Synthesis, antimicrobial and DFT studies of novel fused thiazolopyrimidine derivatives

Introduction

Heterocyclic compounds are highly ranked among pharmaceutically important natural and synthetic materials. The remarkable ability of heterocyclic nuclei to serve as both biomimetics and active pharmacophores has largely contributed to their unique value as traditional key elements of numerous drugs. The thiazolidinone system frequently appears in the structure of various natural products, notably thiamine, compounds possessing cardiac and glycemic benefits such as troglitazone [1] and many metabolic products of fungi and primitive marine animals. Pioglitazone [2] and rosiglitazone are therapeutic agents for the treatment of diabetes [3].

The 4-thiazolidinone system is a core structure in many synthetic compounds exhibiting broad pharmacological spectrum and affinity for various biotargets [4, 5]. Some derivatives are peroxisome proliferator-activated receptor γ (PPAR-γ) agonists showing hypoglycemic activity [6], aldose reductase inhibitors guarding against diabetic complications [7] and cartilage degradation inhibitors [8]. Additionally, 4-thiazolidinone derivatives are reported as anticancer [9, 10], antibacterial [11], anti-HIV [12], anticonvulsant [13], antidiabetic [14] and anti-inflammatory [15] agents.

In continuation of our research program on the synthesis of condensed 4-thiazolidinines [16–18], we report in this paper a novel condensed thiazolo-pyrimidine system. Both thiazole and pyrimidine are bioactive nuclei and it is thought that the condensed system may exhibit enhanced biological activity due to possible synergistic effects. Two novel thiazolo[3,2-a]pyrimidin-3(5H)-one 5 and dihydro-2H-thiazolo[3,2-a]pyrimidine 6 systems are reported. Arylidene derivatives 7 of compounds 5 were also synthesized.

Computational studies

The molecular geometry optimization and ¹H and ¹³C NMR spectra calculations were performed with the Jaguar software package version 6.5112 by using density functional theory (DFT) methods with B3LYP (Becke three parameter Lee-Yang-Parr) exchange correlation functional, which combines the hybrid exchange functional of Becke [19], with the gradient-correlation functional of Lee et al. [20]. The 6–31G** basis set was used for calculations in the gas phase of the structure 5a and its isomer 8a, respectively.

Figure 1

Optimized geometry of compound 5a and its isomer 8a.

As the crystal structures of the molecules are not available, DFT calculations were carried out to predict the geometry of the molecules. The optimized bond lengths and bond angles obtained by geometry optimization at B3LYP/6–31G** level of theory for structure 5a are reported in Table 1. For structure 5a, the optimized bond lengths of C=O and S-C in thiazolidinone ring fall in the range of 1.208 Å and 1.856 Å. The optimized bond angles for N-C-O and N-C-S were observed at 122.6° and 126°, respectively. The optimized configurations of structure 5a and its isomer 8a with atom numbering schemes are shown in Figure 1.

Table 1

Selected bond lengths and bond angles of the optimized structure 5a.

Bond lengths (Å)		Bond angles (°)
Entry	Optimized lengths	Entry	Optimized angles
C(21)-H(36)	1.11	H(36)-C(21)-H(35)	109.4
C(8)-H(29)	1.10	H(36)-C(21)-S(22)	112.2
C(11)-N(10)	1.46	H(36)-C(21)-C(20)	108.8
C(20)-N(10)	1.46	N(10)-C(20)-C(21)	118.0
C(9)-N(10)	1.47	N(10)-C(20)-O(23)	122.6
C(8)-C(9)	1.49	C(21)-C(20)-O(23)	122.5
N(12)-C(7)	1.45	C(19)-C(13)-C(9)	121.4
C(11)-N(12)	1.26	C(7)-N(12)-C(11)	115.0
S(22)-C(11)	1.85	N(10)-C(11)-N(12)	126.0
C(21)-S(22)	1.82	N(12)-C(11)-S(22)	126.0
C(20)-C(21)	1.50	C(11)-N(10)-C(20)	124.0
C(20)-O(23)	1.20	C(11)-N(10)-C(9)	108.0
C(9)-H(14)	1.11	N(10)-C(9)-H(14)	107.5
C(5)-C(7)	1.50	C(8)-C(7)-N(12)	120.0

Shielding tensors of structures 5a and 8a were evaluated by using B3LYP functional with basis set given above. To express the chemical shifts in ppm, the geometry of tetramethylsilane (TMS) and chloroform molecules had been optimized and then their ¹H and ¹³C NMR spectra were calculated by the same method using the same basis set as in the case of the calculations on structures 5a and 8a. The shielding of TMS is 32.3379 for ¹H NMR and 202.8593 for ¹³C NMR. The calculated isotropic shielding constants σ_i were then transformed to chemical shifts relative to TMS by the equation: δ_i = σ_TMS – σ_i.

Figure 2

Plot of the calculated vs. the experimental ¹H NMR chemical shifts (ppm).

Figure 3

Plot of the calculated vs. the experimental ¹³C NMR chemical shifts (ppm).

Experimental and calculated ¹H and ¹³C NMR chemical shifts (ppm) of compounds 5a and its isomer 8a are reported in Tables 2 and 3. The correlation values of proton chemical shifts are found to be 0.9708 for structure 5a and 0.8153 for its isomer 8a (Figure 2). Similarly, the correlation values of carbon chemical shifts of 5a and 8a are found to be 0.9973 and 0.9828, respectively (Figure 3). The theoretical and experimental ¹H and ¹³C data show good correlations for proposed structure 5a and not for 8a. Total energies, zero point energies, HOMO and LUMO energies theoretically obtained for structures 5a and 8a are reported in Table 4. Spectral correlation and energy values support structure 5a for the product and are not consistent for the isomeric structure 8a.

Table 2

Experimental and calculated ¹H NMR chemical shifts (ppm) of compound 5a and its isomer 8a.

Table 3

Experimental and calculated ¹³C NMR chemical shifts (ppm) of compound 5a and its isomer 8a.

Compound 5a				Compound 8a
Entry	Expt. NMR	Calcd shield	Calcd NMR	Entry	Calcd shield	Calcd NMR
C1	132	73.0785	135.64	C1	69.2439	139.65
C3	137	71.5494	137.24	C3	79.3263	129.31
C9	55	151.3189	53.87	C9	145.8602	59.57
C11	161	49.491	160.29	C11	35.2589	175.17
C20	189	26.6823	184.14	C22	5.7554	206.01
C21	39	167.0381	37.44	C21	169.2145	35.16

Table 4

Theoretically computed energies for compounds 5a and 8a.

Parameters	Compound 5a	Compound 8a
HOMO energy (kcal/mol)	-170.01	-165.13
LUMO energy (kcal/mol)	37.88	49.79
Zero point energy (kcal/mol)	186.23	187.71

Results and discussion

Treatment of chalcone 3a, obtained by the reaction of acetophenone and benzaldehyde, with thiourea in alcoholic KOH gave compound 4a (Scheme 1). The unsymmetrical thione 4a on reaction with chloroacetic acid followed by cyclization of the intermediate in situ was likely to be transformed into compound 5a or 8a or both depending on the mode of cyclization. However, the treatment of thione 4a with chloroacetic acid in the presence of anhydrous sodium acetate in absolute ethanol afforded a single product (thin layer chromatography, TLC) 5a or 8a in 78% yield. The appearance of a band at 1744 cm^-1 (C=O) in the IR spectrum, appearance of a peak at δ 189 (C=O) in ¹³C NMR spectrum and the presence of a molecular ion peak at m/z 307 (63%) in the mass spectrum of the TLC pure product suggested that cyclization had indeed taken place. The IR and mass spectral data were of little help in deciding in favor of either structure 5a or 8a. However, the structure 5a was finally assigned to this cyclization product in preference to structure 8a on the basis of ¹H NMR spectral data. Similarly, structure 5b was finally assigned to this cyclization product in preference to structure 8b on the basis of ¹H NMR spectral data.

Scheme 1

Synthesis of condensed 4-thiazolidinones (5) and its arylidene derivatives.

The reaction of compound 4a with 1,2-dibromoethane gave a product which was purified by column chromatography and could be represented by either structure 6a or 9a. In either structure (6a or 9a), the singlet at δ 5.32 in its ¹H NMR spectrum integrating for one proton was assignable to H_A. If the structure 8a is correct for the cyclization product, obtained from thione 4a and chloroacetic acid, and then H_A would resonate in the same region as that of structure 6a (or 9a). By contrast, if structure 5a is correct, H_A will be deshielded by the thiazolidinone ring and consequently H_A will resonate downfield in comparison to H_A in 6a (or 9a). The appearance of a downfield singlet at δ 5.59 for H_A in structure 5a (or 8a) as compared with singlet at δ 5.32 for H_A in structure 6a (or 9a) supports structure 5a and is not consistent with structure 8a for which such a downfield shift would not be expected. The deshielding effect is due to the magnetic anisotropy of the C=O group with a minor contribution from the rest of the ring. Similarly, structure 5b was assigned to the final product (not 8b). The same structural conclusion is supported by comparing the chemical shifts of H_A of thione 4a with that of cyclization product 5a. The H_A proton in thione 4a resonates at δ 5.12, whereas the downfield signal at δ 5.59 (1H, s, H_A) in the cyclized product supports structure 5a in preference to structure 8a. Also, from DFT studies (Table 4) structure 5a is preferred over 8a by energy of 1.48 kcal/mol. Arylidene thiazolidinones 7a–d were prepared by two routes. In the first approach thiazolidinone 5 was condensed with aldehydes to give arylidene thiazolidinones, whereas in the second approach compound 7a was obtained directly by heating compound 4a with chloroacetic acid and benzaldehyde. Structures 7a–d were established by IR and ¹H NMR spectral data. The parent thiazolidinones 5a and 5b exhibit absorption bands at 1744 and 1744 cm^-1 (C=O), but unsaturation at the 2-position being conjugated with the carbonyl group at the 3-position as in arylidene thiazolidinones produces a bathochromic shift [21], as expected; the carbonyl absorption band appears at 1728, 1726 cm^-1 in structures 7a–b and 1728, 1720 cm^-1 in structures 7c–d, respectively. The downfield shift in the position of H_A in ¹H NMR of the product obtained by reaction of 4a with chloroacetic acid may be due to the presence of the imine group in structure 8a. If this is the case then in arylidene derivatives 7a–b H_A proton will not be affected further because of absence of a conjugated carbonyl group. In the case of arylidene derivatives obtained from 5a the conjugated carbonyl function produces further downfield shift in position of H_A in 7a–b, that is, δ 6.28 and δ 6.32. This corroborates with structure 5a and is not consistent with 8a. Structure 6 (not 9) for the product, obtained from the reaction of compound 4 with 1,2-dibromoethane, was assigned based on the analogy with structure 5. In a similar way, the structures for compounds 5b and 7b were established by spectral data.

Experimental

Melting points (capillary, sulfuric acid bath) are uncorrected. TLC was performed on silica gel G plates using petroleum ether/ethyl acetate (4:1) as eluent and iodine as a visualizing agent. IR spectra were recorded on an ABB FTIR spectrometer. ¹H NMR (400 MHz) and ¹³C NMR (100 MHz) spectra were recorded in DMSO-d₆ on a Bruker Advance II 400 NMR spectrometer. Mass spectra were recorded on a TOF MS ES+ 3.56e4 instrument. The elemental analysis of compounds was performed on a Carlo Erba-1108 elemental analyzer. The structures were optimized by molecular mechanics using the PM3 method based on Hyperchem with version 7.5 packages.