The Acid-Free Cyclopropanol-Minisci Reaction Reveals the Catalytic

Letter Cite This: Org. Lett. 2018, 20, 796−799

pubs.acs.org/OrgLett

The Acid-Free Cyclopropanol-Minisci Reaction Reveals the Catalytic Role of Silver−Pyridine Complexes Andrei Nikolaev,§ Claude Y. Legault,† Minhao Zhang,§ and Arturo Orellana*,§ §

Department of Chemistry, York University, 4700 Keele Street, Toronto, Ontario M3J 1P3, Canada Department of Chemistry, University of Sherbrooke, 2500 Boulevard de l’Universite, Sherbrooke, Québec J1K 2R1, Canada

†

S Supporting Information *

ABSTRACT: A well-defined homogeneous silver precatalyst can be utilized for the direct C−H functionalization of a wide range of aromatic nitrogen heterocycles with cyclopropanols under acidfree conditions. This reaction can be conducted on gram-scale and with low catalyst loadings (as low as 1%), which is rare for silver-catalyzed Minisci-type reactions. Moreover, reactivity trends, as well as steric and calculated electronic properties of the heterocycles, strongly suggest that silver−heterocycle complexes formed in situ behave as redox active catalysts and as Lewis acid activators of the heterocycle and that the electronic nature of the heterocyclic substrates tunes the reactivity of the resulting complexes.

N

type nucleophilic radical addition to electron-deficient heterocycles (Figure 1b). One advantage in this approach is that it reveals a ketone function8 that may prove advantageous for further transformations. During the early stages of our work, Lectka9a reported a similar transformation using superstoichiometric amounts of MnO2 and CF3CO2H.10 More recently, Li9b reported a similar silver-catalyzed reaction that is limited in heterocycle scope. In this letter, we report the functionalization of a wide range of electron-deficient heterocycles with cyclopropanol-derived radicals using a well-defined silver precatalyst and under acidfree conditions. Significantly, our studies show that silver pyridine complexes behave as single electron oxidants toward cyclopropanols, as suggested by early observations by Narasaka,7a and as Lewis acids toward heterocycles. We also provide evidence for the ability of the heterocycles to tune the Lewis acidity of the in situ-formed complexes (when a precatalyst is not used). These mechanistic insights may prove meaningful in the context of silver-catalyzed Minisci-type reactions, whose mechanism remains poorly understood and for which no broadly useful catalyst has been developed. Scouting experiments using AgNO3 as the catalyst revealed that some heterocycles undergo substitution as expected, whereas other electronically similar but sterically different heterocycles did not (Figure 2). For example, reaction with 3,5-dimethylpyridine gives a mixture of alkylated products, whereas the electronically similar 2,6-dimethylpyridine does not. The fact that heterocycles bearing a sterically shielded nitrogen do not participate in the reaction suggested that the substrates may serve as ligands in a silver complex that forms under the reaction conditions, and which becomes the catalytically active species; if the heterocycle is sterically demanding, no such

itrogen heterocycles play a central role in drug development, and efficient methods for their synthesis and functionalization are therefore highly desirable. Methods that enable rapid access to decorated pyridines1 and related heterocycles are particularly advantageous in establishing structure−activity relationships.2 One approach to functionalizing pyridines relies on their preactivation, usually by installation of a halide, prior to C−C or C−X bond formation (X = NR2, OR, etc.) by nucleophilic addition (SNAr) or metal-catalyzed crosscoupling reactions.3 A more direct approach involves the conversion of the heterocycle C−H to a C−C bond by addition of a nucleophilic radical and subsequent oxidation of the intermediate radical to restore aromaticity, as exemplified by the Minisci reaction4 (Figure 1a). Recent advances that exploit this

Figure 1. Minisci and cyclopropanol-Minisci reaction.

fundamental reactivity include the addition of aryl radicals generated from boronic5a,b or carboxylic acids,5c,d addition of alkyl radicals generated from alkylsulfinates,5e−g alkenes,5h amino acids,5i alcohols,5j trifluoroborate salts,5k and unactivated alkyl bromides,5l controlled radical generation using photoredox5b,c,j−n and electrochemical5o strategies, and deeper understanding of the factors affecting chemoselectivity5p of radical addition. An inherent limitation in the majority of these reactions is that the functional group necessary for free-radical generation is lost during the reaction. We envisioned a different approach to heterocycle functionalization that integrates the transition metal-catalyzed fragmentation of cyclopropanols6 to β-keto radicals7 with the Minisci© 2018 American Chemical Society

Received: December 18, 2017 Published: January 19, 2018 796

DOI: 10.1021/acs.orglett.7b03938 Org. Lett. 2018, 20, 796−799

Letter

Organic Letters

redox cycle in oxidative silver-catalyzed reactions.11 Furthermore, its bidentate ligands would likely render it less prone to decomposition through ligand dissociation. Eq 5 summarizes the

Figure 2. Initial experiments suggest that the steric environment around the heterocycle nitrogen impacts reactivity.

final optimized conditions and some highlights of the optimization studies (see Supporting Information (SI) for full details). We showed that the loading of Ag(II)(bipy)2·S2O8 can be reduced to 5 mol % without appreciable reduction in yield and that a 1 mol % catalyst loading yields 53% of the coupled product. This is notable because most Minisci-type reactions use high loadings of silver and often require a second loading. Furthermore, control experiments confirmed that silver is required for the reaction. We also showed that the reaction could be conducted in a practical scale providing the coupled product in 72% yield. We next explored the scope of this reaction using 1-npentylcyclopropanol and a range of nitrogen heterocycles under optimized conditions with Ag(II)(bipy)2·S2O8 and AgNO3 as catalysts. The substrates in Figure 3 span a wide range of electron affinities12 and possess varied steric environments around the heterocycle nitrogen as well as synthetically useful functional group content, including acid-labile groups. Not surprisingly, pyridine (1) and methylpyridines (2−4) provided the expected mixture of products alkylated at the 2- and 4-positions in good combined yields (53−66%). The same trend was observed using AgNO3 but in consistently lower yields (48−55%). 2-Phenylpyridine (5) also gave the expected mixture of products in modest combined yields (32−35%). The reaction with 3,5dimethylpyridine (6) provided the expected mixture of products under both conditions, although the use of AgNO3 resulted in lower overall yields. In contrast, 2,6-dimethylpyridine (7) did not give the desired product with AgNO3 or Ag(II)(bipy)2·S2O8. This is surprising given the essentially identical electron affinity of these compounds and suggests that the steric environment around the nitrogen atom of 7 is responsible for its failure to react. 4-Bromopyridine (8) failed to give coupled product under either conditions likely due to catalyst poisoning. As expected given its low electron affinity, 4-dimethylaminopyridine (9) did not give the coupled product using either catalyst. 4Methoxypyridine (10) gave the desired product only when Ag(II)(bipy)2·S2O8 was used. To our knowledge, this is the only example of direct functionalization of 4-methoxypyridine using a Minisci-type reaction.13 As expected, pyridines bearing electronwithdrawing groups (12−15) gave the best yields of products. Isoquinoline (16) provided the expected product in good yield; however, quinoline (17) only gave the desired product when Ag(II)(bipy)2·S2O8 was used. The same catalyst dependence was observed with benzimidazole (18) and thiazole (19). Whereas pyrazine (20) gave the expected product in good yield, pyrimidine (21) gave a complex mixture of products. A substituted pyrimidine (22) gave the coupled product in good yield. In contrast, pyridazine (23) and acridine (24) failed to give the desired product. Notably, this reaction allows the functionalization of substrates bearing acid-labile groups such as a TIPS-protected alcohol (25),

complex can form, and the reaction fails. Alternatively, perhaps Lewis acid activation of the heterocycle by coordination to silver is necessary for nucleophilic radical addition with sterically demanding heterocycles failing to react due to their inability to coordinate to the metal. It is also possible that both effects are important in these reactions. We conducted a series of experiments to gain some insight into these reactions (Scheme 1). When an activated pyridine (methyl Scheme 1. Evidence for Silver−Pyridine Complexes as Active Catalysts in the Cyclopropanol-Minisci Reaction

isonicotinate) was reacted with a cyclopropanol in the presence of AgNO3 as the catalyst and a terminal oxidant, the expected coupled product was formed in 47% yield (1). In contrast, when benzimidazole was subjected to identical reaction conditions, the reaction failed (2). We reasoned that if methyl isonicotinate and AgNO3 combine to form a catalytically active complex in the first experiment, such a complex might enable the functionalization of benzimidazole. This is indeed the case; when benzimidazole is added to a reaction between methyl isonicotinate and cyclopropanol already in progress, the benzimidazole-coupled product is obtained in 37% yield (3). This evidence suggested that a preformed silver−pyridine complex could be a convenient homogeneous precatalyst for these reactions and would be particularly useful when the use of simple silver salts fail. Gratifyingly, when benzimidazole was reacted with cyclopropanol in the presence of Ag(II)(bipy)2·S2O8 as a catalyst, the coupled product was formed in 42% yield. We selected Ag(II)(bipy)2·S2O8 as a catalyst for further reaction optimization because both Ag(I) and Ag(II) bipyridine complexes are known, and it is therefore reasonable to suggest that they will support the commonly proposed Ag(I)−Ag(II) 797

DOI: 10.1021/acs.orglett.7b03938 Org. Lett. 2018, 20, 796−799

Letter

Organic Letters

We were particularly intrigued by the functionalization of 4methoxypyridine (10) using Ag(II)(bipy)2·S2O8 as the catalyst. As shown in Figure 3, no product is observed with this substrate when AgNO3 is used as the catalyst. In addition, this substrate is known to fail in classic Minisci reactions (using a carboxylic acid as the radical source). Given that 4-methoxypyridine is not sterically hindered, it is safe to assume that it can coordinate with Ag in solution. That being the case, these complexes must not be catalytically active, at least under these conditions, and this suggests that the nature of the silver−pyridine complexes formed in silver-catalyzed heterocycle functionalizations is critical for their success. To test this hypothesis, we conducted a series of experiments shown in Scheme 2. 4-Methoxypyridine is not functionalized Scheme 2. Evidence for Lewis Acid Activation of Pyridines by in Situ-Formed Silver−Pyridine Complexes

using AgNO3 under optimized conditions (entry 1). In contrast, adding 4-methoxypyridine to a reaction with pyridine already in progress provides the functionalized 4-methoxypyridine in 14% yield (entry 2). Remarkably, addition of 4-methoxypyridine to a reaction with methyl isonicotinate already in progress gives an improved yield of 24% (entry 3). Although it is impossible to attribute the change in reactivity to redox properties or Lewis acidity of the in situ-formed complexes, the results in Scheme 2 clearly implicate them as active catalysts. This finding is meaningful given the debate regarding the catalytic activity of silver complexes in related reactions14 and will aid in the development of well-defined and broadly useful catalysts for this reactivity manifold. Finally, we have explored the effect of cyclopropanol structure, and thereby the nature of the radical generated, on the functionalization of ethyl isonicotinate (13) (Scheme 3). The reaction tolerates 1-alkyl (a) and 1-aryl (b) monosubstituted cyclopropanols, giving the product of primary radical addition in

Figure 3. Heterocycle scope of the acid-free cyclopropanol-Minisci reaction.

a Boc-protected amine (26), and a dimethyl acetal (27) group, which would not survive under typical Minisci reaction conditions using strong acid (inset).10 Finally, it is worth noting that the yields observed for positional isomers (10 vs 11 and 13 vs 14) correlate with the calculated electron affinity, validating it as a predictive tool for this reaction. Examination of the yields obtained with both catalytic systems, along with the calculated electron affinities and steric environment around the nitrogen of the heterocycles in Figure 3, suggests that silver complexes play a dual catalytic role in these reactions. For instance, the results for isoquinoline (16) and quinoline (17), which have nearly identical electron affinities, using AgNO3 as the catalyst suggest that in the case of quinoline a redox active silver−quinoline complex does not form. This notion is supported by the more sterically shielded nitrogen in 17 and the fact that use of Ag(II)(bipy)2·S2O8 enables functionalization of quinoline. The comparison between quinoline (17) and acridine (24) is more striking. Although acridine has the highest calculated electron affinity, it is not functionalized with AgNO3 or Ag(II)(bipy)2·S2O8 as the catalyst, suggesting that Lewis acid activation of the heterocycle by coordination of Ag to the nitrogen atom is also important and not possible for acridine due to its sterically shielded nitrogen. Comparing pyridines 6 and 7 also supports this conclusion.

Scheme 3. Effect of Cyclopropanol Structure on the Cyclopropanol-Minisci Reaction

798

DOI: 10.1021/acs.orglett.7b03938 Org. Lett. 2018, 20, 796−799

Letter

Organic Letters

Heterocycl. Chem. 1990, 27, 79. (c) Duncton, M. A. J. MedChemComm 2011, 2, 1135. For a seminal contribution to palladium-catalyzed direct functionalization of pyridines, see: (d) Campeau, L.-C.; Rousseaux, S.; Fagnou, K. J. Am. Chem. Soc. 2005, 127, 18020. (5) (a) Seiple, I. B.; Su, S.; Rodriguez, R. A.; Gianatassio, R.; Fujiwara, Y.; Sobel, A. L.; Baran, P. S. J. Am. Chem. Soc. 2010, 132, 13194. (b) Li, G.-X.; Morales-Rivera, C. A.; Wang, Y.; Gao, F.; he, G.; Liu, P.; Chen, G. Chem. Sci. 2016, 7, 6407. (c) Kan, J.; Huang, S.; Lin, J.; Zhang, M.; Su, W. Angew. Chem., Int. Ed. 2015, 54, 2199. (d) Garza-Sanchez, R. A.; Tlahuext-Aca, A.; Tavakoli, G.; Glorius, F. ACS Catal. 2017, 7, 4057. (e) Ji, Y.; Brueckl, T.; Baxter, R. D.; Fujiwara, Y.; Seiple, I. B.; Su, S.; Blackmond, D.; Baran, P. S. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 14411. (f) Gianatassio, R.; Kawamura, S.; Eprile, C. L.; Foo, K.; Ge, J.; Burns, A. C.; Collins, A. C.; Baran, P. S. Angew. Chem., Int. Ed. 2014, 53, 9851. (g) Fujiwara, Y.; Dixon, J. A.; O’Hara, F.; Daa Funder, E.; Dixon, D. D.; Rodriguez, R. A.; Baxter, R. D.; Herle, B.; Sach, N.; Collins, M. R.; Ishihara, Y.; Baran, P. S. Nature 2012, 492, 95. (h) Ma, X.; Herzon, S. B. J. Am. Chem. Soc. 2016, 138, 8718. (i) Mai, D. N.; Baxter, R. D. Org. Lett. 2016, 18, 3738. (j) Jin, J.; MacMillan, D. W. C. Nature 2015, 525, 87. (k) Matsui, J. K.; Primer, D. N.; Molander, G. A. Chem. Sci. 2017, 8, 3512. (l) McCallum, T.; Barriault, L. Chem. Sci. 2016, 7, 4754. (m) DiRocco, D.; Dykstra, K.; Krska, S.; Vachal, P.; Conway, D. V.; Tudge, M. Angew. Chem., Int. Ed. 2014, 53, 4802. (n) Huff, C. A.; Cohen, R. D.; Dykstra, K. D.; Streckfuss, E.; DiRocco, D. A.; Krska, S. W. J. Org. Chem. 2016, 81, 6980. (o) O’Brien, A. G.; Maruyama, A.; Inokuma, Y.; Fujita, M.; Baran, P. S.; Blackmond, D. G. Angew. Chem., Int. Ed. 2014, 53, 11868. (p) O’Hara, F.; Blackmond, D. G.; Baran, P. S. J. Am. Chem. Soc. 2013, 135, 12122. (6) For recent reviews on the transition metal-catalyzed reactions of cyclopropanols, see: (a) Murakami, M.; Ishida, N. J. Am. Chem. Soc. 2016, 138, 13759. (b) Nikolaev, A.; Orellana, A. Synthesis 2016, 48, 1741. (7) (a) Chiba, S.; Cao, Z.; El Bialy, S. A. A.; Narasaka, K. Chem. Lett. 2006, 35, 18. (b) Chiba, S.; Kitamura, M.; Narasaka, K. J. Am. Chem. Soc. 2006, 128, 6931. (c) Ilangovan, A.; Saravanakumar, S.; Malayappasamy, S. Org. Lett. 2013, 15, 4968. (d) Wang, C.-Y.; Song, R.-J.; Xie, Y.-X.; Li, J.-H. Synthesis 2016, 48, 223. (e) Zhao, H.; Fan, X.; Yu, J.; Zhu, C. J. Am. Chem. Soc. 2015, 137, 3490. (f) Ishida, N.; Okumura, S.; Nakanishi, Y.; Murakami, M. Chem. Lett. 2015, 44, 821. (g) Huang, F.-Q.; Xie, J.; Sun, J. G.; Wang, Y.-W.; Dong, X.; Qi, L.-W.; Zhang, B. Org. Lett. 2016, 18, 684. (8) For an alternative approach to β-keto pyridines, see: Aycock, R. A.; Wang, H.; Jui, N. T. Chem. Sci. 2017, 8, 3121. (9) (a) Lectka, T.; Bume, D. D.; Pitts, C. R.; Lectka, T. Eur. J. Org. Chem. 2016, 2016, 26. (b) Lu, S.-C.; Li, H.-S.; Xu, S.; Duan, G.-Y. Org. Biomol. Chem. 2017, 15, 324. (10) The use of strong acid to promote Minisci reactions is common, see: Tauber, J.; Imbri, D.; Opatz, T. Molecules 2014, 19, 16190. (11) (a) Joshaghani, M.; Bahadori, M.; Rafiee, E.; Bagherzadeh, M. ARKIVOC 2007, 260. (b) Bowmaker, G. A.; Effendy; Marfuah, S.; Skelton, B. W.; White, A. H. Inorg. Chim. Acta 2005, 358, 4371 It remains unclear at this point if a Ag(I)−Ag(II) or Ag(0)−Ag(I) catalytic cycle is involved in this reaction.. (12) The adiabatic electron affinities were calculated at the M06-2X/ Def2-TZVPP level. See SI for additional information. (13) For activation of electron-rich pyridines as the corresponding Noxides, see: (a) Suresh, R.; Kumaran, R. S.; Senthilkumar, V.; Muthusubramanian, S. RSC Adv. 2014, 4, 31685. (b) Suresh, R.; Kumaran, R. S.; Muthusubramanian, S.; Manickam, G. Asian J. Org. Chem. 2014, 3, 604. (14) For recent mechanistic studies on the borono-Minisci reaction, see: (a) Baxter, R. D.; Liang, Y.; Hong, X.; Brown, T. A.; Zare, R. N.; Houk, K. N.; Baran, P. S.; Blackmond, D. G. ACS Cent. Sci. 2015, 1, 456. (b) Patel, N. R.; Flowers, R. A., II J. Am. Chem. Soc. 2013, 135, 4672.

practical yields. As expected, a 1,2-dialkyl-substituted cyclopropanol (c) gives the product arising from addition of a secondary radical in good yield. A cyclopropanol bearing aryl substitution at the 2-position (d) did not give the desired product due to substrate instability under these reaction conditions. Both a cyclic secondary radical and a tertiary radical, arising from the corresponding cyclopropanols (e and f, respectively), failed to give the desired product. This is surprising because both radical types participate in standard Minisci reactions (where the radical is derived from a carboxylic acid) and points to subtle mechanistic differences between these reactions. In summary, we have developed the silver-catalyzed direct functionalization of electron-deficient heterocycles with cyclopropanol-derived radicals under acid-free conditions and using a well-defined catalyst at low loadings. This reaction enables the functionalization of a wide range of aza-heterocycles, including substrates bearing acid-labile functional groups. In addition, our work strongly supports the notion that silver pyridine complexes are involved in the homolytic cleavage of cyclopropanols and also activate the heterocyclic substrates toward nucleophilic radical addition by coordination of silver to the heterocyclic nitrogen.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.7b03938. Experimental procedures, tabulated structural data, 1H-, 13 C, and 19F-NMR spectral data, and calculation of electron affinities (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Claude Y. Legault: 0000-0002-0730-0263 Arturo Orellana: 0000-0002-8372-2222 Author Contributions

A.O. conceived this study and designed experiments. A.N. conducted most of the experimental work. C.Y.L. performed all calculations. M.Z. conducted some substrate scope experiments in Figure 3 to address reviewers’ comments. Funding

We gratefully acknowledge support of this work through a Discovery grant from the Canadian Natural Sciences and Engineering Council of Canada (NSERC). Computational resources were provided by Calcul Québec and Compute Canada. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Vitaku, E.; Smith, D. T.; Njardarson, J. T. J. J. Med. Chem. 2014, 57, 10257. (2) (a) Ming, Y.; Lo, J. C.; Edwards, J. T.; Baran, P. S. J. Am. Chem. Soc. 2016, 138, 12692. (b) Cernak, T.; Dykstra, K. D.; Tyagarajan, S.; Vachal, P.; Krska, S. W. Chem. Soc. Rev. 2016, 45, 546. (3) Eicher, T.; Hauptmann, S.; Speicher, A. The Chemistry of Heterocycles, 3rd ed.; Wiley-VCH: Weinheim, Germany, 2012. (4) (a) Minisci, F.; Bernardi, R.; Bertini, F.; Galli, R.; Perchinummo, M. Tetrahedron 1971, 27, 3575. (b) Minisci, F.; Fontana, F.; Vismara, E. J. 799

DOI: 10.1021/acs.orglett.7b03938 Org. Lett. 2018, 20, 796−799