Ruthenium-Catalyzed Aminocarbonylation with Isocyanates Through Weak Coordinating Groups

Graphical Abstract

Under mild conditions: A redox-neutral and base-free protocol for the directed ruthenium-catalyzed C−H aminocarbonylation is developed through high-throughput experimentation. The mild conditions allowed the introduction of privileged amide moieties to a wide range of anilides bearing various substitution patterns and different weakly Lewis basic directing groups.

Abstract

Introducing amide functional groups under mild conditions has growing importance owing to the prevalence of such moiety in biologically active molecules. Herein, we disclose a mild protocol for the directed ruthenium-catalyzed C−H aminocarbonylation with isocyanates as the amidating agents developed through high-throughput experimentation (HTE). The redox-neutral and base-free reaction is guided by weakly Lewis basic functional groups, including anilides, lactams and carbamates to access anthranilamide derivatives. The synthetic utility of this transformation is reflected by large-scale synthesis and late-stage functionalization.

Introduction

Introducing amide functional groups under mild conditions has growing importance due to the ubiquity of such structures in biologically active molecules such as peptides and proteins. In the pharmaceutical industry, amide bond formation is one of the most common reactions that is performed and amide bonds are found in approximately a quarter of drugs on the market (Figure 1A).¹ In addition, such moiety can be commonly found in the backbone of a range of natural and synthetic polymers.² Thus, how to access the amide functionality efficiently and selectively has been a long-standing goal for synthetic chemists. In contrast to the usual amide bond, we decided to focus on directly installing an aminocarbonyl group (−CONH−) with reversed connectivity, since it exhibits different and sometimes more favorable biological properties due to a change in the location of hydrogen bond donor and acceptor.³

Figure 1

Medicinally relevant compounds and ruthenium-catalyzed directed C−H aminocarbonylation.

Transition-metal-catalyzed C−H activation has emerged as an effective tool to install such pivotal motif with reduced chemical steps and minimized waste generation.⁴ By harnessing the intrinsic Lewis basicity of some functional groups, directed C−H activation can be achieved in a highly position-selective fashion.⁵ Such functional groups can coordinate to the metal catalyst, bringing the reactive metal center into proximity of the desired site of functionalization. Therefore, ortho- aromatic C−H functionalization is the primary site of reactivity.⁶ Among all the directing groups, monodentate auxiliaries can be categorized as L (strong) or X (weak) type. C−H functionalization using L type directing groups, such as pyridines, pyrazoles and oxazolines has been extensively explored.⁷ In 2011, Shi and co-workers reported a rhodium catalyzed C−H addition to p-tolylsulfonyl-benzaldimine.⁸ The method achieves direct nucleophilic addition of an inert C−H bond onto a C=N double bond without reactive Grignard reagents. However, the use of expensive rhodium catalyst limited the synthetic utility of this methodology. In addition, although arylpyridines are highly useful for the coordination with the metal center during the C−H activation step, they are difficult to remove and might not be desired in the final product.

More recently, the Ackermann group developed a protocol which provides access, via a single step, to synthetically challenging phthalimides from benzamides.⁹ The method utilized an X type auxiliary as the directing group, which are less developed due to the weakly coordinating nature.¹⁰ The first step of the transformation employs ruthenium-catalyzed C−H activation assisted by the inherent amide directing group, followed by intramolecular cyclization. These examples clearly showed the challenging nature of achieving C−H aminocarbonylation with a weakly coordinating directing group. Herein, we disclose a ruthenium-catalyzed methodology allowing rapid access to anthranilamides via C−H aminocarbonylation under mild conditions (Figure 1B). Isocyanate was chosen as the coupling partner due to its wide application as an amide precursor.¹¹ The protocol shows broad tolerance to a range of acetanilides and was extended to other directing groups such as benzanilide, pyrrolidone and carbamate. Examples of late-stage functionalization of marketed pharmaceuticals also demonstrate the synthetic utility of this transformation. Derivatizations of such drugs were achieved with significantly improved step counts compared to de-novo synthesis. In addition, deuterium labelling experiments and DFT calculations were performed to probe the mechanism.

Results and Discussion

Reaction optimization and substrate versatility

We initiated our studies by investigating the C−H aminocarbonylation of acetanilide (1 a) with isocyanate 2 in the presence of ruthenium pre-catalyst and silver salt as a halogen scavenger. High-throughput experimentation (HTE) was used as the primary approach for reaction optimization. HTE has the advantage of generating a large set of data points whilst minimizing reagent consumption, which is in line with our goal of achieving a more sustainable chemistry.¹² Adding to this is the access of more well-rounded data since the dependence between different variables, such as catalysts, solvents and additives etc., can be investigated via multi-parameter optimization. Initial screening of different combinations of catalysts and silver salts afforded the desired ortho-functionalized product 3 a with excellent mono-selectivity (Figure 2). Silver salt with bulky counterions generally delivered higher conversion to the amino product 3 a. After considering factors such as toxicity and cost, we excluded the use of AgSbF₆ and [Ru₂Cl₃(p-cymene₎₂][PF₆] since the former is much more toxic than the other silver salts and the latter requires one more step to prepare, which is not in accordance with the pursuit of green chemistry. Combining those factors and the efficiency of the catalyst-salt pair, [Ru(benzene)Cl₂]₂ and AgPF₆ were chosen to carry on forward. Of note, other ruthenium catalysts – silver salts combination also gave the product with satisfying conversion, showcasing the robustness of the transformation. Control experiments demonstrated that both the ruthenium catalyst and the silver salt were essential for the reaction (Row H & Column 12). Interestingly, the cationic catalyst [Ru₂Cl₃(p-cymene)₂][PF₆] still required activation by a silver salt, indicating the need of a strongly electron-poor metal center for the reaction to proceed. Then, a set of individual experiments were set up to evaluate other variables including equivalents of isocyanate, molarity and temperature (Table S1).

Figure 2

Evaluation of ruthenium catalysts and Ag salts under HTE format. [a] Reactions were carried out using 1 a (0.025 mmol), 2 (0.028 mmol), ruthenium catalyst (10 mol %), Ag salt (20 mol %) in DCE (0.1 M) at 60 °C under N₂ for 20 h. Conversion of 1 a was calculated based on LCMS analysis. [b] Reactions were carried out using ruthenium catalyst (5 mol %).

With the optimized method in hand, we evaluated the scope of the C−H aminocarbonylation with respect to anilides (Scheme 1). Acetanilide was converted to the desired product 3 a in 75 % yield, while halving [Ru(benzene)Cl₂]₂ and AgPF₄ loadings did not diminish the yield. Products derived from acetanilides with electron-donating groups were delivered in good yields (3 b and 3 c). Similarly, electron-neutral acetanilides successfully furnished the desired products (3 d–3 f). Acetanilides bearing electron-withdrawing groups also gave the desired products, albeit in slightly lower yields (3 g and 3 h). Meta-substituted acetanilides showed similar trends (3 i–3 k) but overall higher yields than the para-substituted examples. This can likely be ascribed to the increased electron density on the activated para-C−H bond relative to the substituent. To our delight, the 3,5-dimethoxyacetanilide also proceeded in good yield to furnish anthranilamide 3 l. However, ortho-substituted acetanilide did not give any product 3 m. which could be a result of inhibited C−H activation step due to steric hinderance. Interestingly, 1-acetamidonaphthalene gave 5 % C−H functionalized product 3 n, despite bearing only one available ortho C−H bond. By contrast, 2-acetamidonaphthalene furnished the desired mono-functionalized product 3 o in 81 % yield. In addition to the acetamide group, other amido moieties were tested under the optimized conditions. Other alkyl amido directing groups gave similar or better yields compared to the standard substrates (3 p–3 r). Heterocycles, such as furan was also tolerated under the reaction conditions and delivered product 3 s in 70 % yield. When moving away from anilides as directing groups, benzanilide afforded exclusively product 3 t, while the C−H_b bond remained intact. This could be explained with a combination of an electronic and directing group effect, where the aniline being more electron rich, and the formation of a six-membered ruthenacycle upon C−H activation at the ortho-position (see Figure 3). In addition, pyrrolidone 1 u and carbamate 1 v were also shown to be effective directing groups, which gave 76 % and 11 % of the corresponding products. Lastly, ruthenium-catalyzed C−H aminocarbonylation enabled late-stage functionalization of Trametinib and Tamibarotene. Mono-functionalized products were observed exclusively in both cases, highlighting the selectivity for the activation of the least sterically hindered ortho-C−H bond, as well as the functional group tolerance of the reaction conditions. Of note, Tamibarotene with a benzanilide moiety embedded was installed with the aminocarbonyl group selectively at the aniline aromatics, which was consistent with the product obtained from substrate 1 t. The success of late-stage functionalization showcased the potential of diversifying complex molecules without lengthy and resource-intensive de-novo synthesis. Here, the mass balance accounted for the unreacted drug.

Scheme 1

Directing group scope for C−H aminocarbonylation. [a] Reactions were carried out using 1 (0.25 mmol), 2 (0.45 mmol), [RuCl₂(benzene)]₂ (5.0 mol %), AgPF₆ (20 mol %) in 1,2-DCE (0.5 M) at 60 °C under N₂ for 20 h; yields represent isolated materials. [b] The reaction was carried out using [RuCl₂(benzene)]₂ (2.5 mol %), AgPF₆ (10 mol %). The reaction was carried out in DCE (0.1 M) and at 80 °C.

Figure 3

Computed relative Gibbs free energies (ΔG_333.15) in kcal mol⁻¹ for ruthenium(II)-catalyzed C(sp²)−H aminocarbonylation between C−H activation and proto-demetalation steps at the PBE0-D4/def2-TZVP-SDD(Ru)-SMD(DCE)//PBE0-D3(BJ)/def2-SVP-SDD(Ru) level of theory. In the transition state structures, nonrelevant hydrogens were omitted for clarify and emphasized distances are given in Å.

We next turned our attention to the scope of the isocyanates (Scheme 2). Replacing Cbz protected piperidine 2 with other secondary amines yielded 46 % and 37 % of desired products 5 a and 5 b respectively. The sterically encumbered adamantyl isocyanate also delivered product 5 c, albeit in a lower yield.

Scheme 2

Isocyanate scope for C-H aminocarbonylation. [a] Reactions were carried out using 1 a (0.25 mmol), 4 (0.45 mmol), [RuCl₂(benzene)]₂ (5.0 mol %), AgPF₆ (20 mol %) in DCE (0.1 M) at 60 °C under N₂ for 20 h.

Interestingly, once we moved away from secondary or tertiary isocyanates towards primary or aromatic isocyanates, a mixture of products was observed. Competing N−H functionalization led to the formation of side products 6 and 7.¹³ Both electron-rich or -poor isocyanates delivered the desired products (5 e–5 g) with synthetically useful yields. Finally, isocyanate bearing the strongly coordinating pyridine group was also tolerated, giving a mixture of three products.

Thereafter, scale up and derivatization of the reaction products were performed. Increasing the scale 10-fold to 2.5 mmol provided 62 % of aminocarbonylated acetanilide 3a (Scheme 3A). The Cbz protecting group of the aminocarbonylated product was then cleaved under standard hydrogenation conditions, affording the free amine product 8 a in 99 % yield (Scheme 3B). Pleasingly, the deacetylation of the amide directing group proceeded efficiently affording 9 a using Schwartz reagent, opening the possibility for ortho-functionalization of free aniline (Scheme 3C). Successful deprotection at both sites allows the product to be further conjugated with other molecules, which could potentially be used as a linker to access PROTAC.¹⁴

Scheme 3

Scale-up and derivatization reactions.

Mechanistic investigation

To shed light on the mechanism of the ruthenium-catalyzed C−H aminocarbonylation, deuterium labelling experiments were carried out. When deuterated acetanilide 3 a–d₅ was subjected to the standard reaction conditions, a significant amount of deuterium exchange was observed at the ortho-positions of the unreacted acetanilide (Scheme 4A). A similar result was obtained from the aminocarbonylated product (40 % H). This observation suggested that the initial C−H activation step is likely to be reversible and not rate-determining. To confirm the hypothesis, a competition reaction was conducted using an equimolar amount of deuterated and non-deuterated acetanilides (Scheme 4B). The product ratio was 1.4 : 1 favoring the formation of non-deuterated acetanilide derived product, which could be rationalized by a secondary kinetic isotope effect.

Scheme 4

Deuterium labelling studies.

Next, density functional theory (DFT) calculations were performed to elucidate the mechanism of ruthenium-catalyzed C(sp²)−H aminocarbonylation reaction at the PBE0-D4/def2-TZVP-SDD(Ru)-SMD(DCE)//PBE0-D3(BJ)/def2-SVP-SDD(Ru) level of theory (Figure 3). These were conducted between C−H activation and proto-demetalation elementary steps. Our studies commenced with the O-substrate-coordinated complex int1 which undergoes C−H activation via a four-membered ring transition state TS2 with a barrier of 21.0 kcal mol⁻¹ with the carbonyl group of the second substrate acting as a base. Six-memberded ring transition state TS2’ was also considered with an energy barrier of 27.0 kcal mol⁻¹. For the alternative N-assisted pathway, C−H activation was proven to be energetically disfavored with an energy barrier of 31 kcal mol⁻¹ (Figure S5). The formation of int3 was proven to be endothermic. After release of the acetimidic acid, int4 undergoes coordination of the isocyanate followed by migratory insertion via transition state TS6 with an energy barrier of 14.2 kcal mol⁻¹. Afterwards, proto-demetalation occurs with the assistance of the previously generated acetimidic acid via transition state TS9, which was proven to be the rate-limiting step with an energy barrier of 26.0 kcal mol⁻¹. Finally, after the exchange of the ortho amino product 3 a with acetanilide, cyclometalated int1 is regenerated for the next catalytical cycle.

Based on our experimental and computational findings, a plausible catalytic cycle is proposed (Figure 4). Initially, [Ru(benzene)Cl₂]₂ reacts with AgPF₆ and acetanilide to generate di-substrate coordinated cationic ruthenium complex A, followed by the C−H cleavage to generate ruthenium species B. Upon ligand exchange, migratory insertion of isocyanate into the ruthenium-carbon bond takes place affording intermediate D. With the assistance of the generated acetimidic acid, intermediate D undergoes proto-demetalation to deliver the desired product 3 a.

Figure 4

Proposed catalytic cycle.

Conclusions

In summary, we have developed a redox-neutral and base-free strategy for the ruthenium-catalyzed ortho-C(sp²)−H aminocarbonylation of anthranilamide. The optimized conditions were quickly identified with the aid of high-throughput experimentation. A wide range of anilides bearing various substitution patterns, different directing groups such as pyrrolidone and carbamate, as well as a range of isocyanates were tolerated and gave anthranilamide derivatives in moderate to good yields. We also showcased the synthetic utility of the protocol by late-stage functionalization of drug molecules. The catalytic mechanism was probed with deuterium labelling studies and DFT calculations.

Experimental Section

General procedure of ruthenium-catalyzed C−H amino-carbonylation: The reactions were setup using an oven-dried microwave vial using 0.25 mmol of substrate. On the benchtop or in a glovebox under nitrogen, the vials were charged with amide (0.25 mmol, 1.0 equiv.), isocyanate (0.45 mmol, 1.8 equiv.), [Ru(benzene)Cl₂]₂ (0.01 mol, 5.0 mol %), AgPF₆ (0.05 mol, 20 mol %) and DCE (1.25 mL). The microwave vials were then sealed, taken out of the glovebox and heated to 60 °C under stirring (500 rpm) in an appropriate metal block. After 20 h the reaction mixtures were allowed to cool to room temperature and were purified by either automated flash column chromatography or preparative reverse phase HPLC to furnish the respective anthranilamides.

Supporting Information

The authors have cited additional references within the Supporting Information.^{10a, 15-17, 18, 19, 20, 21}

Acknowledgments

The authors gratefully acknowledge support from the DFG Gott-fried-Wilhelm-Leibniz award (L.A.), the European Union's Horizon 2020 research and innovation program (Marie Skłodowska-Curie Grant Agreement No. 860762 to E.Y.L.), and the CSC (fellowship to B.Y.). The authors thank Pernilla Korsgren for HRMS analysis. Thanks to Lucas Guillemard and Daniele Antermite for helpful discussions and for proofreading this manuscript.

Conflict of interest

The authors declare no conflict of interest.

Open Research

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.