Reagent-dictated site selectivity in intermolecular aliphatic C–H functionalizations using nitrogen-centered radicals

Intermolecular aliphatic C–H functionalizations using nitrogen-centered radicals with reagent-dictated site selectivities are described, including reactions targeting primary C–H bonds across a range of simple and complex substrates.


Introduction
Recent advances in intermolecular aliphatic C-H functionalization have provided unique tools for the synthesis and latestage derivatization of complex molecules. 1 The site selectivity of these reactions is critical to their application owing to the number of different C-H bonds present in many substrates. The intrinsic reactivity of the substrate C-H bonds oen dictates the site selectivity of a given functionalization. 2 Among the factors inuencing the relative reactivity of C-H bonds are steric accessibility, electron richness, and participation in hyperconjugation. Sole reliance on substratedictated selectivity limits the scope of intermolecular C-H functionalization, as many substrates contain C-H bonds with similar reactivity, thus affording poor discrimination between sites. Moreover, functionalizations at less inherently reactive C-H bonds are impossible without additional elements of control.
Reagents or catalysts that override inherent substratedictated site selectivity hold signicant promise in expanding the capabilities of intermolecular C-H functionalization. Noteworthy advances in this area have been achieved via the use of modied ligand architectures in catalytic C-H functionalizations proceeding via high-valent metal oxo or metal carbenoid intermediates. 3 Protein engineering has also led to enzymatic aliphatic C-H functionalizations with altered regioselectivities. 4 We have previously developed practical systems for aliphatic C-H functionalization using readily accessed  In those studies, moderate site selectivities were achieved primarily through substrate control. We anticipated that the highly modular nature of our reagents would offer unique opportunities in introducing reagent-dictated site selectivity (Fig. 1). Herein, we report our studies towards this goal using an easily prepared, modied amide to provide high levels of site selectivity in C-H functionalizations of diverse substrates. Computational studies of both the amidyl radical structures and representative C-H abstraction transition states were performed to shed light on the unique, reagent-dictated site selectivities observed.

Results and discussion
We commenced our studies with the functionalization of trans decalin, a benchmark substrate used in studies of intermolecular, aliphatic C-H functionalizations. 6 The difficulty of achieving steric discrimination between the methylene sites is evident upon surveying the site selectivities of several previously reported functionalizations (Table 1, entries 1-4). For example, iron-catalyzed oxidations of trans decalin with bulky tetradentate ligands poorly discriminated between these sites (entries 1 and 2). 6b Manganese-catalyzed C-H chlorination using a hindered tetramesitylporphyrin (TMP) system achieved good selectivity (4.0 : 1), although tertiary functionalization was also observed (entry 3). 6c The C-H chlorination using our reported N-chloroamide 3 proceeded with no selectivity between the methylene sites (entry 4). 5b We hypothesized that increasing the steric demand of the Nchloroamide would offer improved site selectivities. We thus evaluated a set of N-chloroamides containing ortho tri-uoromethyl substitution of the benzamide, which are readily available on scale (>10 grams). Reagents 4 and 5 containing a single ortho substituent increased the selectivity for the less hindered, distal methylene site, although the ratio of 1 : 2 remained close to 2.0 : 1 (entries 5 and 6). We next sought to prepare the highly hindered 2,6-bistriuoromethyl reagent 6, which was obtained via addition of the aryl Grignard to tBu isocyanate. 7 The C-H chlorination of trans decalin with 6 proceeded with higher site selectivity than the other N-chloroamides or metal-catalyzed systems, with a 6.5 : 1 ratio of products (entry 7). Performing the reaction with 1 equiv. of substrate provided the products in a 7.2 : 1 ratio in 76% combined yield (entry 8). To the best of our knowledge, this is the most site-selective secondary C-H functionalization of trans decalin known using substrate as limiting reagent. [8][9][10] We next evaluated the site selectivities of the C-H chlorination using reagent 6 across a range of substrates. As shown in Fig. 2, reagent 6 is a free-owing white powder that is readily available on scale. The C-H chlorination of these substrates with N-chloroamide 6 also proceeded with unique site selectivities, and notably includes radical functionalizations selective for primary sites. Site selectivities with our rst-generation reagent 3 are shown for direct comparison. For example, the chlorination of 3-methylpentane (7) with reagent 3 showed a preference for the 2-position (64%) over the 1-position (27%). However, reagent 6 provided a remarkable reagent-dictated switch in site selectivity, favoring the 1-chloro isomer (69%, 69% combined yield). Free radical species demonstrating such selectivity for the strongest, primary C-H bond over secondary or tertiary C-H bonds in synthetic alkane functionalizations are unknown to the best of our knowledge. 11 Our efforts continued with the C-H chlorination of diverse compounds (Fig. 2). The chlorination of 2,2-dimethylpentane (8) and 2,2-dimethylbutane (9) demonstrate that fully substituted carbon centers enable steric discrimination between C-H sites, with 89% primary C-H selectivity in the reaction of substrate 9 with N-chloroamide 6. There was an increase in the level of primary functionalization of n-hexane (10) with reagent 6, however the 2-position remained the preferred site of functionalization.
The presence of electron-withdrawing functionality disfavors reaction at proximal methylene sites and increased the site selectivity for the primary positions to 88% and 89% for substrates 11 and 12, respectively. The functionalization of 2phenylbutyl benzoate (13) remarkably demonstrated a high selectivity for the primary position (89%) even in the presence of a benzylic C-H site. The fully substituted carbon centers in acetal 14 and ester 15 lead to a clear reagent-dictated change in selectivity between reagents 3 and 6; the primary C-H chlorination product is favored with up to 75% selectivity with ester substrate 15.
Phthalimide-protected amines undergo intermolecular C-H functionalizations using amidyl radicals with high levels of electronically-dictated site selectivities. 5 We surveyed several substrates containing this electron-withdrawing group. With the lack of major steric discrimination between the C-H sites, the functionalization of substrate 16 with reagent 6 only increased the level of primary functionalization to a modest 33%. In the case of substrate 17 containing chain branching, a higher level of primary functionalization was observed (45%). The shorter aliphatic chain in the norleucine derivative 18 electronically disfavored methylene functionalization, and the Reactions of the N-chloroamides were performed in PhCF 3 at 65 C using 10 mol% benzoyl peroxide as initiator with 5 equiv. of substrate, 0.5 equiv. of Cs 2 CO 3 , and 1 equiv. N-chloroamide. Yields and selectivities of the methylene functionalizations were determined by 1 H NMR analysis, and those of the tertiary products were determined by GC with dodecane as the internal standard. b X ¼ Cl unless otherwise noted. c Reaction performed using 1 equiv. of substrate.
use of hindered N-chloroamide 6 resulted in good primary selectivity (80%). The reagent-dictated site selectivity provided by N-chloroamide 6 is also evident in reactions with cyclic substrates 19-21. Each of these structures contain eight methylene and either three or six methyl C-H bonds for chlorination. In all cases C-H chlorination with N-chloroamide 3 was highly methylene selective. Switching to reagent 6 for the chlorination of these substrates signicantly increased the levels of methyl group functionalization. This unique selectivity prole enabled the 1 C-H functionalizations of methyl groups in complex substrates (vide infra).
We next sought to capitalize on the unique selectivity prole of reagent 6 in the aliphatic C-H functionalization of complex natural products (Fig. 3). The late-stage functionalization of the antimalarial artemisinin has been targeted in several studies. 3b,6e,12,13 Most commonly, the reactive site is the methine of the methyl-bearing carbon in the cyclohexyl ring. Recently, groundbreaking studies from the White group (catalyst-dictated selectivity) 3b and the Fasan group (enzyme-dictated selectivity) 13 have achieved transformations capable of overcoming this reactivity prole.
We questioned whether the use of reagent 6 would enable C-H functionalization of the methyl group of the cyclohexyl ring. To date, only enzymatic catalysts have successfully hydroxylated this methyl site. 13 In the event, chlorination with N-chloroamide 6 provided the desired product with moderate site selectivity (1.9 : 1) in 59% combined yield (Fig. 3). The C-H chlorinations of cedryl acetate and tetrahydrolinalool benzoate also demonstrated the ability of reagent 6 to target primary C-H sites. Switching from reagent 3 to reagent 6 brings about a signicant increase in methyl functionalization, most notably with 89% primary selectivity in the functionalization of 24.
In prior work, we have developed protocols for C-H bromination and xanthylation using functionalized amides. 5a,c As these transformations are likewise synthetically valuable, 14 we sought to demonstrate that the reagent-dictated selectivity herein was transferrable to those reactions. We prepared the Nbromo and N-xanthylamides 25 and 27 to address this question. The C-H functionalizations of trans decalin with these reagents provided similar methylene site selectivities and reaction efficiencies as compared to the chlorination (Fig. 4). 15 Thus, tuning the amidyl radical responsible for dictating the site selectivity of the C-H abstraction is a general approach to introducing reagent control to this class of C-H functionalizations.
The structure of the N-chloroamide 6 provides some insight into the basis for increased site selectivity (Fig. 5). 16 Non-bonded interactions between the ortho triuoromethyl groups and the amide bond dictate a conformation where the aromatic plane and the amide are nearly orthogonal. This is clear from the C 6 C 1 C 7 O 1 dihedral angle, which is 96.6 . The ortho substituents positioned above and below the plane of the amide are expected to shield the nitrogen atom and the resulting amidyl radical, leading to an increased selectivity for sterically less hindered C-H sites. This unique feature of reagent 6 is not present in Nchloroamides 3-5, which have more planar structures.
In order to gain further insight regarding the observed site selectivities, we performed computational studies of the Reactions were performed in PhCF 3 at 65 C using (2 Â 10 mol%) benzoyl peroxide as initiator with 1 equiv. of substrate, 0.5 equiv. of Cs 2 CO 3 , and 1-1.5 equiv. N-chloroamide. Yields and selectivities were determined by GC with dodecane as an internal standard or 1 H NMR analysis. See ESI † for further details regarding the distribution of minor products. a 3 equiv. of substrate used. b 5 equiv. of substrate used. c Secondary site selectivity includes all secondary chlorination products. structures of the amidyl radicals involved, as well as the hydrogen atom abstractions of select substrates. The relevant transition states of the H-atom abstractions were optimized, and the associated energy changes were evaluated. A DFT approach was used, with the uB97XD functional, standard 6-31G(d,p) basis set, and Gaussian16 soware (see ESI † for details). Optimizations were initially carried out in vacuo, then solvent effects were included through single-point calculations at the same level of theory with the SMD model and adopting the standard parameters for CH 2 Cl 2 (a computationally convenient alternative to PhCF 3 , 17 the solvent used in this work).
The calculated structures of amidyl radicals 3 0 and 6 0 are shown in Fig. 6, along with their spin density plots and some relevant parameters. Both species are essentially N-centered radicals; the nitrogen atoms have sp 2 -like hybridization, and the unpaired electron lies in a pure p orbital perpendicular to the plane consisting of the substituents attached to the nitrogen atom. The presence of the tert-butyl group on the nitrogen atom is known to impart a twisted structure to amidyl radicals, decreasing the conjugative delocalization of the unpaired electron from N to O. 18 This is also apparent in the case of amidyl radicals 3 0 and 6 0 , where the OC 7 NC 10 dihedral angles markedly deviate from zero. The environments around the radical centers are also signicantly affected by the substitution patterns of the phenyl ring, and are markedly different for the two amidyl radicals, however. For example, the aromatic ring of 3 0 is fully conjugated with the amide carbonyl, and they lie in nearly the same plane (only a 7 deviation is observed). On the other hand, the aromatic ring of 6 0 is out of the carbonyl plane (54 dihedral angle) owing to the steric hindrance of the two ortho CF 3 substituents. As a result, several hydrogen atoms of the tertbutyl group are within a <3Å distance of the uorine atoms in 6 0 , consistent with a highly congested chemical environment around the radical center.
We next modeled the transition states (TS) of the H-atom abstractions by 3 0 and 6 0 involving select substrates, with the goal of identifying important factors that account for the observed site selectivities. We studied a diverse set of cyclic and acyclic substrates, including trans decalin, 3-methylpentane (7) and 3-pentyl acetate-a model for the benzoate substrate 12. Representative TS for the abstractions are shown in Fig. 7 for the reactive primary C-H bonds of 3-methylpentane (7) involving amidyl radicals 3 0 (le) and 6 0 (right). The reactive site is much more crowded in the case of 6 0 (Fig. 7, right), as highlighted by the yellow lines in the TS forming a smaller angle than in the case for 3 0 (Fig. 7, le). The primary C-H bond of 3methylpentane approaches amidyl radical 6 0 through a relatively tight pocket. On the other hand, the substrate engages with the nearly at conformation of 3 0 from the top. It is worth noting that the two amidyl radicals retain the geometric parameters from Fig. 6 in the TS.
The Gibbs free energy barriers (DG ‡ ) and changes (DG) for the C-H abstractions evaluated are listed in Table 2. As expected, all C-H abstractions occur with an overall energy gain (negative DG values), regardless of the substrate, position, or amidyl radical considered. Comparisons of the energy barriers involved indicate different trends for the two reagents, however. In particular, the last column of Table 2 reports DDG ‡ values quantifying the energetic difference between the two amidyl radicals for reaction at a given C-H site. For example, amidyl radical 6 0 is more reactive than 3 0 in the abstraction at the   methylene C-H site distal from the ring fusion of trans decalin, as indicated by the negative DDG ‡ value (for the observed experimental results see Table 1). On the other hand, virtually no difference is observed at the proximal methylene site, with a small DDG ‡ of +0.35 kcal mol À1 .
In the case of acyclic substrates, amidyl radical 6 0 is consistently more reactive than 3 0 at primary C-H sites, while little difference is observed at secondary positions. Notably, C-H abstraction at the reactive primary position of 3-methylpentane (7) shows the largest (in absolute terms) calculated DDG ‡ value (À4.80 kcal mol À1 ), which is consistent with experimental results; functionalization at this position involves the largest selectivity change upon switching from 3 0 to 6 0 (27% to 69%, see Fig. 2). Overall, our computational ndings are in good agreement with the experimental results, as reagent 6 consistently provided higher selectivities for the least hindered positions of substrates-either primary or distal-throughout our studies.

Conclusions
In conclusion, we have developed an approach to introduce reagent-dictated site selectivity to aliphatic C-H functionalizations using tuned nitrogen-centered radicals. The use of hindered N-functionalized amides has led to C-H functionalization of diverse substrates that are among the most siteselective intermolecular C-H transformations known. This strategy has also enabled the rst radical-mediated aliphatic C-H functionalizations targeting primary C-H bonds across a wide range of simple and complex substrates. Computational studies revealed the different chemical environments of the reacting amidyl radicals, and established a direct correlation between the steric hindrance at the reacting site and the preference for functionalization of the least hindered substrate C-H bonds. Future studies will continue to develop unique reagents that offer increased levels of site selectivity in intermolecular C-H functionalization for applications in chemical synthesis.

Conflicts of interest
There are no conicts to declare.

Acknowledgements
Financial support was provided by the ACS PRF (55108-ND1) and Award R01 GM 120163 from the National Institute of General Medical Sciences. D. R. thanks the MIUR for nancial support (SIR Project "Organic Synthesis via Visible Light Photocatalytic Hydrogen Transfer"; Code: RBSI145Y9R). We thank Dr Brandie Ehrmann (UNC) for mass spectrometry assistance and Dr Roger Sommer (NCSU) for X-ray crystallographic analysis. Calculations were carried out at the CINECA Supercomputer Center (Italy), with computer time granted by ISCRA projects (code: HP10CSWOH1). We thank the Fasan group (University of Rochester) for a sample of artemisinin chlorination product S19.   a Energy values expressed in kcal mol À1 at the SMD(CH 2 Cl 2 )-uB97XD/6-31G(d,p) level of theory have been reported (see Table S41 in the ESI for details). b Value obtained by taking the difference of DG ‡ (6 0 ) À DG ‡ (3 0 ). A negative value indicates a lower energy barrier for 6 0 than for 3 0 . c The functionalizations of the axial and equatorial C-H bonds have been considered, but only the parameters of the most stable TS have been reported (see Table S41).