4-ones 4 could be rationalized with an initial reduction of the O-N bond, to amidine I accompanied by an intramolecular cyclization into intermediate II, followed by the elimination of ammonia, to obtain compound 4 (Scheme 4) [35].

Inhibition Test of Dipeptidyl-Peptidase IV (DPPIV) Enzyme
The synthesized compounds were screened for their in vitro DPPIV inhibitor activity The DPPIV enzyme is a therapeutic target in diabetic disease, and its inhibition increase the glucagon-like peptide levels with a consequent decrease of postprandial glycemia [36] DPPIV activity was determined by cleaving the substrate to yield a fluorescent produc (λex = 360/λem = 460), proportional to the enzymatic activity. As shown in Figure 1, th tested compounds showed an important inhibitor activity at 100 µM, in particular, th alkaloid Glicosine 4d [37] and its derivatives. The guanidine derivatives 2k, 2n and 2e inhibit the enzyme at the same concentration ( Table 2). The other tested compounds were not active or gave spectral interference with the assay. For those compounds with a good inhibitor activity (>83%), the IC50 was determined. Table 2 shows the IC50 values for the most active compounds, revealing a good potency in the sub-micromolar range for all compounds and an IC50 value for Glicosine of the same order of that of reference drug Sitagliptin.

Inhibition Test of α-Glucosidase Enzyme
The α-glucosidase enzyme catalyzes the hydrolysis of various carbohydrate substrates into monosaccharides that are easily absorbed through the small intestine [37]. The inhibition of α-glucosidase slows down the glycemic levels and represents a therapeutic strategy for diabetic patients. Compounds were evaluated in vitro for their alpha-glucosidase inhibitory activity. The assay is based on enzyme catalyzed hydrolysis of 4-nitrophenyl β-D-glucopyranoside (pNPG) to 4-nitrophenol measured for absorbance at 490 nm. The inhibitor Acarbose was used as a positive control. The tested compounds gave absorbance values similar to that of the uninhibited enzyme (data not shown); therefore, no inhibitory activity was envisaged.

Cytotoxicity Assay and Evaluation of Intracellular Redox State
The toxicity of several derivatives was assessed on a human neuroblastoma cell line SH-SY5Y (ATCC: CRL-2266 TM ) and on human adenocarcinoma alveolar basal epithelial cells A549 (CCL-185) (results not shown). Cells were exposed for 24 h to four concentrations of the compounds. The absorbance values measured in the MTS assay were proportional to the cell viability. The results were similar for both cell lines. At the lower concentrations of 12.5 and 25 µM, the compounds showed a good viability percentage comparable with that of untreated cells ( Figure 2). Some compounds exerted a toxic effect at the concentrations of 50 and 100 µM. The observation of the cell morphology under the electron microscope showed crystals formation for the compounds 4i and 3m.
The inhibition of α-glucosidase slows down the glycemic levels and represents a therapeutic strategy for diabetic patients. Compounds were evaluated in vitro for their alpha-glucosidase inhibitory activity. The assay is based on enzyme catalyzed hydrolysis of 4-nitrophenyl β-D-glucopyranoside (pNPG) to 4-nitrophenol measured for absorbance at 490 nm. The inhibitor Acarbose was used as a positive control. The tested compounds gave absorbance values similar to that of the uninhibited enzyme (data not shown); therefore, no inhibitory activity was envisaged.

Cytotoxicity Assay and Evaluation of Intracellular Redox State
The toxicity of several derivatives was assessed on a human neuroblastoma cell line SH-SY5Y (ATCC: CRL-2266 TM ) and on human adenocarcinoma alveolar basal epithelial cells A549 (CCL-185) (results not shown). Cells were exposed for 24 h to four concentrations of the compounds. The absorbance values measured in the MTS assay were proportional to the cell viability. The results were similar for both cell lines. At the lower concentrations of 12.5 and 25 µM, the compounds showed a good viability percentage comparable with that of untreated cells ( Figure 2). Some compounds exerted a toxic effect at the concentrations of 50 and 100 µM. The observation of the cell morphology under the electron microscope showed crystals formation for the compounds 4i and 3m. Figure 2. Viability percentage of SH-SY5Y cells exposed for 24 h to serial concentrations of the compounds. Data were compared by one-way ANOVA followed by Dunnett's multiplecomparison test: *, p < 0.05; **, p < 0.01 as compared to control (C) group.
The assay with 2′,7′-dichlorofluorescin-diacetate (DCFH-DA) was performed to evaluate the effect of the compounds on the intracellular redox potential of SH-SY5Y cells and A549 cells (results not shown). DCFH-DA is a non-fluorescent and cell-permeant . Viability percentage of SH-SY5Y cells exposed for 24 h to serial concentrations of the compounds. Data were compared by one-way ANOVA followed by Dunnett's multiple-comparison test: *, p < 0.05; **, p < 0.01 as compared to control (C) group.

of 16
The assay with 2 ,7 -dichlorofluorescin-diacetate (DCFH-DA) was performed to evaluate the effect of the compounds on the intracellular redox potential of SH-SY5Y cells and A549 cells (results not shown). DCFH-DA is a non-fluorescent and cell-permeant molecule used as an indicator for reactive oxygen species (ROS). Inside the cell, the DCFH-DA is hydrolyzed by intracellular esterase with cleavage of the acetate group.
In the presence of ROS, the molecule is oxidized to the highly fluorescent 2 ,7dichlorofluorescin (DCF). The emitted fluorescence is proportional to the quantity of the intracellular ROS. The treated cells gave fluorescence values similar to the untreated cells (C) (Figure 3). This result suggests that the tested compounds do not stimulate ROS production and, therefore, do not give oxidative stress.

Discussion
The reduction reactions are one of the most investigated transformations in organic synthesis, and there is a growing interest to identify new reagents and methodologies for that purpose. Here, we report a new reduction method of 1,2,4-oxadiazole involving readily available and cheap ammonium formate as a hydrogen source and Pd/C as catalyst. This method allowed us to obtain, using mild conditions, acylamidine, acylguanidine and diacylguanidine without hydrolysis to amidines. The reduction of 5-(2′-aminophenyl)-1,2,4-oxadiazoles gave a reductive rearrangement into 1alkylquinazolin-4(1H)-ones.
Particularly, the reduction of O-N bond gave the N-acylamidines that cyclize in situ to produce quinazolin-4-ones in excellent yields. Usually, quinazolinone compounds can be synthetized by oxidative cyclization of benzamide or benzonitrile [38,39]. Our method, compared to the classical one, does not involve an oxidizing agent and it allowed to obtain quinazolinone derivatives alkylated on nitrogen at position 1. This feature is of interest considering that the alkylation of quinazolinones usually occurs at N(3) or gives mixtures of different alkylation products.
Considering the assessed hypoglycemic activity of Glicosine analogues of the hypoglycemic alkaloid glycosine [37] and the similar activity of similar compounds, we decided to test the potential application of synthesized compounds as antidiabetic drugs. Considering that the mode of action of Glicosine is still unknown, we tested major target of drugs with hypoglycemic activity against DPPIV and α-glucosidase. The obtained compounds were inactive against α-glucosidase; however, four quinazolinones 4b,d,e,l were able to inhibit DPPIV in the sub-micromolar range.
These results confirm previously reported docking studies that proposed the DPPIV enzyme as a biological target for Glicosine [37]. Considering recent evidence regarding the pro-oxidant and neurotoxic activity of hypoglycemic drug Metformin [40], we also tested the pro-oxidant activity of selected compounds. As shown by MTS and dichlorofluorescein assays, the derivatives have a good cytocompatibility in terms of cell viability and pro-oxidant activity.

Materials
All solvent and reagents were obtained from commercial sources and were used

Discussion
The reduction reactions are one of the most investigated transformations in organic synthesis, and there is a growing interest to identify new reagents and methodologies for that purpose. Here, we report a new reduction method of 1,2,4-oxadiazole involving readily available and cheap ammonium formate as a hydrogen source and Pd/C as catalyst. This method allowed us to obtain, using mild conditions, acylamidine, acylguanidine and diacylguanidine without hydrolysis to amidines. The reduction of 5-(2 -aminophenyl)-1,2,4-oxadiazoles gave a reductive rearrangement into 1-alkylquinazolin-4(1H)-ones.
Particularly, the reduction of O-N bond gave the N-acylamidines that cyclize in situ to produce quinazolin-4-ones in excellent yields. Usually, quinazolinone compounds can be synthetized by oxidative cyclization of benzamide or benzonitrile [38,39]. Our method, compared to the classical one, does not involve an oxidizing agent and it allowed to obtain quinazolinone derivatives alkylated on nitrogen at position 1. This feature is of interest considering that the alkylation of quinazolinones usually occurs at N(3) or gives mixtures of different alkylation products.
Considering the assessed hypoglycemic activity of Glicosine analogues of the hypoglycemic alkaloid glycosine [37] and the similar activity of similar compounds, we decided to test the potential application of synthesized compounds as antidiabetic drugs. Considering that the mode of action of Glicosine is still unknown, we tested major target of drugs with hypoglycemic activity against DPPIV and α-glucosidase. The obtained compounds were inactive against α-glucosidase; however, four quinazolinones 4b,d,e,l were able to inhibit DPPIV in the sub-micromolar range.
These results confirm previously reported docking studies that proposed the DPPIV enzyme as a biological target for Glicosine [37]. Considering recent evidence regarding the pro-oxidant and neurotoxic activity of hypoglycemic drug Metformin [40], we also tested the pro-oxidant activity of selected compounds. As shown by MTS and dichlorofluorescein assays, the derivatives have a good cytocompatibility in terms of cell viability and prooxidant activity.

Materials
All solvent and reagents were obtained from commercial sources and were used without purification. The reactions were monitored by thin layer chromatography (TLC). The synthesized compounds were purified by silica flash chromatography using silica gel (0.040-0.063 mm) and a mixture of ethyl acetate and petroleum ether (fraction boiling in the range of 40-60 • C) in various ratios as the eluent. The melting points were determined on a Kofler apparatus. FTIR spectra (Nujol or CH 2 Cl 2 ) were determined with a Cary 630, Agilent Technologies instrumentation (Agilent Technologies, Santa Clara, CA, USA). 1 H-NMR, 13 C-NMR and HPLC/MS were utilized to verify the structure and purity of synthesized compounds. 1 H-NMR and 13 C-NMR were recorded at 300 MHz for 1 H and 62.5 MHz for 13 C; DMSO-d 6 or CDCl 3 were used as solvent and TMS as an internal standard. HRMS spectra were recorded in positive or negative mode with HPLC/MS (6540 UHD Accurate Mass Q-TOF LC/MS-Agilent Technologies, Santa Clara, CA, USA) and Dual AJS ESI source. The compounds 1a [41], 1b [42], 1c [43], 1d [44], 1e,j,o [45], 1f [46], 1g,h [47], 1i,k,n [48], 1l,m [49] and 1p [50] were obtained as previously reported. The enzyme α-glucosidase Type I from Saccharomyces cerevisiae, acarbose and 4-nitrophenil-α-D-glucopyranoside have been bought by Sigma Aldrich. DPPIV Activity Fluorometric Assay Kit (Sigma Aldrich, St. Louis, MO, USA) was used for inhibition test of dipeptidyl peptidase. Both inhibition assays were conducted in 96-well plates using a 96-well microplate reader (BioTek, Winooski, VT, USA)

General Procedure for the Synthesis of N-Acylamidines, Quinazolin-4-(1H)-One, Acyl Guanidine and Diacyl Guanidine
To a solution of appropriate oxadiazole (1 mmol) in MeOH (20 mL), NH 4 CO 2 H (3 mmol) was added as a reducing agent and Pd/C 5% (10 mg) was added as a catalyst. The reaction mixture was heated to 60 • C for 1 h (monitored with TLC). The solution was filtered to remove palladium and the filtrate was evaporated under vacuum. The residue was treated with water and extracted with ethyl acetate. If necessary, the product was further purified by column chromatography using petroleum ether and ethyl acetate (1:1) as eluent.   Method A: To a solution in ethanol (30mL) of appropriate amidoxime (1 mmol) isatoic anhydride (1 eq) was added. The reaction mixture was heated to reflux for about 5 h and monitored by TLC. The mixture was evaporated, and the oxadiazole was purified by column chromatography using a mixture of petroleum ether and ethyl acetate (20:1) as eluent.

Method B:
A solution of isatoic anhydride (1 mmol) in DMSO (20 mL) was treated with K 2 CO 3 (1 mmol) and the appropriate alkyl halide (1.1 mmol) and stirred at room temperature for 12 h. The reaction was monitored by TLC. The solvent was evaporated under a vacuum, and the residue was treated with water and extracted with ethyl acetate. The organic layer was dried with Na 2 SO 4 and evaporated under a vacuum. The solid, without further purification, was solubilized in toluene and heated to reflux with the amidoxime (1.5 eq). The reaction was monitored with TLC for about 3 h, and then the solvent was removed under vacuum. The desired product was isolated by chromatography using a mixture of petroleum ether and ethyl acetate (20:1) as eluent.
Method C: 5-(2-fluorophenyl)-1,2,4-oxadiazole (1 mmol) [60] was reacted with methylamine or ethylamine (10 mmol) in EtOH (10 mL) into a pressure tube. The tube was stirred in an oil bath at the temperature of 80 • C for 12 h monitoring the reaction by TLC. The desired product was isolated by column chromatography using a mixture of petroleum ether and ethyl acetate (20:1) as eluent.

DPPIV Activity Assay
The DPPIV activity assay was performed using DPPIV Activity Fluorometric Assay Kit (Sigma Aldrich, St. Louis, MO, USA) in accordance with the manufacturer's instructions. The compounds dissolved in DMSO were diluted to initial concentration of 10 mM and tested at 100 µM. The enzyme inhibition test was conducted in a 96-well flat for fluorescence assay. To the wells have been added 25 µL of tested compound (400 µM), 49 µL of assay buff.
After, an enzymatic reaction mix (23 µL of assay buffer and 2 µL of substrate) was added to each well. The DPPIV activity was determined by cleaving the substrate to yield a fluorescence product (λ ex = 360/λ em = 460) proportional to the enzymatic activity. The measure of fluorescence was made in kinetic mode for 30 min at 37 • C. The fluorescence values were taken every minute. Sitagliptin was used as a positive control. The fluorescence was plotted against time, and the enzyme inhibition percentage was calculated as follows: % Relative Inhibition = (Slope EC-Slope SM)/Slope EC × 100. Slope SM = the slope of the compound's curve. Slope EC = the slope of the Enzyme Control curve. The IC 50 values have been calculated evaluating the inhibition percentage in the presence of 100, 10, 1, 0.1 and 0.01 µM of the compounds.

α-Glucosidase Inhibition Test
An α-Glucosidase inhibition assay was performed according to the method previously reported in the literature [61]. The compounds were prepared in DMSO at a concentration of 10 mM and were tested at five concentrations (100, 10, 1, 0.1 and 0.01 µM). The inhibition test was conducted in a 96-well flat. In each well, we added 70 µL of Na 2 HPO 4 buffer (50 mM) pH 6.8, 10 µL of the tested compounds and 10 µL of enzyme solution in buffer (0.0234 U). The plate was pre-incubated for 10 min at 37 • C and pre-read at 400 nm in absorbance, using a 96-well microplate reader (BioTek, Winooski, VT, USA). The substrate p-nitrophenyl glucopyranoside was added and the plate was incubated at 37 • C for another 30 min. The absorbance, due to the formation of p-nitrophenol, was measured at 400 nm. Acarbose was used as a positive control. The percentage of inhibition was found as follows: % inhibition = abs of control − abs of test compound/abs of control × 100.

MTS Assay
The MTS assay (Promega Italia, S.r.l., Milan, Italy) was performed in accordance with the manufacturer's instructions. The cells were plated on a 96-well plate up to a concentration of ≈6 × 10 5 cell/mL. After 24 h, the cells were treated with the compounds at the concentration of 12.5, 25, 50 and 100 µM. A solution of MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (20 µL) was added to each well and the plate was incubated for 4 h at 37 • C under atmosphere of CO 2 /5%. The absorbance at 490 nm was read with a microplate reader WallacVictor 2 1420 Multilabel Counter (PerkinElmer, Monza, Italy). The results are reported as the absorbance values of treated cells compared to those of the control: % cell viability = (OD compound/OD Control) × 100.

Analysis of Reactive Oxygen Species (ROS)
The cells were plated on a 96-well plate up to a concentration of 6 × 10 5 cell/mL and were treated with the compounds at the concentration of 50 µM and 100 µM. After the treatment, a 1 mM solution of DCFH-DA in PBS was added to the wells. The plate was incubated for 10 min at room temperature and away from light. The cells were washed with PBS and the fluorescence (λ ex = 485 nm/λ em = 530 nm) was monitored by a microplate reader GloMax fluorimeter (Promega Corporation, Madison, USA) and a fluorescence microscope Zeiss Axio Scope 2 (Zeiss, Jena, Germany)

Conclusions
This study highlights an ammonium formate-Pd/C system as a valid reducing agent for O-N bond of 1,2,4-oxadiazoles, allowing us to obtain amidines, guanidines and other derivatives. This strategy also obtained quinazolinones through a new reduction-induced ring-rearrangement of ortho-arylamino-oxadiazole compounds. This strategy make available the alkaloid Glicosine, here identified as an effective DPPIV inhibitor, with potential applications as an antidiabetic drug also in light of the good cytocompatibility and the lack of induction of oxidative stress. Glicosine is a promising lead compound that is selective toward DPPIV and will be further optimized in the future.