Regiospecific synthesis, antibacterial and anticoagulant activities of novel isoxazoline chromene derivatives

1 Introduction

The coumarin skeleton is found in many natural products and is also used as a synthetic intermediate for the preparation of numerous heterocyclic compounds, possessing a wide spectrum of biological activities (Paramjeet et al., 2012; Mutalik and Phaniband, 2011) such as antibacterial (Olayinka and Obinna, 2010), antiviral, antitumor (Hua et al., 2008), anti-HIV and anti-inflammatory properties (More and Mahulikar, 2011). Main representatives of this class are the 4-hydroxy-2H-chromen-2-one and the 4-mercapto-2H-chromene-2-thione, which also have shown good anticoagulant activity combined with low side effects, little toxicity (Manolov et al., 2006) and very useful for the synthesis of other coumarin derivatives. Until now, an enormous number of heterocyclic compounds with fused isoxazoline and dihydrothiophene have been excellently reviewed (Desai et al., 2008; Majumdar et al., 2010). Particularly, isoxazoline is a five-membered heterocyclic which is a versatile lead compound for designing potent bioactive agents (Kaur et al., 2010; Baseer et al., 2012). It is interesting to note from the chemical literature that 2-isoxazoline derivatives showed various biological activities such as anti-stress (Rakesh et al., 2011), antinociceptive (Karthikeyan et al., 2009), anticonvulsant (Hemlata et al., 2010), analgesic (Habeeb et al., 2001) and anti-influenza effects (Kai et al., 2001). Also they showed a good potency in animal models of thrombosis (Pinto, 2001) and played a crucial role in the theoretical development of heterocyclic chemistry.

The 1,3-dipolar cycloaddition has been the subject of intense research over the last decade, due to its great synthetic value (Nair and Suja, 2007) and was the most effective process to the synthesis of five-membered heterocycles which are difficult to be prepared with other means. Furthermore, it gives access to several substances with pronounced biological activities (Gaonkar et al., 2007; Jadhav, 2010). As a part of our search for five-membered heterocyclic compounds, we proposed to investigate the behavior of the 4-(allyloxy)-2H-chromen-2-one 2a among other analogs 2b–d, all prepared from the natural synthon, the 4-hydroxy-2H-chromen-2-one 1 toward different arylnitrile oxides as a dipole (Pellissier, 2007). The reaction was regiospecific and led to a series of new 4-((3-aryl-4,5-dihydroisoxazol-5-yl)methoxy)-2H-chromen-2-ones 3a–f. Next, the 4-(2-thioxo-2H-chromen-4-ylthio)-2H-chromen-2-one 5 and the 4,4′-thiobis (2H-chromen-2-one) 6 have been synthesized from allylic halides and 4-mercapto-2H-chromene-2-thione 4 by the application of the phase-transfer catalysis condition. Similarly, the S-alkylation of 4 under classical conditions using K₂CO₃ and acetone was also investigated for the preparation of new 4-(allylthio)-2H-chromene-2-thiones 7a–d. Following the same way, we have developed the 1,3-cycloaddition between various arylnitrile oxides and 4-(allylthio)-2H-chromene-2-thione 7a, however this specific addition provided, via the Claisen rearrangement, the 2-methyl-2,3-dihydrothieno[3,2-c]chromen-4-one 8. The evaluation of the anticoagulant and antibacterial activities of all the synthesized heterocycles has been studied and reported here.

2 Results and discussion

The requisite starting materials for this study, the 4-(allyloxy)-2H-chromen-2-ones 2a–d, were prepared by a classical alkylation (Avetisyan and Alvandzhyan, 2006) by condensing the 4-hydroxy-2H-chromen-2-one 1 in equimolar amounts with different allylic halides in refluxing anhydrous acetone in the presence of potassium carbonate for 20 h to give a number of 4-(allyloxy)-2H-chromen-2-ones 2a–d (Scheme 1).

Close

Scheme 1

Synthesis of derivatives 3a–f.

Export to PPT

The structures of compounds 2a–d were established on the basis of their spectroscopic data. Thus, compound 2a (as an example) was obtained as a white solid, its ES-MS spectrum gave a pseudo-molecular ion peak [M + H]⁺ at m/z 203 which is consistent with the molecular formula C₁₂H₁₀O₃. The ¹H NMR spectrum of compound 2a run at 300 MHz in CDCl₃ indicates the presence of characteristic signals of coumarin skeleton which can be, according to their chemical shifts and multiplicity, readily assigned to H-3 (δ_H 5.69, s, 1H), H-7 (δ_H 7.55, m, 1H), H-5 (δ_H 7.85, dd, 1H, J = 8.1 Hz, J = 1.5 Hz) and H-6,8 (δ_H 7.29, m, 2H). In addition to the signals corresponding to the protons introduced by 4-hydroxy-2H-chromen-2-one, we revealed the appearance of a signal at δ_H 4.69 (d, 2H, J = 6 Hz) relative to the protons H-2′. The two doublet of doublets signal at δ_H 5.42 (dd, 1H, J = 10.5 Hz, J = 0.9 Hz) and δ_H 5.52 (dd, 1H, J = 17.1 Hz, J = 0.9 Hz) was attribuated to the terminal ethylenic protons H-4′a and H-4′b. The multiplet at δ_H 6.07 (m, 1H) was assigned to the ethylenic proton H-3′. On the other hand, C-2′ (69.8 ppm), C-4′ (119.5 ppm), C-3′ (130.7 ppm) and aromatic carbons: C-5, C-6, C-7 and C-8 were readily assigned from the ¹³C NMR spectrum.

Encouraged by our results and in conjunction with the pharmaceutical importance known for fused heterocyclic incorporating a coumarin moiety, we chose to explore the terminal olefin moiety in order to synthesize new isoxazoline chromenes. Compound2a was treated according to the 1,3-dipolar cycloaddition with various arylnitrile oxides as a dipole (Liu et al., 1980) in refluxing anhydrous toluene for 6 h, that provided the corresponding heterocyclic compounds 3a–f (Scheme 1).

Compound 3b was obtained as a white powder. Its positive ES–MS showed a pseudo-molecular ion peak [M + H]⁺ at m/z 336 compatible with the molecular formula C₂₀H₁₇NO₄. The structure was evidenced by the disappearance in the ¹H NMR spectrum of the signal at δ_H 5.46 relative to the terminal ethylenic protons, the appearance of a doublet of doublets signal at δ_H 3.50 (dd, 2H, J = 16.8 Hz, J = 6.6 Hz) attributable to the protons H-4′ and the presence of a singlet at δ_H 2.39 (s, 3H) corresponding to the methyl protons. A characteristic AA′BB′ pattern for aromatic hydrogens was observed in the ¹H NMR spectrum. Examination at 300 MHz offered an excellent resolution with a doublet at δ_H 7.25 (d, 2H, J = 8.1 Hz, H-3′′,5′′) and a second doublet at δ_H 7.60 (d, 2H, J = 8.1 Hz, H-2′′,6′′). The ¹³C NMR spectrum confirmed the above spectral data by the observation of signals at 126.0 (C-1′′), 126.7 (C-2′′,6′′), 129.5 (C-3′′,5′′) and 140.8 (C-4′′) ppm relative to carbons of the p-substituted aromatic system. The same spectrum showed signals at δ_C 77.4 and 156.2 attributable to the bearing oxygen carbon C-5′ and C-3′ of the imine function, respectively. In addition, a whole set of linkages confirming the molecular structure of compound 3b was reinforced by the HMBC spectrum which showed correlations between the proton H-4′ and C-1′′, C-3′, C-5′ and C-6′ as well as correlations between H-2′′,6′′ and C-1′′, C-3′ and C-4′′. Moreover, the regiochemistry was confirmed by the NOE observed between the protons H-4′ and the aromatic protons H-2′′,6′′.

The regioselectivity of this cycloaddition reaction generally leads to a mixture of 1,4 and 1,5-regioisomers (Loránd et al., 2009). Although, in this study the novel 1,2,4-isoxazoline derivatives 3a–f were formed as unique products, indicating the regiospecificity of the reaction. Indeed, the non formation of the other 1,5-regioisomer may be explained by a possible steric crowding and by electronic factors.

On the other hand, the preparation of the substrates 4-(allylthio)-2H-chromene-2-thione 7a–d was tested according to the literature (Majumdar and Ghosh, 2002) by phase-transfer catalyzed alkylation of 4-mercapto-2H-chromene-2-thione 4, prepared according to the previously reported method (Ibrahim, 2006), with a number of different allylic halides using tertiary butyl ammonium chloride (TBAC) in aqueous sodium hydroxide and chloroform. Coincidentally the PTC did not afford the desired products 7a–d, but provided two new heterocyclic compounds 4-(2-thioxo-2H-chromen-4-ylthio)-2H-chromen-2-one 5 and 4,4′-thiobis (2H-chromen-2-one) 6 (Scheme 2).

Close

Scheme 2

Synthesis of derivatives 7a–d and products 5, 6.

Export to PPT

Compounds 5 and 6 were established on the basis of their spectroscopic data. Their positive HRMS-ESI showed pseudo-molecular ion peaks [M + Na]⁺ at m/z 360.9954 and at m/z 345.0189, respectively, which were consistent with the molecular formulas C₁₈H₁₀O₃NaS₂ and C₁₈H₁₀O₄NaS, respectively. Then the structural assignment of the products 5 and 6 was made on the basis of ¹H and ¹³C NMR spectral analysis and by comparison of their spectroscopic data with those reported for related systems. In particular, the ¹H NMR spectrum of compound 5 exhibited two singlets at δ_H 6.39 (s, 1H) and at δ_H 7.24 (s, 1H) attributable to two ethylenic protons H-3 and H-3′, respectively. In addition, four doublet of doublets was observed at δ_H 7.42 (dd, 1H, J = 8.1 Hz, J = 1.6 Hz), δ_H 7.54 (dd, 1H, J = 8.1 Hz, J = 1.6 Hz), δ_H 7.69 (dd, 1H, J = 8.4 Hz, J = 1.5 Hz) and δ_H 7.82 (dd, 1H, J = 8.4 Hz, J = 1.5 Hz) corresponding to the aromatic protons H-8, H-8′, H-5 and H-5′, respectively. Also we observed four multiplets at δ_H 7.36 (m, 1H), δ_H 7.38 (m, 1H), δ_H 7.65 (m, 1H) and δ_H 7.69 (m, 1H) assignable to the aromatic protons H-6, H-6′, H-7 and H-7′, respectively.

Compared to that of compound 5, the ¹H NMR spectrum of 6 indicates the presence of characteristic signals of coumarin skeleton which can be, according to their chemical shifts and multiplicity, readily assigned to H-3 (6.42 ppm, s, 1H), H-6 (7.34 ppm, m, 1H), H-8 (7.40 ppm, dd, 1H, J = 8.4 Hz, J = 1.5 Hz), H-7 (7.64 ppm, m, 1H) and H-5 (7.80 ppm, dd, 1H, J = 7.8 Hz, J = 1.5 Hz). A complete ¹³C assignment of compounds 5 and 6 was evidenced by the persistence of each of them similarly signals at δ_C 158.4 relative to the lactone carbonyl and the presence of a signal at δ_C 194.1 attributed to the thione of compound 5.

To solve the problem unambiguously caused by dimerization, it sufficed to condense compound 4 with various allylic halides, in refluxing anhydrous chloroform in the presence of potassium carbonate (Majumdar et al., 1992) for 24 h providing the corresponding heterocyclic compounds 7a–d (Scheme 2). Whole of the combined spectral data conclusively agrees with the confirmed structures, compared to those of compounds 2a–d. However, the relative deshielding of the aromatic protons in compounds 7a–d is due to the sulfur atom.

Furthermore, this reaction exhibits two sites; a novel terminal olefin system and the thione function as dipolarophiles, which normally react in 1,3-dipolar cycloaddition reactions with arylnitrile oxides. We performed this reaction as shown in Scheme 3.

Close

Scheme 3

Synthetic pathway of 2-methyl-2,3-dihydrothieno[3,2-c]chromen-4-one 8.

Export to PPT

This reaction did not give the suitable heterocycles but it has led, via the Claisen rearrangement (Majumdar and Biswas, 2004), to the formation of the 2-methyl-2,3-dihydrothieno[3,2-c]chromen-4-one 8. The structure 8 was ascertained from its spectroscopic data. For example, its ¹H NMR spectrum showed a singlet at δ_H 1.45 (s, 3H) corresponding to methyl protons, two doublet of doublets at δ_H 2.97 (dd, 1H, J = 16.5 Hz, J = 5.4 Hz) and δ_H 3.42 (dd, 1H, J = 16.5 Hz, J = 8.7 Hz) attributable to the diastereotopic protons H-3a and H-3b, respectively, together with a multiplet signal at δ_H 4.09 (m, 1H) relative to the proton H-2. The ¹³C NMR spectrum of 8 was also a good support for the proposed structure which exhibited characteristic signals at δ_C 23.0, 41.4 and 45.7 corresponding to carbon of the methyl group, C-3 and C-2 of dihydrothiophene moiety, respectively. Also, the quaternary carbons (C-3a, and C-9b) appeared at δ_C 116.7 and 153.1.

We have noticed that the arylnitrile oxides reacted neither with the terminal ethylenic system nor with the thione function. Hence, the products 8 and 9a–c were obtained as a result from a mono-condensation of the dipole on the thione function of the dipolarophile leading to an unstable spiro-compound which underwent an intramolecular rearrangement to give exclusively the 2-methyl-2,3-dihydrothiophene fused coumarin 8 and the aryl isothiocyanates 9a–c (Elhazazi et al., 2003).

3 Antibacterial activity

Compounds 3a–f were screened for their antibacterial activity against two Gram-negative bacteria (Escherichia coli and Pseudomonas aeruginosa) and two other Gram-positive bacteria (Staphylococcus aureus and Enterococcus faecalis). The obtained data revealed that most of the compounds showed moderate to excellent activities against the used microorganisms (Table 1). Gentamicin was used as a reference drug for the determination of antibacterial activities. DMSO was used as a blank exhibited no activity against any of the used strains.

It has been observed that some of the compounds exhibited interesting antibacterial activities. Indeed, compounds 3a, 3b and 3d showed effective activity against P. aeruginosa (MIC = 0.03 mg/mL) compared to that of Gentamicin (MIC = 0.5 mg/mL). Moreover, compound 3f with the pyrrol system showed a good activity against E. faecalis (MIC = 0.31 mg/mL) but a little less active than the remaining compounds. An increase of affinity of compound 3a against S. aureus, E. faecalis and E. coli (MIC = 0.62 mg/mL) was observed. This result could be due to the absence of any substituent in the aryl system of compound 3a. A further drop in activity of compound 3d against S. aureus, E. faecalis and E. coli (MIC = 1.25 mg/mL) as well as of compound 3f against E. coli and P. aeruginosa (MIC = 1.25 mg/mL) was noted. Compound 3e displayed poor activity against both Gram-positive and Gram-negative bacteria.

4 Anticoagulant activity

The in vitro anticoagulant activity of the synthesized 4-((3-aryl-4,5-dihydroisoxazol-5-yl)methoxy)-2H-chromen-2-one derivatives 3a–f was assessed by measuring the activated partial thromboplastin time and was compared to that of 4-hydroxy-2H-chromen-2-one 1 which has been used as starting material for their synthesis and as positive control. The results are shown in Table 2. All compounds 3a–f had profound anticoagulant activities as indicated by the significant prolongation of aPTT in a concentration-dependent manner.

As shown in Table 2, the aPTT assay indicated that compounds 3a and 3f exhibited the highest anticoagulant activity. The potency of the compound 3e was similar to that of coumarin-derived product 3d whereas 3c and 3b had much lower anticoagulant activities.

Introduction of 3-aryl-4,5-dihydroisoxazol-5-yl (3a) and 3-(1H-pyrrol-2-yl)-4,5-dihydroisoxazole scaffolds (3f) at the coumarin C-4 position resulted in significantly increased anticoagulant potency, which suggested that those chemical groups are responsible for the improvement of their anticoagulant activities. Compounds 3b and 3c both bearing an alkylated aryl moiety were shown to be the less active tested derivatives.

5 Conclusion

In conclusion, we have successfully achieved in this work two important aspects: the methodology by a classical alkylation for the 4-O- and 4-S-allylic chromene derivatives 2a–d and 7a–d, respectively; and a practical method for the synthesis of new isoxazoline chromene derivatives 3a–f and 2-methyl-2,3-dihydrothieno[3,2-c]chromen-4-one 8 via 1,3-dipolar cycloaddition involving arylnitrile oxides. These cycloadditions display appreciable regiospecificity and chemospecificity, giving the products in good yields. It has also been observed that most of the prepared isoxazolines 3a–f exhibited interesting antibacterial activity against some Gram negative and Gram positive strains. Indeed, the compounds 3a, 3b and 3d showed effective activity (MIC = 0.03 mg/mL) toward P. aeruginosa compared to that of Gentamicin (MIC = 0.5 mg/mL) used as a positive control. The biological evaluation showed that isoxazolines 3a–f had profound anticoagulant potency. While introducing 3-aryl-4,5-dihydroisoxazol-5-yl (3a) and 3-(1H-pyrrol-2-yl)-4,5-dihydroisoxazole scaffolds (3f) at the coumarin C-4 position resulted in significantly increased anticoagulant effect.

6 Experimental section

Melting points were taken on a Büchi-510 capillary melting point apparatus. ¹H (300 MHz) and ¹³C (75 MHz) NMR spectra were recorded with AC-300 Bruker spectrometers. Mass spectra were obtained with Micromass LCT (ESI technique, positive mode) spectrometer. HRMS spectra were acquired with an ESI–TOF (LCT Premier XE, Waters) using the reflectron mode in the positive ion mode. Leucine-enkephalin peptide was employed as the LockSpray lockmass. Commercial TLC plates (Silica gel 60, F254, SDS) were used to monitor the progress of the reaction. Column chromatography was performed with silica gel 60 (particle size 40–63 μm, SDS).