English

• Cycloaddition reactions are a universal and straightforward pathway to carbocyclic or heterocarbocyclic organic compounds in which two or more bonds may be formed in a single operation [1]. The photochemical [2 + 2] cycloaddition reaction is the most frequently used cycloaddition reaction to access carbocyclic products with structural moieties of cyclobutane rings [2–12]. Carbocyclic products incorporating fused cyclobutane rings, such as cubane, homocubane, tetraastarane, and pentacycloundecane, continue to attract attention in drug discovery for their higher conformational rigidity and lipophilicity [13–16]. The cyclobutane ring has multiple sp³ scaffolds in the three-dimensional structure, which not only provide a unique chemical space, but are also correlated to improve the pharmacokinetic properties and toxicological benefits [17]. The photochemical [2 + 2] cycloaddition reaction involve two olefins as substrates, one of which is required to be excited by ultraviolet light. There are a wide variety of olefins, such as thymoquinone, α-truxillic acid, enone [18–20]. Typical olefins are the (+)-carvone for carvone camphor, the endo-dimer of 2-bromocyclopentadienone for the key intermediates of the cubane ring system, the Diels-Alder adduct of p-bensoquinone and cyclopentadiene for pentacycloundecanes (Cookson's diketone), and the 1,4-dihydrophthalic anhydride for tetraastarane derivatives, and so on [21]; [22].

  As an important and common structural unit, 1,4-dihydropyridines (1,4-DHPs) are commonly used clinically as calcium antagonists and antimicrobials for the treatment of cardiovascular disease [23–29]. Similar to 1,4-dihydrophthalic anhydride, the two double bonds in 1,4-DHPs enable them to undergo [2 + 2] photocycloaddition to provide diazatetraasteranes [30–32]. Historically, the first [2 + 2] photocycloadditions of 1,4-DHPs were found in the chemical literature, as reported by Eisner in 1970. The [2 + 2] photocyclization of 3,5-dicarbonyl-1,4-DHPs had given 3,9-diazatetraasteranes with a medium-pressure mercury lamp (MPM) as the light source (Scheme 1a) [33], while the [2 + 2] photocyclization of 3-carbonyl-1,4-DHPs had given 3,9-diazatetraasteranes under a high-pressure mercury lamp (Scheme 1b) [31,34-35]. The regioselective [2 + 2] photocycloaddition of 3,5-dicarbonyl-1,4-dihydropyridines had given 3,6-diazatetraasterane via a covalent link in a head-to-head manner by the phthaloyl group (Scheme 1c) [36]. The 3,9-diazatetraasteranes, so as it is the structure analogue of 3,9-dioxotetraasteranes, shows good pharmacological activity, with HIV-1 protease inhibitory activity for C₂ symmetry, and anti-tumor activity by inhibiting epidermal growth factor receptor (EGFR) and so on [37–40]. Therefore, the study of the synthesis and pharmacological activity of 3,9-diazatetraasteranes and their analogs have attracted extensive attention in recent years [41–44].

  Scheme 1

  

  Scheme 1.  [2 + 2] Photocycloadditions of 1,4-DHPs (1).

  DownLoad: Full-Size Img PowerPoint

  To continue research into the [2 + 2] photocycloaddition of 1,4-DHPs, UV LED lamps were used to replace the traditional mercury lamps in order to improve the synthetic efficiency of 3,9-diazatetraasteranes with ethyl 3-carboxylate-1,4-DHPs (1) as olefins. When using 1 as the substrate, the reaction products were found to differ clearly with solvent, regardless of the wavelength of the LED lamp. The 3,9-diazatetraasteranes (P₁) were obtained in tetrahydrofuran, while P₁ and unexpected 3,9-diazatetracyclododecanes (P₂) were observed in methanol. (Scheme 1). In order to elucidate this phenomenon, density-functional theory (DFT) and time-dependent DFT (TDDFT) theoretical calculations were performed.

  Initially, our idea was to improve the synthetic efficiency of 3,9-diazatetraasteranes (P₁) by using UV LED lamps instead of traditional mercury lamps. To optimize the reaction conditions, the factors such as the light source, irradiation wavelength, solvent, atmosphere and substituent of 1,4-DHPs (1), were systematically screened. The photocycloaddition of 1a was selected as the model compound (Table 1). It was worth mentioning that there was more than one product, with both P₁ and unexpected P₂ produced in MeOH solvent. The yields of the products clearly differed with solvent, regardless of wavelength. The synthetic efficiency of 3,9-diazatetraasterane (P₁) was significantly improved, and the reaction time decreased from 8 weeks, as reported in the literature, to about 20 h.

  Table 1

  

  Table 1.  Yields of P_1a and P_2a under different reaction conditions.

  DownLoad: CSV

  As shown in Table 1, whether it was the formation of P_1a or P_2a, the yields under an N₂ atmosphere were higher than those without it. The reason was that oxygen in the air could generate the excited state singlet oxygen (¹O₂) to reduce the yield by combining with carbon radicals. The irradiation wavelengths did not affect the type of product, but only affected the yield of product.

  Under irradiation at 365 nm LED lamp under an N₂ atmosphere, P_1a was obtained in 56% yield in THF, while P_2a was obtained in 52% yield with 12% of P_1a in MeOH (Table 1, entries 1 and 5). The generalizability of the different substituted 1,4-diaryl-3-carboxylate-1,4-DHPs (1) was investigated (Scheme 2). The yields of P₁ were 53%–57%, and those of P₂ were about 50%–53%. The substituents on the 1,4-DHPs had little effect on the yields of P₁ and P₂. The structures of all synthesized P₁ and P₂ were confirmed by NMR and HRMS, especially the structures of P₁ & P₂ were confirmed by single crystal X-ray diffraction (Scheme 2, Fig. 1).

  Scheme 2

  

  Scheme 2.  Structures and yields of 3,9-diazatetraasteranes (P₁) and 3,9-diazatetracyclododecanes (P₂).

  DownLoad: Full-Size Img PowerPoint

  Figure 1

  

  Figure 1.  Structures of P_1d and P_2a obtained by X-ray single-crystal diffraction.

  DownLoad: Full-Size Img PowerPoint

  Based on the [2 + 2] cycloaddition of α, β-enone and alkene derivatives and the mechanism of the 3,5-disubtitued-1,4-DHPs [45–48], the possible mechanisms for the formation of P₁ and P₂ were proposed, as shown in Scheme 3. Considering the substituents had little effect on this reaction, the mechanism was discussed using 1a as a model compound. The formation mechanism of P_1a was the [2 + 2] photocycloaddition of 1,4-DHP (1), which may presumably be divided into two steps. The first step was the intermolecular [2 + 2] photocycloaddition. The second step was the intramolecular [2 + 2] photocycloaddition of the syn-dimer 2a to form P_1a (Path 1a). The formation mechanism of P_2a was the [2 + 2] photocycloaddition of 1,4-DHP (1), similar that of P_1a, containing an intermolecular [2 + 2] and a half intramolecular [2 + 2] photocyclization. The difference was that in the intramolecular [2 + 2] photocyclization of 2a, the first σ-bond was formed at different sites to that of P_1a. The two carbon radicals at the cross position combined with each other, resulting in the inability to form a second σ-bond, but instead the solvent became involved, and the carbon radicals combined with the methoxy radicals provided by the MeOH, to form two C—O σ-bonds in P_a2.

  Scheme 3

  

  Scheme 3.  Proposed formation mechanism of 3,9-diazatetraasteranes (P₁) and 3,9-diazatetracyclododecanes (P₂).

  DownLoad: Full-Size Img PowerPoint

  At the PCM/B3LYP/def2-TZVP//B3LYP/6–31 g(d) level, DFT/ TDDFT calculations were carried out to investigate the favorable [2 + 2] photocycloaddition pathways of the 1,4-DHPs (1). The calculated free energies of the formation pathways in MeOH were discussed, and the reactive sites of 2a (T₁) was qualitatively analyzed using the Fukui function. The formation mechanisms of P₁ and P₂ were the [2 + 2] photocycloaddition of 1,4-DHPs (1), which could be divided into two steps. The first steps were the intermolecular [2 + 2] photocycloaddition to form a syn-dimer 2a. The second steps were the intramolecular [2 + 2] photocycloaddition of the syn-dimer 2a in both cases, with the difference being that P₂ in a half intramolecular [2 + 2] photocyclization involved MeOH (Path 1a and Path 2a).

  According to theoretical research on the [2 + 2] photocycloaddition mechanism of the 1,4-DHPs, the formation of syn-dimer 2a was similar to the first step in previous reports [26]. Therefore, the calculation analysis focused on the second steps in order to elucidate the differences in the formation of P₁ and P₂. In the following discussion, the most stable conformation of 1a was used, and its Gibbs free energy was used as the energy reference point.

  First, the reactive sites of 2a* (syn, T₁) were qualitatively analyzed using the Fukui function, and the isosurface graphs of 2a* (syn, T₁) were displayed in Fig. 2. The isosurface graphs of 2a* (syn, T₁) showed that unpaired electrons were mainly distributed on the C1 and C3 atoms, with a few electrons distributed on the C2 and C4 atoms. The C1 and C3 atoms were more likely to be the sites of photoreaction with the highest reactivity, which was consistent with the experimental results that P_2a formed the C—C single bond at the cross position first.

  Figure 2

  

  Figure 2.  Fukui function f(0) surface atom in 2a* (syn, T₁) (isovalue = 0.002).

  DownLoad: Full-Size Img PowerPoint

  Second, the energy profile of Path 1a and Path 2a in MeOH and their relative Gibbs free energy were calculated (Fig. 3). The energy barrier of the IM2 (T₁) formation through the transition state TS2 (Path 1a) was 53.6 kcal/mol, and that of the IM3 (T₁) formation through the transition state TS3 (Path 2a) was only 16.2 kcal/mol. IM3 (T₁) was more favorable to form than IM2 (T₁). In the conversion of IM2 (T₁) and IM3 (T₁) to IM2 (S₀) and IM3 (S₀), spin-flipping was required, and the minimal energy crossing point (MECP) between T₁/S₀ was successfully localized. The free energies were 1.8 kcal/mol and 2.5 kcal/mol, respectively, so the two spin-flipping processes were readily occurred. In the conversion of IM2 (T₁) to P_1a, the second σ-bond formed between the C2 and C3 radicals (ring-closure) was proven to be barrierless [26]. In the conversion of IM3 (T₁) to P_2a, the form C—O σ-bond involving C2 or C4 with methoxy radicals was speculated to suggest the possible existence of another transition state, but attempted to locate the transition site failed several times. This was finally proven to be a barrierless process by the means of relaxed scanning of the bond distance (C2-O3), wherein energy decreased continuously with the shortening of the bond distance (Fig. 5). Therefore, the rate-determining steps of Path 1a and Path 2a were the first σ-bond formation. The overall energy barrier was 53.6 kcal/mol and 16.2 kcal/mol, respectively. The Gibbs free energy of P_1a and P_2a was −76.6 kcal/mol and −123.4 kcal/mol, respectively. This suggests that Path 2a was the dominant pathway in both thermodynamically and kinetically. This was consistent with the experimental results showing that the yield of P₂ was always much greater than that of P₁ in MeOH.

  Figure 3

  

  Figure 3.  Energy profile of Path 1a and Path 2a and their relative Gibbs free energies (kcal/mol).

  DownLoad: Full-Size Img PowerPoint

  Figure 4

  

  Figure 4.  Optimized structures and significant bond distances of transition states (TS2 and TS3).

  DownLoad: Full-Size Img PowerPoint

  Figure 5

  

  Figure 5.  Relaxed scanning curves for the combination of C₂ and methoxy radical.

  DownLoad: Full-Size Img PowerPoint

  In conclusion, under irradiation at 365 nm LED lamp in an N₂ atmosphere, it was shown that the photocycloaddition reaction of ethyl 1,4-diaryl-1,4-dihydropyridine-3-carboxylate reaction could tolerate different substituents, with yields of P₁ of about 55% in THF, and that yields of P₂ of about 50% in methanol. The formation mechanism of P₁ and P₂ was speculated to occur in two steps; the first step was the intermolecular [2 + 2] photocycloaddition which formed the syn-dimer 2a, the second step was the intramolecular [2 + 2] photocycloaddition of the syn-dimer 2a to form P₁, or a half intramolecular [2 + 2] photocyclization involving MeOH solvent to form P₂. The rate-determining steps in the process of the syn-dimer 2 to P₁ and P₂ (Path 1a and Path 2a) were all the first σ-bond formation in IM2 (T₁) and IM3 (T₁). The calculated values of the energy barrier and the Gibbs free energy also suggested that Path 2a was the dominant pathway both thermodynamically and kinetically, meaning that P_2a was the dominant product, and it was consistent with the experimental results.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgment

  This work was supported by the Beijing Natural Science Foundation (No. 2192004).

  
  
• 1. [1]

     L.R. Domingo, S.R. Emamian, Tetrahedron 70 (2014) 1267–1273. doi: 10.1016/j.tet.2013.12.059

  2. [2]

     S. Poplata, A. Tröster, Y. Zou, T. Bach, Chem. Rev. 116 (2016) 9748–9815. doi: 10.1021/acs.chemrev.5b00723

  3. [3]

     V.M. Dembitsky, J. Nat. Med. 62 (2008) 1–33.

  4. [4]

     S. Piao, Y. Song, W. Jiao, et al., Org. Lett. 15 (2013) 3526–3529. doi: 10.1021/ol400933x

  5. [5]

     R. Medishetty, A. Husain, Z. Bai, et al., Angew. Chem. Int. Ed. 126 (2014) 6017–6021. doi: 10.1002/ange.201402040

  6. [6]

     A.E. Hurtley, Z. Lu, T.P. Yoon, Angew. Chem. Int. Ed. 126 (2014) 9137–9140. doi: 10.1002/ange.201405359

  7. [7]

     C. García-Morales, B. Ranieri, I. Escofet, et al., J. Am. Chem. Soc. 139 (2017) 13628–13631. doi: 10.1021/jacs.7b07651

  8. [8]

     M.A. Ischay, Z. Lu, T.P. Yoon, J. Am. Chem. Soc. 132 (2010) 8572–8574. doi: 10.1021/ja103934y

  9. [9]

     D. Yang, Q. Yan, E. Zhu, J. Lv, W.M. He, Chin. Chem. Lett. 33 (2022) 1798–1816. doi: 10.1016/j.cclet.2021.09.068

  10. [10]

      S. He, X. Chen, F. Zeng, et al., Chin. Chem. Lett. 31 (2020) 1863–1867. doi: 10.1016/j.cclet.2019.12.031

  11. [11]

      W. He, L. Gao, X. Chen, et al., Chin. Chem. Lett. 31 (2020) 1895–1898. doi: 10.1016/j.cclet.2020.02.011

  12. [12]

      Z. Gan, G. Li, X. Yang, et al., Sci. China Chem. 63 (2020) 1652–1658. doi: 10.1007/s11426-020-9811-6

  13. [13]

      A. Sergeiko, V.V. Poroikov, L.O. Hanus, V.M. Dembitsky, Open. Med. Chem. J. 2 (2008) 26–37. doi: 10.2174/1874104500802010026

  14. [14]

      Y. Fan, X. Gao, J. Yue, Sci. China Chem. 59 (2016) 1126–1141. doi: 10.1007/s11426-016-0233-1

  15. [15]

      B. Zhang, J. Chen, Chin. J. Org. Chem. 42 (2022) 3429–3430. doi: 10.6023/cjoc202200058

  16. [16]

      W. Zhou, Y. Jiang, L. Chen, K. Liu, D. Yu, Chin. J. Org. Chem. 40 (2020) 3697–3713. doi: 10.6023/cjoc202004045

  17. [17]

      J.M. Anderson, N.D. Measom, J.A. Murphy, D.L. Poole, Angew. Chem. Int. Ed. 60 (2021) 24754–24769. doi: 10.1002/anie.202106352

  18. [18]

      J. Bertram, R. Kürsten, J. Prakt. Chem. 51 (1895) 316–325. doi: 10.1002/prac.18950510123

  19. [19]

      C.N. Riiber, Ber. Dtsch. Chem. Ges. 35 (1902) 2908–2909. doi: 10.1002/cber.19020350373

  20. [20]

      G. Ciamician, P. Silber, Ber. Dtsch. Chem. Ges. 35 (1902) 1992–2000. doi: 10.1002/cber.190203502148

  21. [21]

      G. Ciamician, P. Silber, Ber. Dtsch. Chem. Ges. 41 (1908) 1071–1080. doi: 10.1002/cber.190804101211

  22. [22]

      G. Büchi, I. Goldman, J. Am. Chem. Soc. 79 (1957) 4741–4748. doi: 10.1021/ja01574a042

  23. [23]

      S. Goldmann, J. Stoltefuss, Angew. Chem. Int. Ed. 30 (1991) 1559–1578. doi: 10.1002/anie.199115591

  24. [24]

      D.J. Triggle, D.A. Langs, R.A. Janis, Med. Res. Rev. 9 (1989) 123–180. doi: 10.1002/med.2610090203

  25. [25]

      F. Bossert, W. Vater, Med. Res. Rev. 9 (1989) 291–324. doi: 10.1002/med.2610090304

  26. [26]

      Q. Fan, H. Tan, P. Li, H. Yan, New J. Chem. 42 (2018) 16795–16805. doi: 10.1039/C8NJ02192A

  27. [27]

      M.S. Saddala, R. Kandimalla, J.A. Pradeepkiran, S.S. Bhashyam, U.R. Asupatri, Sci. Rep. 7 (2017) 45211. doi: 10.1038/srep45211

  28. [28]

      I. Drapak, L. Perekhoda, T. Tsapko, N. Berezniakova, Y. Tsapko, J. Heterocycl. Chem. 54 (2017) 2117–2128. doi: 10.1002/jhet.2837

  29. [29]

      M. Zhang, H. Lan, N. Li, et al., J. Org. Chem. 85 (2020) 8279–8286. doi: 10.1021/acs.joc.0c00963

  30. [30]

      C. Coburger, J. Wollmann, M. Krug, et al., Bioorgan. Med. Chem. 18 (2010) 4983–4990. doi: 10.1016/j.bmc.2010.06.004

  31. [31]

      A. Hilgeroth, M. Wiese, A. Billich, J. Med. Chem. 42 (1999) 4729–4732. doi: 10.1021/jm991115k

  32. [32]

      Y. Liu, H. Tan, H. Yan, X. Song, Chem. Biol. Drug. Des. 82 (2013) 567–578. doi: 10.1111/cbdd.12185

  33. [33]

      U. Eisner, J.R. Williams, B.W. Matthews, H. Ziffer, Tetrahedron 26 (1970) 899–909. doi: 10.1016/S0040-4020(01)97888-3

  34. [34]

      A. Hilgeroth, H. Lilie, Eur. J. Med. Chem. 38 (2003) 495–499. doi: 10.1016/S0223-5234(03)00060-6

  35. [35]

      M. Richter, J. Molnár, A. Hilgeroth, J. Med. Chem. 49 (2006) 2838–2840. doi: 10.1021/jm058046w

  36. [36]

      H. Tan, Z. Zhao, Z. Ma, H. Yan, Tetrahedron 74 (2018) 529–534. doi: 10.1016/j.tet.2017.10.019

  37. [37]

      A. Hilgeroth, U. Baumeister, F.W. Heinemann, Eur. J. Med. Chem. 2000 (2000) 245–249. doi: 10.1002/(SICI)1099-0690(200001)2000:2<245::AID-EJOC245>3.0.CO;2-2

  38. [38]

      A. Hilgeroth, J. Molnár, E. De Clercq, Angew. Chem. Int. Ed. 41 (2002) 3623–3625. doi: 10.1002/1521-3773(20021004)41:19<3623::AID-ANIE3623>3.0.CO;2-V

  39. [39]

      A. Hilgeroth, A. Billich, Arch. Pharm. 332 (1999) 380–384. doi: 10.1002/(SICI)1521-4184(199911)332:11<380::AID-ARDP380>3.0.CO;2-T

  40. [40]

      L. Mao, N. Tian, C. Wei, H. Wang, H. Yan, Russ. J. Gen. Chem. 92 (2022) 446–456. doi: 10.1134/S1070363222030124

  41. [41]

      P. Chen, H. Wang, M. Li, et al., Dis. Markers. 2020 (2020) 5068067.

  42. [42]

      Q. Fan, L. Zhu, H. Ren, H. Lin, G. Wu, Chem. Phys. Lett. 771 (2021) 2207–2213.

  43. [43]

      H. Xin, X. Zhu, H. Yan, X. Song, Tetrahedron Lett. 54 (2013) 3325–3328. doi: 10.1016/j.tetlet.2013.04.016

  44. [44]

      Q. Song, H. Wang, H. Yan, C. Ni, R. Zhong, J. Mol. Struct. 1006 (2011) 489–493. doi: 10.1016/j.molstruc.2011.09.056

  45. [45]

      G.M.J. Schmidt, Pure. Appl. Chem. 27 (1971) 647–678. doi: 10.1351/pac197127040647

  46. [46]

      E. García-Expósito, M.J. Bearpark, R.M. Ortuño, M.A. Robb, V. Branchadell, J. Org. Chem. 67 (2002) 6070–6077. doi: 10.1021/jo026047f

  47. [47]

      P. Jaque, A. Toro-Labbé, P. Geerlings, F.D. Proft, J. Phys. Chem. A 113 (2009) 332–344. doi: 10.1021/jp807754f

  48. [48]

      S. Wilsey, L. González, M.A. Robb, K.N. Houk, J. Am. Chem. Soc. 122 (2000) 5866–5876. doi: 10.1021/ja0006595

• 

  Scheme 1  [2 + 2] Photocycloadditions of 1,4-DHPs (1).

  下载: 全尺寸图片 幻灯片
  

  Scheme 2  Structures and yields of 3,9-diazatetraasteranes (P₁) and 3,9-diazatetracyclododecanes (P₂).

  下载: 全尺寸图片 幻灯片
  

  Figure 1  Structures of P_1d and P_2a obtained by X-ray single-crystal diffraction.

  下载: 全尺寸图片 幻灯片
  

  Scheme 3  Proposed formation mechanism of 3,9-diazatetraasteranes (P₁) and 3,9-diazatetracyclododecanes (P₂).

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Fukui function f(0) surface atom in 2a* (syn, T₁) (isovalue = 0.002).

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Energy profile of Path 1a and Path 2a and their relative Gibbs free energies (kcal/mol).

  下载: 全尺寸图片 幻灯片
  

  Figure 4  Optimized structures and significant bond distances of transition states (TS2 and TS3).

  下载: 全尺寸图片 幻灯片
  

  Figure 5  Relaxed scanning curves for the combination of C₂ and methoxy radical.

  下载: 全尺寸图片 幻灯片

  Table 1.  Yields of P_1a and P_2a under different reaction conditions.

  下载: 导出CSV
•