CHAPTER 1: Asymmetric C–H Bond Insertion Reactions

It has been recognized for over 70 years¹  that C–H bond functionalization can be realized through carbene insertion reactions.²  In recent years, many outstanding works including asymmetric reactions have appeared in this area.³ 

Considering metal carbenoid-induced C–H bond insertions, there exists a general pattern, as shown in Scheme 1.1. The catalytic cycle is initiated by a metal complex via the decomposition of diverse carbene precursors (such as diazo compounds) to deliver a transient metal carbenoid intermediate in situ. Subsequently, the highly reactive metal carbenoid intermediate inserts into the C–H bond to afford the corresponding product and readily regenerates the metal complex to complete the catalytic cycle. Note that the metal atom is not thought to interact with the C–H bond directly. Moreover, since the transient metal carbenoid intermediate is highly reactive, the reaction conditions are typically mild and pH neutral, which renders this method compatible with a range of functional groups, like halides, triflates, and boronates.

However, the extraordinary reactivity of the carbenoid intermediates, in general, also makes them open to many possible reaction scenarios. Thus, reactivity control has been the essential need when working towards a synthetically useful methodology. Crucial breakthroughs are beginning to be made for the intramolecular approach and a handful of reviews have summarized the progresses on this topic.^3a–g  For intermolecular reactions, reactivity control is more challenging. Besides the intrinsic selectivity of diverse substrates, the key to solving this problem is being able to rely on the metal carbene precursors which, in general, determine the reactivity of the generated metal carbenoid intermediates. Along with the involvement of carbene precursors, especially the diazo compounds, the reactivity of metal carbenoid intermediates can be tunable, and thus an intermolecular approach can be achieved. According to the characteristics of the substituents on the carbene precursors, they can be classified into three major groups: acceptor carbenoid, acceptor/acceptor carbenoid, and donor/acceptor carbenoid (Figure 1.1). The electrophilic properties of the substituents at the metallocarbenoid carbon center play a significant role in the reactivity and selectivity of the insertion reaction. Generally, an electron-withdrawing group, typically a carbonyl moiety, causes the carbenoid intermediate to be highly electrophilic and reactive, while an electron-donating group stabilizes the carbenoid intermediate. As far as acceptor carbenoids and acceptor/acceptor carbenoids are concerned, an electron-withdrawing group can both make the carbene precursor too stable to be decomposed by a metal complex and render the carbenoid intermediate highly reactive and susceptible to other competing processes. In that case, it is not so hard to understand they are apt to dimerization or hydride transfer to form zwitterionic intermediates, which can be overcome by intramolecular design. However, the donor/acceptor carbenoids, as late arrivals to the field of metal carbenoid chemistry, revolutionized the situation dramatically and show great potential in highly selective intermolecular C–H bond functionalization; this is because the donating groups present, such as vinyl and aryl, can stabilize the carbenoid through resonance. Meanwhile, the aryl and vinyl groups also make the diazo precursors stable. In this case, highly active catalysts are required to effectively decompose this type of diazo compounds.^3h 

Rh and Cu complexes are commonly employed in metal carbenoid-involved asymmetric C–H bond functionalization while chiral catalysts based on Ir, Ru, and Fe as well as Lewis acids came into this area recently. This section aims to introduce the recent developments in asymmetric C–H bond functionalization achieved by metal carbenoids.

Rh-catalyzed asymmetric C–H bond functionalization via a carbene insertion reaction was extensively documented in the early days, especially the intramolecular reactions. Thanks to enormous efforts from the groups of Davies and Doyle, asymmetric intramolecular C–H bond insertion by Rh carbenoids has become a reliable methodology and has been employed frequently in the total synthesis of complex natural products.^3c–e,g,4 

One important advantage of the intermolecular carbene insertion reactions is that simple starting materials can be employed and accordingly there is no need for the construction of complex substrates in advance. However, the intermolecular process requires a delicate balance between electronic and steric effects for metal carbenoids.^3b  On the other hand, there are several obstacles to be overcome, including chemo-, regio-, and enantioselectivity. Fortunately, great efforts have been devoted in the past decade and a series of carbene precursors and chiral Rh catalysts have been developed, so satisfactory yields and ee can be obtained in some catalytic systems. Generally, suitable carbene precursors, such as donor/acceptor diazo compounds, could reduce the chance of side product formation due to carbene dimerization. On the other hand, the dirhodium bridge caged within a “lantern” structure is thought to be essential to the success of dirhodium complexes in which two rhodium atoms are surrounded by four ligands in a nominal D₄ symmetry. Both computational studies and characterization of dirhodium carbenoid intermediates suggested that the intermediate adopts a Rh–RhC framework. In another word, two rhodium atoms are bound to one carbene center, and the bonding scenario obeys the three-center orbital paradigm.⁵  As such, metal carbenoids derived from chiral Rh^II complexes and donor/acceptor diazo compounds are routinely utilized.

Based on initial findings from the studies on achiral catalysts, chiral Rh^II complexes which have been developed for enantioselective C–H bond functionalization can be classified into four categories: Rh^II carboxylates, Rh^II carboxamidates, Rh^II phosphates, and ortho-metallated arylphosphine Rh^II complexes (Figure 1.2).⁶  Among various Rh^II complexes, Rh₂(DOSP)₄ has proven to be the most effective and versatile catalyst for an array of reactions with a broad range of substrates. In order to understand the outstanding performance on stereocontrol, detailed research (including computational calculations⁷ ) has been carried out to construct a prediction model.^3h 

Furthermore, competition experiments were conducted utilizing donor/acceptor carbenoids to provide a picture of the relative reactivities among a series of substrates (Figure 1.3).⁸  It showed that regioselectivity of different C–H bonds is a balance between steric and electronic factors. Considering the C–H bond insertion reaction, the most reactive sites are those that can mostly stabilize a build-up of positive charge in the transition state. Therefore, the more nucleophilic site, e.g. tertiary C(sp³)–H bond, allylic C–H bond, or C–H bond α to a heteroatom, is more preferred to undergo insertion for electronic reasons. Meanwhile, if the tertiary C(sp³)–H bond is comparatively congested and the Rh^II complex is sterically hindered, the steric effect becomes a significant issue and the reaction will take place at the less crowded C–H bond. In addition, the size of the substituents on the carbenoid source can also impact the regioselectivity due to the steric cause.^3c,h,8 

Since the Rh-catalyzed asymmetric carbene insertion reactions are relatively well developed methodologies, here we only discuss selected recent successful intermolecular examples, which are classified according to the type of the C–H bond which is functionalized via carbene insertion.

The first enantioselective intermolecular C–H bond insertion which could be of practical synthetic application was reported in 1997 by Davies and Hansen (Scheme 1.2).⁹  In the presence of a variety of relatively unreactive cycloalkane solvents 2 (compared with C–H bonds at the allylic and benzylic positions, as well as C–H bonds α to a heteroatom), the Rh complex Rh₂(S-DOSP)₄, a privileged catalyst derived from l-proline, was found capable to catalyze the decomposition of aryldiazoacetates 1, inducing the functionalization of cycloalkanes 3 (Scheme 1.2, eqn (1)). Circumventing chemoselectivity and regioselectivity by substrate design, yields and enantioselectivity can reach up to 96% and 93% ee, respectively, by conducting the reactions under refluxing conditions. Further improvements in enantioselectivity (88–96% ee) without considerable decrease in yield were realized by performing the reaction at 10 °C in degassed solvent.⁸  It is noteworthy that the key to the success of this intermolecular process is the unusual reactivity and selectivity of aryl- and vinyl-diazoacetates. These donor/acceptor carbenoids are significantly more stabilized and, accordingly, are more versatile than the more traditional carbenoids derived from acceptor and acceptor/acceptor diazoacetates. While other alkanes (4, 6, and 8) were subjected to the reaction conditions, the general trend described before (Figure 1.3) was exemplified vividly. Tertiary C–H bonds are much more activated, which can be supported by the formation of the sole corresponding tertiary C–H sites insertion products 5 and 7 in the reactions of 4 and 6 (Scheme 1.2, eqn (2) and (3)), respectively. While 2-methylpentane 8 bearing a sterically demanding tertiary C–H bond was exposed to the reaction conditions, the tertiary C–H bond insertion product 9 and the secondary C–H bond insertion product 9′ were afforded in similar yields (Scheme 1.2, eqn (4)). Based on an array of experiments, this catalytic system was shown for the first time to be outstandingly chemoselective and the reactivity of different C–H bonds was expected to follow the sequence: tertiary ≈ secondary >> primary C–H bonds.⁸ 

Intriguingly, the dirhodium(ii) catalyst Rh₂(S-PTAD)₄ can be synthesized via the C–H bond insertion of adamantane 6 by the carbenoid derived from vinyldiazoacetate 10 and Rh₂(S-DOSP)₄ (Scheme 1.3).¹⁰ 

Apart from the utilization of aryl- and vinyl-diazoacetates that can achieve the moderate to high chemo-, regio-, and enantioselectivity in intermolecular asymmetric C–H bond insertion reactions, N-sulfonyl-1,2,3-triazole 11 was found to be able to function as an alternative carbene precursor for diverse transformations (Scheme 1.4). One advantage for using the N-sulfonyl-1,2,3-triazole is that it could be easily prepared by the Cu^I-catalyzed azide–alkyne cycloaddition (CuAAC) reaction,¹¹  and in some cases, delicately designed reactions can be conducted in a one-pot procedure starting from alkynes and sulfonyl azides. Moreover, since there exists an inherent equilibrium between diazoimines 11′ and closed 1,2,3-triazole 11 which favors the latter, the application of N-sulfonyl-1,2,3-triazole removes the need for special slow addition techniques that are often required to guarantee a low concentration of the highly reactive carbenes and parent diazo compounds in the reaction mixture. Accordingly, side products due to carbene dimerization could be largely compressed. Furthermore, the generated metal-bound imino carbenes 11″, which are inaccessible from traditional donor/acceptor carbene chemistry, are synthetically useful intermediates and can perform various transformations, such as cycloaddition, ring expansion, ylide formation, and direct heterocycle synthesis.¹² 

In 2011, the Fokin group¹³  exploited the asymmetric C–H bond insertion reaction between the azavinyl carbenoids derived from triazoles 11 and Rh₂(S-NTTL)₄ or Rh₂(S-PTAD)₄ and a series of inert alkanes 12 under mild conditions (Scheme 1.5). The direct insertion products were converted into the corresponding β-chiral amines 13 via subsequent LiAlH₄ reduction in up to 97% ee. With respect to the earlier report (Scheme 1.2),⁹  higher regioselectivity (5 : 1) of 13f was observed, suggesting that the insertion reaction favors tertiary C–H bonds over secondary C–H bonds, which can be ascribed to the lower steric requirement of the sulfonyl imine group compared with the ester group in diazoacetate. It was proposed that the Rh^II catalyst facilitates the ring-chain isomerization of the triazole and the following diazo decomposition.¹⁴ 

It is well-established that olefins tend to undergo cyclopropanation while exposed to metal carbenoids.¹⁵  Meanwhile, olefins also impact the adjacent C–H bonds and enhance the reactivity towards C–H bond insertion reaction by metal carbenoids. It was shown that the chemoselectivity between the intermolecular C–H bond insertion vs. cyclopropanation can be tuned by proper choice of diverse metals, ligands, and carbene precursors.¹⁶  Generally, the donor/acceptor carbenoids (aryldiazoacetate and vinyldiazoacetate) are far more prone towards C–H bond insertion than cyclopropanation, which is opposite to acceptor and acceptor/acceptor carbenoids.^3c 

At first, various chiral dirhodium complexes were examined. Enantioselectivity was only moderate while cyclohexene 14a was used as a model substrate and DCM as solvent (entries 1–5, Table 1.1).¹⁶  When the reaction was conducted in 2,2-dimethylbutane (DMB) at −50 °C, the C–H bond insertion product could be obtained in 58% yield and 93% ee (entry 6, Table 1.1).¹⁷  Notably, poor diastereoselectivity is typically observed for intermolecular C–H bond insertion of simple alkenes.

Similarly poor diastereoselectivity was also observed in 1-substituted cyclohexenes 14 and acyclic trisubstituted olefin substrates 17 (Tables 1.2 and 1.3). Despite the excellent chemoselectivity and enantioselectivity that Rh₂(S-DOSP)₄ carbenoids generally achieved, the diastereoselectivity was unsatisfactory. It is worth mentioning that the reaction occurs preferentially at the sterically less hindered allylic site.

It was proposed that considerable size differentiation between the two potential reaction sites for the C–H bond insertions (e.g. C3 in Table 1.4) would help to improve the diastereoselectivity. Under such a hypothesis, installation of a bulky vinylsilane group at the C1 position becomes a wise choice because it not only provides the desired stereo differentiation but also eliminates cyclopropanation as a side reaction for steric reasons. Indeed, while TBDPS was introduced (Table 1.4, entry 2), good diastereoselectivity could be achieved.

While silyl enol ethers 21 and 23 were subjected to similar reaction conditions (Tables 1.5 and 1.6), the allylic C–H bond could also be functionalized by metal carbenoids to afford silyl-protected 1,5-dicarbonyls 22 and 24 respectively, which can be viewed as an equivalent of an asymmetric Michael reaction.¹⁸  Although the double bond is highly electron-rich and readily undergoes cyclopropanation in the presence of most other metal carbenoids, by using aryldiazoacetates 1 as carbene precursors, cyclic silyl enol ethers 21 were readily transformed into their corresponding allylic C–H bond insertion products 22 (22′) in excellent yields, excellent ee and moderate de (Table 1.5). Noticeably, while acyclic silyl enol ethers 23 were subjected to the reaction, excellent diastereoselectivity (>90% de) was obtained, which shows great potential in synthetic applications (Table 1.6).

When 1,4-dienes are utilized, the activity of allylic C–H bonds to undergo carbene insertion reactions is further enhanced. For instance, a range of chiral Rh^II complexes is able to convert 1,4-cyclohexadiene into the corresponding C–H bond insertion product in a high yield and ee in the presence of methyl phenyldiazoacetate 1a (Scheme 1.6, eqn (1)).^16,19  The synthetic utility of this methodology was further demonstrated by total synthesis of natural products (+)-indatraline (Scheme 1.6, eqn (2)) and (+)-cetiedil (Scheme 1.6, eqn (3)).²⁰ 

For other cyclopolyenes, while 1,3-cyclohexadiene^19,20b  and 1,3-cycloheptadiene²¹  reacted less selectively than 1,4-cyclohexadiene, 1,3,5-cycloheptatriene also proceeded highly site selective allylic C–H bond insertion with an array of aryldiazoacetates 1 in moderate yields and excellent ee (Table 1.7).²¹ 

The asymmetric allylic C–H bond insertion reaction of 1,4-cyclohexadiene was further improved by Denton and Davies in 2009.²²  Using different donor/acceptor carbenoids derived from α-aryl-α-diazoketones 25 and a chiral dirhodium complex Rh₂(S-PTAD)₄ instead of methyl phenyldiazoacetate 1a and Rh₂(S-DOSP)₄ in previous work (Scheme 1.6, eqn (1)), the corresponding C–H bond insertion products 26 could be obtained in up to 90% yield and 89% ee in refluxing DMB (Scheme 1.7a). Later, the catalytic efficiency was significantly enhanced by conducting the reaction under solvent-free conditions.²³  Since donor/acceptor carbenoids are more stable and less prone to catalyst decay and carbene dimerization, they are suitable for reactions with low catalyst loadings. In that case, Rh₂(S-DOSP)₄-catalyzed asymmetric carbenoid insertion reaction of 1,4-cyclohexadiene proceeded smoothly at 0 °C using 1,4-cyclohexadiene as the solvent under merely 0.01 mol% catalyst loading to generate product 27 in 96% yield with 81% ee (Scheme 1.7b). It is presumed that the absence of solvent favors the productive bimolecular pathway for desired product formation over the unimolecular process for metal carbenoid destruction.

Recently, as a continuing interest in the synthetic application of α-alkyl-α-diazo ester 28, Hashimoto and co-workers reported the first Rh-catalyzed intermolecular asymmetric C–H bond insertion reaction of α-alkyl-α-diazo ester (Scheme 1.8).²⁴  Despite only moderate yields and ee obtained by employing Rh₂(S-TFPTTL)₄ or Rh₂(S-TCPTTL)₄, excellent results were obtained in a previous report on the asymmetric cyclopropanation of alkyne.²⁵ 

Intriguingly, while vinyldiazoacetates are utilized as carbene precursors in the allylic C–H bond insertion reaction catalyzed by Rh^II complexes, the combined C–H bond functionalization/Cope rearrangement occurs readily. This has emerged as a reliable methodology for the construction of 1,5-diene compounds bearing two vicinal stereogenic centers and will be discussed in detail in Section 1.1.2.6.

Analogous to allylic C(sp³)–H bonds, benzylic C(sp³)–H bonds are activated due to additional stabilization of the build-up of positive charge provided by the phenyl group. However, intermolecular benzylic C(sp³)–H bond insertion reactions often suffer from competitive cyclopropanation which can be suppressed completely by introducing a para-substituent on the phenyl group.²⁶  In general, the most favorable site for C–H bond insertion is the methylene benzylic C(sp³)–H bond (31, 34, and 35) due to the subtle interplay of steric and electronic elements (Scheme 1.9).

It is worth noting that this methodology has been applied in the concise synthesis of (+)-imperanene and (−)-α-conidendrin (Scheme 1.10).^26b 

In 2005, the Davies group discovered that the generally reliable catalyst Rh₂(DOSP)₄ failed to induce high levels of chiral control when benzyl silyl ether derivatives 37 were employed as reaction components.²⁷  Fortunately, Hashimoto’s Rh₂(S-PTTL)₄ catalyst showed a superb performance, delivering C–H bond functionalization products 38 in prominent levels of diastereo- and enantioselectivity (up to >95% de and 98% ee) (Scheme 1.11).

Moreover, due to collaboration with the Yu group, who developed a new method for rapidly constructing dihydrobenzofurans via Pd-catalyzed C–H functionalization/C–O cyclization,²⁸  the Davies group showed that the C–H bond functionalization products 38 could be further transformed into complex 2,3-dihydrobenzofurans in high enantiopurity via sequential C–H bond functionalization (Scheme 1.12).²⁹  Starting from relatively simple substrates, and metal carbenoids derived from Rh₂(R-PTTL)₄ and α-aryl-α-diazo esters 1, the insertion into the C–H bond of secondary benzylic ethers was accomplished to deliver the corresponding products 39 in good yields and excellent diastereo- and enantioselectivity (up to 92% yield, >97 : 3 dr and 99% ee). After subsequent removal of the TBS group, the Pd-catalyzed C–H functionalization/C–O cyclization was carried out to afford 2,3-dihydrobenzofurans 40 with no loss of enantiomeric purity. Moreover, the oxidative Heck coupling reaction³⁰  was exploited to further transform the benzofuran derivatives 41. It is a vivid example that demonstrates the promising future of C–H bond functionalization in the streamlined synthesis of complex targets.

As mentioned above, the most reactive C–H bond that can participate in Rh-catalyzed carbene insertion reactions is usually the most nucleophilic C–H bond in a molecule. Accordingly, the C–H bond α to a heteroatom is highly reactive since the lone pair electrons on the heteroatom provide additional stabilization of the build-up of positive charge in the transition state. Complementary to classic disconnection strategies in organic synthesis (Scheme 1.13), insertion into the C(sp³)–H bond α to a heteroatom by Rh carbenoids has been developed into a highly regio-, diastereo-, and enantioselective methodology (Figure 1.4).

With respect to the insertion reaction into the C–H bond α to an oxygen atom, tetrahydrofuran (THF) was the first successful prototypical substrate. In the presence of a catalytic amount of Rh₂(S-DOSP)₄, the reaction proceeded smoothly with a series of aryldiazoacetates 1, delivering the corresponding 2-substituted tetrahydrofurans 43 in moderate yields (56–74%) and diastereoselectivity (1.6–4.0 : 1 dr), and outstanding enantioselectivity (95–98% ee).^8,9  Moreover, acyclic substrates 44³¹  and 46³²  could also be applied in this transformation, affording their corresponding C–H bond insertion products 45 and 47 respectively in excellent diastereo- and enantioselectivity. It is worth mentioning that the C–H bond insertion products 45 and 47 are typically obtained from aldol reactions, and relatively low catalyst loading (0.25 mol%) is demanded in this Rh-catalyzed C–H bond functionalization reaction to achieve superb chiral control (Scheme 1.14).

Similar to the ether substrates mentioned above, both cyclic^20b,33  and acyclic³⁴  N-protected amines could be converted into their corresponding ortho-C–H bond functionalization products via the insertion of metal carbenoids derived from Rh^II catalysts and donor/acceptor diazo compounds. Taking N-Boc-protected pyrrolidine 48 as an example (Scheme 1.15), Rh₂(S-DOSP)₄ catalyzed the decomposition of methyl phenyldiazoacetate 1a and converted the N-Boc-pyrrolidine 48 into the corresponding C–H bond insertion product 49 in 72% yield, 94% ee, and 92% de. Furthermore, the C₂-symmetric amine 50 could be formed in 78% yield and 97% ee under altered reaction conditions. Further investigation demonstrates that this intermolecular C–H bond insertion could also be applied in the kinetic resolution of 2-substituted pyrrolidine 51. The corresponding C–H bond insertion reaction proceeded smoothly, and subsequent treatment with TFA delivered the deprotected product 52 in high diastereo- and enantioselectivity (45% yield, 91% ee, >94% de).^33c 

A subsequent study from the Davies group illustrated the synthetic utility of this Rh^II-catalyzed intermolecular C–H bond insertion approach by the asymmetric synthesis of antidepressant venlafaxine (Scheme 1.16).^34c 

Quite recently, Davies and co-workers developed a new class of sterically demanding dirhodium tetracarboxylate catalysts, especially Rh₂(R-BPCR)₄, that changed the site selectivity of the C(sp³)–H bond insertion reaction.³⁵  In the presence of catalytic amount of Rh₂(R-BPCR)₄, the primary C–H bond is the preferred reaction site of various substrates containing primary benzylic C–H bonds, allylic C–H bonds, or C–H bonds α to oxygen, which is complementary to Rh₂(R-DOSP)₄ which favors secondary C–H bonds (Scheme 1.17a–c). Moreover, the use of this methodology was further proved by the selective C–H bond functionalization of complex molecules such as (−)-α-cedrene (Scheme 1.17d).

Indoles are among the most common heterocyclic scaffolds in a wide range of naturally occurring alkaloids. In this regard, diverse methodologies have been developed to enantioselectively functionalize this type of motif.³⁶ 

It is well established that the reaction of carbenoids with N-alkylindoles delivers zwitterionic intermediates.³⁷  The reason why this scenario is favored can be ascribed to the fact that the positive charge of the intermediate is stabilized by the electron-rich indole while the negative charge is stabilized by the carbenoid component. In other words, the site of C3 is highly reactive in metal carbenoid insertion reactions. In 2010, Lian and Davies described such a process in their seminal work on Rh-catalyzed [3 + 2] annulation of indoles. In the presence of 1,2-dimethylindole 53, Rh₂(S-DOSP)₄ induced the decomposition of methyl α-phenyl-α-diazoacetate 1a and C–H bond insertion of indole, providing the C3 functionalization product 54 in 95% yield but negligible asymmetric induction (<5% ee).³⁸  It is proposed that the poor chiral induction in the formation of C–H bond insertion product 54 can be attributed to the rapid proton transfer from the zwitterionic intermediate A to the achiral enol B, which can further tautomerize into the observed C–H bond insertion product 54 (Scheme 1.18).

In 2011, Fox and co-workers accomplished highly enantioselective C–H bond functionalization of indoles with α-alkyl-α-diazo esters 56 by using Rh₂(S-NTTL)₄ as catalyst (Scheme 1.19).³⁹  Under the essential low temperature (−78 °C), prominent yields and ee were achieved for an array of the N-aryl and N-alkyl indole derivatives 55 bearing small substituent (H or Me) at the indole C2 position. Noticeably, a methyl group at the indole C2 position generally led to a higher level of enantioselectivity. The cyclopropanation/fragmentation pathway was excluded by control experiments and further density functional theory (DFT) calculations. The latter also suggested that a Rh–ylide intermediate might be involved. These results also supported the presumption that Rh₂(S-NTTL)₄ adopts the “chiral crown” conformation in which the four phthalimide groups are projected on the same face of the complex during the reaction.⁴⁰  The outstanding performance on chiral control could be explained in that the indole approaches the Si-face of the Rh-carbene, followed by aromatization and stereoretentive protonation. Alternatively, the enantioselectivity-determining step might involve the dynamic kinetic resolution of a Rh-enolate intermediate.

Not long before, the Hashimoto group reported another successful case of asymmetric C–H bond functionalization of indoles (Scheme 1.20).⁴¹  In the presence of a different diazo ester 59 and catalytic amount of Rh₂(R-PTTEA)₄ bearing the exceptionally bulky triethylmethyl group which was essential for the chiral induction, N-MOM-protected 2,3-unsubstituted indoles 58 were converted into C3-adorned chiral indoles 60, which could be employed in the asymmetric synthesis of acremoauxin A, a potent plant-growth inhibitor.⁴²  It is particularly noticeable that lowering the temperature only influences the enantioselectivity marginally and the 2-methylindole derivative is not a suitable substrate here, which is in sharp contrast to Fox’s outcome.³⁹  Moreover, a MOM protecting group on the indole nitrogen is essential for this reaction. It is not only an insurance of the high levels of enantio control but also a synthetic advantage as the N-MOM group can be effectively removed under Fujioka conditions.⁴³ 

In 2012, Lian and Davies described an alternative metal carbenoid approach to functionalize indoles in an enantioselective fashion.⁴⁴  Utilizing carbenoid sources derived from E-vinyldiazoacetates 62 and Rh₂(S-biTISP)₂, a series of C2-substituted indoles 61 and pyrroles 64 were converted into their corresponding C3 functionalization products 63 and 65, respectively, in up to 86% yield and 95% ee. Noticeably, E-vinyldiazoacetates 62, which exist as a equilibrating mixture of s-trans or s-cis conformers, are essential to the success of the asymmetric C–H bond insertion of indoles. Presumably, while a sterically demanding nucleophile was used, the vinylogous reactivity of the E-vinyldiazoacetates 62 will be enhanced, and can be further increased via reinforcing the s-trans conformation of the carbenoid through the employment of highly sterically hindered catalysts. Indeed, the usage of such an extremely bulky catalyst Rh₂(S-biTISP)₂ enabled the effective C–H bond insertion, delivering the desired indole derivatives in up to 86% yield and 95% ee (Scheme 1.21). Moreover, the substrate scope was expanded into pyrroles at a slightly elevated temperature (−20 °C).

As depicted in Scheme 1.18, it is proposed that the zwitterionic intermediate A, which is generally involved in the C–H bond insertion reactions with diazo compounds, usually undergoes rapid proton transfer to generate achiral intermediate B and subsequently tautomerizes to the C–H bond functionalization product, which might be responsible for the poor chiral control in Davies’ observation in 2010. However, while the zwitterionic intermediates were trapped promptly by imines which were activated by chiral phosphoric acid,⁴⁵  the asymmetric catalysis could be realized via a cascade reaction and it was achieved by Hu and co-workers in 2012 (Scheme 1.22).⁴⁶  In their work, the zwitterionic intermediates C generated from the formal intramolecular C–H bond insertion by Rh carbenoids derived from α-methyl-N-methyl-N-phenyl diazoacetamides 66 were trapped by chiral phosphoric acid (S)-C1 activated N-aryl imines 67 to deliver polyfunctionalized oxindole derivatives 68 bearing two consecutive chiral centers in outstanding diastereo- and enantioselectivity (up to 99 : 1 dr and 98% ee). Subsequently, analogous tactics were applied in the capture of intermolecularly generated zwitterionic intermediates D by an array of N-aryl imines 67 activated by chiral phosphoric acid (S)-C1. Under optimized conditions, the three-component reactions proceeded smoothly to yield indole derivatives 71 bearing two successive chiral centers in conspicuous diastereo- (>20 : 1 dr) and enantioselectivity (up to 99% ee). The integration of Rh catalyst and chiral phosphoric acid in a dual catalytic system⁴⁷  provides a novel approach to access highly functionalized complex targets in exceptional efficiency, diastereo- and enantioselectivity.

Since it was discovered by the Davies group in 1998, the combined C–H functionalization/Cope rearrangement (CHCR) has developed into a reliable methodology and been applied in an extensive range of organic synthesis.⁴⁸  The CHCR reaction occurs between Rh-bound vinylcarbenoids and cyclic or acyclic substrates containing allylic C–H bonds and generally delivers products bearing two newly generated stereocenters in exceptional diastereo- and enantioselectivity while chiral dirhodium catalysts are employed.⁴⁹  In some cases, the CHCR reactions have performed as synthetic equivalents of the Michael reaction,⁵⁰  the vinylogous Mukaiyama aldol reaction,⁵¹  the tandem Claisen rearrangement/Cope rearrangement,⁵²  or the tandem aldol reaction/siloxy-Cope rearrangement.⁵³  Moreover, the CHCR reaction has already been exploited in the total synthesis of natural products such as (+)-sertraline,¹⁹  (+)-erogorgiaene,^49a  (+)-colombiasin A, and (+)-elisapterosin B^49b  (Scheme 1.23).

After the seminal discovery of the CHCR reaction of 1,3-cyclohexadiene, some representative cyclic olefins 72 and vinyldiazoacetates 73⁵²  were subjected to the CHCR reaction in the presence of Rh₂(S-DOSP)₄, leading to their corresponding CHCR products 74 with moderate yields (44–87%), superb levels of diastereo- and enantioselectivity (>30 : 1 dr and 96% ee), together with minor competing direct C–H bond insertion products, which could be circumvented by the usage of acyclic olefins 75 as substrates. This is because that, in most cases such as eqn (2) in Scheme 1.24, the direct C–H bond insertion product 76′ is no longer the thermodynamic product and is susceptible to a siloxy-Cope rearrangement into the CHCR product 76 once 76′ is subjected to heating or microwave conditions (Scheme 1.24).⁵³ 

Originally, it was proposed that the CHCR reaction proceeded in a concerted asynchronous process. The CHCR reaction was speculated to be triggered by a C–H bond insertion event at the allylic site which was interrupted by a subsequent [3,3]-rearrangement via a chairlike transition state (Scheme 1.25a).^{19,49a,b,50,52,54}  In 2011, detailed DFT calculations were conducted to deepen the understanding of the mechanism.⁵⁵  This research supported the idea that the initial C–H bond functionalization might be a hydride transfer event, affording the charged transition state TS-HT which could then bifurcate to the CHCR product PA or the direct C–H bond insertion product PB (Scheme 1.25b). This might be the explanation to why, in some cases, both products were observed and the level of chiral control of the two products were often very close, if not identical.⁵³ 

Intriguingly, the computational investigation revealed that a chairlike transition state was not as energetically favorable as presumed earlier. In principle, considering that the CHCR reaction could proceed through a chairlike or boatlike transition state while the rhodium-bound vinylcarbenoid could react in an s-cis or s-trans orientation, all four transition states are all energetically accessible, which would lead to different stereochemical outcomes. In another word, the diastereoselectivity could be potentially switched by using appropriate substrates that either favor the s-cis/chair approach or the s-cis/boat approach. Elaborately, the substrates bearing large substituents at the R² position would adopt the s-cis/boat approach preferentially. On the contrary, the substrates bearing large substituents at the R¹/R⁴ positions would prefer the s-cis/chair approach (Scheme 1.26a). This hypothesis was further verified by Davies and co-workers.⁵⁶  The application of cyclohexene 77 and cyclopentene 79 as the substrates in CHCR reaction respectively (Scheme 1.26b) led to the designed β-keto-α-diazoacetates 78 and 80 with high stereoselectivity but in the opposite diastereomeric series after subsequent transformations, which further supported the working models depicted in Scheme 1.26a.

Subsequently, based on the evaluation of the s-cis/boat transition state model, acyclic trisubstituted vinyl ethers were proposed to be ideal substrates for the CHCR reaction, which was further demonstrated in 2011 by Lian and Davies (Scheme 1.27).⁵¹  In the presence of a catalytic amount of Rh₂(S-DOSP)₄, a wide range of vinyl ethers, readily accessible from the corresponding ketones via the Wittig reaction, were transformed into the corresponding CHCR products in extremely high enantio- (>98% ee in the most of the cases) and diastereoselectivity (observed as single diastereomer in each case). Further investigation revealed that the absolute configuration of 82 was found to be accordance to the s-cis/boat approach. Moreover, it was observed that only the secondary C–H bond trans to the methoxy group in E-81a was cleaved when a 1 : 1 mixture of E/Z-81a was used, which further demonstrated that the CHCR reactions are very sensitive to steric and electronic effects. In addition, the CHCR reactions have also been applied in the kinetic resolution of the racemic vinyl ether 81b. It is worth noting that the CHCR reaction of vinyl ethers could serve as a surrogate for the vinylogous Mukaiyama aldol reactions.⁵⁷ 

Besides Rh, Cu is another transition metal widely used to catalyze carbene transformations. Early on, intramolecular insertions of carbenoids have been realized by copper complexes ligated with chiral bis(oxazolines) and related ligands in relatively moderate to low yields, chemoselectivity, and enantioselectivity.⁵⁸  The situation was changed in 2010 by Maguire and co-workers who reported the first example of Cu-catalyzed intramolecular C–H bond insertion reactions of α-diazosulfones 83 (Scheme 1.28a).⁵⁹  The combination of a catalytic amount of CuCl and the chiral BOX ligand (R,R)-L1 led to the formation of the favorable six-membered ring cis-thiopyrans 84 in moderate yields (up to 68%) and superb enantioselectivity (up to 98% ee). If the cyclization to generate the preferred six-membered ring was impossible, the insertion reaction was compelled to give the trans five-membered ring products 85. It is note worthy that rhodium acetate was not suitable for these substrates 83. Following this seminal study, Slattery and Maguire reported another successful example of Cu-catalyzed intramolecular C–H bond insertion reactions.⁶⁰  The employment of the analogous chiral Cu–BOX complex converted α-diazo-β-keto sulfones 86 into enantioenriched cyclopentanones 87 in up to 82% ee. In the condition optimization process, it was unveiled that the non-coordinating counterion NaBARF or KBARF (BARF = tetrakis[3,5-bis(trifluoromethyl)phenyl]borate), as an additive, is critical for gaining high level of chiral control for both reactions. Subsequent mechanistic research uncovered that partial or complete chloride abstraction contributed to the formation of the naked alkali metal cations which could make a significant alteration to the characteristic of the Cu catalyst.⁶¹ 

However, intermolecular C–H bond insertion reactions accomplished by a chiral Cu catalyst still remained a great challenge. The breakthrough was achieved by Fraile and co-workers in 2007.⁶²  They discovered that the Cu complexes with bis(oxazoline) (BOX) and azabis(oxazoline) (azaBOX) ligands succeeded in catalyzing the asymmetric insertion of α-diazo compounds 1a into the C–H bonds of cyclic ethers including THF, THP, 1,4-dioxane, and 1,3-dioxolane. When the chiral complex (S,S)-C2 was immobilized with laponite clay or supported on silica or silica–alumina via electrostatic interactions, not only was the catalytic performance enhanced considerably in terms of conversion and stereoselectivity, but also the heterogeneous catalyst could be recovered and reused several times with similar results to those obtained in the first run (Scheme 1.29). Very recently, Jiménez-Osés, Fraile, and co-workers unambiguously determined the absolute configuration of the conformationally flexible products of the Cu-catalyzed C–H bond insertion reactions between 1a and cyclic ethers using vibrational circular dichroism (VCD) spectroscopy in combination with high level quantum mechanics calculations.⁶³ 

Apart from the extensively documented dirhodium catalysts, several chiral Ir complexes have lately also been shown to be catalytically efficient in the asymmetric C–H bond insertion reactions of metal carbenoids and they offer complementary reactivity profiles. In general, 1,4-cyclohexadienes and THF are two commonly used model substrates. Several groups, including Katsuki,⁶⁴  Che,⁶⁵  and Musaev, Davies, and Blakey,⁶⁶  have realized the intermolecular C–H bond insertion reaction of 1,4-cyclohexadienes and THF by diverse chiral Ir complexes (Figure 1.5).

With regard to 1,4-cyclohexadiene (Table 1.8), all these above groups realized the intermolecular C–H bond insertion reaction with the Ir-carbenoids derived from α-diazo esters 88 to achieve the corresponding adducts 89 in excellent yields and enantioselectivity. It is worth noting that, in Katsuki’s work, the methyl α-diazopropionate (88, R¹ = Me) was recognized as a feasible metallocarbene precursor in this reaction. Historically, intermolecular reactions with alkyl-substituted α-diazo esters have been complicated by competing β-hydride elimination, which is not observed in this case. As for Che’s report, the catalyst TON reached to 9600 in a scale-up reaction. The recent development achieved by Musaev, Davies, Blakey, and co-workers made the reactions more operative under ambient conditions without the demand for slow addition of the diazo ester or low temperature. Moreover, substituted 1,4-cyclohexadienes were also subjected to the reactions and the corresponding α,α-biarylacetates could be obtained in good yields (64–98%) with excellent enantioselectivity (90–99% ee) after the subsequent oxidation with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ). Further DFT calculations disclosed that, in the active catalytic species, the coordination of the carbene segment and the Ir center tends to be perpendicular to the plane of the phebox ligand. A predictive model was also proposed to illustrate the experimentally observed stereoinduction in metal complexes of oxazoline-X-oxazoline tridentate ligands.

With respect to THF (Table 1.9), both the Katsuki group and the Che group accomplished the intermolecular C–H bond insertion reactions by utilizing their own chiral Ir complexes to afford the corresponding 2-substituted tetrahydrofuran 90 with satisfactory yields and ee. Similarly, the methyl α-diazopropionate (88, R¹ = Me) was also a suitable carbene precursor in this transformation in Katsuki’s work. The main difference between these two reports is that the major isomers 90 obtained in the C–H bond insertion reactions are opposite. In Katsuki’s reports, the syn-90 is the major diastereoisomer, which is contrary to Che’s outcomes that the anti-90 is the major one.

Che and co-workers expanded the scope of the Ir–porphyrin complex catalyzed asymmetric C–H bond insertion reaction to intramolecular variants. A catalytic amount of (+)-C5 (L = H₂O or solvent) converted an array of benzyl α-aryl-α-diazo esters 91 into cis-β-lactones via intramolecular asymmetric C–H bond insertion reaction in good yields (up to 87%) and enantioselectivity (up to 78% ee) (Scheme 1.30).⁶⁷  It is noteworthy that common dirhodium carboxylate catalysts such as Rh₂(S-PTAD)₄ or Rh₂(S-MEPY)₄ could not catalyze this reaction effectively.

Very recently, Che and co-workers designed, synthesized, and characterized a series of bis(NHC)Ru^II–porphyrin complexes, which displayed extremely high reactivity in alkene cyclopropanation and aziridination, carbene insertion into C–H, N–H, S–H, and O–H bonds, and nitrene C–H bond insertion.⁶⁸  Intriguingly, while a chiral element was introduced into the porphyrin scaffold, the chiral catalyst, Ru(D₄-Por)(IMe)₂, analogous to (−)-C5 was able to catalyze the asymmetric C–H bond insertion reactions between 1a and 1,4-cyclohexadiene, giving the desired product 27 in 80% yield and 92% ee (Scheme 1.31). It was supported by DFT calculations that the strong σ-donor strength of the trans NHC ligand leads to a lower activation barrier for the decomposition of diazo compounds by stabilizing the metal–carbene reaction intermediate.

Fe-catalyzed asymmetric reactions have captured enormous interest and been studied extensively in recent years because of iron’s cheapness, non-toxicity, and abundance in the Earth’s crust.⁶⁹  Zhou and co-workers achieved the asymmetric C–H bond functionalization reactions of indoles 93 by Fe-catalyzed insertion of a carbenoid derived from α-aryl-α-diazo esters 1 (Scheme 1.32).⁷⁰  It is worthy of note that the reaction of these substrates could not give satisfactory enantioselectivity while chiral dirhodium catalyst was utilized.³⁸  The elaborate screening of the reaction parameters unveiled that the combination of spiro bisoxazoline (S,S,R_a)-L3 as ligand and Fe(ClO₄)₂ as Fe precursor could give the optimal results (92% yield, 73% ee). The corresponding α-aryl-α-indolylacetates 94 were obtained in up to 94% yield and 78% ee, which could be readily purified by recrystallization in most cases. This study suggested a promising future for the application of sustainable and environmentally friendly catalysts in carbene transformations.

While transition metal-catalyzed asymmetric C–H bond functionalizations with diazo compounds via intra- and intermolecular patterns have been studied profoundly during the past decade, chiral Lewis acids were also discovered to have the capability to promote the transformation, which was exemplified by the chiral oxazaborolidinium salts. The chiral oxazaborolidinium ions, as powerful Lewis acids, are generated from the corresponding oxazaborolidines by protonation with strong Brønsted acids and have proven to be effective catalysts for asymmetric Diels–Alder reactions, cyanosilylations, tandem Michael–aldol reaction, cyclopropanation, and the Roskamp reaction. In 2013, the group of Hwang and Ryu demonstrated they could also be exploited in enantioselective C–H bond functionalization via asymmetric carbene insertion reactions into β-vinyl C–H bonds of cyclic enones.⁷¹  In the presence of chiral oxazaborolidinium salts (S)-C7 or (S)-C8 as the catalysts and α-alkyl-α-diazo esters 96 as carbene precursors, various cyclic enones 95 could be transformed into β-substituted cyclic enones 97 in high yields with outstanding chiral control (Scheme 1.33).⁷²  Noticeably, the catalyst loading could be reduced to 5 mol% in some cases. In addition, while racemic substrate 95 (n = 1, R¹ = 6-Me) was subjected into the reaction conditions, efficient kinetic resolution was realized. Moreover, this transformation was applied in the formal synthesis of a natural sesquiterpene (+)-epijuvabione,⁷³  which further demonstrated the synthetic utility of this methodology.

Nitrenes are the nitrogenous isoelectronic analogs of carbenes. In the context of developing new protocols for asymmetric functionalization of C–H bonds, amination via metal nitrenoids species is direct and significantly important. In the past decade, remarkable progress has been made in this area. This section provides an overview of these developments. The reactions will be presented according to the type of the metal involved in these transformations.

Of the metal complexes that can catalyze C–H bond nitrene insertion reactions, rhodium catalysts are the ones which are most studied. The reactions exhibit features analogous to the corresponding Rh-catalyzed carbene transfer reactions. Because of its importance, it has attracted the attention of many groups.

The racemic insertion reaction of metal nitrenoids into C–H bonds emerged as early as the 1980s, pioneered by Breslow and Mansuy.⁷⁴  In 1997, the first enantioselective C–H bond insertion reaction involving Rh nitrenoids was realized by Müller and co-workers (Scheme 1.34).⁷⁵  They developed an intermolecular C–H bond insertion reaction, which possibly goes through a direct nitrene insertion pathway. Preliminary experiments showed that the reaction can be enantioselective. When using Ikegami’s dirhodium Rh₂(S-PTPA)₄ as the catalyst and NsNIPh as the nitrene precursor,⁷⁶  indane was transformed to the amidated product 98 with 7% ee. When changing the catalyst to Pirrung’s dirhodium Rh₂(R–BNP)₄,⁷⁷  the enantioselectivity could be enhanced to 31% ee. Although the results were not ideal, they opened the door to the advances in this specific area.

In 2002, Hashimoto and co-workers developed a new chiral Rh^II catalyst Rh₂(S-TCPTTL)₄ for this process (Scheme 1.35).⁷⁸  By tuning the electronic property of the catalyst, they found that this amidation reaction of a variety of substrates 99 bearing either a benzylic C–H bond or an allylic C–H bond could proceed smoothly with moderate to good enantioselectivity.

In an extension of this work,⁷⁸  Hashimoto and co-workers tested several chiral dirhodium catalysts for the intramolecular amidation reactions in 2006.⁷⁹  Under the catalysis of Rh₂(S-TFPTTL)₄ as the optimal catalyst, indan-2-yl sulfamate 101 could be transformed to the cyclization product 102 in 98% yield, albeit with only 48% ee (Scheme 1.36).

In 2001, Du Bois and co-workers developed the intramolecular amination of diverse C–H bonds, including aliphatic ones, by employing the commercially available Rh-catalyst [Rh₂(OAc)₄] and oxidant PhI(OAc)₂.⁸⁰  Inspired by this work, Müller and co-workers realized intra- and intermolecular amidation reactions by in situ formed nitrene species.⁸¹  Employing the chiral dirhodium Rh₂(S-NTTL)₄ developed by their group,⁸²  cyclic sulfamidates 104 were synthesized in good yields and stereocontrol with sulfamates 103 and PhI(OAc)₂ as the nitrene precursors (Scheme 1.37). Subsequently, they expanded the reaction to intermolecular versions by using indane as the substrate. The enantioselectivity was just modest, with only 30% ee, albeit the yield was high (97%). Later, the same group studied this intramolecular amidation reaction in detail.⁸³  However, the challenge is still the stereocontrol and the highest ee value is only 66%.

In 2006, Müller, Dodd, Dauban, and co-workers designed an efficient diastereoselective intermolecular C–H bond amination reaction under the catalysis of Rh.⁸⁴  Based on their previous reports,⁸⁵  they used chiral sulfonimidamide 107 as the nitrene precursor which could generate highly reactive nitrene intermediates. After optimizing the reaction conditions, they found this reaction is suitable for a wide range of substrates bearing active C–H bonds (Scheme 1.38). For the substrates bearing benzylic C–H bonds, the reaction went smoothly to give the products with excellent de values. However, for the substrates bearing allylic C–H bonds, the stereocontrol was just moderate. For the simple alkanes, the desired C–H amination products were also obtained in high yields. Deprotection of the corresponding aminated product under reductive conditions gave the corresponding enantiopure amine, which is of high interest for synthetic chemists.

In 2008, Müller, Dodd, Dauban, and co-workers further studied their previously reported⁸⁴  stereoselective C–H bond amination reactions in detail.⁸⁶  With this method, various substrates could be employed in this reaction, including compounds bearing either benzylic C–H bonds or allylic C–H bonds and simple alkanes. Besides using a routine chiral reagent controlled strategy, kinetic resolution also worked well (Scheme 1.39). By using excess amounts of racemic sulfonimidamide 107, the kinetic resolution reaction proceeded to yield the desired product 108 with high diastereoselectivity (98% de) which parallels the result obtained from (S)-107. As shown in Scheme 1.39, this reaction is considered to be a sequence of three steps involving: (1) in situ generation of iminoiodane from PhI(OCO^tBu)₂ and the nitrogen-containing substrate, (2) formation of a Rh–nitrene intermediate following oxidation of the Rh^II complex by the iminoiodane, and (3) insertion of the resulting metallanitrene into the C–H bond.

In 2006, Davies and co-workers demonstrated dirhodium tetracarboxylate, Rh₂(S-TCPTAD)₄ as an efficient catalyst to enable the intermolecular and intramolecular (not shown here) C–H bond amination reactions, respectively.¹⁰  Utilizing this catalyst, the aminated product 98 was obtained in excellent levels of yield (95%) and enantioselectivity (94% ee) by employing indane as the substrate (Scheme 1.40). However, the enantioselectivity was just moderate for other substrates including the substituted indane derivatives. With the aminated product 98 in hand, an anti-Parkinson agent rasagiline 109 was synthesized via an N-propargylation/deprotection sequence.

Later, Du Bois and co-workers reported a highly efficient Rh-catalyzed enantioselective C–H bond amination reaction.⁸⁷  They developed a new chiral dirhodium catalyst Rh₂(S-NAP)₄ which contains a hydrogen bond between the N–H moiety and the carbonyl oxygen of the amide and found that it has an excellent performance in the C–H bond amination reaction. As shown in Scheme 1.41, by employing the in situ generated nitrene species methodology, compounds bearing either benzylic C–H bonds (110) or allylic C–H bonds (112) underwent the intramolecular C–H bond amination reactions, producing the corresponding products (111 and 113) in high levels of enantioselectivity. The reactions involving benzylic C–H bonds are generally more effective than those involving allylic C–H bonds.

In 2010, by employing asymmetric C–H bond amination reactions developed by Du Bois,⁸⁷  Kang and Lee reported a highly efficient and enantioselective synthesis of (S)-dapoxetine, which is a potent selective serotonin reuptake inhibitor.⁸⁸  As shown in Scheme 1.42, this synthesis commenced with a nucleophilic substitution reaction, producing the nitrene precursor 114. Subsequently, sulfamate ester 115 was obtained through the Du Bois asymmetric C–H bond amination reaction in high levels of yield and enantioselectivity. Methylation of the sulfamate ester afforded the cyclized intermediate 116, which led to the free secondary amine 117 by reacting with 1-naphthol in the presence of NaH. Finally, a reductive amination reaction took place to afford (S)-dapoxetine. During the course of this study, the absolute configuration of the enantiomer of 115 was determined to be R and not S as reported originally by Du Bois when the same chiral catalyst was used.⁸⁷ 

Different from their previous reports,⁸⁹  Hashimoto and co-workers developed a novel reaction in 2009.⁹⁰  They found that allylic C–H bond amination reaction occurs when silyl enol ether 120 derived from cyclohexanone is used. This interesting insertion reaction is ascribed to the fact that silyl enol ether 120 can just be E-shaped. On the other hand, when Z-shaped silyl enol ethers 118 were employed (Scheme 1.43, top), the aziridination reactions occurred under similar conditions. After treating with trifluoroacetic acid, the α-aminated carbonyl products 119 were obtained in high levels of enantioselectivity. Starting from the β-aminated carbonyl compound 121, they accomplished the formal synthesis of (−)-pancracine (Scheme 1.43, bottom).

A new chiral nitrene precursor 123 was introduced into the stereoselective aziridination and C–H bond amination reactions by Lebel and co-workers in 2011 (Scheme 1.44).⁹¹  It was found that monosubstituted alkenes 122 can produce aziridinated products 124 while di- or tri-substituted alkenes 125 yield aminated products 126 at the allylic positions. In all cases, high levels of yield and stereocontrol were obtained under mild conditions. Another advantage associated with this method is that the Ph-Troc protecting group can be easily cleaved by using zinc in acetic acid. Treating the Ph-Troc allylic amine 126b with lithium methoxide produced the methyl carbamate 128 in good yield without notable erosion of the enantiomeric purity. Moreover, the chiral alcohol 129 can be recovered with high ee value, suggesting great potential for practical utilization with this method. Later, by using the same strategy, the same group expanded the substrate scope to include compounds bearing either a benzylic C–H bond or a propargylic C–H bond with excellent stereoselectivity.⁹² 

The previous excellent results obtained by combining a chiral amide with a chiral catalyst encouraged Dauban’s group to further explore the chemoselective functionalization of more complex molecules with the aim to better predict the selectivity of intermolecular C–H bond amination reactions.⁹³  In 2012, they reported the site-selective C–H bond amination of complex substrates including terpenes, enol ethers, and alkanes (Scheme 1.45, top). Employing this method gives access to enantiopure aminated derivatives, which are not easily obtained by classical organic synthesis. Recently, they employed this strategy to synthesize a series of octahydroindole derivatives with high efficiency (Scheme 1.45, bottom).⁹⁴ 

In the context of C–H bond insertion reactions involving metal nitrenoids, Mn and Ru catalysts hold important positions. These catalysts often contain multidentate ligands, such as porphyrins and salen ligands. With these catalysts, the enantioselectivity can be well controlled in many cases.

In 1999, Che and co-workers reported the first asymmetric Mn- and Ru-catalyzed asymmetric amidation reactions of saturated C–H bonds.⁹⁵  As shown in Scheme 1.46, the chiral porphyrin with metal complexes C9 and C10 were prepared according to the procedures reported previously.⁹⁶  Subsequently, amidation reactions were employed to test the efficiency of these two catalysts. Pleasingly, the amidated products were obtained with moderate yields and enantioselectivity. Under the same conditions, the manganese catalyst C10 affords the product in a considerably higher yield than the ruthenium catalyst C9. Interestingly, the enantiocontrol has a similar trend, the manganese catalyst C10 being more effective. Efforts have also been made to ascertain the active species in these reactions. They found that complex C11 likely functions as the active species responsible for the stereocontrol.

In 2002, Che and co-workers expanded their previous intermolecular amidation reactions to intramolecular versions.⁹⁷  This time, ruthenium porphyrin complex C9 displayed great selectivity. The corresponding amidated products were easily prepared with good enantioselectivity by using indan-2-yl sulfamate 101 or phenethyl 130 (phenpropyl 132) sulfamates as the starting materials (Scheme 1.47). Soon afterwards, they studied the reactions catalyzed by C9 in detail, including intramolecular amidation and aziridination reactions.⁹⁸  In both cases, good conversions and stereocontrols were observed. It was also surmised that amidation reactions occur through an intramolecular hydrogen atom abstraction process.

(Salen)manganese complexes were introduced into this area in 2001 by Katsuki and co-workers.⁹⁹  After screening various ligands, they uncovered a manganese salen complex containing an electron-withdrawing group (Br) which gives the best results. As shown in Scheme 1.48, under the catalysis of C12, aminated indane compound 134 was produced in 63% yield and 66% ee by using PhINTs as the nitrene precursor.

In 2005, Che and co-workers tested the intramolecular amidation reactions by employing chiral manganese Schiff base complexes.¹⁰⁰  After extensive investigations, C12′ was found to be the best catalyst (Scheme 1.49, top). To shed light on the rate-limiting step, competition experiments were performed with different groups on the arene core. The preliminary results suggested that electron-donating groups accelerate the insertion reaction whereas electron-withdrawing groups retard this process (Scheme 1.49, bottom). Meanwhile, it was suggested that this reaction is stereospecific as the enantiopure substrate gives the desired product with the identical absolute configuration.

A highly effective Ru-catalyzed asymmetric intramolecular amination of benzylic and allylic C–H bonds was developed in 2008 by Blakey and co-workers.¹⁰¹  Ru–Pybox complexes were readily prepared by using Nishiyama’s method.¹⁰²  Initially, the reaction was tested without any additives. However, the desired product 133 could be produced in only modest yields and ee values (Table 1.10, entries 1–3). Based on the studies by Du Bois,¹⁰³  the authors rationalized that a cationic catalyst would be more reactive than the neutral one. Thus, the catalyst prepared from halide abstraction by AgOTf gave improved yields and selectivities (Table 1.10, entries 4–6). Further screening of ligands, solvents, and temperature led to the optimal conditions: reaction of sulfamate ester 132 with 5 mol% of C16, 110 mol% of PhI(OAc)₂, and 230 mol% of MgO in benzene at 4 °C (Table 1.10, entries 7–14). Under these conditions, a series of five- and six-membered-ring products were obtained with up to 71% yield and 92% ee. Notably, the cis-olefin substrate 135 could also participate in the reaction (Scheme 1.50). Under the standard conditions, the corresponding six-membered ring product 136 was obtained and the aziridination product was not observed.

In 2013, Katsuki and co-workers designed and developed a highly efficient Ru catalyst for the asymmetric amination reactions.¹⁰⁴  They have achieved a highly enantio- and regioselective intermolecular benzylic and allylic C–H bond amination by using a new Ru(CO)–salen complex C17 as the catalyst and SESN₃, [2-(trimethylsilyl)ethanesulfonyl] azide (caution: this compound is potentially explosive), as the nitrene precursor (Scheme 1.51). Under these conditions, a lot of substrates which contain benzylic or allylic C–H bonds were transformed to their corresponding aminated products with excellent enantioselectivity. It is noteworthy that only methyl, ethyl, and cyclic methylene groups in the benzylic or allylic positions can be aminated. In addition, the amination product of aliphatic linear Z-olefin derivatives is not observed.

Copper nitrenoids that participate in C–H bond insertion reactions have also been studied for a long time. In 1997, the first asymmetric reaction was reported by Katsuki and co-workers.¹⁰⁵  Their design was derived from the Kharash–Sosnovsky reaction (Scheme 1.52, top),¹⁰⁶  the Cu-catalyzed allylic oxidation. An asymmetric amination was examined by employing C18 as the catalyst and simple indane as the substrate. However, the amination product was obtained in only 4% yield and 28% ee (Scheme 1.52, bottom).

In 2010, Nicholas and co-workers studied the enantioselective benzylic amination reactions.¹⁰⁷  As shown in Table 1.11, they tested different kinds of ligands including (S)-histidine, (S)-proline, diimine ligands, and chiral phenanthroline. Although the corresponding product can be prepared with high yields, the enantioselectivity is low. The preliminary results from the mechanistic studies support a stepwise C–H bond insertion process, most likely through the intermediacy of carbon-centered radicals. Subsequently, they expanded this reaction to an intramolecular version with up to 18% ee.¹⁰⁸ 

In 2011, Katsuki and co-workers achieved a highly enantioselective intramolecular C–H bond amination with azide compounds as the nitrene precursors for the first time.¹⁰⁹  They synthesized a series of Ir–salen complexes by a modular route. A survey of these catalysts revealed that the cyclization of 2-ethylbenzenesulfonyl azides 137 proceeded smoothly only at the benzylic positions in high levels of enantioselectivity. Surprisingly, when acyclic substrates having a homobenzylic methylene carbon atom such as 139 were employed, six-membered sultams 141 were produced (Scheme 1.53). It was inferred that the control of regio-, diastereo- and enantioselection largely depends on whether the substrate can adopt a conformation that permits an appropriate orbital interaction between the C–H moiety and Ir-nitrenoid bonds. This study demonstrates that Ir-catalyzed nitrenoid insertion is a promising approach by using an atom-economical nitrene precursor.

Nature has always been a source of inspiration for scientists. While most of research developments are centered on simulating nature, it would be worthwhile to challenge nature’s ingenuity to mimic the synthetic reactions. To this end, catalytic C–H bond amination is an excellent platform for the development of a non-natural enzymatic reaction. Whereas enzymes are capable of inserting oxygen atoms into even inactivated C–H bonds, the sites into which nitrogen atoms can be incorporated are more constrained. Therefore, advancements in this area have been prevented for a long time.

In 2013, Arnold and co-workers revisited the possibility of engineering an enzyme catalyst for this useful transformation.¹¹⁰  Combining ortho-substituted benzenesulfonyl azides 142 with engineered P450 enzymes which include a reduced Fe²⁺ center resulted in an efficient intramolecular benzylic C–H bond amination reaction. The desired aminated products 143 were produced in 73% ee with a total turnover number (TTN) of 383. In addition, expression of the catalyst in E. coli makes the reaction proceed smoothly on a 50 mg scale with high ee value (Scheme 1.54).

Recently, Fasan and co-workers provided an alternative enzyme catalyst for this process.¹¹¹  The optimal enzyme was found to be P450_BM3FL#62, which afforded the products 143 with up to 390 TTN and 91% ee (Scheme 1.55). Meanwhile, this transformation is feasible on a preparative scale (30 mg, 42% yield). Soon after, Arnold and co-workers developed two engineered variants of P450_BM3 that provide divergent regioselectivity for C–H bond amination reactions.¹¹²  By employing different types of enzyme variants, either α- or β-amination of 2,5-disubstituted benzene sulfonyl azides occurred, respectively.

Development of an efficient catalyst for selective oxygenation is an important objective in synthetic organic chemistry. However, the asymmetric C–H bond oxidation reactions by metal oxo species are still challenging due to the over oxidation of the newly formed C–O bonds. Metalloenzymes always accomplish highly efficient and selective oxygenation of organic molecules under mild conditions. To mimic these systems, a few artificial catalysts have been developed for this process.

The initial examples of Fe-oxo species catalyzed C–H bond insertion reactions employed chiral iron porphyrin complexes. As early as the 1980s, Viski and co-workers demonstrated the utility of a chiral iron porphyrin catalyst C20 in the asymmetric hydroxylation reaction (Scheme 1.56).¹¹³  Under the catalysis of C20, benzylic alcohols 145 were prepared in the presence of stoichiometric amounts of oxidant PhIO. At the same time, they also tested a manganese porphyrin complex, which also led to the hydroxylated product, albeit with lower efficiency.

In 2005, Itoh and co-workers developed a new dinucleating ligand L13.¹¹⁴  Combining L13 with FeCl₃ gave a dinuclear Fe^III complex. With this catalyst, hydroxylation of benzylic C–H bonds was realized. Among these cases, one example was the asymmetric hydroxylation of tetrahydronaphthalene, affording the desired oxidized product 146 with only 9.9% ee (Scheme 1.57).

The search for general and selective C–H bond oxidation reactions never stops due to its importance in synthetic chemistry. A breakthrough in this area has been realized in the realm of Fe-catalyzed oxidation reactions reported by White and co-workers since 2007.¹¹⁵  They developed an electrophilic iron catalyst C21 with a bulky ligand framework that utilized H₂O₂, an inexpensive and environmentally friendly oxidant, to enable highly selective oxidations of inactivated C(sp³)–H bonds over a broad range of substrates. For example, after three consecutive additions of catalyst C21 (5 mol%), HOAc (0.5 equiv.), and H₂O₂ (1.2 equiv.) over a period of 30 min, the corresponding hydroxylated product 148 was isolated in a preparative yield (51%) (Scheme 1.58, top). Selective oxidation of methylene moieties to ketones could be achieved by the same catalytic system if no suitable tertiary C–H bonds existed.¹¹⁶  It was also shown that sites that are proximal to electron-withdrawing groups are deactivated towards oxidation, while sterically hindered C–H bonds are shielded from the bulky oxidant.¹¹⁷  Subsequently, they found that carboxylic acids can overcome a range of substrate bias (in terms of electronic, steric, and stereoelectronic properties) in C–H bond oxidation reactions mediated by the non-heme iron catalyst C21.¹¹⁸  The possible reason is that the carboxylic acid group can act as the ligand for the iron complex. When racemic carboxylic acid 149 was exposed to the standard conditions, kinetic resolution of 149 occurred and moderate enantioselectivity was obtained for both the starting material and the product (Scheme 1.58, middle). Furthermore, the catalyst control of site selectivity was achieved by changing the catalyst to the CF₃-modified complex C22.¹¹⁹  Artemisinin 151 was transformed to C10-oxidized hydroxyl artemisinin 152 under the catalysis of C21. However, when C22 was used, C9-oxidized 9-oxo-artemisinin 153 was obtained as the main product. Furthermore, the development of quantitative structure-based catalyst reactivity models could predict the ratio of the site selectivity (Scheme 1.58, bottom). This discovery should inspire and guide future catalyst design.

In 1994, Jacobsen and co-workers demonstrated that stereoselective oxidation of benzylic C–H bonds is possible utilizing readily available chiral Mn(salen) complexes.¹²⁰  They studied the kinetic resolution of 1,2-dihydronaphthalene oxide via an asymmetric C–H bond hydroxylation reaction (Scheme 1.59). During the course of experiments on the asymmetric epoxidation of 1,2-dihydronaphthalene with C24, it was observed that the enantiomeric excess of the desired epoxide product 154 increases with time at the expense of product yield. This behavior is suggestive of a secondary kinetic resolution process. Then, they investigated the scope of this process, and the efficient kinetic resolution of 154 (s = 4.8) indicated that high asymmetric induction is possible in C–H bond hydroxylation reactions catalyzed by manganese catalysts.

In 1996, Katsuki and co-workers studied asymmetric benzylic oxidation reactions by using Mn(salen) complexes.¹²¹  However, the best result they obtained was just 29% yield and 64% ee when 1,1-dimethylindane was used as the substrate and C25 as catalyst (Scheme 1.60). Later, they further enhanced the enantioselective control to 90% ee by changing the catalyst to C26, albeit the yield is only 24.5%.¹²² 

In 1997, the same group disclosed an example of highly enantioselective C–H bond oxidation reactions.¹²³  As shown in Scheme 1.61, conformationally fixed 3-oxa-bicycle[3.3.0]-octane 156 was oxidized at the α position under the catalysis of the Mn catalyst C25 ligated by a BINOL-derived salen ligand. Subsequently, the same group further expanded the scope to a meso-pyrrolidine derivative (up to 76% ee) and a meso-tetrahydrofuran derivative (up to 90% ee).¹²⁴  Later, by using a different Mn(salen) catalyst, Murahashi and co-workers realized the direct oxidation of an alkane 158, furnishing the optically active ketone 159 with up to 22% ee (Scheme 1.61).¹²⁵ 

A novel manganese complex C29 bearing a chiral strapping unit was developed and used in the asymmetric hydroxylation reactions by Murahashi and co-workers (Scheme 1.62).¹²⁶  It was found that this complex is effective for C–H bond hydroxylation of the indane derivatives yielding the hydroxylated products with up to 34% ee.

Recently, an efficient and enantioselective hydroxylation reaction of C–H bonds using an environmentally benign and easily accessible oxidant (H₂O₂) catalyzed by a water-soluble chiral manganese porphyrin complex C30 was developed by Simonneaux and co-workers.¹²⁷  The corresponding secondary alcohols 161 were delivered as the major products, albeit only moderate enantiomeric excess was obtained (Scheme 1.63). Notably, imidazole was found to play an important role in this catalytic system, working as an axial ligand to the Mn center.

In 1999, Che and co-workers designed and synthesized a D₄-symmetric chiral ruthenium porphyrin complex C31.¹²⁸  With this catalyst, hydroxylation reactions occurred smoothly at the benzylic positions (Scheme 1.64). In this way, secondary alcohols 161 can be prepared with up to 76% ee from the simple ethylbenzene derivatives 160. It was postulated that the production of enantioenriched products 161 may arise from the preferential collapse of the benzylic radical on the pro-S face versus the pro-R face in the oxygen atom rebound step.

In the past decade, asymmetric C–H bond insertion reactions, especially the ones involving carbenoids generated by dirhodium complexes and diazo compounds, have been investigated marvelously and applied to a wide range of substrates, even in the synthesis of highly complex molecules and natural products. An array of carbene precursors and mature catalytic systems derived from chiral dirhodium catalysts are readily available, rendering asymmetric C–H bond insertion reactions highly selective, tunable, and efficient. With the advent of donor/acceptor carbenoids, intermolecular C–H bond insertion has evolved into a powerful strategy in C–H bond functionalization reactions. Particularly, the combined C–H functionalization/Cope rearrangement (CHCR) was surveyed extensively and has become a reliable method in organic synthesis. Meanwhile, computational studies have also provided detailed mechanism and prediction models in asymmetric C–H bond insertion reactions, which renders this methodology more integrated. On the other hand, asymmetric C–H bond insertion by metal nitrenoids and metal oxo species have also witnessed intensive development in recent years. Quite a few competent transition metal-based chiral catalytic systems that are suitable for these transformations have emerged.

Nevertheless, there still exist some obstacles to be overcome. Generally, the diazo compounds as carbene precursors require a slow addition technique to avoid potential side reactions such as dimerization. For C–H bond insertion by metal nitrenoids, and metal oxo species, only a handful of substrates can lead to products in satisfactory yield and enantioselectivity. Under these circumstances, albeit in limited cases, new carbene precursors are still in great demand to make the reaction more operative. Since Rh catalysts are not so cheap, reduction in the catalyst loading, recovery of catalyst, simplified synthetic routes for the catalyst, and other cheap and efficient metal catalysts might be the potential solutions to solve this issue. Furthermore, C–H bond insertion by metal nitrenoids, and metal oxo species are largely unexplored. The discovery of brand new precursors and suitable combinations of metals and ligands could be conducive to bring about a highly selective and efficient catalytic system for diverse substrates.