Alkyl Isocyanates via Manganese-Catalyzed C–H Activation for the

Article Cite This: J. Am. Chem. Soc. 2017, 139, 15407-15413

pubs.acs.org/JACS

Alkyl Isocyanates via Manganese-Catalyzed C−H Activation for the Preparation of Substituted Ureas Xiongyi Huang,† Thompson Zhuang,† Patrick A. Kates, Hongxin Gao, Xinyi Chen, and John T. Groves* Department of Chemistry, Princeton University, Princeton, New Jersey 08544, United States S Supporting Information *

ABSTRACT: Organic isocyanates are versatile intermediates that provide access to a wide range of functionalities. In this work, we have developed the first synthetic method for preparing aliphatic isocyanates via direct C−H activation. This method proceeds efficiently at room temperature and can be applied to functionalize secondary, tertiary, and benzylic C−H bonds with good yields and functional group compatibility. Moreover, the isocyanate products can be readily converted to substituted ureas without isolation, demonstrating the synthetic potential of the method. To study the reaction mechanism, we have synthesized and characterized a rare MnIV−NCO intermediate and demonstrated its ability to transfer the isocyanate moiety to alkyl radicals. Using EPR spectroscopy, we have directly observed a MnIV intermediate under catalytic conditions. Isocyanation of celestolide with a chiral manganese salen catalyst followed by trapping with aniline afforded the urea product in 51% enantiomeric excess. This represents the only example of an asymmetric synthesis of an organic urea via C−H activation. When combined with our DFT calculations, these results clearly demonstrate that the C−NCO bond was formed through capture of a substrate radical by a MnIV−NCO intermediate.

■

phosgenation of amines,8 oxidation of organic isonitriles,9 reductive carbonylation of nitro groups,10 and transition metalcatalyzed cross-couplings.11 Of particular interest is the Curtius rearrangement, whose sole purpose is to convert acyl azides into isocyanates to enable further reactivity with nucleophiles.12 This reaction involving the intermediacy of an isocyanate was instrumental in the total synthesis of the key antiviral drug Oseltamivir.13 Needless to say, a method providing direct access to the super reactive isocyanate group based on C−H activation would be highly desirable. Such a method would harness the synthetic utility of isocyanates and the ubiquity of C−H bonds in organic molecules to enable more stepeconomical routes to compounds that are difficult to prepare by conventional methodology.14 Such a transformation would also greatly facilitate the late-stage diversification of organic compounds, providing ready access to a large family of derivatives from a single parent molecule.15 Although the construction of aliphatic C−N bonds via C−H activation has been well established for azide and amine syntheses,16 a method based on direct C−H activation is notably absent from the current repertoire of reactions for isocyanate synthesisa testament to the latter’s difficulty. Very recently, our group developed a series of novel aliphatic C−H functionalization reactions in which complex molecules

INTRODUCTION Organic isocyanates are important synthetic intermediates that have wide applications in organic synthesis, pharmaceuticals, and materials science.1 The broad utility of this class of compounds is built upon the versatile reactivity of the isocyanate group,2 which includes cycloadditions with unsaturated substrates containing CO, CN, CS, CC, or CC bonds to afford various heterocyclic structures, as well as insertions into metal−carbon, metal−hydrogen, metal−nitrogen, and acidic C−H bonds to yield amidation products. Perhaps the most well-known transformations of organic isocyanates are those with nucleophiles such as amines, alcohols, and thiols, which provide an important route to ureas, carbamates, and thiocarbamates and have irreplaceable roles in diverse industrial implementations.1b,c,3 Although the isocyanate group is isoelectronic with the azide functionality, this perceived similarity can be deceptive. In particular, isocyanates are potent carbonyl sources, rendering them highly electrophilic. As such, isocyanates participate in a much broader range of nucleophilic reactions as compared to their azide analogs and, importantly, directly provide a reactive carbonyl handle for ready access to a range of carbonyl-based structures. Given the importance of organic isocyanates, there have been continuous efforts to develop synthetic methods to install the isocyanate functionality.4 Typical methods for isocyanate synthesis include rearrangement reactions with betaxamic acids5 or primary amides,6 decomposition of carbamates,7 © 2017 American Chemical Society

Received: July 21, 2017 Published: October 4, 2017 15407

DOI: 10.1021/jacs.7b07658 J. Am. Chem. Soc. 2017, 139, 15407−15413

Article

Journal of the American Chemical Society containing multiple aliphatic C−H bonds can be selectively functionalized.17 All of these reactions share common mechanistic features in which substrates are activated via hydrogen-atom abstraction by oxo-MnV intermediates and incipient substrate radicals recombine with heteroatom-ligated MnIV−X intermediates (X  OCl, F, or N3) to form C−X bonds (Scheme 1). These successes led us to consider an

Table 1. Screening of Reaction Parameters

Scheme 1. Concept of Isocyanate Synthesis via MnCatalyzed C−H Activation

1 2 3 4 5 6 7 8 9 10 11

catalyst

NCO source

solvent

yielda

Mn(TMP)Cl Mn(TMP)Cl Mn(TMP)Cl Mn(TMP)Cl Mn(TMP)Cl Mn(TMP)Cl Mn(TPFPP)Cl Mn(salen)Cl Mn(TMP)F Mn(TMP)F Mn(TMP)F

NaOCN (aq.) TBA−OCN Me3Si−NCO Me3Si−NCO Me3Si−NCO Me3Si−NCO Me3Si−NCO Me3Si−NCO Me3Si−NCO Me3Si−NCO Me3Si−NCO

CH2Cl2 CH2Cl2 CH2Cl2 CH3CN acetone EtOAc EtOAc EtOAc EtOAc EtOAc EtOAc

5% N.R. 3% trace trace 12% 2% trace 29% 35%b 55%b,c

a

All reported yields were determined by GC-MS. bWith 0.25 eq. AgF additive. cWith 2 eq. PhIO.

analogous route for preparing organic isocyanates via C−H activation. However, simple substitution of isocyanate for azide in our published protocols failed to work. By extending this chemical logic, we have now developed the first synthetic method to generate alkyl isocyanates via C(sp3)−H bond activation. This reaction utilizes trimethylsilyl isocyanate (Me3Si−NCO) as the isocyanate source and is performed under mild conditions. The crude reaction mixture can be used for isocyanate derivatization without further purification. In this way, various N,N′-di- and N,N,N′-trisubstituted ureas were accessed in a one-pot preparation, clearly demonstrating the synthetic potential of the method. In direct contrast to C−H azidation, this is the first and only method for installing the high-value isocyanate group at unactivated aliphatic C−H sites. Furthermore, several prominent features such as the requirement of the unique Me3Si−NCO as isocyanate source, distinguish this reaction from a simple translocation of our previous C−H azidation method.

cantly increased the isocyanation yield to 29%. Addition of 0.25 equiv of AgF was also beneficial to the reaction, further improving yield to 35%. Since only 1 equiv of PhIO oxidant was used in the above conditions, serial addition of PhIO further increased reaction yield. Gratifyingly, addition of 2 equiv of PhIO resulted in a 55% yield of cyclooctyl isocyanate from cyclooctane. Isocyanation of C(sp3)−H Bonds. These reaction-screening results established the basis for the optimized reaction conditions. As shown in Table 2, this method was applied to secondary, tertiary, and benzylic C−H bonds with yields up to 60%. Functional groups including ketones, esters, ethers, and halogens are well tolerated. This method can be applied to compounds containing common structural motifs found in bioactive molecules, such as tetrahydronaphthalene, Indane, tetralone, and adamantane. The

■

RESULTS AND DISCUSSION Reaction Development. We set out to develop an efficient catalytic method for isocyanate synthesis via Mn-catalyzed C− H activation. With cyclooctane as lead substrate, several exploratory reaction conditions were evaluated (Table 1). Initial attempts utilized aqueous sodium cyanate as the isocyanate source, in direct analogy with our previous azidation reaction. Unfortunately, we obtained only a 5% yield and observed substantial decomposition of the isocyanate product over time, presumably due to hydrolysis. We subsequently turned to organic-soluble isocyanate sources and found that trimethylsilyl isocyanate (Me3Si−NCO) resulted in a 3% yield of the desired alkyl isocyanate. Use of ethyl acetate (EtOAc) as solvent improved the yield to a promising 12%. A manganese salen (salen = (R,R)- or (S,S)-(−)-N,N′-bis(3,5-di-tert-butylsalicylidene)-1,2cyclohexanediamino) catalyst was not reactive for cyclooctane and the more electron-withdrawing Mn(TPFPP)Cl catalyst yielded predominantly oxygenation products with only a 2% yield of isocyanate. Chlorinated products were observed with the Mn(TMP)Cl (TMP = 5,10,15,20-tetramesitylporphinato) catalyst, suggesting insufficient ligand exchange between Mn(TMP)Cl and Me3Si-NCO. Changing the precatalyst to Mn(TMP)F completely suppressed chlorination and signifi-

Table 2. Substrate Scope of Isocyanate Synthesis via MnCatalyzed C−H Activationa

a Reactions were performed under N2. PhIO was added in portions (0.35 equiv/portion). Yields were determined by GC-MS. The identities of the isocyanate products were further confirmed by their derivatized ureas. b25 mol % Mn(salen)F used as catalyst. cIsolated yields from gram-scale vacuum distillation.

15408

DOI: 10.1021/jacs.7b07658 J. Am. Chem. Soc. 2017, 139, 15407−15413

Article

Journal of the American Chemical Society manganese salen catalyst was also reactive for substrates containing benzylic C−H bonds, albeit with lower catalytic efficiency. For norbornane, the less sterically hindered exo-2isocyanatonorbornane (1n) was obtained as the predominant product (exo/endo = 7). This large exo/endo ratio suggests the involvement of a catalyst-bound isocyanate intermediate in the isocyanate transfer step.17b When 4-ethyltoluene was used as substrate, isocyanation occurred on the ethyl group with a yield of 44% compared to only 5% at the methyl position (1m). For adamantane, both tertiary and secondary isocyanates (1l) were observed with a 3°/2° ratio of 1.2:1 (3.6:1 when normalized by the number of hydrogen atoms of each type). This ratio is larger than that of C−H fluorination (3°/2° = 1:1.4),17b but smaller than that of C−H azidation (3°/2° = 1.5:1).17d The difference in regioselectivity suggests that the axial ligand can affect the hydrogen-abstraction capability of the oxo-MnV complex. Synthesis of Substituted Ureas via Direct C(sp3)−H Bond Activation. A major advantage of this C−H functionalization reaction is that the organic isocyanate products can be readily derivatized to a variety of functionalities. To demonstrate this potential, we set out to test whether we could trap the isocyanates with amines to afford ureas, which are common motifs found in bioactive molecules including the chemotherapeutic agents sorafenib and lomustine.1b Currently, there is no method for the synthesis of organic ureas through direct activation of inert aliphatic C−H bonds. Due to the instability of organic isocyanates, our initial attempts to isolate the isocyanate products by column chromatography were unsuccessful. Nevertheless, subsequent investigation revealed that the desired organic ureas could be obtained by simply adding the reaction mixture into a toluene solution containing the amine. Furthermore, removing the EtOAc solvent and redissolving the crude reaction mixture in toluene prior to amine coupling substantially increased urea yields. Using this protocol, we successfully obtained a variety of substituted ureas through direct C−H activation. As shown in Table 3, the conversion of isocyanates to ureas ranged from 65% to 95%, consistent with yields previously reported for reactions between aliphatic isocyanates and amines.9b,18 Both N,N′-di- and N,N,N′-trisubstituted ureas were readily formed in comparable yields. A diverse range of nucleophilic amines were tolerated, including primary and secondary amines and amines bearing heteroaromatic groups (2i). The urea derivative of the common anti-inflammatory drug ibuprofen was obtained in 30% overall yield (2k). For the complex molecule methyl O-methylpodocarpate, which has multiple aliphatic C−H bonds, the weaker benzylic C−H bond was selectively activated and afforded the urea product in 26% isolated yield (2q). Some of the urea products shown in Table 3 have already been studied as bioactive compounds. For example, the adamantyl urea derivative 2r is a potent inhibitor of soluble epoxide hydrolase (sEH) with demonstrated efficacy for the treatment of hypertension and type 2 diabetes.19 Ordinarily, synthesis of 2r uses commercially available 1adamantyl isocyanate as the starting material.19c Using our method, we were able to simultaneously obtain both 1- and 2adamantyl isocyanate, which is not readily commercially available, and their ureas in a parallel, one-pot reaction. Further, norbornyl urea derivatives have also shown promise as antihypertensive agents and 2,3-dihydro-1-indenyl urea is the core structure of a series of vanilloid receptor subtype 1 (VR1) inhibitors.20 These results clearly demonstrate the significant

Table 3. Synthesis of Substituted Ureas by Trapping Isocyanates with Aminesa

a

Conditions: EtOAc was removed and reaction mixture was redissolved in toluene and slowly added into a toluene solution containing the amine. bOverall yield reported includes both 3° and 2° isomers in ratio of 0.9:1. Conversions of isocyanates to ureas are reported with final isolated yields of ureas shown in parentheses.

potential of our method for pharmaceutical discovery and synthesis. Manganese-Catalyzed Isocyanation vs Azidation. Subsequently, we conducted a number of experiments to probe the isocyanation reaction mechanism. Although use of azide anion allowed for Mn-catalyzed C−H azidation in our previous study,17d use of cyanate anion for this reaction in either aqueous- or organic-soluble forms (Table 1, entries 1 and 2) resulted in low or no yields, highlighting an important difference between these strategies. To examine the nature of this difference, we titrated the N3− and NCO− aqueous anions against MnIII(TMP)X (X = Cl or F) precatalyst. With a solution of Mn(TMP)Cl in EtOAc, we obtained Kd values of 0.19 mM and 0.39 mM for aqueous N3− and NCO−, respectively (Table 4 and Figure S1 of the Supporting Information, SI). Interestingly, despite the affinity of isocyanate anion for Mn(TMP)Cl, use of NaOCN as isocyanate source resulted in low yields, presumably due to the known tendency of cyanate anions to hydrolyze in aqueous media.21 In stark contrast, [Me3Si−NCO]50, denoting the Me3Si− NCO concentration at which half of the precatalyst has been converted to Mn(TMP)NCO, was measured to be 38 mM (Table 4). This very poor conversion of Mn(TMP)Cl to Mn(TMP)NCO accounts for our observation of chlorinated 15409

DOI: 10.1021/jacs.7b07658 J. Am. Chem. Soc. 2017, 139, 15407−15413

Article

Journal of the American Chemical Society Table 4. Ligand Affinities for Mn(TMP)Cl

1 2 3

Scheme 2. Monophasic Manganese-Catalyzed C−H Azidation Under Isocyanation Conditions

ligand source X

Kd

NaN3 (aq.) NaOCN (aq.) Me3Si−NCO

0.19 mM 0.39 mM 38 mMa

Scheme 3. (a) Dependence of Regioselectivity on Catalyst Ligand Structure. (b) KIE Study of the Reaction

[Me3Si−NCO]50 is shown. Conditions: 5 μM Mn(TMP)Cl in 2 mL EtOAc solvent. Ligands were solvated in aqueous (NaN3 and NaOCN) or EtOAc (Me3Si−NCO) solvent. Final ligand concentrations varied from 1 nM to 0.5 M. a

products with Mn(TMP)Cl as catalyst (Table 1, entry 6). Moreover, the difference of 2 orders of magnitude between these values for NCO− aqueous anion and Me3Si−NCO strongly suggests that the latter does not act as a simple source of isocyanate anion, in contrast to the role of azide anion in manganese-catalyzed C−H azidation. Instead, we hypothesize that Me3Si-NCO participates directly in ligand transfer to the metal center and acts as a reservoir of protected isocyanate for controlled release, which is crucial to the success of this method. With a solution of Mn(TMP)F in EtOAc, [Me3Si-NCO]50 was found to be in the single micromolar range, indicating much more facile ligand transfer and suggesting that MnIII− NCO complex formation is driven thermodynamically by the formation of a strong Si−F bond. This also suggests a rationale for the unique reactivity of Me3Si−NCO. That is, the MnIV− OH complex formed after hydrogen-atom transfer can readily undergo ligand exchange with Me3Si-NCO to form the competent MnIV−NCO rebound intermediate, a process driven by the formation of a very strong Si−O bond. Indeed, when using NaOCN as the isocyanate source under our reaction conditions, the major products observed were oxygenation products deriving from rebound to a MnIV−OH intermediate. Upon substitution of Me3Si−NCO as isocyanate source, the oxygenation-to-isocyanation ratio reversed to favor isocyanation products, indicating predominant radical rebound to a MnIV−NCO complex. Our previous manganese-catalyzed C−H azidation protocol necessitated a biphasic (aqueous and organic) reaction mixture to facilitate dissolution of solid sodium azide.17d Although that reaction achieved high yields and selectivity, discovery of a single-phase protocol is desirable from a scalability and process perspective. Applying these new insights regarding the facility of ligand exchange with trimethylsilyl species, we attempted to utilize Me3Si−N3 as azide source under conditions similar to those of our isocyanation reaction. Using those conditions, we obtained only a 9% yield of cyclooctyl azide without further optimization efforts (Scheme 2). Current studies are now directed toward optimizing this monophasic reaction for C−H azidation. Mechanistic Studies. We then studied the initial C−H abstraction step by probing its regioselectivity using adamantane. The 3°/2° ratio of isocyanate products increased from 1.2 to 1.9 when changing the precatalyst to less-sterically hindered Mn(DTBPP)F (Scheme 3a). The dependence of regioselectivity on catalyst ligand structure suggests the involvement of a

catalyst-based intermediate, presumably an oxo-MnV species, in hydrogen abstraction. Oxo-MnV involvement was further supported by the large competitive intermolecular kinetic isotope effect (KIE; 4.5) observed for the isocyanation of a 1:1 mixture of cyclooctane and cyclooctane-d16 (Scheme 3b). A putative MnIV intermediate with at least one ligating isocyanate is generated after C−H activation. The feasibility of isocyanate transfer from this MnIV−NCO intermediate to alkyl radicals to afford alkyl isocyanates is pivotal to the success of this reaction. Previously, we synthesized a rare trans-difluoromanganese(IV) complex, MnIV(TMP)F2, and demonstrated its capacity for transferring a fluorine atom to alkyl radicals to yield fluorinated products.17b Here, a similar trans-diisocyanatomanganese(IV) complex, MnIV(salen)(NCO)2, was successfully synthesized by ligand exchange between MnIV(salen)F2 and Me3Si−NCO (Figure S2). Its crystal structure showed that both isocyanate ligands are coordinated to manganese via nitrogen with Mn−N bond lengths of 1.944 Å and Mn−N−C angles of 138.6° and 127.7° (Figure 1a). The Mn−N bond lengths are slightly longer than those reported for a previously characterized MnIV−NCO complex with a porphyrin ligand (1.918 and 1.934 Å).22

Figure 1. (a) X-ray crystal structure of trans-MnIV(salen)(NCO)2 drawn at 50% probability with selected bond lengths (Å) and angles (deg). H atoms are omitted for clarity. (b) Trapping of alkyl radicals by MnIV(salen)(NCO)2 through isocyanate transfer. Yield based on MnIV(salen)(NCO)2. 15410

DOI: 10.1021/jacs.7b07658 J. Am. Chem. Soc. 2017, 139, 15407−15413

Article

Journal of the American Chemical Society With this well-characterized MnIV(salen)(NCO)2 in hand, we turned to testing its ability to transfer its isocyanate ligand to alkyl radicals. A solution containing MnIV(salen)(NCO)2 in toluene and the radical precursor tert-butyl-2-phenylpropaneperoxoate was irradiated with UV light at room temperature (Figure 1b). The reaction was completed within 10 min as confirmed by the disappearance of the MnIV absorption in the UV−vis spectrum. GC-MS analysis revealed a 20% yield of the isocyanate product. This result clearly demonstrates the feasibility of isocyanate transfer from MnIV−NCO complexes to alkyl radicals. After the successful synthesis and characterization of these MnIV species, we turned to the direct detection of MnIV intermediates under catalytic conditions. By freezing the isocyanation reaction mixture in liquid nitrogen, we were able to directly observe the Mn species in situ via electron paramagnetic resonance (EPR) spectroscopy at 10 K (Figure 2). Furthermore, we synthesized the MnIV(TMP)(NCO)2

5/2 55 Mn nucleus was resolved for the authentic compound at both g = 3.93 and g = 2.01. The manganese porphyrin-catalyzed isocyanation reaction mixture (Figure 2a, red trace) contains similar features and g values (3.91, 2.01) as compared to both the authentic MnIV(TMP)(NCO)2 and the previously reported MnIV(TPP)(NCO)2,22 which is consistent with the intermediacy of a MnIV(TMP)(NCO)2 complex during the isocyanation reaction. The spectral features at g = 3.91 and 2.01 are diagnostic of the existence of a high-spin, d3 MnIV species with an axially symmetric environment.22,23 Hyperfine splitting of the I = 5/2 55 Mn nucleus of the catalyst is clear at g = 2.01, while the six small features at g = 3.91 can be resolved with derivatives (Figure S3). In addition, the manganese salen-catalyzed isocyanation reaction mixture (salen = (R,R)- or (S,S)-(−)-N,N′-Bis(3,5di-tert-butylsalicylidene)-1,2-cyclohexanediamino) also displays an EPR spectrum with g values of 5.56, 2.67, and 1.70, implicating the intermediacy of a MnIV complex in manganese salen-catalyzed isocyanation (Figure 2b). These g values are also characteristic of a d3 MnIV with S = 3/2.24 The signals at g = 5.56, 2.67, and 1.70 arise from the ms = ± 1/2 Kramer doublets of a S = 3/2 spin system with rhombic symmetry. The signal at g = 5.56 also displays hyperfine splitting due to the I = 5/2 55 Mn nucleus. The spectral features due to the presence of a d3 MnIV center in our reaction contain marked similarities to those found in a previously characterized nitrogen-bound MnIV(salen)(NO3)2, which has g values of 5.04, 2.96, and 1.83.25 DFT Calculations of Isocyanate Rebound. We further studied the isocyanate transfer step with DFT calculations using an unsubstituted MnIV−NCO salen compound and tolyl radical as model reactants (Figure 3). The MnIV−NCO complex was

Figure 2. (a) X-band EPR spectra of manganese porphyrin-catalyzed isocyanation reaction mixture (red) and authentic MnIV(TMP)(NCO)2 (gray) in EtOAc. (b) X-band EPR spectrum of manganese salen-catalyzed isocyanation reaction mixture. Spectrometer parameters: temperature (10 K), microwave frequency (9.4 GHz), microwave power (2 mW). See SI for experimental details.

Figure 3. DFT calculations for (top) MnIV−NCO/OCN complexes and (bottom) isocyanate transfer step (kcal/mol at 298 K).

found to be 10 kcal/mol more stable than the MnIV−OCN complex, which is consistent with our crystal structure in which only MnIV−NCO binding was observed. There are four possible pathways through which the tolyl radical can interact with the MnIV−NCO rebound intermediate: attacking either the nitrogen or oxygen atoms of the MnIV−NCO complex on either triplet or quintet energy surfaces. As shown in Figure 3, the lowest energy barriers were found on quintet surfaces. Isocyanate transfer through oxygen (Oquintet) and nitrogen (Nquintet) show comparable energy barriers. However, the Nquintet pathway is more thermodynamically favored than the Oquintet pathway by 25 kcal/mol, due to the instability of organic cyanates. Indeed, aliphatic cyanates are

complex using literature methods to facilitate direct comparison.17b,22 Both EPR spectra exhibit features characteristic of high-spin d3 complexes, thus confirming the oxidation state of Mn in both samples as MnIV (Figure 2a). The authentic MnIV(TMP)(NCO)2 complex (gray) has g values (3.93, 2.01) close to those reported for a similar MnIV(TPP)(NCO)2 complex (g values of 3.92 and 2.01; TPP = 5,10,15,20tetraphenylporphinato),22 confirming the successful synthesis of the di-isocyanate compound. Hyperfine coupling of the I = 15411

DOI: 10.1021/jacs.7b07658 J. Am. Chem. Soc. 2017, 139, 15407−15413

Article

Journal of the American Chemical Society known to spontaneously isomerize into isocyanates.26 Alkyl isocyanates have also been previously shown to be 20−27 kcal/ mol more stable than the corresponding cyanates.27 These results are consistent with our experimental observations, as no organic cyanate products were observed. The Nquintet transition state showed a C−N bond distance of 2.64 Å and Mn−N− Cbenzylic angle of 103.4°. The tolyl radical is in close proximity to the 3 position of the salen phenyl group (3.19 Å), which suggests the plausibility of asymmetric induction if a chiral salen catalyst and pro-chiral substrate is employed. These mechanistic studies and our observation of MnIV intermediates under catalytic conditions provide direct experimental support for a heteroatom-rebound catalysis (HRC) reaction mechanism (Scheme 4), in which the substrate Scheme 4. Proposed Reaction Mechanism

Figure 4. (a) Asymmetric synthesis of substituted ureas. (b) Linear relationship between ee of catalyst and ee of product. The product detected by chiral HPLC is the organic urea resulting from trapping the isocyanate product with aniline. (R,R)-(−)-N,N′-Bis(3,5-di-tertbutylsalicylidene)-1,2-cyclohexanediaminomanganese(III) fluoride catalyst was employed as the excess enantiomer. R2 = 0.99.

■

CONCLUSIONS In summary, we have developed the first method for synthesizing organic isocyanates via direct aliphatic C−H activation. The reaction employs Me3Si−NCO as an isocyanate source and operates at room temperature. The isocyanate products can be readily converted into substituted ureas without isolation. Furthermore, a 51% ee was detected in the urea product of celestolide when a chiral manganese salen catalyst was employed, representing the first example of an enantioselective synthesis of an organic urea via direct C−H activation. Our mechanistic and EPR studies directly implicate the involvement of a MnIV-bound rebound intermediate in our reaction. Therefore, we postulate the intermediacy of a reactive trans-diisocyanato-manganese(IV) complex in C−H isocyanation. We then successfully synthesized a rare MnIV(salen)(NCO)2 complex and demonstrated its capacity for isocyanate transfer to alkyl radicals. These results contribute to our mechanistic understanding of heteroatom-rebound catalysis and fully demonstrate the premise of HRC as a general paradigm for the development of novel C−H functionalization reactions.

is activated via hydrogen atom abstraction by an oxo-MnV intermediate. The incipient substrate radical is then captured by a MnIV−NCO intermediate to form the isocyanate product and regenerate the catalyst. Asymmetric Synthesis of Substituted Ureas. By employing a manganese catalyst with a chiral salen ligand, (R,R)-(−)-N,N′-Bis(3,5-di-tert-butylsalicylidene)-1,2-cyclohexanediamine, we observed a 50.7 ± 0.6% enantiomeric excess (ee) in the final urea product derived from celestolide (Figure 4a). This represents the first and only example of an enantioselective synthesis of a urea via direct aliphatic C−H activation. This significant enantiomeric excess also implicates the involvement of a Mn-bound isocyanate intermediate in the isocyanate transfer step. In order to gain further insight into the enantioinductive isocyanate rebound step, we attempted to correlate the ee of the chiral salen ligand with the ee of the final urea product. By varying the enantiomeric composition of the manganese salen catalyst, we observed a clear linear effect in the Kagan plot for the isocyanation of celestolide (Figure 4b), indicating that a monomeric MnIV species is operant in the stereoinductive rebound step.28 The species that transfers isocyanate is most likely the MnIV(salen)(NCO)2 complex we synthesized above, whose capacity for performing isocyanate transfer to a carbon-centered radical was earlier demonstrated. This study also suggests that the enantioselectivity of our C−H isocyanation method can be tuned by modulating the properties of the asymmetric ligand, which is solely responsible for asymmetric induction. Such efforts are currently under way in our laboratory.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b07658. Detailed experimental procedures, spectroscopic data for all new compounds, and details for DFT calculation (PDF) Crystal structure data (CIF)

■

AUTHOR INFORMATION

Corresponding Author

*[email protected] 15412

DOI: 10.1021/jacs.7b07658 J. Am. Chem. Soc. 2017, 139, 15407−15413

Article

Journal of the American Chemical Society ORCID

(19) (a) Anandan, S.-K.; Webb, H. K.; Chen, D.; Wang, Y.-X.; Aavula, B. R.; Cases, S.; Cheng, Y.; Do, Z. N.; Mehra, U.; Tran, V.; Vincelette, J.; Waszczuk, J.; White, K.; Wong, K. R.; Zhang, L.-N.; Jones, P. D.; Hammock, B. D.; Patel, D. V.; Whitcomb, R.; MacIntyre, D. E.; Sabry, J.; Gless, R. Bioorg. Med. Chem. Lett. 2011, 21, 983. (b) Chen, D.; Whitcomb, R.; MacIntyre, E.; Tran, V.; Do, Z. N.; Sabry, J.; Patel, D. V.; Anandan, S. K.; Gless, R.; Webb, H. K. J. Clin. Pharmacol. 2012, 52, 319. (c) Jones, P. D.; Tsai, H.-J.; Do, Z. N.; Morisseau, C.; Hammock, B. D. Bioorg. Med. Chem. Lett. 2006, 16, 5212. (20) (a) Moriguchi, S.; Kimura, Y.; Yokoo, H.; Furutoku, I.; Ebisawa, H. Japanese Patent JPS6023313(A), Feb. 2, 1985. (b) Gomtsyan, A.; Bayburt, E. K.; Koenig, J. R.; Lee, C.-H. U.S. Patent 2004/0254188A1, Dec 16, 2004. . (21) Wen, N.; Brooker, M. H. Can. J. Chem. 1994, 72, 1099. (22) Camenzind, M. J.; Hollander, F. J.; Hill, C. L. Inorg. Chem. 1983, 22, 3776. (23) Konishi, S.; Hoshino, M.; Imamura, M. J. Phys. Chem. 1982, 86, 4537. (24) (a) Kurahashi, T.; Kikuchi, A.; Tosha, T.; Shiro, Y.; Kitagawa, T.; Fujii, H. Inorg. Chem. 2008, 47, 1674. (b) Kurahashi, T.; Fujii, H. Inorg. Chem. 2008, 47, 7556. (c) Bryliakov, K. P.; Babushkin, D. E.; Talsi, E. P. J. Mol. Catal. A: Chem. 2000, 158, 19. (d) Adam, W.; MockKnoblauch, C.; Saha-Möller, C. R.; Herderich, M. J. Am. Chem. Soc. 2000, 122, 9685. (e) Campbell, K. A.; Lashley, M. R.; Wyatt, J. K.; Nantz, M. H.; Britt, R. D. J. Am. Chem. Soc. 2001, 123, 5710. (25) Kurahashi, T.; Hada, M.; Fujii, H. J. Am. Chem. Soc. 2009, 131, 12394. (26) (a) Jensen, K. A.; Due, M.; Holm, A. Acta Chem. Scand. 1965, 19, 438. (b) Jensen, K. A.; Due, M.; Holm, A.; Wentrup, C. Acta Chem. Scand. 1966, 20, 2091. (27) (a) Pasinszki, T.; Westwood, N. P. C. J. Phys. Chem. 1995, 99, 1649. (b) Pasinszki, T.; Havasi, B.; Kovacs, A. J. Phys. Chem. A 2003, 107, 1720. (28) (a) Puchot, C.; Samuel, O.; Dunach, E.; Zhao, S.; Agami, C.; Kagan, H. B. J. Am. Chem. Soc. 1986, 108, 2353. (b) Girard, C.; Kagan, H. B. Angew. Chem., Int. Ed. 1998, 37, 2922.

John T. Groves: 0000-0002-9944-5899 Author Contributions †

These authors contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research was supported by the US National Science Foundation award CHE-1464578. X.H. thanks the Howard Hughes Medical Institute for fellowship support. The authors thank Prof. A. G. Doyle for access to chiral HPLC and Dr. Phil Jeffrey for assistance in crystal data collection.

■

REFERENCES

(1) (a) Six, C.; Richter, F. In Ullmann’s Encyclopedia of Industrial Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2000. (b) Gallou, I. Org. Prep. Proced. Int. 2007, 39, 355. (c) Hensarling, R. M.; Rahane, S. B.; LeBlanc, A. P.; Sparks, B. J.; White, E. M.; Locklin, J.; Patton, D. L. Polym. Chem. 2011, 2, 88. (2) (a) Saunders, J. H.; Slocombe, R. J. Chem. Rev. 1948, 43, 203. (b) Braunstein, P.; Nobel, D. Chem. Rev. 1989, 89, 1927. (c) Hesp, K. D.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc. 2011, 133, 11430. (d) Allen, A. D.; Tidwell, T. T. Chem. Rev. 2013, 113, 7287. (e) Hummel, J. R.; Ellman, J. A. Org. Lett. 2015, 17, 2400. (3) Delebecq, E.; Pascault, J. P.; Boutevin, B.; Ganachaud, F. Chem. Rev. 2013, 113, 80. (4) Kreye, O.; Mutlu, H.; Meier, M. A. R. Green Chem. 2013, 15, 1431. (5) (a) Yale, H. L. Chem. Rev. 1943, 33, 209. (b) Bauer, L.; Exner, O. Angew. Chem., Int. Ed. Engl. 1974, 13, 376. (6) Wolff, M. E. Chem. Rev. 1963, 63, 55. (7) (a) Butler, D. C. D.; Alper, H. Chem. Commun. 1998, 2575. (b) Uriz, P.; Serra, M.; Salagre, P.; Castillon, S.; Claver, C.; Fernandez, E. Tetrahedron Lett. 2002, 43, 1673. (8) Slocombe, R. J.; Hardy, E. E.; Saunders, J. H.; Jenkins, R. L. J. Am. Chem. Soc. 1950, 72, 1888. (9) (a) Feuer, H.; Rubinstein, H.; Nielsen, A. T. J. Org. Chem. 1958, 23, 1107. (b) Le, H. V.; Ganem, B. Org. Lett. 2011, 13, 2584. (10) Paul, F. Coord. Chem. Rev. 2000, 203, 269. (11) Vinogradova, E. V.; Fors, B. P.; Buchwald, S. L. J. Am. Chem. Soc. 2012, 134, 11132. (12) (a) Shioiri, T.; Yamada, S.; Ninomiya, K. J. Am. Chem. Soc. 1972, 94, 6203. (b) Scriven, E. F. V.; Turnbull, K. Chem. Rev. 1988, 88, 297. (13) Ishikawa, H.; Suzuki, T.; Hayashi, Y. Angew. Chem., Int. Ed. 2009, 48, 1304. (14) Cernak, T.; Dykstra, K. D.; Tyagarajan, S.; Vachal, P.; Krska, S. W. Chem. Soc. Rev. 2016, 45, 546. (15) Hartwig, J. F. J. Am. Chem. Soc. 2016, 138, 2. (16) (a) Sharma, A.; Hartwig, J. F. Nature 2015, 517, 600. (b) Davies, H. M. L.; Long, M. S. Angew. Chem., Int. Ed. 2005, 44, 3518. (c) Davies, H. M. L. Angew. Chem., Int. Ed. 2006, 45, 6422. (d) Che, C.-M.; Lo, V. K.-Y.; Zhou, C.-Y.; Huang, J.-S. Chem. Soc. Rev. 2011, 40, 1950. (e) Roizen, J. L.; Harvey, M. E.; Du Bois, J. Acc. Chem. Res. 2012, 45, 911. (17) (a) Liu, W.; Groves, J. T. J. Am. Chem. Soc. 2010, 132, 12847. (b) Liu, W.; Huang, X. Y.; Cheng, M. J.; Nielsen, R. J.; Goddard, W. A.; Groves, J. T. Science 2012, 337, 1322. (c) Huang, X.; Liu, W.; Ren, H.; Neelamegam, R.; Hooker, J. M.; Groves, J. T. J. Am. Chem. Soc. 2014, 136, 6842. (d) Huang, X. Y.; Bergsten, T. M.; Groves, J. T. J. Am. Chem. Soc. 2015, 137, 5300. (e) Liu, W.; Groves, J. T. Acc. Chem. Res. 2015, 48, 1727. (18) (a) Carnaroglio, D.; Martina, K.; Palmisano, G.; Penoni, A.; Domini, C.; Cravotto, G. Beilstein J. Org. Chem. 2013, 9, 2378. (b) Thalluri, K.; Manne, S. R.; Dev, D.; Mandal, B. J. Org. Chem. 2014, 79, 3765. (c) Spyropoulos, C.; Kokotos, C. G. J. Org. Chem. 2014, 79, 4477. 15413

DOI: 10.1021/jacs.7b07658 J. Am. Chem. Soc. 2017, 139, 15407−15413