[3 + 2] Cycloaddition reactions of thioisatin with thiazolidine-2-carboxylic acid: a versatile route to new heterocyclic scaffolds

Abstract

A facile synthesis of azabicycloadducts is described by 1,3-dipolar cycloaddition reactions of thioisatin with thiazolidine-2-carboxylic acid in the presence of various electron rich and electron deficient dipolarophiles. Theoritical calculations have been performed to study the regioselectivity of products. The geometrical and energetic properties have been analyzed for the different reactants, transition states and cycloadducts formed.

Background

The construction of sophisticated molecules requires viable, selective and highly reliable reactions as potent synthetic tools [1–3]. The 1,3-dipolar cycloaddition [4–9] has also become one of the most important legation method in biology and material chemistry. Thioisatin derivatives [10] have received the attention of biochemists because of their therapeutic and biological activities. Similarly thiazolidine-2-carboxylic acid [11, 12] exhibit strong antioxidant properties. Therefore any heterocyclic scaffold containing these two moieties might be expected to have considerably enhanced biological activities.

The 1,3-dipolar cycloaddition reaction of an azomethine ylide with an alkene leads to the formation of pyrrolidine [13, 14] derivatives. Recently, we have reported the results on azomethine ylides derived from 9,10-phenanthrenequinone and some secondary cyclic α-amino acids with different dipolarophiles [15]. Herein we report the reactivity and regioselectivity of 1,3-dipolar cycloaddition reactions of azomethine ylides derived from benzo[b]thiophene-2,3-dione with thiazolidine-2-carboxylic acid in the presence of various acetylenic and ethylenic dipolarophiles. Besides synthetic work, a systematic and comprehensive theoretical study at Gaussian 03 [16] suite of programs has been carried out to address the mechanism as well as regio- and stereochemical course of the reactions.

2. Result and discussion

The reaction of thioisatin 1 with thiazolidine-2-carboxylic 2 acid was carried out in equimolar ratio in refluxing dry acetonitrile in the presence of diphenylacetylene as dipolarophile to afford a diastereomeric mixture of {(5R,8R)-spiro-6,7-diphenyl-1aza-4thia-bicyclo [3,3,0]-6-octene-8,3'}-benzo[b]thiophene-2'-one} 8 as cis/trans isomers. Analogous reactions of thioisatin with other dipolarophiles viz methyl acrylate, phenylacetylene, phenylpropyne and ethyl phenyl propiolate produced diasteroisomeric mixtures of cycloadducts 5-9 in 75%- 63% yield. The mechanism for the formation of the cycloadducts 5-9 involve the initial formation of an iminium species 3 followed by the loss of CO₂ via stereospecific 1,3-cycloreversion [17] to azomethine ylides 4. Subsequent [3 + 2] cycloaddition with various dipolarophiles then produce novel azabicycloadducts (Scheme 1).

Scheme 1

Reaction of thioisatin with thiazolidine-2-carboxylic acid in presence of different dipolarophiles produced regioisomeric mixtures of cycloadducts.

The structure of all the cycloadducts has been ascertained from their spectral data. Thus the IR spectrum of a typical diphenylacetylene cycloadducts 8 showed characteristic bands at 1715, 1420 and 690 for > C = O, C-N and C-S stretching vibrations respectively. Its ¹H NMR spectrum showed a triplet at δ 2.51 (J = 2.7Hz) for 3H protons, another triplet at δ 2.60 (J = 3.0Hz) was associated with 2H protons, a singlet at δ 4.15 appeared for 5H and the aromatic protons resonated between δ 6.79- δ 8.12 ppm. Its ¹³C NMR spectrum showed a signal at 183.54 for C-2' carbonyl carbon, aromatic carbons appeared in the range δ 146.41-δ131.32 ppm, the olefenic carbons (C-6, C-7) resonated at δ 127.32 and 126.54, spiro carbon (C-8) was noticed at δ 86.23, C-5 at δ 46.72, C-2 at δ 45.43, and C-3 at 34.56 ppm respectively. Additional evidence was gathered from the mass spectrum of cycloadduct 8. The molecular ion and the base peaks were present at m/z 413(32%) and 235(100%) respectively; another peak at m/z 385(39%) corresponded to [M-CO]^.+ whereas the peak at m/z 108(35%) was assigned to [C₆H₄S]^.+.

2.1 Theoretical calculations: Regioselectivity of cycloadducts 5-9

The stereochemical course of the cycloaddition was examined by AM1 calculations. To calculate the relevant activation and stabilization energies, minimized geometries of the reactants, products and transition states are required. The molecular geometry of the simplest azomethine ylide (abbreviated as am y) 4 derived from thioisatin and thiazolidine-2-carboxylic acid has been optimized on Gaussian 03 program at AM1 level.

Geometry optimization showed that amy 4 has almost planar structure (Figure 1). Instead of having an envelope shape the thiazolidine ring is planar and lies in the same plane as that of thioisatin ring. It may exist in two isomeric forms, one in which > C = O group and C-H of the dipole are syn to each other, 4 _syn , and the other in which these two groups are anti, 4 _anti (Figure 2).

Figure 1

Optimized geometry of amy 4 _anti.

Figure 2

Mode of attack of dipolarophile (diphenylacetylene) on amy 4.

Methyl acrylate may approach either of the amy with the formation of products having three chiral centers. Therefore a total 8 + 8 = 16 isomers 5a-5p are possible (Figure 3).

Figure 3

Possible isomers of cycloadduct 5 and its regioselectivity.

Attack of methyl acrylate on syn amy results in the inward movement of thiazolidine ring towards thioisatin nucleus which imposes steric hindrance and makes it unstable. In fact we failed to locate the transition state in any such case 5i-5p. The remaining 8 isomers may be obtained by the attack of methyl acrylate on anti amy. Out of these 8 possibilities only four 5a-5d have concerted mechanism. We have optimized the geometries of all the four isomers. Results show that all isomers have almost same ΔH _f, indicating that thermodynamically all are nearly equally stable.

We have carried out transition state calculations on all the four isomers but have been successful in locating the transition state for only two isomers 5a-5b (Figure 4).

Figure 4

Transition states of cycloadducts 5a and 5b.

The transition state of the concerted 1,3-dipolar cycloadditions is usually controlled by frontier molecular orbitals of dipolarophiles and dipole (azomethine ylide). The favoured path involves HOMO_dipole and LUMO_{dipolarophile}. The ΔH _f , HOMO, LUMO energies and HOMO-LUMO energy gaps of azomethine ylides 4 with dipolarphiles are given in Table 1.

Table 1 ΔH_f, HOMO, LUMO, energies and H-L And L-H energy gaps

From the Table, it may be concluded that HOMO_dipole-LUMO_{dipolarophile} energy gap is lower than the LUMO_dipole-HOMO_{dipolarophile} gap and therefore the dominant FMO approach is HOMO_dipole-LUMO_{dipolarophile}. Both the HOMO and the LUMO of the dipole show uneven distribution of the electron density along the C-N-C dipole. In the HOMO case, the orbital coefficient is larger at C₁ (0.288) than at C₂ (-0.163). Similarly in the LUMO of methyl acrylate the atomic orbital coefficient are (0.611) and (-0.521) respectively (Figure 3). Thus it may be concluded that there is better overlap when -COOMe group lies towards thioisatin ring, giving two possibilities 5a and 5b. Out of these 5a is formed in diastereomeric excess probably due to the endo approach of -COOMe.

2.2 Regioselectivity of the addition of symmetrical and unsymmetrical acetylenes

Parallel to methyl acrylate, diphenylacetylene can also attack either of the azomethine ylide (4 _syn or 4 _anti ) with the formation of products having two chiral centers. Therefore a total of 4 + 4 = 8 isomers (4 pairs of enantiomers) could be possible 8a-8h (Figure 5).

Figure 5

Possible isomers of cycloadduct 8.

Attack of diphenylacetylene on syn-azomethine ylide 4 _syn results in the inward movement of the thiazolidine ring towards the thioisatin nucleus and the transition state could not be located even in a single case (Figure 2) ruling out the possibility of the formation of products 8e-8h. Thus it leaves the possibility of attack on the anti azomethine ylide 4 _anti and hence only four isomers 8a-8d are left for consideration. We have optimized the geometry of all the four isomers. Results show that all isomers have almost same ΔH _f, indicating that thermodynamically all are nearly equally stable.

Of remaining four possibilities 8c and 8d where N and H atoms on the adjacent carbon atoms do not lie on the same side, the transition state could not be located because concerted mechanism is not possible in such a situation. This leaves only two isomers 8a-8b for consideration. Out of these two isomers we could optimize the transition state in case of 8a only (Figure 6). This can be explained using the FMO approach along with the endo approach of the phenyl ring (Figure 7) as discussed above.

Figure 6

Transition state of diphenylacetylene cycloadduct 8.

Figure 7

Atomic orbital coefficients and overlapping of dipole( amy ) with diphenylacetylene.

Similarly attack of unsymmetrical dipolarophile, such as ethyl phenyl propiolate, on syn or anti azomethine ylide may produce a cycloadduct having two chiral centres and therefore a total of 4 + 4 = 8 stereoisomers 9a-9h could be possible (Figure 8) and it was concluded that isomer 9a is formed preferentially (Figure 9). The energy profile diagrams for azabicycloadducts 5-9 are presented in (Figure 10).

Figure 8

Possible isomers of cycloadduct 9.

Figure 9

Atomic orbital coefficients and overlapping of amy with ethylphenyl propiolate.

Figure 10

Energy profile diagrams of azabicycloadducts 5-9 (all values in Kcal/mol).

ΔH _{f -R}, ΔH _{f -Ts}, ΔH _{f -P}, Ea(activation energy) and stabilization energy of amy with different dipolarophiles have been tabulated in Table 2.

Table 2 ΔH_f-R, ΔH_f-Ts, ΔH_f-P, Ea. and stabilization energy of amy with different dipolarophiles

3. Conclusions

From the above discussions, it may be concluded that:

• a. The azomethine ylide exits in two conformations; 4 _syn and 4 _anti .

• b. The dominant FMO approach is HOMO_dipole-LUMO_{dipolarophile} as this energy gap is lower than the LUMO_dipole-HOMO_{dipolarophile} gap.

• c. Azomethine ylide is stabilized by the delocalization of dipolar charge into thioisatin nucleus, thus increasing the activation energy for the reaction path.

4. Experimental

The uncorrected melting points were taken in open glass capillaries. The IR spectra were recorded on a Nicolet Magna IR Spectrometer Model 550 in KBr pellets and band positions are reported in wave numbers (cm^-1). The ¹H NMR spectra and ¹³C NMR spectra have been recorded on a Bruker DRX-300 MHz and 75.47 MHz model respectively in CDCl₃ and DMSO using tetramethylsilane as an internal standard. The chemical shifts are given in δ ppm values. The mass spectra were recorded on a JEOL-SX 102 (FAB). Most of the spectra were recorded at CDRI, Lucknow, India. Elemental analyses were performed on a Perkin Elmer Series C, H, N, S Analyzer 2400. The solvents were purified by standard procedures [18, 19]. Acetonitrile was dried by refluxing with anhydrous calcium chloride for 5-6 h and then distilling it. Column chromatography was performed on silica gel 60 (Merck).

Methods

Synthesis of (5S,7R,8R)-spiro-{7-methoxycarbonyl-1-aza-4-thia-bicyclo [3, 3, 0]-octane 8, 3'}-benzo[b]thiophene-2'-one (5)

An equimolar mixture of thiosatin 1 (0.36 gm, 2.0 mmol), thiazolidine-2-carboxylic acid 2 (0.26 gm, 2.0 mmol) and methyl acrylate (0.32 gm, 2.0 mmol) in dry acetonitrile (50 ml) was refluxed for 22 h. The reaction was monitored by TLC until the consumption of the reactants. The reaction mixture was filtered, solvent evaporated and the residue was subjected to column chromatography. The petroleum ether/chloroform (4:1) fraction afforded the desired azabicycloadduct 5 as pale brown solid.

Pale brown solid, yield: 0.33g (70%), mp: 105-107°C. IR (KBr): 1710(C = O),1450(C-N), 715(C-S) cm^-1. ¹H NMR (CDCl3, δ ppm): = 2.29 (1H, t, J = 3.0Hz 7-CH), 2.46(2H, dd, J ₁ = 4.2Hz, J ₂ = 3.3Hz 6-CH₂), 2.50(2H, t, J = 1.2Hz 3-CH₂), 2.68(2H, t, J = 2.1Hz 2-CH₂), 3.10(3H, s, OCH₃), 4.15(1H, t, J = 3.54 Hz, 5-CH), 7.36-7.54(4H, m, ArH). ¹³C NMR (CDCl3, δ ppm): 35.43(CH₂S), 44.58(CH₂), 46.34(CH), 85.99(C-N), 125.86-124.32(C = C), 142.58-131.36(CHaro), 175.67(O = C-O), 184.32(C = O). EI-MS: m/z (%) = 321(M+ 38), 275(100), 261(20), 284(15), 234(18). Anal. Calcd. For C₁₅H₁₅NO₃S₂: C, 56.05%; H, 4.70%; N, 4.36%. Found: C, 56.45%; H, 4.78%; N, 4.76%.

(5R,8R)-spiro-{7-phenyl-1-aza-4-thia-bicyclo [3,3,0]-6-octene-8,3'}-benzo[b] thiophene-2'-one (6)

Coffee brown powder, yield: 0.31g (65%), mp: 95-97°C. IR (KBr): 1720(C = O), 1445(C-N), 690(C-S) cm^-1. ¹H NMR (CDCl3, δ ppm): = 2.50(2H, t, J = 3.0 Hz, 3-CH₂), 2.68(2H, t, J = 3.3Hz, 2-CH₂), 4.12(1H, d, J = 2.4Hz, 5-CH), 4.32(1H, d, J = 6.6Hz, 6-CH), 7.32-7.54(9H, m, ArH). ¹³C NMR (CDCl3, δ ppm): 31.51(CH₂S), 36.54(CH₂), 45.46(CH), 85.95(C-N), 128.81-124.49(C = C), 143.87-132.36(CHaro), 180.79(C = O). EI-MS: m/z (%) = 337(M+ 36), 203(100), 309(20), 108(34). Anal.Calcd for C₁₉H₁₅NOS₂: C, 67.62%; H, 4.48%; N, 4.15%. Found: C, 67.78%; H, 4.53%; N, 4.32%.

(5R,8R)-spiro-{7-phenyl-6-methyl-1-aza-4-thia-bicyclo [3,3,0]-6-octene-8,3'}- benzo[b]thiophene-2'-one (7)

Brownish solid, yield: 0.32g (68%), mp: 120-122°C. IR (KBr): 1725(C = O), 1420(C-N), 678(C-S) cm^-1.¹H NMR (CDCl3, δ ppm): 2.14(3H, d, J = 7.2 Hz, Me), 2.50(2H, t,J = 3.0Hz 3-CH₂), 2.68(2H, t,J = 3.2Hz 2-CH₂), 2.84(1H, q, J = 3.6Hz,5-CH), 7.59-8.45(9H, m, ArH). ¹³C NMR(CDCl3, δ ppm): 32.33(CH₂S), 35.43(CH₂), 44.45(CH), 86.54(C-N), 124.23-122.86(C = C), 143.41-130.63(CHaro), 184.34(C = O). Anal.Calcd for C₂₀H₁₇NOS₂: C, 68.34%; H, 4.87%; N, 3.98%. Found: C, 68.55%; H, 4.93%; N, 4.13%.

(5R,8R)-spiro-{6,7-diphenyl-1-aza-4-thia-bicyclo [3,3,0]-6-octene 8, 3'}- benzo[b]thiophene-2'-one (8)

Shiny brown powder, yield: 0.29g (63%), mp: 135-133°C. IR (KBr): 1715(C = O), 1420(C-N), 6905(C-S) cm^-1. ¹H NMR (CDCl3, δ ppm): 2.51(2H, t,J = 2.7Hz, 3-CH₂), 2.60(2H, t, J = 3.0Hz, 2-CH₂), 4.15(1H, s, 5-CH) 6.79-8.12(14H, m, ArH). ¹³C NMR (CDCl3, δ ppm): 34.54(CH₂S), 45.43(CH₂), 46.72(CH), 88.23(C-N), 127.32-126.54(C = C), 146.41-132.32(Caro), 183.54(C = O). EI-MS: m/z (%) = 413(M⁺ 32), 385(39), 235(100), 108(35). Anal. Calcd for C₂₅H₁₉NOS₂: C, 72.61%; H, 4.63%; N, 3.39%. Found: C, 72.73%; H, 4.78%; N, 3.86%.

(5R,8R)-spiro-{6-ethoxycarbonyl-7-phenyl-1-aza-4-thia-bicyclo [3,3,0]-6-octene-8, 3'}-benzo[b]thiophene-2'-one (9)

Dark brown power, yield: 0.31g (69%), mp: 110-112°C. IR (KBr): 1710(C = O), 1410(C-N), 690(C-S) cm^-1. ¹H NMR (CDCl3, δ ppm): 1.25(3H, t, J = 2.4Hz Me), 2.61(2H, t, J = 2.9Hz 3-CH₂), 2.69(2H, t, J = 3.1Hz, 2-CH₂), 2.67(2H, t, CH₂), 3.75(2H, q, J = 6.3Hz, OCH₂),4.12 (1H. s,5-CH) 7.49-7.94(4H, m, ArH). ¹³C NMR (CDCl₃, δ ppm): 33.85(CH₂), 44.96(CH), 46.32(CH₂), 84.36(C-N), 124.32-122.93(C = C), 140.14-130.12(CHaro), 180.25(O = C-O), 185.44(C = O). EI-MS: m/z (%) = 321(M+ 38), 275(100), 261(20), 284(15), 234(18). Anal. Calcd for C₂₂H₁₉NO₃S₂: C, 64.52%; H, 4.68%; N, 3.42%. Found: C, 64.68%; H, 4.79%; N, 3.54%.