Report

Arylsulfonylacetamides as bifunctional reagents for alkene aminoarylation

See allHide authors and affiliations

Science  28 Sep 2018:
Vol. 361, Issue 6409, pp. 1369-1373
DOI: 10.1126/science.aat2117

Arenes and amides from a single source

Pharmaceutical synthesis often requires the formation of adjacent carbon-carbon and carbon-nitrogen bonds. Monos et al. present a method that delivers the carbon and nitrogen components in a single reagent, specifically, an aryl ring tethered through sulfur dioxide to an amide. A light-activated catalyst primes an olefin to react with the nitrogen, which in turn leads to migration of the aryl ring and loss of the sulfur bridge. The efficient room-temperature process is applicable to a variety of different arenes, including heterocycles.

Science, this issue p. 1369

Abstract

Alkene aminoarylation with a single, bifunctional reagent is a concise synthetic strategy. We report a catalytic protocol for the addition of arylsulfonylacetamides across electron-rich alkenes with complete anti-Markovnikov regioselectivity and excellent diastereoselectivity to provide 2,2-diarylethylamines. In this process, single-electron alkene oxidation enables carbon-nitrogen bond formation to provide a key benzylic radical poised for a Smiles-Truce 1,5-aryl shift. This reaction is redox-neutral, exhibits broad functional group compatibility, and occurs at room temperature with loss of sulfur dioxide. As this process is driven by visible light, uses readily available starting materials, and demonstrates convergent synthesis, it is well suited for use in a variety of synthetic endeavors.

The arylethylamine motif is conserved in dopamine, serotonin, and many opioid receptor drugs responsible for modulating pain sensation and treating neurobehavioral disorders (Fig. 1A) (1, 2). In light of the opioid epidemic, the climate surrounding opioid pain medications is conflicted. It is noteworthy that frontline medications treating opioid addiction contain such arylethylamine substructures (naltrexone and buprenorphine) (35). With this rationale, continued drug development in the arylethylamine chemical space is necessary for general hit-to-lead exploration and the discovery of new and safer medicines. Conventional methods to synthesize arylethylamines use multistep homologation and reductive amination sequences. Alternatively, alkene aminoarylation, particularly of anethole and other biomass-derived alkenes, allows for direct access to this medicinally desirable functionality. The development of methodologies to rapidly construct two new bonds (C–C and C–N) in a single operation from feedstock chemicals can improve and expedite the discovery of new arylethylamine-based small-molecule therapeutics.

Fig. 1 Strategies to access arylethylamines.

(A) Arylethylamines as valuable motifs. (B) Current approaches toward arylethylamines with transition metal catalysis. (C) The Smiles-Truce rearrangement as an aryl migration strategy. (D) The method proposed herein. Me, methyl; Ph, phenyl; AIBN, azobisisobutyronitrile; HSnBu3, tributyltin hydride.

Alkene aminoarylation has been demonstrated with palladium (6, 7), copper (810), nickel (11, 12), and gold (13), in which alkenes are activated by the transition metal to facilitate a stereoselective amine cyclization, followed by a two-electron metal-mediated arylation event (Fig. 1B). The metals used in these aminoarylation platforms control stereoselectivity and activate the alkene for reactivity while suppressing protodemetallation or β-hydride elimination pathways that hinder desired C–C bond formation. Amides and amines are more nucleophilic than the alkene coupling partner; thus, elevated temperatures are often necessary to facilitate ligand substitution to unite the reactants in the initial amination event (14). Despite robust investigation, these methods are generally limited by the need for directing groups and intramolecular reaction designs that restrict the products to pyrrolidine and piperidine structures. Recently, transformations effecting intermolecular aminoarylation and carboamination have been accomplished in which the alkene is decoupled from the arylation and amination reagents. In one case, Lin and Liu demonstrated an enantioselective copper(I)-catalyzed aminoarylation of vinyl arenes relying upon preoxidized sulfonamide reactants (N-fluoro-N-methylbenzenesulfonamide) (Fig. 1B) (9). Separately, Rovis and Piou demonstrated an intermolecular carboamination using N-enoxyphthalimides and Rh(III) catalysis (15).

Photocatalysis and radical-based chemistry have proven similarly influential in alkene difunctionalization. The simplest strategy is Meerwein aminoarylation, a Markovnikov-selective reaction that begins with the reductive generation of a radical from a suitable precursor (arene diazonium salt or diaryliodonium salt) followed by radical-polar crossover and carbocation trapping with acetonitrile solvent (16). These reactions are regioselective but are devoid of stereoselectivity. Photocatalytic anti-Markovnikov–selective alkene hydro- and carboamination reactions were recently demonstrated by Knowles (1721) and Nicewicz (2224). These approaches represent contrasting C–N bond formation strategies while using a common catalytic cycle. Knowles and co-workers have demonstrated both aminium radical cation and amidyl radical generation for the addition to olefins. In both cases, nitrogen-centered radicals couple with π-systems to generate β-amino radicals that are rapidly trapped with an H-atom transfer reagent. Successful H-atom transfer reagents are minimally nucleophilic to prevent thiol-ene reactivity. Nitrogen radical–based chemistry is particularly challenging because both alkene addition and allylic H-atom abstraction are kinetically competitive processes (25); thus, success often requires excesses of the alkene component or intramolecular amino-cyclization. Additionally, amine and amide oxidation generate a more reactive, but not a more nucleophilic, nitrogen atom. In contrast, Nicewicz and co-workers have targeted alkene single-electron oxidation, a process approximately as rapid as amide or amine oxidation. This approach benefits from converting the alkene to a more electrophilic species in solution, necessitating lower equivalents of the nitrogen nucleophile to conduct alkene difunctionalization.

To contrast the widely investigated field of transition metal–mediated aminoarylation and build on the successes of photocatalytic alkene difunctionalization chemistry, we were inspired by the possibility of a radical Smiles-Truce rearrangement to provide alkene aminoarylation products in a diastereoselective fashion. Traditionally, the Truce variant of the Smiles rearrangement is a nucleophilic aromatic substitution effected by benzylic lithiation of ortho-tolyl-arylsulfones (26). The rearrangement is more broadly applicable to ipso-substitution reactions with aryl sulfides, sulfoxides, sulfones, and amides. Pennell and Motherwell furthered the utility of this transformation by demonstrating that aryl radicals are also capable of the same arene transposition (27) (Fig. 1C). Although there are numerous intramolecular examples of radical Smiles-Truce reactions (2833), many of these reactions use net reductive conditions, generate a stoichiometric amount of waste, and rely on a substrate design that tethers the radical precursor to the aryl-sulfonate derivative. Realizing that this intramolecular tether can be formed via in situ oxidation of an alkene and subsequent nucleophilic trapping with an arylsulfonylacetamide (34), we sought to design a photocatalyzed radical Smiles-Truce reaction that showcases the utility of arylsulfonylacetamides as capable reagents for both C–N and C–C bond formation in aminoarylation (Fig. 1D).

A general catalytic cycle was postulated to begin with an oxidation event between a photoexcited catalyst (*IrIII) and an alkene (I) (Fig. 2A) (23, 24). Single-electron oxidation of the alkene would enable nucleophilic addition of an arylsulfonylacetamide (II) to afford the desired β-amino-alkyl radical intermediate (III) (3537). This radical is poised for regioselective cyclization onto the ipso-position of the appended arene to generate IV (38). Lastly, an entropically favored desulfonylation can proceed via two plausible pathways to generate the aminoarylation product, VII: (i) rapid radical desulfonylation from IV to generate nitrogen-centered radical V followed by catalyst turnover, or (ii) homolytic fragmentation of the CAr–S bond to furnish VI, which can turn over the catalyst and undergo desulfonylation to VII. Exploiting both the electronic activation of the sulfonylated arene unit and the tunable nucleophilicity of the nitrogen motif allows for this photoredox catalysis platform to promote both the C–N and C–C bond-forming events with arylsulfonylacetamides.

Fig. 2 Proposed reaction design for aminoarylation with arylsulfonylacetamides.

(A) Proposed reaction mechanism. (B) Initial reaction evaluation. PMP, p-methoxyphenyl.

To realize the proposed aminoarylation reactivity, we first conducted reaction optimization with vinyl anisole (1) [Ep/2 = 1.6 V versus saturated calomel electrode (SCE)] (39) and 1-naphthylsulfonylacetamide (2) (table S1). A potent photooxidant, [Ir(dF(CF3ppy)2)(5,5′-CF3-bpy)]PF6 (3) (IrIII*/II = 1.68 V versus SCE in MeCN) (40) was initially selected for alkene radical cation formation (Fig. 2B). Early optimization experiments lent evidence to the chemoselectivity of this reaction; excess loading of arylsulfonylacetamide and base were unnecessary (table S1). Nearly equivalent stoichiometry between 1 and 2 afforded the highest yield for the optimization product 4. A base screen revealed potassium acetate, benzoate, and tribasic phosphate as superior bases to the less basic potassium trifluoroacetate and potassium phosphate (mono- or dibasic). The reaction was incompatible with pyridine or with stronger alkoxide bases, as photocatalyst decomposition was observed. Reaction dilution past 0.1 M slowed the rate of product formation, whereas reaction concentrations greater than 0.1 M inhibited product formation. Further optimization proved that less oxidizing photocatalysts such as [Ru(bpy)3]Cl2 (RuII*/I = 0.77 V versus SCE in MeCN), [Ir(dF(CF3ppy)2)(dtbbpy)]PF6 (IrIII*/II = 0.89 V versus SCE in MeCN), [Ir(ppy)2(dtbbpy)]PF6 (IrIII*/II = 0.31 V versus SCE in MeCN) (41), were unable to catalyze this transformation. Use of Fukuzumi’s catalyst (PC*/PC = 1.88 V versus SCE in MeCN) (42) did produce 4 in 13% yield. Finally, H-atom donor additives such as 1,4-cyclohexadiene and isopropanol did not improve on the established conditions for the optimization product 4. Exclusion of either light or photocatalyst failed to promote aminoarylation (table S1). With the proof of concept established, we identified the acyl group, among a range of amides and carbamates, as the optimal activating group for the sulfonamide reagent in this transformation (Fig. 3, 47). We reasoned that the acidity and the steric encumbrance of the sulfonamide activating group control the nucleophilicity of the arylsulfonylacetamide.

Fig. 3 Exploration of substrate scope.

All yields are isolated yields. Relative configurations of products were assigned by analogy to 15 and 23. (A) Evaluation of scope of aryl group. (B) Scope of cyclic trans-PMP alkenes. (C) Scope of a compatible cyclic cis-PMP alkene. Ac, acetyl; Boc, tert-butyloxycarbonyl; nPent, n-pentyl; Ts, tosyl. *2:1 mixture of E/Z alkene diastereomers.

A substantial increase in aminoarylation was observed when using 1,2-disubstituted p-methoxyphenyl alkenes in comparison to 1 (Fig. 3A). This substitution allowed us to realize the aryl transfer of several groups including 1-naphthyl (46, 810, 21, 22), 2-napthyl (11), 3-thiophenyl (12, 13), 2-thiophenyl (14, 15, 18), 2- furanyl (16), 8-quinolino (17), 2-benzothiazole (19), and β-styrene (20) all in greater than 20:1 diastereoselectivity. X-ray crystallographic analysis of 15 was found to show a syn-configuration between the 5-bromothiophene and the acetamide groups supporting the stereochemical assignment. Use of cyclic (E)-alkenes allowed for the synthesis of cyclic arylethylamines (2326) containing two contiguous stereocenters, one of which is quaternary (Fig. 3B). Furthermore, the cis-diastereomer 27 can be formed when a cyclic (Z)-alkene is used as the oxidizable alkene substrate partner (Fig. 3C). Preparation of arylethylamine 21 containing an N-tosyl amide showcases the chemoselective nature of this aminoarylation, and the successful isolation of 22 suggests that nucleophiles tethered to the alkene are well tolerated under the reaction conditions. The current aminoarylation conditions are not amenable to benzenesulfonylacetamides, likely as a result of the increased enthalpic barrier for dearomatization during the initial radical cyclization (fig. S3).

To provide mechanistic insight, we carried out several studies to understand the efficiency and high diastereoselectivity of this transformation. We hypothesized that both acyclic (Z)- and (E)-alkenes would convert to the same trans-aminoarylation diastereomer as a result of bond rotation outcompeting cyclization of intermediate III. Notably, performing the title aminoarylation with (Z)-anethole afforded a nearly identical yield of 9 (72%), in comparison to (E)-anethole (82%), whereas diastereomer 9′ was not observed (Fig. 4A). Reaction progress analysis by 1H nuclear magnetic resonance spectroscopy of (Z)-anethole aminoarylation revealed that (E)-anethole is generated during the reaction (Fig. 4B). On the basis of this observation, we examined the rates of isomerization for each anethole isomer to the photostationary state (Fig. 4C). This revealed a photostationary state of 1.4:1 (Z:E), with the initial rate of (Z)-anethole isomerization being much faster than (E)-anethole isomerization (figs. S4 and S8 to S10) (43). Furthermore, initial rate analysis of aminoarylation shows alkene consumption to be slower (fig. S6) than (Z)-anethole isomerization (figs. S8 and S9). These data suggest that the diastereoselectivity arises from either (i) a kinetically favored generation of (E)-anethole radical cation and subsequent aminoarylation, or (ii) a thermodynamic preference of radical intermediate III to adopt an anti-periplanar conformation between the para-methoxyphenyl (PMP) and methyl substituents prior to cyclization (Fig. 4D). One other competing possibility is that the (E)-anethole radical cation reacts with 2 faster than does the (Z)-anethole radical cation.

Fig. 4 Experiments to probe reaction mechanism.

(A) Aminoarylation with (Z)-anethole (28). (B) Tracking reaction progress for aminoarylation with (Z)-anethole (28) [triangles, 2; red squares, (Z)-anethole; diamonds, 9; purple squares, (E)-anethole]. (C) Determination of the photostationary state for anethole isomers catalyzed by 3. (D) Favored and disfavored conformations for intermediate III.

Overall, these mechanistic details describe how the combination of a Smiles-Truce aryl transfer and radical cation chemistry can be combined into a highly diastereoconvergent alkene aminoarylation. Given the current availability of sulfonamide building blocks along with the ubiquity of alkenes as feedstock substrates, we view the method to be a highly enabling platform for research efforts synthesizing the arylethylamine pharmacophore diastereoselectively in a single operation.

Supplementary Materials

www.sciencemag.org/content/361/6409/1369/suppl/DC1

Materials and Methods

Figs. S1 to S11

Tables S1 to S5

References (4454)

References and Notes

Acknowledgments: We thank J. W. Kampf for assistance with x-ray crystallographic analyses. Funding: Supported by National Institute of General Medical Sciences grant R01-GM096129, the Camille Dreyfus Teacher-Scholar Award Program, and the University of Michigan. This material is based on work supported by a NSF Graduate Research Fellowship (grant DGE 1256260) (R.C.M.). Author contributions: T.M.M. and R.C.M. performed the experiments; T.M.M., R.C.M., and C.R.J.S. designed the experiments; T.M.M., R.C.M., and C.R.J.S. wrote the manuscript. Competing interests: The authors declare no competing financial interests. Data and materials availability: X-ray data for compounds 15 and 23 are available free of charge from the Cambridge Crystallographic Data Centre under CCDC 1572215 and 1572214, respectively.
View Abstract

Stay Connected to Science

Navigate This Article