Download PDF
Synlett 2022; 33(04): 357-360
DOI: 10.1055/s-0040-1706014
cluster
Late-Stage Functionalization

Dual Ligand-Enabled Late-Stage Fujiwara–Moritani Reactions

Carlos Santiago
,
Hao Chen
,
Arup Mondal
,
Manuel van Gemmeren  ∗

Financial support has been obtained from the DFG (Emmy Noether Programme) and the Westfälische Wilhelms-Universität Münster.
› Further Information
 
 PDF Download Permissions and Reprints All articles of this category


Abstract

In this study, we describe the use of dual ligand-based palladium catalysts for the late-stage olefination of arenes. Building upon a method previously developed for simple arenes, a variety of complex arene substrates were functionalized. Importantly, the method uses the arene as a limiting reactant and is therefore suitable for valuable starting materials that cannot be used in excess. The regioselectivity of the transformation is controlled by the steric and electronic properties of the substrate, providing access to regioisomers that would be challenging to prepare through other synthetic approaches.


#

Key words

C–H bond activation - late-stage functionalization - palladium catalysis - arenes - Fujiwara–Moritani reaction - olefination

Methods for the so-called late-stage functionalization (LSF) of complex organic compounds play an important role in various fields of research.[1] For example, derivatives of drug candidates can be accessed without having to repeat the compound synthesis for every desired variant.[2] Likewise, structural diversity can be introduced into complex scaffolds during the generation of compound libraries. In natural-product synthesis, streamlining can be achieved because key late-stage functionalization steps permit the use of simpler precursors and intermediates in the preceding synthesis, for example by obviating the need for protecting groups.[3] It is therefore not surprising that substantial efforts continue to be directed toward LSF of important scaffolds such as aromatic cores. Such methods can be enabled through various mechanistic approaches, ranging from modern methods for electrophilic aromatic substitution, through radical-based methods (which can, for example, also be photoredox catalyzed or based on electrochemistry), to transition metal-enabled C–H activation processes.[1] [4]

When focusing on methods based on C–H activation,[5] two fundamentally different approaches can be adopted. The first strategy relies on the use of directing groups (DGs)[6] that coordinate to the catalyst and thereby render the decisive C–H activation intramolecular in nature. This precoordination induces reactivity and controls the regio­selectivity of the process. The use of DGs has permitted a plethora of synthetically highly useful transformations for late-stage modification. One often-invoked disadvantage of this strategy has been the need to develop such reactions individually for each functionalization to be achieved and for each directing group on the substrate. Recent studies have shown, however, that it is possible to develop methods for LSF that are suitable for a broad range of DGs under identical conditions and/or that are able to use naturally occurring functional groups as DGs.[7]

The second approach toward late-stage C–H activation is to develop nondirected methods, in which no donor functionality is required on the substrate for the reaction to occur.[8] In contrast to directed methods that typically deliver a single regioisomer, nondirected reactions often deliver a mixture of regioisomeric products. This aspect can be considered a disadvantage, but can, in fact, often be exploited, because these mixtures frequently contain regioisomers that would otherwise only be accessible through tedious multistep syntheses. Therefore, in conjunction with modern separation techniques, the generation of regioisomeric mixtures through late-stage nondirected C–H activation can be a valuable tool, for example in the synthesis of compound libraries.[2] Historically, many methods for nondirected C–H activation have required an excess of the substrate to overcome the lack of a directing group. This need for an excess of the substrate is inherently incompatible with the use of such methods in LSF. In this context, our group introduced a dual ligand-based catalyst design that permits Pd-catalyzed nondirected C–H activation to occur efficiently, with the arene as the limiting reactant (Scheme [1]).[9]

Our first report on these dual ligand-enabled Pd-catalysts centered on the C–H olefination of arenes. In later studies, this design principle was also applied to further valuable transformations. In 2019, our group and that of Ritter reported dual ligand-enabled methods for the nondirected cyanation of arenes.[10] Furthermore, we described an analogous alkynylation later that year.[11]

Zoom Image
Scheme 1 Methods for late-stage C–H activation/functionalization of arenes enabled by dual ligand-based Pd catalysts

Importantly, all of these methods used the arene as the limiting reagent and can therefore, in principle, be used for late-stage functionalization. In our study on the Fujiwara–Moritani reaction, we hypothesized that this should be possible, but the suitability of our protocol for the late-stage olefination of arenes was not experimentally evaluated. It should be noted that for the cyanation reaction, Ritter and co-workers demonstrated a broad utility for late-stage cyanation.[10b] Likewise, we demonstrated the late-stage alkynylation of arenes.[11] Despite these findings, which implied that the late-stage olefination of arenes should indeed be possible, we envisaged that an experimental assessment would be required to encourage practitioners in the various fields that rely on LSF to adopt this technology. Here we report our studies on the late-stage nondirected Fujiwara–Moritani reaction with dual ligand-based palladium catalysts.

To obtain a representative picture regarding the applicability of our protocol that would allow practitioners to predict the results for various product classes under standard conditions,[9] we decided to study a variety of complex substrates with a broad range of electronic properties without individually fine-tuning the reaction conditions for the separate cases. The results of these experiments are summarized in Scheme [2].

Zoom Image
Scheme 2 Scope studies for the nondirected late-stage olefination of arenes. The regioisomeric ratios in this scheme were determined by NMR spectroscopy. When required, 2D-NMR spectra were used to assign the regioisomers. Note: Because regioisomeric mixtures were obtained, the structures of the respective starting materials are shown. The molecule numbers, yields, and ratios correspond to the respective products. HFIP = 1,1,1,3,3,3-Hexafluoro-2-propanol.

We began with comparably highly reactive electron-rich substrates. First, the methyl ether derived of the narcotic propofol reacted to give product 2a in 61% yield as a single regioisomer. Next, a series of substrates were tested that, due to their substitution pattern, featured no sterically unhindered positions. The synthetic estrogen estrone methyl ether was olefinated to give 2b in a moderate yield of 28%. The lipid-lowering agents clofibrate and gemfibrozil (as its methyl ester) gave 2c and 2d, in yields of 48 and 28%, respectively. These products were obtained as regioisomeric mixtures derivatized at the sterically most accessible positions. These examples showed that, due to the steric sensitivity of our catalyst system, yields remain limited when no unhindered C–H bonds were available in the substrate. We therefore continued to explore electron-rich substrates in which such positions occur. The methyl ether of guaifenesin delivered 2e in a good overall yield of 63% as a mixture of the two expected regioisomers. When we used the pesticide carbofuran as a substrate, the product 2f was again obtained in good yield. We observed that the carbamate group acts as a directing group, leading to ortho substitution in the major product.

Next, a substrate bearing both electron-rich and electron-deficient positions was tested. Substrate 1g, an intermediate in the synthesis of the chemotherapeutic sonidegib, was olefinated to give 2g in 55% overall yield with two major regioisomers, the olefination occurring in the sterically unhindered position and the position ortho to the trifluoromethyl ether. The menthol ester of benzoic acid was tested as a representative of an electron-deficient substrate. The product 2h was obtained in 42% yield with good meta selectivity. We next tested 1i, an intermediate in the synthesis of palonosetron, a medication used against chemotherapy-induced nausea. The product 2i was obtained in a synthetically useful yield and with a high selectivity toward the β-regioisomer.

The methyl ester of nateglinide, a blood-glucose-lowering agent for the treatment of diabetes, containing both ester and amide groups, was olefinated to give 2j in 64% yield. Analogously, the olefination of Evans-type reagent 1k gave 2k in 51% yield. The olefination product of chromanone, 2l was obtained in 46% yield with the two electronically preferred positions leading to the major isomers. In agreement with a report in the literature regarding a related catalyst system, the α′-position was favored in this reaction.[9c] Analogously, two suberone derivatives led to 2m and 2n with electronic selectivity between the sterically accessible positions. Finally, we confirmed that the protocol can be used to combine structural complexity in both the arene substrate and the olefin coupling partner. When a fenchol-derived olefin was combined with propofol methyl ether as substrate, product 3 was obtained in 39% yield.

As expected on the basis of literature reports on the Fujiwara–Moritani reaction, all products were obtained with an E-configuration of the double bond.[9] The regioselectivities shown in Scheme [2], in line with our previous studies,[9a] [10] [11] can serve as a basis for predicting the reaction outcome with future substrates. In essence, the electrophilic palladium catalyst prefers to activate electron-rich positions of the arene substrate. At the same time, the catalyst is comparatively encumbered and therefore avoids sterically hindered positions whenever possible. The reaction thus proceeds under a combined steric and electronic control. Finally, strongly Lewis-basic functional groups can interrupt the nondirected C–H activation, leading to ortho isomers through a directed pathway.

The results summarized in Scheme [2] show that our initial working hypothesis – the suitability of our protocol for the nondirected C–H olefination of arenes in the context of late-stage functionalization – was indeed verified. By using the protocol previously developed for simple arenes, a wide variety of complex substrates could be functionalized. As expected on the basis of related transformations and our mechanistic data, the reactions proceed with a mixture of steric and electronic control, providing access to the desired products in moderate to good yields. Importantly, on the basis of literature studies,[9a] [c] , [10] [11] [12] individual yields could probably be increased substantially by fine-tuning the reaction conditions. In this study we provide evidence that our protocol can be used to introduce olefins into complex molecular scaffolds.[13] These results are expected to encourage practitioners in the various fields that rely on late-stage functionalization to include dual ligand-based palladium catalysts in their toolboxes as valuable tools that often prove complementary to established technologies.


#

Acknowledgment

We thank the members of our analytical departments for their excellent service. Furthermore, we are indebted to Professor F. Glorius for his generous support.

Supporting Information

    Supporting information for this article is available online at https://doi-org.accesdistant.sorbonne-universite.fr/10.1055/s-0040-1706014.
  • Supporting Information

  • References and Notes

  • 4 Mondal A, Wedi P, van Gemmeren M. Remote C–H Bond Functionalization: Methods and Strategies in Organic Synthesis . Maiti D, Guin S. Wiley-VCH; Weinheim: 2021. Chap. 7, 191
    CrossrefSearch in Google Scholar
  • 5 Shilov AE, Shul’pin GB. Chem. Rev. 1997; 97: 2879
    CrossrefPubMedSearch in Google Scholar
  • 11 Mondal A, Chen H, Flämig L, Wedi P, van Gemmeren M. J. Am. Chem. Soc. 2019; 141: 18662
    CrossrefPubMedSearch in Google Scholar
  • 12 Chen H, Farizyan M, Ghiringhelli F, van Gemmeren M. Angew. Chem. Int. Ed. 2020; 59: 12213
    Search in Google Scholar
  • 13 Ethyl (2E)-3-(3,5-Diisopropyl-4-methoxyphenyl)acrylate (2a); Typical Procedure An oven-dried 10 mL Schlenk tube was charged with Pd(OAc)2 (4.5 mg, 0.020 mmol, 10 mol%), L1 (9.4 mg, 0.040 mmol, 20 mol%), N-acetylglycine (7.0 mg, 0.060 mmol, 30 mol%), AgOAc (100.2 mg, 0.6000 mmol, 3 equiv), propofol methyl ether (38.5 mg, 0.200 mmol, 1 equiv), and HFIP (2 mL). The mixture was stirred at rt for 2 min. Ethyl acrylate (0.600 mmol, 3 equiv) was added, and the reaction vessel was tightly sealed and placed in an aluminum block with a tightly fitting recess on a magnetic stirrer at 90 °C. The mixture was stirred at 90 °C for 24 h, then allowed to cool to rt, filtered through silica, transferred into a 100 mL round-bottomed flask, and concentrated under reduced pressure. The crude product was purified by column chromatography [silica gel, pentane–EtOAc (80:1 to 60:1)] to give a colorless solid; yield: 35.5 mg (61%). 1H NMR (400 MHz, CDCl3): δ = 7.58 (d, J = 16.0 Hz, 1 H), 7.20 (s, 2 H), 6.29 (d, J = 16.0 Hz, 1 H), 4.19 (q, J = 7.1 Hz, 2 H), 3.67 (s, 3 H), 3.25 (hept, J = 6.9 Hz, 2 H), 1.27 (t, J = 7.1 Hz, 3 H), 1.17 (d, J = 6.9 Hz, 12 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 167.2, 156.7, 144.9, 142.4, 130.7, 124.3, 116.8, 62.3, 60.4, 26.5, 23.9, 14.4 ppm. HRMS (ESI+): m/z [M + Na]+ calcd for C18H26NaO3 +: 313.1780; found: 313.1771.

Corresponding Author

Manuel van Gemmeren
University of Münster, Organisch-Chemisches Institut
Corrensstraße 36, 48149 Münster
Germany   

Publication History

Received: 02 December 2020

Accepted after revision: 23 December 2020

Article published online:
18 January 2021

© 2021. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany

  • References and Notes

  • 4 Mondal A, Wedi P, van Gemmeren M. Remote C–H Bond Functionalization: Methods and Strategies in Organic Synthesis . Maiti D, Guin S. Wiley-VCH; Weinheim: 2021. Chap. 7, 191
    CrossrefSearch in Google Scholar
  • 5 Shilov AE, Shul’pin GB. Chem. Rev. 1997; 97: 2879
    CrossrefPubMedSearch in Google Scholar
  • 11 Mondal A, Chen H, Flämig L, Wedi P, van Gemmeren M. J. Am. Chem. Soc. 2019; 141: 18662
    CrossrefPubMedSearch in Google Scholar
  • 12 Chen H, Farizyan M, Ghiringhelli F, van Gemmeren M. Angew. Chem. Int. Ed. 2020; 59: 12213
    Search in Google Scholar
  • 13 Ethyl (2E)-3-(3,5-Diisopropyl-4-methoxyphenyl)acrylate (2a); Typical Procedure An oven-dried 10 mL Schlenk tube was charged with Pd(OAc)2 (4.5 mg, 0.020 mmol, 10 mol%), L1 (9.4 mg, 0.040 mmol, 20 mol%), N-acetylglycine (7.0 mg, 0.060 mmol, 30 mol%), AgOAc (100.2 mg, 0.6000 mmol, 3 equiv), propofol methyl ether (38.5 mg, 0.200 mmol, 1 equiv), and HFIP (2 mL). The mixture was stirred at rt for 2 min. Ethyl acrylate (0.600 mmol, 3 equiv) was added, and the reaction vessel was tightly sealed and placed in an aluminum block with a tightly fitting recess on a magnetic stirrer at 90 °C. The mixture was stirred at 90 °C for 24 h, then allowed to cool to rt, filtered through silica, transferred into a 100 mL round-bottomed flask, and concentrated under reduced pressure. The crude product was purified by column chromatography [silica gel, pentane–EtOAc (80:1 to 60:1)] to give a colorless solid; yield: 35.5 mg (61%). 1H NMR (400 MHz, CDCl3): δ = 7.58 (d, J = 16.0 Hz, 1 H), 7.20 (s, 2 H), 6.29 (d, J = 16.0 Hz, 1 H), 4.19 (q, J = 7.1 Hz, 2 H), 3.67 (s, 3 H), 3.25 (hept, J = 6.9 Hz, 2 H), 1.27 (t, J = 7.1 Hz, 3 H), 1.17 (d, J = 6.9 Hz, 12 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 167.2, 156.7, 144.9, 142.4, 130.7, 124.3, 116.8, 62.3, 60.4, 26.5, 23.9, 14.4 ppm. HRMS (ESI+): m/z [M + Na]+ calcd for C18H26NaO3 +: 313.1780; found: 313.1771.

   Permissions and Reprints All articles of this category
Zoom Image
Scheme 1 Methods for late-stage C–H activation/functionalization of arenes enabled by dual ligand-based Pd catalysts
Zoom Image
Scheme 2 Scope studies for the nondirected late-stage olefination of arenes. The regioisomeric ratios in this scheme were determined by NMR spectroscopy. When required, 2D-NMR spectra were used to assign the regioisomers. Note: Because regioisomeric mixtures were obtained, the structures of the respective starting materials are shown. The molecule numbers, yields, and ratios correspond to the respective products. HFIP = 1,1,1,3,3,3-Hexafluoro-2-propanol.