Subscribe to RSS
DOI: 10.1055/s-0041-1738454
Regioselective Ring Opening of Oxetanes Enabled by Zirconocene and Photoredox Catalysis
This work was supported by JSPS KAKENHI Grants Nos. JP21H05213 (Digi-TOS) (to J.Y.) and JP20K15290 (to E.O.), the Sumitomo Foundation (to E.O.), the Satomi Foundation (to E.O.), and Daiichi Kigenso Kagaku Kogyo (to E.O.). This work was also partly supported by JST ERATO Grant No. JPMJER1901 (to J.Y.).
Abstract
Oxetanes are frequently utilized in organic synthesis, both as target products and as fairly reactive intermediates. Whereas ring cleavage of oxetanes through polar mechanisms has been extensively investigated, their radical-based counterparts remain underexplored. We used zirconocene and photoredox catalysis to open an oxetane ring in a radical manner. In our protocol, the reaction selectively delivers the more-substituted alcohols via putative less-stable radicals. This method not only affords the corresponding hydrogenated products, but also provides unique benzylidene acetal products.
#
Key words
zirconocene - photoredox catalysis - oxetanes - reverse regioselectivity - C–O bond homolysis - radicalsOxetanes are intriguing yet fairly frequently encountered structural motifs in various natural products and biologically active compounds (Figure [1]A).[1] In medicinal chemistry, the small size and polar nature of oxetanes can endow improved ‘drug-like’ physicochemical properties to a molecule, for example, when replacing a lipophilic gem-dimethyl group. A variety of protocols to construct oxetanes,[2] [3] or to introduce substituents onto an oxetane ring[4] have been described.


Moreover, oxetanes can also be used as building blocks in organic synthesis. Oxetanes exhibit a high ring-strain energy (25.2 kcal/mol),[5] permitting favorable ring opening in a similar manner to oxiranes (26.8 kcal/mol). Accordingly, numerous protocols for opening oxetane rings have been explored. Generally, C–O bond cleavage of oxetanes occurs by polar mechanisms using nucleophiles, where the oxetane acts as an electrophile.[6] [7] In contrast, homolytic cleavage of these C–O bonds would permit the use of oxetanes as nucleophilic carbon radicals, leading to different products to those available through a polar mechanism. This mode of ring opening by a radical mechanism is less well explored, whereas functionalization of oxetanes through a polar mechanism has been extensively reported (Figure [1]B).[1] In 2008, Gansäuer and co-workers reported that titanocene(III) can cleave the C–O bond by a single electron transfer.[8] However, this reaction only affords a relatively stable tertiary alkyl radical or benzyl radical. The Okamoto group reported a ring opening of oxetanes by using a titanatrane complex with Mg as a stoichiometric reductant.[9] This reaction permits the opening of other substituted types of oxetanes. Later, the same research group developed a ring-opening functionalization of oxetanes using an iron catalyst.[10] The iron catalyst causes a C–O bond homolysis, affording a carbon radical that is then trapped by the iron catalyst. The ionic intermediate formed by transmetalation of the iron complex with an organomagnesium reagent can be converted into alcohols by treatment with a variety of electrophiles. In addition, Gryko and co-workers recently reported a Giese addition and a Ni-catalyzed cross-coupling reaction combined with the ring opening of oxetanes employing a vitamin B12 catalyst.[11] In this reaction, TMSBr primarily converted an oxetane into a bromo alcohol that underwent subsequent functionalization by cobalt catalysis.
Recently, our group has developed a ring-opening reaction of epoxides by using zirconocene and photoredox catalysis (Figure [1]C).[12] This reaction showcases the power of zirconocene-mediated photocatalysis to afford a less-stable alkyl radical in a regioselective manner. In this reaction, the formation of a strong Zr–O bond might lead to an early transition state for the ring opening, resulting in a unique regioselectivity. Building on these results, we hypothesized that a reverse regioselectivity might also be exhibited in the ring opening of oxetanes, due to the similarity of their ring-strain energy to that of oxiranes. Here, we report a ring opening of oxetanes to afford the less-stable radical (Figure [1]D). This catalytic method not only furnishes hydrogenated products, but also leads to the formation of the corresponding acetals.
Our experimental investigations began with the reaction conditions from our previous work[12] (Table [1]). Treatment of oxetane 1A with Ir(4-MeOppy)3 (4-MeOppy = 2-(4-methoxyphenyl)pyridine, Cp2Zr(OTf)2·THF, N-methyl-N′-phenylthiourea (TU1), and cyclohexa-1,4-diene (CHD) in PhCF3 gave the secondary alcohol 2A and the primary alcohol 3A in yields of 9 and 2%, respectively (Table [1], entry 1). The reaction efficiency was strongly influenced by solvent, and THF gave 2A in 78% yield (entries 2 and 3). Cp2Zr(OTf)2·THF gave the best results, whereas a change to Cp2ZrCl2 or Cp2Zr(OTs)2 decreased the yields of 2A and 3A, which were obtained in a ~1:1 ratio (entries 4 and 5). The substituents on the thiourea additive also affected the product ratio. N,N′-Dimethylthiourea (TU2) provided the alcohol products in a good yield but with slightly diminished regioselectivity (entry 6), whereas the use of N,N′-diphenylthiourea (TU3) resulted in decreased product formation (entry 7). A higher concentration (0.10 M) resulted in consumption of the remaining starting materials, and afforded 2a in a better yield (entry 8). Control experiments conducted without visible light, zirconocene, or thiourea resulted in significant reductions in the product yield (entries 9–11). In entry 11, a significant amount of a polyether product formed, presumably through polymerization of the THF solvent due to the strong Lewis acidity of the zirconocene.[13] The thiourea might be beneficial in reducing the Lewis acidity of the zirconocene catalyst and improving the regioselectivity through complexation.[12] These results confirmed that irradiation, zirconocene, and thiourea are all required to achieve the observed reactivity.
a NMR yield.
b 0.10 M.
c No irradiation.
With these optimized conditions in hand, we evaluated the generality of the reaction scope (Scheme [1]). Benzoate (1A), benzyl ether (1B), chloroalkyl ether (1C), and silyl ether (1D) groups were all tolerated, giving the corresponding secondary alcohols in moderate to good yields. Surprisingly, the chloroalkane remained intact, whereas a dechlorinative hydrogenation occurs in a similar zirconocene-catalyzed reaction.[14] A 2,2-disubstituted oxetane containing an amide moiety (1E) also afforded the more-substituted alcohol product in a moderate yield. Furthermore, natural-product-derived oxetanes from xylose (1F) and estrone (1G) produced the desired alcohols. In the reaction of 1F, a secondary alcohol was obtained with complete regioselectivity (>19:1) in contrast to the Ti-catalyzed method.[9] These results confirmed that this zirconocene catalysis is complementary to ring opening using titanium.


Next, we performed acetal formation with 3,3-disubstituted oxetanes bearing a benzylic ether (Scheme [2]). This transformation involves a C–O bond cleavage, followed by a 1,5-hydrogen atom transfer (HAT).[12] [15] Presumably, the oxetane initially coordinates to Zr(III) and C–O bond homolysis then occurs. The resulting carbon radical abstracts a hydrogen atom at the benzylic position. Finally, C–O bond formation affords the corresponding acetals and regenerates ZrIII. In this acetal formation, benzyl (1H), p-methoxybenzyl (1I), and p-(trifluoromethyl)benzyl (1J) ethers were successfully transformed into the corresponding benzylidene acetals along with small amounts of the corresponding hydrogenated products. This acetal formation proceeded smoothly regardless of the electron density on the aryl group. In contrast to the previous titanium catalysis, which forms the corresponding hydrogenated alcohols as the major products from these benzylic ether substrates,[9] the current catalytic system permits the formation of the acetals.


In conclusion, we have successfully developed a ring opening of oxetanes to form hydrogenated products or acetals.[16] This catalytic protocol involves the synergistic merger of zirconocene and photoredox catalysts, permitting a reversal in the regioselectivity of oxetane ring opening compared with titanocene catalysis. Further mechanistic investigations and explorations of other transformations with zirconocene and photoredox catalysis are currently underway in our laboratory.
#
Conflict of Interest
The authors declare no conflict of interest.
Acknowledgment
The Materials Characterization Central Laboratory of Waseda University is acknowledged for their support of the NMR and HRMS measurement.
Supporting Information
- Supporting information for this article is available online at https://doi-org.accesdistant.sorbonne-universite.fr/10.1055/s-0041-1738454.
- Supporting Information
-
References and Notes
- 1 Bull JA, Croft RA, Davis OA, Doran R, Morgan KF. Chem. Rev. 2016; 116: 12150
- 2a Flores DM, Schmidt VA. J. Am. Chem. Soc. 2019; 141: 8741
- 2b Rykaczewski KA, Schindler CS. Org. Lett. 2020; 22: 6516
- 2c Zheng J, Dong X, Yoon TP. Org. Lett. 2020; 22: 6520
- 2d Osato A, Fujihara T, Shigehisa H. ACS Catal. 2023; 13: 4101
- 3 For another recent example of the construction of an oxetane ring, see: Qi D, Bai J, Zhang H, Li B, Song Z, Ma N, Guo L, Song L, Xia W. Green Chem. 2022; 24: 5046
- 4a Ravelli D, Zoccolillo M, Mella M, Fagnoni M. Adv. Synth. Catal. 2014; 356: 2781
- 4b Jin J, MacMillan DW. C. Angew. Chem. Int. Ed. 2015; 54: 1565
- 5 Eigenmann HK, Golden DM, Benson SW. J. Phys. Chem. 1973; 77: 1687
- 6 Ahmad S, Yousaf M, Mansha A, Rasool N, Zahoor AF, Hafeez F, Rizvi SM. A. Synth. Commun. 2016; 46: 1397
- 7 For a recent enantioselective protocol, see: Strassfeld DA, Wickens ZK, Picazo E, Jacobsen EN. J. Am. Chem. Soc. 2020; 142: 9175
- 8 Gansäuer A, Ndene N, Lauterbach T, Justicia J, Winkler I, Mück-Lichtenfeld C, Grimme S. Tetrahedron 2008; 64: 11839
- 9 Takekoshi N, Miyashita K, Shoji N, Okamoto S. Adv. Synth. Catal. 2013; 355: 2151
- 10 Sugiyama Y, Heigozono S, Okamoto S. Org. Lett. 2014; 16: 6278
- 11 Potrząsaj A, Ociepa M, Chaładaj W, Gryko D. Org. Lett. 2022; 24: 2469
- 12 Aida K, Hirao M, Funabashi A, Sugimura N, Ota E, Yamaguchi J. Chem 2022; 8: 1762
- 13 Seto R, Yamada S, Matsumoto K, Endo T. Polym. Bull. (Heidelberg, Ger.) 2019; 76: 3355
- 14 Okita T, Aida K, Tanaka K, Ota E, Yamaguchi J. Precis. Chem. 2023; 1: 112
- 15 Funk P, Richrath RB, Bohle F, Grimme S, Gansäuer A. Angew. Chem. Int. Ed. 2021; 60: 5482
- 16 2-Hydroxybutyl Benzoate (2A); Typical Procedure 1,4-CHD (58.6 μL, 0.60 mmol, 3.0 equiv) and oxetane 1A (0.20 mmol, 1.0 equiv) were added to a solution of Ir(4-MeOppy)3 (4.5 mg, 6.0 μmol, 3.0 mol%), Cp2Zr(OTf)2·THF (5.9 mg, 10 μmol, 5.0 mol%), and TU1 (20.0 mg, 120 μmol, 60 mol%) in THF (2.0 mL), and the resulting mixture was irradiated with a 456 nm LED (Kessil) for 12 h. The mixture was then passed through a short silica gel pad with EtOAc as an eluent, and the filtrate was concentrated in vacuo. The residue was purified by preparative TLC (hexane/EtOAc = 2:1, then CHCl3/acetone = 9:1) to afford a colorless oil [Run 1; yield: 32.0 mg (84%), Run 2; yield: 24.4 mg (65%)]. 1H NMR (400 MHz, CDCl3): δ = 8.08–8.04 (m, 2 H), 7.60–7.55 (m, 1 H), 7.48–7.43 (m, 2 H), 4.41 (dd, J = 11.6, 3.2 Hz, 1 H), 4.25 (dd, J = 11.6, 6.4 Hz, 1 H), 3.96–3.89 (m, 1 H), 2.08 (d, J = 4.8 Hz, 1 H), 1.68–1.56 (m, 2 H), 1.04 (t, J = 7.2 Hz, 3 H). 13C NMR (101 MHz, CDCl3): δ = 166.7, 133.1, 129.9, 129.6, 128.4, 71.4, 68.8, 26.4, 9.8. The spectra were in accordance with those reported in the literature.17
- 17 Iwasaki F, Maki T, Onomura O, Nakashima W, Matsumura Y. J. Org. Chem. 2000; 65: 996
For recent [2+2]-cycloaddition methods and MHAT examples, see:
For recent C–H functionalizations, see:
Corresponding Authors
Publication History
Received: 01 July 2023
Accepted after revision: 14 August 2023
Article published online:
27 October 2023
© 2023. Thieme. All rights reserved
Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany
-
References and Notes
- 1 Bull JA, Croft RA, Davis OA, Doran R, Morgan KF. Chem. Rev. 2016; 116: 12150
- 2a Flores DM, Schmidt VA. J. Am. Chem. Soc. 2019; 141: 8741
- 2b Rykaczewski KA, Schindler CS. Org. Lett. 2020; 22: 6516
- 2c Zheng J, Dong X, Yoon TP. Org. Lett. 2020; 22: 6520
- 2d Osato A, Fujihara T, Shigehisa H. ACS Catal. 2023; 13: 4101
- 3 For another recent example of the construction of an oxetane ring, see: Qi D, Bai J, Zhang H, Li B, Song Z, Ma N, Guo L, Song L, Xia W. Green Chem. 2022; 24: 5046
- 4a Ravelli D, Zoccolillo M, Mella M, Fagnoni M. Adv. Synth. Catal. 2014; 356: 2781
- 4b Jin J, MacMillan DW. C. Angew. Chem. Int. Ed. 2015; 54: 1565
- 5 Eigenmann HK, Golden DM, Benson SW. J. Phys. Chem. 1973; 77: 1687
- 6 Ahmad S, Yousaf M, Mansha A, Rasool N, Zahoor AF, Hafeez F, Rizvi SM. A. Synth. Commun. 2016; 46: 1397
- 7 For a recent enantioselective protocol, see: Strassfeld DA, Wickens ZK, Picazo E, Jacobsen EN. J. Am. Chem. Soc. 2020; 142: 9175
- 8 Gansäuer A, Ndene N, Lauterbach T, Justicia J, Winkler I, Mück-Lichtenfeld C, Grimme S. Tetrahedron 2008; 64: 11839
- 9 Takekoshi N, Miyashita K, Shoji N, Okamoto S. Adv. Synth. Catal. 2013; 355: 2151
- 10 Sugiyama Y, Heigozono S, Okamoto S. Org. Lett. 2014; 16: 6278
- 11 Potrząsaj A, Ociepa M, Chaładaj W, Gryko D. Org. Lett. 2022; 24: 2469
- 12 Aida K, Hirao M, Funabashi A, Sugimura N, Ota E, Yamaguchi J. Chem 2022; 8: 1762
- 13 Seto R, Yamada S, Matsumoto K, Endo T. Polym. Bull. (Heidelberg, Ger.) 2019; 76: 3355
- 14 Okita T, Aida K, Tanaka K, Ota E, Yamaguchi J. Precis. Chem. 2023; 1: 112
- 15 Funk P, Richrath RB, Bohle F, Grimme S, Gansäuer A. Angew. Chem. Int. Ed. 2021; 60: 5482
- 16 2-Hydroxybutyl Benzoate (2A); Typical Procedure 1,4-CHD (58.6 μL, 0.60 mmol, 3.0 equiv) and oxetane 1A (0.20 mmol, 1.0 equiv) were added to a solution of Ir(4-MeOppy)3 (4.5 mg, 6.0 μmol, 3.0 mol%), Cp2Zr(OTf)2·THF (5.9 mg, 10 μmol, 5.0 mol%), and TU1 (20.0 mg, 120 μmol, 60 mol%) in THF (2.0 mL), and the resulting mixture was irradiated with a 456 nm LED (Kessil) for 12 h. The mixture was then passed through a short silica gel pad with EtOAc as an eluent, and the filtrate was concentrated in vacuo. The residue was purified by preparative TLC (hexane/EtOAc = 2:1, then CHCl3/acetone = 9:1) to afford a colorless oil [Run 1; yield: 32.0 mg (84%), Run 2; yield: 24.4 mg (65%)]. 1H NMR (400 MHz, CDCl3): δ = 8.08–8.04 (m, 2 H), 7.60–7.55 (m, 1 H), 7.48–7.43 (m, 2 H), 4.41 (dd, J = 11.6, 3.2 Hz, 1 H), 4.25 (dd, J = 11.6, 6.4 Hz, 1 H), 3.96–3.89 (m, 1 H), 2.08 (d, J = 4.8 Hz, 1 H), 1.68–1.56 (m, 2 H), 1.04 (t, J = 7.2 Hz, 3 H). 13C NMR (101 MHz, CDCl3): δ = 166.7, 133.1, 129.9, 129.6, 128.4, 71.4, 68.8, 26.4, 9.8. The spectra were in accordance with those reported in the literature.17
- 17 Iwasaki F, Maki T, Onomura O, Nakashima W, Matsumura Y. J. Org. Chem. 2000; 65: 996
For recent [2+2]-cycloaddition methods and MHAT examples, see:
For recent C–H functionalizations, see:







