Abstract
Enoldiazoimides, a new subclass of enoldiazo compounds, generate enol-substituted carbonyl ylides whose reactions with sulfur ylides enable an unprecedented formal [4+2] cycloaddition. The resulting multifunctionalized indolizidinones, which incorporate sulfur, are formed in good yields under mild reaction conditions. The uniqueness of this transformation stems from the role of the silyl-protected enol, since the corresponding acetyldiazoimide failed to provide any cross-products in metal-catalyzed reactions with sulfur ylides. This copper-catalyzed cycloaddition is initiated with the generation of enol-substituted carbonyl ylides and sulfur ylides from enoldiazoimides and sulfonium salts, respectively, and proceeds through stepwise six-membered ring formation, C–O and C–S bond cleavage, and silyl and acetyl group migration.
Keywords: copper, cycloaddition, diazo compounds, sulfonium salts, ylides
Ylides, such as carbonyl and sulfur ylides, are readily accessible reagents (or intermediates) that have been efficiently utilized in various organic transformations, especially cycloaddition reactions.[1] Carbonyl ylides (R2C=O+–C–R2) are conveniently formed by transition-metal-catalyzed intra or intermolecular reactions of carbonyls (R2C=O) with diazo compounds (N2=CR2), and they are widely employed in 1,3-dipolar cycloadditions as (+C–O–C–)-type dipolar components.[2] For example, 1,3-oxazolium-4-oxide derivatives (isomgnchnones), generated in situ by metal-catalyzed (especially rhodium) intramolecular cyclization of α-diazoimides (1a and 2a; Scheme 1a), have been developed for the construction of a series of oxa- and aza-polycyclic ring systems.[3] Sulfur ylides (R2S+ –C–R2; e.g., 3a, Scheme 1a) have been easily prepared from their corresponding sulfonium salts (R2S+–CHR2·X–) under basic conditions, and they usually serve as one-carbon synthons (the ylide carbon) in cyclopropanation, epoxidation, aziridination, and other [n +1] cycloaddition processes with release of the corresponding sulfides (R2S),[4] including a recently reported [3+1] cyclo-addition reaction with enoldiazoacetates.[5]
Scheme 1.
Transition-metal-catalyzed reactions of α-diazoimides with sulfur ylides and the development of enoldiazoimides.
Given that both carbonyl ylides and sulfur ylides have exhibited remarkable efficiencies in their respective cyclo-addition reactions,[2–4] we were intrigued by the possibility that a combination of these two subclasses of ylides could lead to new cycloaddition outcomes. Therefore, reactions between in situ generated isomgnchnone intermediates and the sulfur ylide 3a were initially investigated, but no cross-products were obtained from the reactions initiated by dinitrogen (N2) extrusion from either the α-diazoimide 1a or 2a (Scheme 1a).[6] Inspired by recent advances in the cycloaddition reactions of enoldiazo compounds,[7] especially by those studies in which novel reactivities were achieved,[8] we constructed enoldiazoimides, a new subclass of enoldiazo compounds, by a simple silyl-enolization of acetyldiazoimides (e.g., 2a → 4a, Scheme 1b).[9] To our delight, the cross-coupled product 5a, which is an unpredicted formal [4+2] cycloadduct, was obtained in 74% yield along with methyl acetate (MeOAc) from a copper-catalyzed reaction of the enoldiazoimide 4a with 3a (Scheme 1c). Notably, substituted indolizidinones are the core structure of numerous biologically active natural products and pharmaceuticals,[10] and the development of this methodology would allow straightforward and efficient access to multifunctionalized indolizidinones.
To further simplify the process by using the sulfonium salt 6a as the reactant, we performed the reaction with triethylamine (TEA) to form sulfur ylide 3a in situ, expecting that triethylamine might inhibit the catalytic activity by occupying coordination sites of the catalyst (Table 1). However, the subsequent formal [4+2] cycloaddition not only proceeded smoothly under the same reaction conditions but also furnished a higher yield of 5a (83%; entry 1). A selection of commercially available transition-metal complexes was then evaluated as catalysts: copper(I) hexafluorophosphate [Cu-(MeCN)4PF6] gave results similar to those obtained with copper(I) tetrafluoroborate [Cu(MeCN)4bF4] (entries 1 and 2), whereas rhodium(II) acetate [Rh2(OAc)4] afforded the identical cycloadduct with a slightly diminished yield (entry 3). Other tested metal catalysts, as well as alternative solvents and bases, failed to further improve the product yield (see Table S1 in the Supporting Information). However, conducting the reaction at 0°C led to a 91% yield of isolated 5a (entry 4), although further lowering the temperature to −20°C inhibited cross-reactivity (entry 5).[11]
Table 1:
Transition-metal-catalyzed formal [4+2] cycloaddition of enoldiazoimide 4a with sulfonium salt 6a: Optimization of reaction conditions.[a]
![]() | ||||
|---|---|---|---|---|
| Entry | Catalyst | x | T[°C] | Yield [%][b] |
| 1 | Cu(MeCN)4BF4 | 5 | 23 | 83 |
| 2 | Cu(MeCN)4PF6 | 5 | 23 | 81 |
| 3 | Rh2(OAc)4 | 2 | 23 | 74 |
| 4 | Cu(MeCN)4BF4 | 5 | 0 | 91 |
| 5 | Cu(MeCN)4BF4 | 5 | −20 | <5 |
Reaction conditions: catalyst/4a/6a/TEA= 0.002x:0.24:0.2:0.4 (mmol), in 3 mL of CHCl3 at specified temperature for 16 h.
Yields are of the products isolated after flash column chromatography. See Section 3 of the Supporting Information for more details.
With the optimized reaction conditions in hand (Table 1, entry 4), we sought to explore the substrate scope of this cycloaddition reaction, the results for which are presented in Table 2 and Scheme 2. Changing the silyl protecting group of enoldiazoimides from triisopropylsilyl (TIPS; 4a) to tert-butyldimethylsilyl (TBS; 4b) did not affect the efficiency of this process (Table 2, 5a–d), and the structure of 5d was further confirmed by single-crystal X-ray diffraction (SCXRD) analysis (see Figure S1).[12] Sulfonium salts bearing para- and meta-substituted phenyl groups generally performed well in this reaction (5c–j). A variety of functional groups, including nitro, cyano, bromo, and chloro, were well tolerated, although a diminished yield was obtained with the electron-donating para-methoxy group. Furthermore, sulfonium salts that carry naphthyl and thienyl rings smoothly underwent reactions to furnish the formal [4+2] cycloadducts 5k and 5l, respectively. It should be noted that sulfonium salts bearing an ortho-substituted phenyl group (e.g., 2-methylphenyl) only afforded trace amounts of their respective cycloaddtion products, suggesting a steric limitation to the process. In addition, the chiral enoldiazoimide 4c, synthesized from commercially available optically pure methyl (S)-2-pyrrolidone-5-carboxylate, was also successfully employed in this formal [4+2] cycloaddition reaction, thus producing its chiral derivative 5m in 66% yield and suggesting access to other chiral indolizidinones. Moreover, treating this cyclo-adduct with tetra-n-butylammonium fluoride (TBAF) led to the formation of 7m in an yield of 80% (Table 2).
Table 2:
Copper-catalyzed formal [4+2] cycloaddition of enoldiazoimides 4 with sulfonium salts 6: Substrate scope.[a,b]
|
Reaction conditions: Cu(MeCN)4 BF/4/6/TEA =0.01:0.24:0.2:0.4 (mmol), in 3 mL of CHCl3 at 0°C for 16 h.
Yields are of the products isolated after flash column chromatography. See Section 4 of the Supporting Information for experimental details.
See Section 5 of the Supporting Information for experimental details. TBAF= tetra-n-butyl-ammonium fluoride.
Scheme 2.
Copper-catalyzed formal [4+2] cycloaddition of 4a with the sulfonium salts 6k–m or sulfur ylides 3l,m. See Section 4 of the Supporting Information for experimental details.
Besides dimethyl-substituted sulfonium salts (6a–j), the monomethyl sulfonium salt 6k, which was easily prepared from thioanisole (PhSMe), also underwent copper-catalyzed formal [4+2] cycloaddition with 4a. Only cleavage of the sulfur–methyl (S–Me) bond was observed, furnishing the phenylthio-substituted product 5n (Scheme 2a) and suggesting loss of the S-methyl group by a nucleophilic substitution process. In confirmation of this mode of alkyl-group removal from sulfur, the diphenylsulfide-derived (Ph2S) analogue of 6k failed to provide any cross-products under otherwise identical reaction conditions. The cyclic-sulfide-derived sulfonium salts 6l and 6m were then investigated, and the terminal bromo substituents in the products 5o and 5p, respectively, indicated that bromide anions (Br–) from the sulfonium salts facilitated cleavage of the sulfur–methylene (S–CH2) bonds (Scheme 2b). Interestingly, when the corresponding sulfur ylides 3l and 3m were employed as reactants instead of the sulfonium bromides, the terminal acetoxy (AcO) substituted products 5q and 5r, respectively, were obtained (Scheme 2c), and is consistent with the formation of methyl acetate from the reaction between 3a and 4a (Scheme 1c).
The complexity of this transformation is surprising. Rather than a simple cycloaddition, this reaction involves the cleavage of a C–S bond, an O-to-O transfer of a silyl group, and formation and removal of an acetyl group from the original silyl enolate. Based on these results and previous reports,[13,14] we propose a plausible mechanism for this formal [4+2] cycloaddition. As illustrated in Scheme 3, the copper catalyst [M] facilitates dinitrogen extrusion from 4 to form the metallo-enolcarbene I, which undergoes the well-known intramolecular metallocarbene–carbonyl cyclization to furnish the 5-enol-1,3-oxazolium-4-oxide derivative IIA. Although dissociation of the ligated metal to form carbonyl ylides is common, IIA is a vinylogous system which can exist in equilibrium with the intermediate IIB by 1,3-allylic isomerization.[13] Subsequent nucleophilic addition of 3 (either pre- or in situ generated from sulfonium salt 6 with triethylamine), followed by proton transfer to the vinylogous position produces the intermediate IV which proceeds through a tandem C–O bond cleavage/silyl group migration/six-membered ring-closure sequence to deliver the intermediate V.[14] This pathway is consistent with our early finding that the ortho-substituted phenyl group of sulfonium salts dramatically lowered reactivity, since the ortho substituent can hinder the ring closure (see IV). Finally, in cases where the sulfonium bromide salt is directly used as the reactant, the bromide anion facilitates C–S bond cleavage (V → VI), and subsequent acetyl group migration (VI → VII) followed by elimination of the resulting acetoxy group (VII → 5) to complete the transformation. However, if the sulfur ylide is employed as the reactant, the acetoxy group, formed by intramolecular acetyl migration from V, promotes C–S bond cleavage (see VIII) to provide 5. Additionally, an alternative explanation that the carbonyl ylide generated from diazoimides initially undergoes a direct [3 + n] cycloaddition with the sulfur ylide is unlikely given that neither diazoimide 1a nor 2a produced any cross-products in their reactions with 3a (Scheme 1a).
Scheme 3.
Proposed mechanism for copper-catalyzed formal [4+2] cycloaddition of enoldiazoimides. Curved arrows depict the flow or movement of electrons in a concerted or stepwise manner.
In summary, we have discovered a novel approach to the synthesis of multifunctionalized indolizidinones and it proceeds in one synthetic step by the combination of enol-substituted carbonyl ylides and sulfur ylides, generated in situ from enoldiazoimides and sulfonium salts, serving as the initial electrophile and nucleophile, respectively. This formal [4+2] cycloaddition also involves C–O bond cleavage, silyl group migration, acetyl group migration, and C–S bond cleavage along the reaction pathway. The essential role of the silyl-protected enoldiazoimide is seen in the complete absence of a cross-product when the corresponding acetyldiazoimide was employed. We hope this simple strategy (acyl → enol) will create new possibilities for further method ology developments in the field.
Supplementary Material
Acknowledgements
Support for this research from the National Science Foundation (CHE-1464690) is gratefully acknowledged. The HRMS used in this research was supported by a grant from the National Institutes of Health (G12MD007591). The National Science Foundation supported the acquisition of a NMR spectrometer used in this study (CHE-1625963). L.A. Massey acknowledges support from the National Institute of General Medical Sciences branch of the National Institutes of Health (MARC-U*STAR GM07717).
Footnotes
Supporting information and the ORCID identification number(s) for the author(s) of this article can be found under: https://doi.org/10.1002/anie.201805323.
Conflict of interest
The authors declare no conflict of interest.
References
- [1].For reviews, see:Nitrogen, Oxygen and Sulfur Ylide Chemistry: A Practical Approach in Chemistry (Ed.: Clark JS), Oxford UnIVersity Press, New York, 2002;Modern Ylide Chemistry: Applications in Ligand Design, Organic and Catalytic Transformations (Ed.: Gessner VH), Springer, New York, 2018.
- [2].For reviews, see:Hodgson DM, Pierard FYTM, Stupple PA, Chem. Soc. Rev 2001, 30, 50;Padwa A, Helv. Chim. Acta 2005, 88, 1357;Hodgson DM, Labande AH, Muthusamy S, Org. React 2013, 80, 133.
- [3].For reviews, see:Padwa A, Weingarten MD, Chem. Rev 1996, 96, 223;Padwa A, Chem. Soc. Rev 2009, 38, 3072;Padwa A, Tetrahedron 2011, 67, 8057.
- [4].For reviews, see:Li A-H, Dai L-X, Aggarwal VK, Chem. Rev 1997, 97, 2341;McGarrigle EM, Myers EL, Illa O,Shaw MA, Riches SL, Aggarwal VK, Chem. Rev 2007, 107, 5841;Lu L-Q, Li T-R, Wang Q, Xiao W-J, Chem. Soc. Rev 2017, 46, 4135.
- [5].Deng Y, Massey LA, Zavalij PY, Doyle MP, Angew. Chem. Int. Ed 2017, 56, 7479; Angew. Chem. 2017, 129, 7587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Both rhodium and copper catalysts were tested in reactions of either α-diazoimides 1a or 2a with the sulfur ylide 3a.
- [7].For reviews, see:Cheng Q-Q, Deng Y, Lankelma M, Doyle MP, Chem. Soc. Rev 2017, 46, 5425;Cheng Q-Q, Yu Y, Yedoyan J, Doyle MP, ChemCatChem 2018, 10, 488.
- [8].For selected examples, see:Wang X, Xu X, Zavalij PY,Doyle MP, J. Am. Chem. Soc 2011, 133, 16402;Cheng Q-Q,Yedoyan J, Arman H, Doyle MP, J. Am. Chem. Soc 2016, 138, 44;Cheng Q-Q, Yedoyan J, Arman H, Doyle MP, Angew. Chem. Int. Ed 2016, 55, 5573; Angew. Chem. 2016, 128, 5663;Cheng Q-Q, Lankelma M, Wherritt D, Arman H, Doyle MP, J. Am. Chem. Soc 2017, 139, 9839;Deng Y, Massey LA, Rodriguez Núñez YA, Arman H, Doyle MP, Angew. Chem. Int. Ed 2017, 56, 12292; Angew. Chem. 2017, 129, 12460.
- [9].For experimental details, see Section 2 of the Supporting Information.
- [10].For reviews, see:Pizzolato JF, Saltz LB, Lancet 2003, 361, 2235;Pommier Y, Nat. Rev. Cancer 2006, 6, 789;Verma RP, Hansch C, Chem. Rev 2009, 109, 213;Pommier Y, Chem. Rev 2009, 109, 2894;Pommier Y, Marchand C, Nat. Rev. Drug Discovery 2012, 11, 25.
- [11].Only a trace amount of 5a was formed after 4a was fully consumed.
- [12].CCDC 1821831 (5d) contains the supplementary crystallo-graphic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre.
- [13].a) Shintani R, Takatsu K, Takeda M, Hayashi T, Angew. Chem. Int. Ed 2011, 50, 8656; Angew. Chem. 2011, 123, 8815; [DOI] [PubMed] [Google Scholar]; b) Bertz SH, Hardin RA, Murphy MD, Ogle CA, Richter JD, Thomas AA, J. Am. Chem. Soc 2012, 134, 9557; [DOI] [PubMed] [Google Scholar]; c) Xie W, Chang S, Angew. Chem. Int. Ed 2016, 55, 1876; Angew. Chem. 2016, 128, 1908; [DOI] [PubMed] [Google Scholar]; d) Pulis AP, Yeung K, Procter DJ, Chem. Sci 2017, 8, 5240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].a) Padwa A, Brodney MA, Marino JP, Sheehan SM, J. Org. Chem 1997, 62, 78; [DOI] [PubMed] [Google Scholar]; b) Padwa A, Harring SR, Semones MA,J. Org. Chem 1998, 63, 44. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.




