Cyanomethylation of Substituted Fluorenes and Oxindoles with Alkyl Nitriles

Abstract

The first example of metal-free cyanomethylenation from alkyl nitriles of sp³ C-H bonds to afford quaternary carbon centers is described. This oxidative protocol is operationally simple and features good functional group compatibility. This method provides a novel approach to highly functionalized fluorene and oxindole derivatives, which are commonly used in material and pharmaceutical areas. Control experiments provide evidence for a radical reaction process.

Graphical Abstract

Nitriles are versatile functional groups in organic synthesis. In addition to being a very useful functional group in biologically active compounds, they can be readily converted into amines, carboxylic acids, ketones and even heterocycles.¹ Compounds containing nitrile groups are frequently employed as building blocks in drug discovery programs. ² Introduction of a nitrile group by activation of the α-hydrogen of simple aliphatic nitriles is one of the most efficient and environmentally benign entries to this class of structure. For example, generation and reaction of stable α-nitrile anions is well-explored with a range of electrophiles.³ The nitrile group also stabilizes radicals arising from α-hydrogen atom abstraction, and these radicals permit bond disconnections⁴ complementary to those available from the α-nitrile anions. Amongst these reactions, there are relatively few involving C-H activation of a second component, that is oxidative fragment coupling. Building off of our prior efforts in dual C-H activation of two components (Scheme 1a),⁵ the use of acetonitrile was investigated with oxindoles and fluorenes. To the best of our knowledge, there are no examples of cross-coupling between sp³ C-H bonds and alkyl nitriles under metal-free conditions. Herein, we communicate our efforts culminating in facile, metal-free oxidative cyanomethylenation of oxindoles and fluorenes with alkyl nitriles through C(sp³)-H oxidative radical functionalization using t-BuOOt-Bu as the oxidant (Scheme 1b–c).

Scheme 1.

Dual C-H Activation of Oxindoles and Fluorenes with Toluenes or Alkylnitriles to Generate Quaternary Centers

As an important structural unit, 2-oxindoles with a quaternary carbon center at the C3 position are a class of heterocycles existing in many natural products, pharmaceuticals and drug candidates (Figure 1a).⁶ Their importance has prompted considerable interest in developing new construction methods.

Figure 1.

Structures of Bioactive Compounds with a) Oxindole or b) Fluorene Moieties

Similarly, fluorenes have attracted much attention for a variety of applications involving advanced materials, including those used in semiconductors,⁷ optoelectronics,⁸ and solar cells. ⁹ In addition, fluorene derivatives have also been playing an increasing role in pharmaceuticals and biochemistry¹⁰ Pharmaceutical compounds also can incorporate fluorene moieties (Figure 1b).¹¹ As a consequence, the development of practical synthetic methods for the construction of functionalized fluorenes is in demand.

For example, the asymmetric cyanomethylenation of 3-substituted oxindoles using prefunctionalized cyanomethyl halides have been reported by different groups (Scheme 2a).¹² Recently, many studies focused on the atom-transfer radical addition reactions of nitriles with olefins.¹³ Among them, the Zhu Group has made significant contributions to this research field.¹⁴ In 2016, Ge group reported the first example of the palladium-catalyzed cross-coupling of sp³ C-H bonds with acetonitrile. ¹⁵ Thereafter, Wu¹⁶ and Shen¹⁷ independently reported the direct oxidative cyanomethylenation reactions by adding acetonitrile to 1,3-dicarbonyls and tetrahydroisoquinolines respectively, however other alkyl nitrile coupling partners were unsuccessful.

Scheme 2.

Approaches to Cyanoalkyl Oxindoles and Fluorenes

In 2015, the Liu group developed a simple and efficient synthesis of 9-arylfluorenes via metal-free reductive coupling of arylboronic acids and N-tosylhydrazones.¹⁸ Also, extensive attention has been paid to generating fluorenes via transition-metal catalyzed cyclizations or direct dehydrogenative aryl-aryl coupling via C-H bond activation.¹⁹ In 2016, the Ji group developed a copper-mediated radical alkylarylation of unactivated alkenes with acetonitrile leading to methylene disubstituted fluorenes, which are not easily accessed by conventional methods (Scheme 2b).²⁰ Nonetheless, this transformation still suffered from some drawbacks, such as the use of transition metal and ligand, narrow substrate scope (acetonitrile only), and only access to 9-alkyl substituted fluorenes, thus limiting its further applications.

With a clear need for alternative strategies to construct highly functionalized oxindoles, we initiated our investigations by screening various metal source (Cu, Fe, Pd, Co, Mn and Sc) and reaction conditions (see the Supporting Information) for the t-BuOOt-Bu-mediated coupling of 3-substituted oxindoles with acetonitrile (eq 1). Notably, the protocols employed by Zhu and Li^{14a, 21} involving metal catalysts to generate acetonitrile radicals for additions to alkenes were not effective in these couplings. After extensive investigation of the reaction conditions (see the Supporting Information), control reactions revealed that the metal catalyst was unnecessary leading to a very straightforward oxidative method for introducing the cyanomethylene functionality to an oxindole. The optimum reaction conditions entailed heating a solution of 1a in acetonitrile (0.1 M) in the presence of t-BuOOt-Bu (4 equiv) at 130 °C for 24 h which provided 3aa in 48% yield.

Subsequently, a range of 3-monosubstituted oxindoles were explored for the cyanomethylenation at the C3-position with acetonitrile (Scheme 3). Electron-neutral (3aa), electron-donating (3ba, 3ca, 3ea, 3fa), and electron-withdrawing (3da) substituents on the phenyl ring were all well tolerated under the optimal reaction conditions. Substituents at the different positions did not affect the yields significantly. The 6-chloro-oxindole also gave the corresponding product 3ga in 63% yield. The N-benzyl substituted oxindole also exhibited good reactivity providing 3ha in 52% yield.

Scheme 3.

The Reaction of 3-Monosubstituted Oxindoles with Acetonitrile^a

^aConditions: 1 (0.15 mmol), 2a (1.5 mL, 0.1 M), t-BuOOt-Bu (4 equiv), 130 °C, under Ar, 24 h.

The cyanomethylenated products derived from 3-substituted oxindoles are versatile intermediates in organic synthesis and can be readily converted into other important building blocks including phenyl substituted pyrroloindolines.²² To show the utility of this method in producing useful precursors, a further transformation was carried out on product 3aa (Scheme 4). First, the 2 mmol scale synthesis of product 3aa proceeded successfully, delivering 3aa in 45% yield. Reductive cyclization of oxindole 3aa using LiAlH₄ provided pyrroloindoline 4 in 52% yield. Overall, this route provides comparable or better efficiencies relative to other routes for generating target 4 with aryl substitution at the angular carbon.²³

Scheme 4.

Scale-Up and Synthetic Transformation of the Cyanomethylenated Product 3aa

Application of the above conditions to the coupling of 9-phenyl-9H-fluorene 5a²⁴ and acetonitrile 2a provided product 6aa (eq 2) in 49% yield (see SI). Further experimentation (see SI) ultimately revealed that carboxylic acid additives enhanced the outcome. The optimum conditions were t-BuOOt-Bu (6 equiv) with PivOH (2 equiv) at 125 °C for 23 h which provided 6aa in 65% yield (Scheme 5).

Scheme 5.

The Reaction of Various 9-Substituted Fluorenes with Alkyl Nitriles^a

^aConditions: 5 (0.15 mmol), 2 (1.5 mL, 0.1 M), t-BuOOt-Bu (6 equiv) PivOH (2 equiv), 125 °C, under Ar, 23 h.

With these conditions in hand, the scope of the reaction with respect to the fluorene component and alkyl nitrile was evaluated (Scheme 5). First, different para-substituted aryl groups at C9 on the fluorene were explored. With either electron-donating or electron-neutral substituents, the products were formed in good yields (6aa-6da). Those bearing electron-withdrawing chloro, fluoro, trifluoromethyl or phenyl group gave the corresponding products in a slightly lower yields (6ea-6ha). C9-Aryl groups with either methoxy or fluoro groups at the meta-position reacted smoothly with 2a to give 6ia and 6ja in 51% and 61% yield, respectively. A range of bulkier aryl groups could be tolerated at C9 of the fluorene, including acetal derived, naphthyl, para-carbazolylphenyl affording the corresponding products in 34–49% yields (6ka-6ma). Of particular note, 9-butylfluorene can also react with acetonitrile to afford 6na in 51% yield.

Notably, functional groups such as methoxy, halogen, and nitro can be employed in different positions on the fluorene component (6oa-6qa). 9-Phenyl-9H-xanthene 5r also successfully reacted with acetonitrile, albeit with 20% yield (6ra). Next, other alkyl nitriles, such as propionitrile, n-butyronitrile, n-valeronitrile, and 2-methoxyacetonitrile, were discovered to be effective in this reaction, affording the corresponding fluorenes in 42–75% yield (6ab-6ae). The steric hindrance of these compounds is manifest as judged by the proton and carbon NMR spectra where the phenylfluorene is desymmetrizated from hindered rotation. Tertiary nitriles, such as isobutyronitrile 2f and cyclohexanecarbonitrile 2g smoothly underwent oxidative C-H activation at the α-position to give 6af and 6ag in 42% and 61% yields, respectively. Notably, these adducts arise from the approach of two hindered tertiary centers and give rise to compounds with two adjacent quaternary centers. However, some other nitriles were unreactive including: cyanocyclopropane, 2-methoxypropionitrile, bromoacetonitrile, and ethyl cyanoacetate.

Some control experiments were carried out to gain a better understanding of the mechanism (Scheme 6). The cyanomethylenation reaction was completely inhibited when 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO) or 2,6-di-tert-butyl-4-hydroxytoluene (BHT) was added into the reaction system (Scheme 6a). Moreover, the corresponding adducts 7 and 8 were detected in the reaction mixture by ESI-MS (see the Supporting Information). In addition, we found the compound 10 was isolated in 33% yield from (1-cyclopropylvinyl)benzene 9 (Scheme 6b). This adduct arises from sequential ring opening of a cyclopropylmethyl radical intermediate and cyclization,^13c and this intermediate presumably arises from addition of a cyanomethylenyl radical to the alkene. Together, the above experiments suggest that the current reaction is triggered by a free-radical process. Moreover, all the experiments point to formation and reaction of a cyanomethylenyl radical. Next, an intermolecular kinetic isotopic effect (KIE) experiment was performed in a mixture of acetonitrile (0.75 mL) and acetonitrile-d₃ (0.75 mL). As a result, a k_H/k_D = 6.7 was obtained (Scheme 6c), indicating that the acetonitrile C-H bond cleavage is involved in a product-determining step.

Scheme 6.

Control Experiments

The lack of adducts from either the fluorene or oxindole with any of the radical traps described above (Scheme 6a–b), implies that these stabilized radicals are less reactive than the cyanomethyl radical. It is likely that the resting state of the fluorenyl or oxindole radicals are the dimers as we^5,25 and others²⁶ have observed previously under oxidative conditions. Integrating the formation of dimer with reports of related systems,^13f,16,27 we propose the mechanism outlined in Scheme 7. First, t-BuOOt-Bu decomposes to give the tert-butoxyl radical (A) at high temperature. The oxindole²⁷ or fluorene undergoes facile hydrogen atom abstraction due to the weak C-H bonds (71 and 72 kcal/mol, respectively)^5b forming tert-butanol and the corresponding radical B which is in equilibrium with its dimer C. Substrates lacking the 9-phenyl groups (e.g. fluorene) were not reactive, presumably due to the greater barrier to formation of the corresponding radical C, consistent with this hypothesis. In addition, the dimers of 1a (C’)^5b and 5a (C)²⁸ both gave rise to the product under the reaction conditions (see SI). At this stage the excess t-BuOOt-Bu may cause the alkyl nitrile (CH bond dissociation energy = 96 kcal/mol)²⁹ to undergo hydrogen atom abstraction to generate the radical. Subsequent recombination with the oxindole or fluorene radical or dimer would generate the product (e.g. 6aa in Scheme 7). Alternately, the dimer (C) may react directly with the nitrile to generate one equivalent of product (6aa) and one equivalent of starting material (5a). Regardless, very hindered forms of the radical B are not expected to be able to react, which was supported by the lack of reactivity with 9(2’-methylphenyl)fluorene.

Scheme 7.

Proposed Mechanism

In summary, we have developed a novel and efficient metal-free method to activate the C(sp³)-H bond of alkyl nitriles for the synthesis of highly functionalized fluorene and oxindole derivatives. Based on the control experiments, the transformation is proposed to proceed via radical process. None of the compounds described herein have been reported previously illustrating the absence of methods to generate such hindered nitrile-derived structures. In particular, there are few examples in the literature of any nitrile derived fluorenes.^18–20 Thus, this method contributes to new chemical space as well as providing a means to generate highly hindered quaternary centers, including compounds with adjacent quaternary/tertiary or quaternary/quaternary centers.