DOI: 10.1002/ajoc.201300016

Synthesis of Pyrroles, Indoles, and Carbazoles through Transition-Metal-Catalyzed C�H Functionalization

Naohiko Yoshikai*[a] and Ye Wei[b]

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim466

FOCUS REVIEW

Abstract: Pyrroles, indoles, and carbazoles are among the most important families of nitrogen-containing heterocycles that occur frequently in natural products, pharmaceuticals, agrochemi-cals, and other functional molecules. Consequently, improved syntheses of these compounds con-tinue to interest synthetic chemists. This Focus Review describes recent advances in syntheticmethods for producing these privileged heterocycles that feature transition-metal-catalyzed C�Hactivation approaches. Because of the common five-membered pyrrole core, some of the C�Hactivation approaches are applicable to two or more of the pyrrole, indole, and carbazole skele-tons. The reactions discussed here not only serve as atom- and step-economical alternatives tothe existing synthetic methods, but also show the latest developments in organometallic chemis-try and homogeneous catalysis.

Keywords: C�H activation · heterocycles · homogeneous catalysis · palladium · rhodium

1. Introduction

Pyrrole and its benzo-fused analogues indole and carbazoleare extremely common structural units in biologicallyactive natural and unnatural compounds, as well as in dyes,pigments, and other functional materials (Scheme 1).[1] Con-

sequently, the development of synthetic methods for pro-ducing these privileged heterocyclic scaffolds has receivedconsiderable attention from synthetic chemists.[2,3] Whereasclassical synthetic methods, such as the Paal–Knorr pyrrolesynthesis and the Fischer indole synthesis, have been impor-tant methods for more than a hundred years, transition-metal-catalyzed reactions have emerged and evolved intopractical alternatives over the past several decades. Thepreparation of a tryptamine-based gonadotropin-releasinghormone (GnRH) antagonist through Larock, Fischer, andCastro indole syntheses is an illustrative example thatshows the practical utility of such well-established syntheticmethods (Scheme 2).[4]

A common problem in many of the existing syntheticmethods is, however, the requirement for “prefunctional-ized” starting materials. For example, with respect to theabove-mentioned indole syntheses, Fischer and Larock/Castro methods require carcinogenic aryl hydrazines andrather expensive 2-iodoaniline derivatives, respectively. Fur-

[a] Prof. N. YoshikaiDivision of Chemistry and Biological ChemistrySchool of Physical and Mathematical SciencesNanyang Technological UniversitySingapore 637371 (Singapore)Fax: (+65) 6791-1961E-mail : nyoshikai@ntu.edu.sg

[b] Prof. Y. WeiCollege of PharmacyThird Military Medical University, (China)

Scheme 1. Examples of biologically active and other functional mole-cules that contain pyrrole, indole, or carbazole moieties.

Scheme 2. Preparation of a tryptamine-based GnRH antagonist throughLarock, Fischer, and Castro indole syntheses.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim467

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

thermore, the diversity of commercially available aryl hy-drazines and 2-haloanilines is relatively limited and muchnarrower than that of simple anilines. This limited availabil-ity poses a serious obstacle when one wishes to access a di-verse array of indoles with various substituents on the ben-zene ring.

In the above context, over the last several years, remark-able progress has been made in the syntheses of pyrroles,indoles, and carbazoles through transition-metal-catalyzedC�H bond functionalization, which is summarized in thisFocus Review.[5] Herein, such emerging synthetic methodsare classified based on the key bond disconnection(Scheme 3). Thus, transition metal-catalyzed C�H activa-

tion has proven effective for forging C�C bonds throughdehydrohalogenation or dehydrogenation (disconnection a),C�C and C�N bonds simultaneously through annulationwith an alkyne (disconnections a and b), or C�N bondsthrough oxidative amination or (formal) nitrene insertion(disconnection b or c).

2.1. C�H/C�X Coupling

In the early 1980s, Ames and Opalko reported that diarylethers and diaryl amines with a 2-bromo or 2-iodo substitu-ent underwent palladium-catalyzed dehydrohalogenativecyclization to afford dibenzofurans and carbazoles, respec-tively.[6] This seminal work on biaryl synthesis through C�H

arylation,[7] however, was scarcely revisited for a longtime.[8] In 2002, Bedford and Cazin reported a palladium-catalyzed one-pot synthesis of N-alkylcarbazole from 2-chloro-N-alkylaniline and aryl bromides by sequentialBuchwald-Hartwig amination and C�H arylation(Scheme 4).[9] Modification of the original catalytic system

later allowed the synthesis of N�H carbazoles from unpro-tected 2-chloroanilines in either a one-pot or stepwisemanner.[10] Indole derivatives were also synthesized from 2-chloroanilines and bromoalkenes.

Abstract in Japanese:

Naohiko Yoshikai received his B.Sc. (2000),M.Sc. (2002), and Ph.D. degrees (2005)from the University of Tokyo under theguidance of Professor Eiichi Nakamura.Then he was appointed as an Assistant Pro-fessor of Chemistry at the same institute. In2009, he moved to Singapore to join the Di-vision of Chemistry and Biological Chemis-try, School of Physical and MathematicalSciences, at Nanyang Technological Univer-sity as a Nanyang Assistant Professor/Sin-gapore National Research Foundation Re-search Fellow. His research interests are fo-cused on the development and mechanisticstudy of new transition-metal-catalyzed re-actions and their synthetic applications.

Ye Wei was born in 1982 in China. Heearned his B.Sc. degree in 2005 from Hua-qiao University and his Ph.D. degree in2010 from Fujian Institute of Research onthe Structure of Matter, Chinese Academyof Sciences, under the guidance of Prof.Weiping Su. After a two-year postdoctoralappontiment with Prof. Naohiko Yoshikaiat Nanyang Technological University, hewent back to China to join the College ofPharmacy at Third Military Medical Uni-versity as an Associate Professor. His cur-rent research interests are primarily focusedon atom- and step-economic synthetic

methods, with an emphasis on the transition-metal-mediated organictransformations and exploiting metal–organic cooperatively catalyzed re-actions.

Scheme 3. Possible disconnections of pyrrole, indole, and carbazole, andthe corresponding C�H functionalization strategies.

Scheme 4. Carbazole synthesis through sequential C�N cross-coupling/intramolecular direct C�H arylation.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim468

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

In 2004, Fagnou and co-workers published the first majorcontribution of their group in the area of C�H functionali-zation, that is, palladium-catalyzed intramolecular biarylsynthesis through direct arylation, which allowed efficientformation of six- and seven-membered rings.[11] The powerof their Pd�PCy3 catalytic system was also demonstratedfor the synthesis of carbazole derivatives.[12] For example,the carbazole natural product Mukonine was readily syn-thesized with the intramolecular C�H arylation as the keystep (Scheme 5).

In 2007, Ackermann and Althammer reported an alterna-tive strategy for one-pot carbazole synthesis through se-quential C�N and C�C bond formations (Scheme 6). Thus,

a variety of carbazole derivatives, including the naturalproduct Murrayafoline A, were assembled from simple ani-line derivatives and o-dihalogenated arenes with the aid ofa Pd�PCy3 catalyst.[13] The reaction also allowed the synthe-sis of an annulated indole from cyclic 1,2-dibromoalkenes.

Urabe and co-workers developed the nucleophilic addi-tion of sulfonamides to bromoalkynes to afford olefins withsulfonamide and bromo groups and cis stereochemistry.[14]

With an aryl group on the nitrogen atom, the adduct wassuitable for palladium-catalyzed intramolecular C�H/C�Brcoupling to furnish an N-sulfonylindole (Scheme 7).

2.2. C�H/C�H Coupling

Oxidative linkage of two C�H bonds is one of the idealstrategies for C�C bond formation.[15] This strategy be-comes particularly attractive if the process is catalytic andrequires only a harmless and inexpensive oxidant. In 1975,�kermark et al. found that a stoichiometric amount of Pd-ACHTUNGTRENNUNG(OAc)2 effected cyclization of diarylamines to carbazoles(Scheme 8 a).[16] Subsequent studies by the groups of �ker-

mark and Knçlker significantly improved the reaction anddemonstrated its utility in the synthesis of naturally occur-ring carbazole alkaloids.[17] Most importantly, the reactionwas made catalytic in Pd ACHTUNGTRENNUNG(OAc)2 by using an oxidant such asCuACHTUNGTRENNUNG(OAc)2 or molecular oxygen (Scheme 8 b). On the otherhand, the acidic reaction medium (acetic acid) and harshreaction conditions often caused undesirable side reactionsand, thus, limited the substrate scope. Notable modifica-tions have been made, including the one-pot N-arylation/oxidative coupling protocol by Fujii and co-workers,[18] theuse of pivalic acid as a superior solvent by Fagnou et al. ,[19]

and the use of Pt/C instead of a palladium catalyst in hy-drothermal water by Yamamoto and Matsubara.[20]

Scheme 5. Synthesis of Mukonine by using Fagnou and co-workers�direct arylation protocol. Tf = trifluoromethanesulfonyl.

Scheme 6. Carbazole/indole synthesis from anilines and dihalogenatedarenes/olefins.

Scheme 7. Preparation of N-sulfonylindole through addition of sulfona-mides to bromoacetylenes with subsequent intramolecular C�H alkeny-lation. DMF =N,N-dimethylformamide.

Scheme 8. (a) Pd ACHTUNGTRENNUNG(OAc)2-mediated cyclization of diarylamine to give car-bazole. (b) Pd ACHTUNGTRENNUNG(OAc)2-catalyzed, Cu ACHTUNGTRENNUNG(OAc)2-mediated cyclization of aryla-minoquinone to carbazolequinone, and carbazole alkaloids synthesizedthrough �kermark–Knçlker cyclization.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim469

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

In 2008, Glorius and co-workers brought about a signifi-cant breakthrough in the scope of the oxidative palladiumcatalysis by introducing a new substrate class and modifiedreaction conditions. They showed that N-aryl enamines,which can be readily prepared from anilines and b-dicar-bonyl compounds, are efficiently cyclized into indoles in thepresence of catalytic Pd ACHTUNGTRENNUNG(OAc)2 and a stoichiometricamount of CuACHTUNGTRENNUNG(OAc)2 with K2CO3 and DMF as the base andthe solvent, respectively (Scheme 9).[21] A variety of indolederivatives with electron-withdrawing substituents on theC3 position were synthesized and the reaction tolerateda remarkably broad range of functional groups.

The follow-up study by the same group led to an im-proved catalytic system with K3PO4 instead of K2CO3,which proved effective on a preparatively useful scale.[22]

The reaction was proposed to begin with electrophilic palla-dation of the enamine and subsequent deprotonation,which led to a vinylpalladium(II) species (Scheme 10). Thisspecies then undergoes intramolecular aromatic C�H acti-vation, and a subsequent reductive elimination affords theproduct and palladium(0), which is then reoxidized to palla-

dium(II) with the aid of CuACHTUNGTRENNUNG(OAc)2. Based on the results ofmechanistic studies, which included a large H/D kinetic iso-tope effect (KIE), the intramolecular C�H activation stepwas proposed to involve s-bond metathesis or base-assisteddeprotonation[23] rather than electrophilic palladation.

Glorius and co-workers’ discovery of the enamine cycli-zation was followed by the development of several alterna-tive reaction systems for the same type of transformation.Cacchi and co-workers developed a copper-based systemwith air as an oxidant (Scheme 11 a).[24] Zhao and co-work-

ers reported a metal-free reaction system with iodobenzenediacatate as a stoichiometric oxidant (Scheme 11 b).[25, 26]

Liang and co-workers reported a catalytic system compris-ing FeCl3 as a catalyst and CuACHTUNGTRENNUNG(OAc)2·CuCl2 as an oxidant(Scheme 11 c).[27]

Building on the chemistry pioneered by �kermark,Knçlker, and Glorius, in 2012 we disclosed a palladium-cat-alyzed aerobic oxidative cyclization reaction of N-arylimines to indoles (Scheme 12).[28, 29] A variety of N-arylimines derived from anilines and ketones, including aceto-phenones, 2-arylacetophenones, and aliphatic methyl ke-tones, for example, cyclopropyl methyl ketone, were effi-

Scheme 9. Palladium(II)-catalyzed, copper(II)-mediated cyclization ofN-arylenamines to give indoles. EWG = electron-withdrawing group;Piv =pivaloyl.

Scheme 10. Proposed mechanism for palladium(II)-catalyzed indole for-mation from an enamine.

Scheme 11. Several reaction systems for oxidative cyclization of N-arylenamines to give indoles. DCE =1,2-dichloroethane.

Scheme 12. Palladium(II)-catalyzed aerobic cyclization of N-aryl iminesto give indoles.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim470

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

ciently converted into indole products in the presence ofcatalytic Pd ACHTUNGTRENNUNG(OAc)2 and molecular oxygen (1 atm.) withBu4NBr and dimethyl sulfoxide (DMSO) as the crucial ad-ditive and solvent, respectively, under mild conditions.

The catalytic cycle shown in Scheme 13 was proposed.Palladation of the N-aryl enamine generated through tauto-merization of the imine and subsequent elimination of

HOAc gives an a-palladated imine.[30] This intermediate un-dergoes intramolecular aromatic C�H activation then re-ductive elimination to afford the 3H-indole and palladi-um(0). The former quickly tautomerizes to the indole prod-uct and the latter is oxidized to palladium(II) by molecularoxygen and HOAc. The mechanism of the intramolecularC�H activation step appears similar to that involved inGlorius and co-workers� enamine cyclization because theH/D KIE is similar in magnitude.

As a major limitation of the imine cyclization reaction,imines with b-hydrogen atoms underwent dehydrogenationrather than oxidative cyclization, presumably through rapidb-hydride elimination of the putative a-palladated imine(Scheme 14). This reactivity could be used for the prepara-tion of arylamines from amines and cyclohexanonesthrough dehydrogenative aromatization of the imine inter-mediates.[31,32]

2.3. C�H/N�H/Alkyne Annulation

Annulation of o-haloaniline derivatives and alkynes, knownas the Larock indole synthesis, offers a highly reliablemethod for the synthesis of indoles that is compatible withvarious functional groups. On the other hand, this methodsuffers from the high cost and limited availability of o-hal-oanilines. In this context, the rhodium ACHTUNGTRENNUNG(III)-catalyzed oxida-tive annulation reaction disclosed by Fagnou and co-work-ers in 2008, is a highly attractive alternative(Scheme 15).[33, 34] Thus, a cationic rhodium ACHTUNGTRENNUNG(III) catalyst,

which is generated from [Cp*RhCl2]2 and AgSbF6, in com-bination with a stoichiometric copper(II) oxidant, alloweddirect coupling of N-acetyl anilines and internal alkynes toform a variety of N-acetyl indoles.

A subsequent investigation by the same group led toa significant improvement of the catalytic system. Witha preformed cationic rhodium ACHTUNGTRENNUNG(III) complexACHTUNGTRENNUNG[Cp*Rh ACHTUNGTRENNUNG(MeCN)3]ACHTUNGTRENNUNG[SbF6]2 as a catalyst, the reaction wasmade much milder and catalytic in copper(II) with molecu-lar oxygen (1 atm.) as a terminal oxidant.[35] The improvedcatalytic system enabled a concise synthesis of Paullone(Scheme 16 a) and extended the scope of the reaction to thesynthesis of pyrroles from enamides and alkynes(Scheme 16 b). The regioselectivity of the reaction with un-symmetrical alkynes was primarily governed by electroniceffects, and moderately affected by steric factors.

Extensive mechanistic studies led to the proposal of thecatalytic cycle outlined in Scheme 17, which involves depro-tonative ortho-rhodation assisted by the amide-oxygenatom, insertion of the alkyne into the Rh�aryl bond thendeprotonation to form a six-membered rhodacycle, andC�N reductive elimination to generate the indole productand a rhodium(I) species that is oxidized to rhodium ACHTUNGTRENNUNG(III)with the aid of CuACHTUNGTRENNUNG(OAc)2/O2. Whereas the ortho-rhodationstep is intrinsically reversible, its reversibility in the annula-tion reaction depends on the rate of the subsequent alkyneinsertion step. Therefore, it is reversible and irreversiblewith the original and the improved catalytic systems, re-spectively. The importance of precoordination of the alkyne

Scheme 13. Proposed mechanism for palladium(II)-catalyzed indole for-mation from N-aryl imines.

Scheme 14. Dehydrogenative aromatization of an N-aryl imine derivedfrom tetralone.

Scheme 15. Rhodium ACHTUNGTRENNUNG(III)-catalyzed, copper(II)-mediated oxidative an-nulation of acetanilides and alkynes to give indoles. Am=amyl.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim471

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

to the rhodiumACHTUNGTRENNUNG(III) catalyst was also indicated but is notshown in Scheme 17.

A significant limitation of the annulation reaction waspoor regioselectivity with unsymmetrical dialkylalkynes.Fagnou and co-workers solved this problem by combiningoxidative annulation of enynes and hydrogenation of theresulting indole/pyrrole products (Scheme 18).[36] Thus, the

regiochemistry of the final dialkylindole/pyrrole productscan be secured by the regioselective annulation of enynes,in which C�N bond formation takes place at the carbonatom that is proximal to the alkenyl group.

Glorius and co-workers discovered a different type ofrhodium ACHTUNGTRENNUNG(III)-catalyzed oxidative annulation reaction toproduce pyrroles (Scheme 19).[37] Enamides with methyl

and ester groups at the a and b positions, respectively, un-derwent annulation with an alkyne at the N�H and methylC�H bonds, presumably through C ACHTUNGTRENNUNG(sp3)�H bond activation.Interestingly, when the b substituent R1 was H, the rhodiumcatalyst appeared to reversibly activate the b-C�H position,but no pyrrole product arising from such a process was ob-tained.

Since a study by Ackermann et al. in 2011,[38] rutheniu-m(II) catalysts have proven to serve as inexpensive alterna-tives to rhodium ACHTUNGTRENNUNG(III) catalysts in a series of oxidative C�Hfunctionalization reactions. They developed an annulationreaction of N-pyrimidylindoles with alkynes to form indoleproducts with a [RuCl2ACHTUNGTRENNUNG(p-cymene)]2 catalyst and CuACHTUNGTRENNUNG(OAc)2

as the oxidant (Scheme 20 a).[39] The pyrimidyl group on theindole product was readily removed by using NaOEt. Theyfurther demonstrated the feasibility of the ruthenium-cata-lyzed annulation reaction for the synthesis of pyrroles fromenamines and alkynes (Scheme 20 b).[40]

Whereas the above annulative indole syntheses requireanilines with directing groups on the nitrogen atom, Jiaoand co-workers achieved oxidative annulation by using un-protected anilines. Thus, simple anilines underwent oxida-tive coupling with dimethyl acetylenedicarboxylate(DMAD) in the presence of catalytic Pd ACHTUNGTRENNUNG(OAc)2 and molec-ular oxygen (1 atm., Scheme 21 a).[41] Although the scope of

Scheme 16. Improved rhodium ACHTUNGTRENNUNG(III) catalytic system for a) concise syn-thesis of Paullone and b) annulation of enamides and alkynes to givepyrroles. Boc= tert-butyloxycarbonyl; TBS = tert-butyldimethylsilyl.

Scheme 17. Proposed catalytic cycle for rhodium ACHTUNGTRENNUNG(III)-catalyzed annula-tion of N-acetyl anilines and alkynes.

Scheme 18. Regiocontrolled synthesis of dialkylpyrroles from enynes.

Scheme 19. Rhodium ACHTUNGTRENNUNG(III)-catalyzed annulation of enamines and alkynesinvolving allylic C ACHTUNGTRENNUNG(sp3)�H activation.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim472

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

the alkyne was largely limited to DMAD and its analogues,the catalytic system allowed some intriguing transforma-tions, such as formation of carbazole from o-silylphenyl tri-flate (Scheme 21 b) and formation of a tricyclic indole de-rivative from tetrahydroquinoline (Scheme 21 c). The reac-tion was thought to involve initial formation of an enamineintermediate then sequential palladation/reductive elimina-tion processes as proposed for Glorius and co-workers� en-amine cyclization (Scheme 10). However, a small intramo-lecular KIE suggested that the aromatic C�H activation oc-curred through electrophilic palladation.

Wang and co-workers recently developed an oxidativeannulation reaction of simple anilines with diarylalkynes byusing PdCl2 and Cu ACHTUNGTRENNUNG(OAc)2 as a catalyst and an oxidant, re-spectively, to afford 2,3-diarylindoles (Scheme 22 a).[42] Withmodifications of the reaction conditions, the same startingmaterials afforded pentaarylpyrroles (Scheme 22 b). Theproposed mechanisms for these reactions involve aminopal-ladation of diarylalkyne as a common initial step.

2.4. Oxidative C�H Amination

In 2005, Buchwald and co-workers reported the first exam-ple of carbazole synthesis through oxidative aromatic C�Hbond amination.[43] In the presence of a Pd ACHTUNGTRENNUNG(OAc)2 catalyst,CuACHTUNGTRENNUNG(OAc)2, and molecular oxygen (1 atm.), 2-acetoaminobi-phenyl derivatives efficiently cyclized to afford N-acetylcarbazoles (Scheme 23). The loading of CuACHTUNGTRENNUNG(OAc)2 could be

reduced to a catalytic amount. This oxidative amination re-action in combination with the Suzuki–Miyaura couplingopened a convenient route to carbazoles starting from 2-haloacetanilides and arylboronic acids. Shortly after thisreport, Matsubara and co-workers reported N�H carbazolesynthesis from 2-aminobiphenyls with a Pt/C catalyst underhydrothermal conditions.[20]

The follow-up studies of Buchwald and co-workers im-proved the catalytic system, expanded the substrate scope,and demonstrated the utility of the reaction in the synthesisof carbazole alkaloids, such as Mukonidine.[44] Notably, thereaction did not require CuACHTUNGTRENNUNG(OAc)2 when DMSO was usedas the solvent instead of toluene. Although the reactionwas initially considered to involve intramolecular aromaticC�H palladation of an amidopalladium(II) species, sucha mechanism was not consistent with the substituent effects.Alternatively, a mechanism involving Heck-like or Wacker-like amidopalladation was suggested (Scheme 24).

In 2008, Gaunt and co-workers reported that the sametype of carbazole-forming reaction could be achieved with2-aminobiphenyls that contain N-alkyl, benzyl, and allylgroups in the presence of catalytic Pd ACHTUNGTRENNUNG(OAc)2 andPhI ACHTUNGTRENNUNG(OAc)2 (Scheme 25).[45] Remarkably, the reaction tookplace at room temperature and tolerated various functional

Scheme 20. Rhodium(II)-catalyzed annulation of a) N-pyrimidylindolesand alkynes for indole synthesis and b) enamines and alkynes for pyrrolesynthesis.

Scheme 21. Palladium(II)-catalyzed oxidative annulation of unprotectedanilines and reactive alkynes/benzynes. N,N,-dimethylacetamide (DMA).

Scheme 22. Palladium(II)-catalyzed, copper(II)-mediated oxidative an-nulation of anilines and diarylalkynes to give a) indoles or b) pyrroles.

Scheme 23. Palladium(II)-catalyzed, copper(II)-mediated cyclization of2-acetaminobiphenyls to give N-acetylcarbazoles.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim473

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

groups on the aryl groups of the starting material. Amino-biphenyls with a protected glycosyl group on the nitrogenatom also underwent the reaction.

The stoichiometric reaction of 2-(N-benzyl)aminobiphen-yl with Pd ACHTUNGTRENNUNG(OAc)2 affords a trinuclear palladium complexthat features two six-membered palladacycle moieties andfour bridging acetate ligands (Scheme 26). The reaction isbelieved to go through oxidation of such a pallada(II)cyclewith the hypervalent iodine oxidant to give a palladium(IV)intermediate, and a C�N reductive elimination follows.

Note that oxidative palladium catalysis also allowed cycliza-tion of 2-hydroxybiphenyls to dibenzofurans with the ap-propriate choice of ligand, oxidant, and other condi-tions.[46, 47]

2-(N-sulfonylamino)biaryls and 1,1-diaryl-2-(N-sulfonyla-mino)ethenes were used as precursors for N-sulfonylcarba-zoles and N-sulfonylindoles through palladium-catalyzedoxidative C�H amination by Youn et al. and Inamoto/Doiet al. , respectively (Scheme 27).[48,49] The palladium(II)/

oxone system of Youn et al. was proposed to involve a palla-dium(II)/palladium(IV) catalytic cycle, whereas the palla-dium(II)/copper(II) system of Inamoto/Doi et al. appearsmechanistically more related to the Buchwald system (seeabove).

More recently, Chang and co-workers devised copper-based and metal-free systems for carbazole synthesisthrough intramolecular oxidative C�H amination of 2-sul-fonylaminobiphenyl and related substrates under mild con-ditions (Scheme 28).[50] Whereas the copper-based system

consisted of CuACHTUNGTRENNUNG(OTf)2 and PhI ACHTUNGTRENNUNG(OAc)2 as the catalyst andoxidant, respectively, the use of a stronger oxidant, phenyliodonium bis(trifluoroacetate), PhIACHTUNGTRENNUNG(OTFA)2, allowedmetal-free cyclization. Mechanistic experiments suggestedthat the reaction involves a radical intermediate, and thatthe copper(II) catalyst activates the hypervalent iodine re-agent.

Whereas the above examples used a free N�H bond andan external oxidant for C�H amination, Tan and Hartwigintroduced a conceptually different strategy for indole syn-

Scheme 24. Suggested mechanism for palladium(II)-catalyzed aminativecyclization.

Scheme 25. Palladium(II)-catalyzed cyclization of 2-alkylaminobiphenylsto give N-alkylcarbazoles by using an iodine ACHTUNGTRENNUNG(III) oxidant.

Scheme 26. Stoichiometric cyclopalladation of 2-benzylaminobiphenyl togive a trinuclear palladium(II) complex

Scheme 27. Palladium(II)-catalyzed oxidative C�H amination with anaminotosyl group to give carbazoles and indoles. Ts=p-toluenesulfonyl.

Scheme 28. Cyclization of 2-sulfonylaminobiphenyls to give N-sulfonyl-carbazoles by using copper-catalyzed or metal-free systems.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim474

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

thesis, where an oxime N�O bond serves as an “internal ox-idant” for C�H amination (Scheme 29).[51,52] Thus, O-acetyloximes derived from 1,1-diarylacetone derivatives anda few other ketones cyclized into indole products in thepresence of PdACHTUNGTRENNUNG(dba)2 catalyst and Cs2CO3.

The reaction was proposed to proceed through a catalyticcycle consisting of four elementary steps, that is, oxidativeaddition of the oxime N�O bond to palladium(0) to forman iminylpalladium(II) species,[53] tautomerization to a palla-dium(II) enamide, intramolecular C�H palladation, andC�N reductive elimination (Scheme 30). The viability of

the first step was confirmed by a stoichiometric experimentwith O-pentafluorobenzoyloxime as a substrate, which al-lowed the first isolation of a product of N�O oxidative ad-dition with palladium(0) (see inset in Scheme 30). This iso-lated complex afforded the indole product upon heatingwith Cs2CO3.

2.5. C�H Amination through (Formal) NitreneInsertion

Intramolecular C�H amination through thermal decompo-sition of azide precursors is one of classical methods for thesynthesis of indoles and carbazoles. This method, however,suffers from harsh conditions and limited scope. This(formal) nitrene insertion chemistry has received renewedattention in recent years.[54] In 2006, Taber and Tian demon-strated the utility of thermal rearrangement of 2H-azirines,which can be readily prepared by a Neber reaction ofoximes derived from a-arylketones (Scheme 31 a).[55] Short-

ly after this report, Narasaka and co-workers reported thatthe same cyclization reaction was efficiently catalyzed bya rhodium(II) catalyst, presumably through a rhodium-ni-trenoid species (Scheme 31 b).[56]

In 2007 and 2008, Driver and co-workers reported rho-dium(II)-catalyzed denitrogenative C�H amination reac-tions for the synthesis of indoles from a-azidocinnamatesand 2-azidostyrenes, respectively (Scheme 32 a and b).[57–59]

These reactions serve as complementary methods for thepreparation of 2-substituted indoles, wherein the substitu-ents at the C2 position are the ester and aryl/alkyl substitu-ents from the former and the latter reactions, respectively.The same amination strategy was also applicable to carba-zole synthesis from 2-azidobiphenyl derivatives.[60] Further-more, ZnI2 alone was shown to efficiently catalyze denitro-genative cyclization of dienyl azides to pyrroles(Scheme 32 c).[61]

The rhodium(II)-catalyzed reactions most likely involverhodium-nitrenoid species as reactive intermediates, whichshould form through denitrogenation of the starting materi-als. Whereas the C�H amination could a priori take placein a concerted or a stepwise manner, pieces of experimentalevidence point to the latter pathway, in which electrophilicattack of the nitrogen atom is followed by C�H bond cleav-age.[62] Notable among these pieces of evidence are the ab-sence of an H/D KIE in intramolecular competition be-tween C�H and C�D cleavage (Scheme 33 a) and the for-mation of 2,3-diphenylindole from 2-azidostyrene havingb,b-diphenyl groups, which can be rationalized by a 1,2-

Scheme 29. Palladium(0)-catalyzed cyclization of O-acetyloximes to giveindoles. dba=dibenyzlidineacetone.

Scheme 30. Proposed catalytic cycle for palladium(0)-catalyzed indolesynthesis by oxime N�O bond cleavage.

Scheme 31. Thermal and rhodium(II)-catalyzed rearrangement of 2H-azirines to give indoles.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim475

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

phenyl shift of the electrophilic amination intermediate(Scheme 33 b).

Driver and co-workers further extended the latter obser-vation to achieve the synthesis of 2,3-disubstituted indolesfrom a series of b,b-disubstituted 2-azidostyrenes(Scheme 34).[63] Most notably, with aryl and alkyl substitu-ents on the b position, the aryl group preferentially under-went 1,2-migration with high selectivity (>95:5). Mechanis-tic studies, including Hammett analysis of the intramolecu-lar competition of the migration of different aryl groups,suggested the formation of a phenonium ion intermediate.

Another notable finding in the rhodium-catalyzed cycli-zation of 2-azidostyrene derivatives is 1,2-migration of anelectron-withdrawing group on the b position(Scheme 35).[64] This phenomenon occurs with nitro, ketone,and sulfonyl groups with high selectivity, whereas no signifi-

cant migration occurs with less electron-withdrawinggroups, such as esters and amides.

Transition-metal-catalyzed reduction of a nitro groupwith carbon monoxide is an alternative method to generatea nitrene species.[65] This chemistry was successfully imple-mented for the synthesis of carbazoles and indoles throughreductive C�H amination. Thus, Smitrovich and Davies aswell as Hsieh and Dong achieved conversion of 2-nitrobiar-yls and 1,1-diaryl-2-nitroethenes into carbazoles and 3-arylindoles, respectively, by using a catalytic system consistingof PdACHTUNGTRENNUNG(OAc)2 and 1,10-phenanthroline (Scheme 36).[66, 67]

2.6. Miscellaneous

Takemoto and co-workers recently developed a palladium-catalyzed reaction between aryl isocyanides that haveortho-methyl groups and aryl halides, which affords 2-arylindole derivatives (Scheme 37).[68] The reaction is thought

Scheme 32. a) and b) Rhodium(II)-catalyzed indole synthesis from al-kenyl- and aryl azides and c) Zinc(II)-catalyzed pyrrole synthesis fromdienyl azides. MS=molecular sieves.

Scheme 33. a) Intramolecular competition of aromatic C�H/C�D cleav-age and b) formation of 2,3-diphenylindole through 1,2-phenyl shift.

Scheme 34. Rhodium(II)-catalyzed synthesis of 2,3-disubstituted indolesby migration of aryl or alkyl groups.

Scheme 35. Rhodium(II)-catalyzed indole synthesis from aryl azides in-volving migration of electron-withdrawing groups. esp =a,a,a,a-tetra-methyl-1,3- benzenedipropanoate.

Scheme 36. Synthesis of carbazoles and indoles through palladium-cata-lyzed reduction of nitro groups with carbon monoxide.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim476

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

to involve oxidative addition of the aryl halide to palladi-um(0), isocyanide insertion into the palladium–aryl bond,benzylic C�H activation with the resulting imidoylpalladi-um species, and reductive elimination. For the benzylicC�H activation to take place, the isocyanide substrate musthave two ortho-methyl groups.

Shibata and co-workers reported an iridium-catalyzed di-rected cyclodehydration reaction of a-arylaminoketones toafford 4-acetylindoles (Scheme 38).[69] The reaction likely

involves aromatic C�H metalation directed by the acetylgroup, intramolecular addition of the resulting aryliridiumspecies to the carbonyl moiety, and dehydration.

3. Conclusion

Synthetic chemists have become fully aware of the powerand potential of using C�H bonds as “functional groups”for organic synthesis.[70] The synthesis of nitrogen-contain-ing heterocycles is no exception. The last several years havewitnessed the development of a variety of new catalytic re-actions for the synthesis of pyrroles, indoles, and carbazolesthat feature transition-metal-mediated C�H activation asthe key step, as summarized in this Focus Review. Thesenew reactions are not only practically attractive as atom-and step-economical synthetic methods, but are also funda-mentally interesting and show the latest innovations in or-ganometallic chemistry as well as homogeneous catalysis.Whereas conventional synthetic methods will play majorroles for the time being, the importance of the C�H activa-tion approaches will continue to grow, with prospective ap-plications in academia as well as in industry.

Acknowledgements

This work was supported by the National Research Foundation, Singa-pore (NRF-RF-2009-05) and Nanyang Technological University.

[1] a) M. d’Ischia, A. Napolitano, A. Pezzella in Comprehensive Heter-ocyclic Chemistry III, Vol. 3 (Eds.: A. R. Katritzky, C. A. Ramsden,E. F. V. Scriven, R. J. K. Taylor), Elsevier, Oxford, 2008, pp. 353 –388; b) H. Fan, J. Peng, M. T. Hamann, J.-F. Hu, Chem. Rev. 2008,108, 264; c) A. J. Kochanowska-Karamyan, M. T. Hamann, Chem.Rev. 2010, 110, 4489; d) H.-J. Knçlker, K. R. Reddy, Chem. Rev.2002, 102, 4303; e) A. W. Schmidt, K. R. Reddy, H.-J. Knçlker,Chem. Rev. 2012, 112, 3193; f) R. J. Sundberg, Indoles, AcademicPress, San Diego, 1996.

[2] a) D. Black in Science of Synthesis, Vol. 9 (Ed.: G. Maas), Thieme,Stuttgart, 2000, pp. 441 – 552; b) J. A. Joule in Science of Synthesis,Vol. 10 (Ed.: E. J. Thomas), Thieme, Stuttgart, 2000, pp. 361 – 692;c) T. Gallagher in Science of Synthesis, Vol. 10 (Ed.: E. J. Thomas),Thieme, Stuttgart, 2000, pp. 693 –744.

[3] a) G. W. Gribble, J. Chem. Soc. Perkin Trans. 1 2000, 1045; b) G. R.Humphrey, J. T. Kuethe, Chem. Rev. 2006, 106, 2875; c) G. Zeni,R. C. Larock, Chem. Rev. 2006, 106, 4644; d) S. Cacchi, G. Fabrizi,Chem. Rev. 2011, 111, PR215; e) M. Platon, R. Amardeil, L. Djako-vitch, J. C. Hierso, Chem. Soc. Rev. 2012, 41, 3929; f) M. Inman,C. J. Moody, Chem. Sci. 2013, 4, 29.

[4] a) T. F. Walsh, R. B. Toupence, F. Ujjainwalla, J. R. Young, M. T.Goulet, Tetrahedron 2001, 57, 5233; b) J. P. Simeone, R. L. Bugiane-si, M. M. Ponpipom, M. T. Goulet, M. S. Levorse, R. C. Desai, Tet-rahedron Lett. 2001, 42, 6459; c) R. N. Farr, R. J. Alabaster, J. Y. L.Chung, B. Craig, J. S. Edwards, A. W. Gibson, G. J. Ho, G. R. Hum-phrey, S. A. Johnson, E. J. J. Grabowski, Tetrahedron: Asymmetry2003, 14, 3503.

[5] Some of the reactions were previously covered in review articles,see: a) P. Thansandote, M. Lautens, Chem. Eur. J. 2009, 15, 5874;b) J. J. Song, J. T. Reeves, D. R. Fandrick, Z. Tan, N. K. Yee, C. H.Senanayake, ARKIVOC 2010, 390.

[6] D. E. Ames, A. Opalko, Tetrahedron 1984, 40, 1919.[7] Selected reviews: a) L. C. Campeau, D. R. Stuart, K. Fagnou, Aldri-

chimica Acta 2007, 40, 35; b) D. Alberico, M. E. Scott, M. Lautens,Chem. Rev. 2007, 107, 174; c) L. Ackermann, R. Vicente, A. R.Kapdi, Angew. Chem. 2009, 121, 9976; Angew. Chem. Int. Ed. 2009,48, 9792.

[8] T. Iwaki, A. Yasuhara, T. Sakamoto, J. Chem. Soc. Perkin Trans.1 1999, 1505.

[9] R. B. Bedford, C. S. J. Cazin, Chem. Commun. 2002, 2310.[10] R. B. Bedford, M. Betham, J. Org. Chem. 2006, 71, 9403.[11] L. C. Campeau, M. Parisien, M. Leblanc, K. Fagnou, J. Am. Chem.

Soc. 2004, 126, 9186.[12] L. C. Campeau, M. Parisien, A. Jean, K. Fagnou, J. Am. Chem. Soc.

2006, 128, 581.[13] a) L. Ackermann, A. Althammer, Angew. Chem. 2007, 119, 1652;

Angew. Chem. Int. Ed. 2007, 46, 1627; b) L. Ackermann, A. Al-thammer, P. Mayer, Synthesis 2009, 3493.

[14] a) M. Yamagishi, K. Nishigai, A. Ishii, T. Hata, H. Urabe, Angew.Chem. 2012, 124, 6577; Angew. Chem. Int. Ed. 2012, 51, 6471.

[15] a) C. S. Yeung, V. M. Dong, Chem. Rev. 2011, 111, 1215; b) C. Liu,H. Zhang, W. Shi, A. W. Lei, Chem. Rev. 2011, 111, 1780; c) A. N.Campbell, S. S. Stahl, Acc. Chem. Res. 2012, 45, 851.

[16] B. �kermark, L. Eberson, E. Jonsson, E. Pettersson, J. Org. Chem.1975, 40, 1365.

[17] H. J. Knçlker, Chem. Lett. 2009, 38, 8, and references cited therein.[18] T. Watanabe, S. Ueda, S. Inuki, S. Oishi, N. Fujii, H. Ohno, Chem.

Commun. 2007, 4516.[19] B. Liegault, D. Lee, M. P. Huestis, D. R. Stuart, K. Fagnou, J. Org.

Chem. 2008, 73, 5022.[20] M. Yamamoto, S. Matsubara, Chem. Lett. 2007, 36, 172.[21] S. W�rtz, S. Rakshit, J. J. Neumann, T. Drçge, F. Glorius, Angew.

Chem. 2008, 120, 7340; Angew. Chem. Int. Ed. 2008, 47, 7230.

Scheme 37. Palladium(0)-catalyzed indole synthesis from isocyanidesand aryl halides involving benzylic C�H bond activation. Ad= adaman-tyl.

Scheme 38. Iridium-catalyzed cyclodehydration directed by an acetylgroup to give 4-acetylindoles.

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim477

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.

[22] J. J. Neumann, S. Rakshit, T. Drçge, S. W�rtz, F. Glorius, Chem.Eur. J. 2011, 17, 7298.

[23] a) D. Lapointe, K. Fagnou, Chem. Lett. 2010, 39, 1118; b) D. Bal-cells, E. Clot, O. Eisenstein, Chem. Rev. 2010, 110, 749; c) L. Acker-mann, Chem. Rev. 2011, 111, 1315.

[24] R. Bernini, G. Fabrizi, A. Sferrazza, S. Cacchi, Angew. Chem. 2009,121, 8222; Angew. Chem. Int. Ed. 2009, 48, 8078.

[25] W. Yu, Y. Du, K. Zhao, Org. Lett. 2009, 11, 2417.[26] See also: a) Z. He, W. Liu, Z. Li, Chem. Asian J. 2011, 6, 1340;

b) Z. He, H. Li, Z. Li, J. Org. Chem. 2010, 75, 4636.[27] Z.-H. Guan, Z.-Y. Yan, Z.-H. Ren, X.-Y. Liu, Y.-M. Liang, Chem.

Commun. 2010, 46, 2823.[28] Y. Wei, I. Deb, N. Yoshikai, J. Am. Chem. Soc. 2012, 134, 9098.[29] Z. Shi, F. Glorius, Angew. Chem. 2012, 124, 9354; Angew. Chem.

Int. Ed. 2012, 51, 9220.[30] a) J. Zhu, J. Liu, R. Ma, H. Xie, J. Li, H. Jiang, W. Wang, Adv.

Synth. Catal. 2009, 351, 1229; b) J. Liu, J. Zhu, H. Jiang, W. Wang,J. Li, Chem. Asian J. 2009, 4, 1712.

[31] A. Hajra, Y. Wei, N. Yoshikai, Org. Lett. 2012, 14, 5488.[32] S. A. Girard, X. Hu, T. Knauber, F. Zhou, M.-O. Simon, G.-J. Deng,

C.-J. Li, Org. Lett. 2012, 14, 5606.[33] D. R. Stuart, M. Bertrand-Laperle, K. M. N. Burgess, K. Fagnou, J.

Am. Chem. Soc. 2008, 130, 16474.[34] For a review on Rh3+-catalyzed oxidative C�C bond forming reac-

tions, see: T. Satoh, M. Miura, Chem. Eur. J. 2010, 16, 11212.[35] D. R. Stuart, P. Alsabeh, M. Kuhn, K. Fagnou, J. Am. Chem. Soc.

2010, 132, 18326.[36] M. P. Huestis, L. N. Chan, D. R. Stuart, K. Fagnou, Angew. Chem.

2011, 123, 1374; Angew. Chem. Int. Ed. 2011, 50, 1338.[37] S. Rakshit, F. W. Patureau, F. Glorius, J. Am. Chem. Soc. 2010, 132,

9585.[38] L. Ackermann, A. V. Lygin, N. Hofmann, Angew. Chem. 2011, 123,

6503; Angew. Chem. Int. Ed. 2011, 50, 6379.[39] L. Ackermann, A. V. Lygin, Org. Lett. 2012, 14, 764.[40] L. Wang, L. Ackermann, Org. Lett. 2013, 15, 176.[41] Z. Shi, C. Zhang, S. Li, D. Pan, S. Ding, Y. Cui, N. Jiao, Angew.

Chem. 2009, 121, 4642; Angew. Chem. Int. Ed. 2009, 48, 4572.[42] X. Chen, X. Li, N. Wang, J. Jin, P. Lu, Y. Wang, Eur. J. Org. Chem.

2012, 4380.[43] W. C. P. Tsang, N. Zheng, S. L. Buchwald, J. Am. Chem. Soc. 2005,

127, 14560.[44] W. C. P. Tsang, R. H. Munday, G. Brasche, N. Zheng, S. L. Buch-

wald, J. Org. Chem. 2008, 73, 7603.

[45] J. A. Jordan-Hore, C. C. C. Johansson, M. Gulias, E. M. Beck, M. J.Gaunt, J. Am. Chem. Soc. 2008, 130, 16184.

[46] B. Xiao, T.-J. Gong, Z.-J. Liu, J.-H. Liu, D.-F. Luo, J. Xu, L. Liu, J.Am. Chem. Soc. 2011, 133, 9250.

[47] Y. Wei, N. Yoshikai, Org. Lett. 2011, 13, 5504.[48] S. W. Youn, J. H. Bihn, B. S. Kim, Org. Lett. 2011, 13, 3738.[49] K. Inamoto, T. Saito, K. Hiroya, T. Doi, Synlett 2008, 3157.[50] S. H. Cho, J. Yoon, S. Chang, J. Am. Chem. Soc. 2011, 133, 5996.[51] Y. Tan, J. F. Hartwig, J. Am. Chem. Soc. 2010, 132, 3676.[52] F. W. Patureau, F. Glorius, Angew. Chem. 2011, 123, 2021; Angew.

Chem. Int. Ed. 2011, 50, 1977.[53] a) M. Kitamura, K. Narasaka, Chem. Rec. 2002, 2, 268; b) K. Nara-

saka, M. Kitamura, Eur. J. Org. Chem. 2005, 4505.[54] T. G. Driver, Org. Biomol. Chem. 2010, 8, 3831.[55] D. F. Taber, W. Tian, J. Am. Chem. Soc. 2006, 128, 1058.[56] S. Chiba, G. Hattori, K. Narasaka, Chem. Lett. 2007, 36, 52.[57] B. J. Stokes, H. Dong, B. E. Leslie, A. L. Pumphrey, T. G. Driver, J.

Am. Chem. Soc. 2007, 129, 7500.[58] M. Shen, B. E. Leslie, T. G. Driver, Angew. Chem. 2008, 120, 5134;

Angew. Chem. Int. Ed. 2008, 47, 5056.[59] For the same type of reaction of a-azidocinnamates with an iron-

porphyrin catalyst, see: Y. Liu, J. Wei, C.-M. Che, Chem. Commun.2010, 46, 6926.

[60] B. J. Stokes, B. Jovanovic, H. Dong, K. J. Richert, R. D. Riell, T. G.Driver, J. Org. Chem. 2009, 74, 3225.

[61] H. Dong, M. Shen, J. E. Redford, B. J. Stokes, A. L. Pumphrey,T. G. Driver, Org. Lett. 2007, 9, 5191.

[62] B. J. Stokes, K. J. Richert, T. G. Driver, J. Org. Chem. 2009, 74,6442.

[63] K. Sun, S. Liu, P. M. Bec, T. G. Driver, Angew. Chem. 2011, 123,1740; Angew. Chem. Int. Ed. 2011, 50, 1702.

[64] B. J. Stokes, S. Liu, T. G. Driver, J. Am. Chem. Soc. 2011, 133, 4702.[65] A. M. Tafesh, J. Weiguny, Chem. Rev. 1996, 96, 2035.[66] J. H. Smitrovich, I. W. Davies, Org. Lett. 2004, 6, 533.[67] T. H. H. Hsieh, V. M. Dong, Tetrahedron 2009, 65, 3062.[68] T. Nanjo, C. Tsukano, Y. Takemoto, Org. Lett. 2012, 14, 4270.[69] K. Tsuchikama, Y. Hashimoto, K. Endo, T. Shibata, Adv. Synth.

Catal. 2009, 351, 2850.[70] J. Yamaguchi, A. D. Yamaguchi, K. Itami, Angew. Chem. 2012, 124,

9092; Angew. Chem. Int. Ed. 2012, 51, 8960.

Received: January 21, 2013Published online: March 13, 2013

Asian J. Org. Chem. 2013, 2, 466 – 478 � 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim478

www.AsianJOC.org Naohiko Yoshikai and Ye Wei.