Download PDF
Synthesis 2024; 56(20): 3119-3130
DOI: 10.1055/a-2335-8799
short review

Modern Dearomative Enlargement of Heteroaromatic Rings

Xavier Moreau
,
Clément Ghiazza  ∗

Financial support for this work was provided by the Centre National de la Recherche Scientifique (CNRS), Université de Versailles Saint-Quentin-en-Yvelines (UVSQ), and Université Paris Saclay (UPSaclay). This work was also supported by the Agence Nationale de la Recherche (ANR), Programmes d’investissements d’avenir (PIA) funding (ANR-20-IDEES-0002) and as part of the France 2030 program (ANR-11-IDEX-0003), awarded by the Graduate School Chemistry of the Université Paris-Saclay.
› Further Information
 
 PDF Download Permissions and Reprints All articles of this category


Abstract

Breaking aromaticity by inserting additional atoms within the skeleton of heteroaromatic rings has gained significant attention over the years. As part of the emerging concept of ‘skeletal editing’, this short review retraces the recent progress made on dearomative enlargement reactions of both five- and six-membered heterocycles.

1 Introduction

2 Dearomative Enlargement of Five-Membered Rings

2.1 Pyrroles, Furans, Thiophenes and Their Fused Analogues

2.2 Pyrazoles, Isoxazoles, Isothiazoles and Their Fused Analogues

3 Dearomative Enlargement of Six-Membered Rings

4 Conclusion and Perspectives


#

Key words

dearomatization - ring expansion - atom insertion - heterocycles - skeletal editing - carbenes
1

Introduction

The modulation of the physico-chemical properties of active substances occupies a pivotal place in drug discovery programs. A widespread approach is to operate slight modifications of the structure without significantly altering the overall geometry in order to preserve the size complementary between the receptor (host) and the drug (guest). Among the leverages in the medicinal chemistry toolbox, a compelling method consists of changing the nature of heterocyclic scaffolds at the core of molecular skeletons.

Zoom Image
Clément Ghiazza obtained his Ph.D. in 2019 under the supervision of Dr. Anis Tlili and Dr. Thierry Billard during which he was involved in the development of new fluoroalkylselenolation methods at the University of Lyon, France. He subsequently earned a postdoctoral fellowship from the Alexander von Humboldt foundation to work in the group of Dr. Josep Cornella at the Max Planck Institut für Kohlenforschung in Mülheim an der Ruhr, Germany. During this period, his research topics were mainly focused on the late-stage modification of highly functionalized heterocycles, the synthesis of hybrid fluorinated isosteres and the development of catalytically active bismuth complexes. In October 2022, Clément was recruited as a CNRS researcher at the Institut Lavoisier de Versailles, France. In a group led by Prof. Xavier Moreau, he is interested in innovative heterocyclic synthesis and asymmetric (organo)catalysis.

Moreover, a direct parallel can be drawn between the potential activity of a molecule and the complexity of its constituent elements. This postulate was formulated in the venerable article ‘Escape from Flatland’ nearly 15 years ago.[1] The complexity of a molecule is defined by its saturation degree, three-dimensionality and by the presence of chiral centers. Therefore, dearomatization by saturation of aromatic and flat substrates has received substantial interest in order to increase the success rate in clinical trials (Scheme [1], top).[2] [3] The overall size and molecular weight of the heterocycle remain practically the same; although the molecule now occupies a three-dimensional space.[4] Several properties such as improved solubility and lower melting points are often observed while reducing the aromatic character of a compound.[5,6] The orientation of the substituents becomes a more precise fingerprint in order to accommodate a receptor with an accurate complementarity.

Zoom Image
Scheme 1 Context of modern dearomative enlargement reactions

In recent years, a cutting-edge concept emerged, aiming at interchanging aromatic rings in a single sequence, termed ‘aromatic skeletal editing’ (Scheme [1], middle).[7] [8] [9] This methodological tool enables the practitioner to produce new analogues from available bioactive substances while saving a precious amounts of time and resources. In this context, several strategies aiming at insertion, deleting or exchanging atoms within an aromatic framework have been reported. Moreover, these examples exhibit wide functional group tolerance and therefore meet the standards of late-stage applications. Furthermore, it is noteworthy that the aromaticity is preserved during the process.

Keeping this statement in mind, several research groups have endeavored to merge these two conceptual approaches: developing selective insertion reactions into heteroaromatic skeletons to provide dearomatized heterocyclic structures (Scheme [1], bottom). The recent efforts discussed in this short review are divided into two sections: the dearomative enlargement of 5-membered rings and 6-membered rings.


# 2

Dearomative Enlargement of Five-Membered Rings

2.1

Pyrroles, Furans, Thiophenes and Their Fused Analogues

Ubiquitous in bioactive molecules, 5-membered heteroaromatic scaffolds, including pyrroles, furans, thiophenes and their fused analogues, are easily accessible and abundant feedstocks. Naturally, efforts were devoted to inserting carbon atoms into their skeletons. The main challenge consists in breaking highly stable X–C, or even C–C, bonds selectively to ensure the internalization of the additional atom or group of atoms.

In this context, the group of Liu and Bi recently reported the silver- and rhodium-catalyzed conversion of indoles into dihydroquinolines via a carbene insertion between the 2- and 3-positions (Scheme [2]).[10] A broad scope is presented, encompassing wide functional group tolerance with good yields, even on millimolar scale. For instance, olefins were preserved under these reaction conditions, as observed for selected substrate 6. Supported by mechanistic experiments and DFT, a plausible reaction pathway involves a metallic carbenoid; the latter results from the reaction of the metallic complex and the N-triftosyl hydrazine derivative. Protection of the free NH of the starting indole with a TBS (tert-butyldimethylsilyl) group disrupts the dearomative enlargement. In fact, the indole derivative undergoes a cyclopropanation reaction with the carbenoid forming a strained tricyclic intermediate. Treating the latter with cesium fluoride restores the negative charge on the nitrogen atom which subsequently triggers the ring enlargement at room temperature.

Zoom Image
Scheme 2 Indoles to dihydroquinolines by Liu and Bi. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate. PMP = p-methoxyphenyl.
Zoom Image
Scheme 3 Pyrroles to dihydropyridines by Liu and Bi. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate.

Shortly after their previous report, the dearomative expansion of pyrroles was also explored by the same group (Scheme [3]).[11] Capitalizing on a similar approach, the authors reported a rhodium-catalyzed protocol to internalize an additional carbon atom into the ring. In this context, disubstituted carbene precursors, namely vinyltrifluoromethyl N-triftosyl hydrazones, were selected as reactants to achieve the transformation. The reaction presents a large scope as many functional groups were interrogated. In addition, biorelevant substrates underwent the ring enlargement to produce the 3-trifluoromethyl-dihydropyridine analogues in good yields.

In 2018, a different strategy was developed by Bakthadoss and co-workers (Scheme [4]).[12] Contrasting with single-atom insertion methods, they demonstrated that symmetrical acetylenedicarboxylates reacted under solvent- and catalyst-free conditions with indoles. The equimolar mixture of the neat reactants is heated to 100 °C under air to yield the corresponding benzoazepines in very good yields. The proposed mechanism suggests a stepwise formal [2+2] addition consisting of Michael addition of the indole on the electron-deficient alkyne, followed by an intramolecular cyclization. Electrocyclic ring-opening of the strained intermediate then delivers the 7-membered analogue.

Zoom Image
Scheme 4 Indoles to benzoazepines by Bakthadoss
Zoom Image
Scheme 5 Benzofurans to benzoxaborins by Yorimitsu. IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene. OA = oxidative addition. TM = transmetalation. RE = reductive elimination. TBS = tert-butyldimethylsilyl.

As part of their ‘aromatic metamorphosis’ concept, Yorimitsu and co-workers reported the first catalytic method to insert boron atoms into aromatic skeletons (Scheme [5]).[13] This dearomative process is initiated by an oxidative addition of a nickel(0) complex into the C–O bond of a series of benzofurans. A base-mediated transmetalation with B2pin2 (bis(pinacolato)diboron) followed by a tandem reductive elimination/cyclization forges the C–B and O–B bonds, respectively. Acidic treatment of the boronates generates the benzoxaborins in good yields. These unusual compounds are very versatile and are easily converted into valuable building blocks.

A year later, the same group developed a manganese-catalyzed process to internalize electrophilic sources of heteroatoms into benzofurans (Scheme [6]).[14] Using an excess of an organolithium reactant, the benzofurans undergo a ring-opening reaction, affording the bislithiated species. This reactive intermediate is intercepted by various dielectrophiles, which gave the expected oxaheterocycles after a simple acidic quench. Using their protocol, the authors could prepare benzoxasilins, oxaborins, oxaphosphins, oxagermins and oxatitins with good yields on average. Furthermore, the UV/Vis absorption and fluorescence spectra were recorded for relevant oxaheterocycles, which testified to a strong influence of the inserted heteroatom on the photophysical properties.

Zoom Image
Scheme 6 Benzofurans to 6-membered oxaheterocycles by Yorimitsu. TMEDA = tetramethylethylenediamine.
Zoom Image
Scheme 7 (Benzo)thiophenes to bridged 8-membered heterocycles by Houk and Glorius

In 2023, a team lead by Houk and Glorius reported a photocatalyzed protocol for the dearomative expansion of (benzo)thiophenes to give bridged 8-membered rings (Scheme [7]).[15] In this context, bicyclobutanes were incorporated into the skeletons of the aromatic derivatives. An acridinium salt, [Acr-Mes2]+[BF4]–, was selected as a photocatalyst at 450 nm. The mild conditions enabled the authors to report more than 50 substrates bearing relevant functional groups with very good yields on average. Noteworthy, the reaction is fairly regioselective and preferentially furnishes 2-substituted monocyclic or 4-substituted benzo-fused products starting from thiophenes or benzothiophenes, respectively.

In order to unveil the operative pathway, the authors performed mechanistic experiments as well as DFT (density functional theory) calculations (Scheme [8]). Stern–Volmer (SV) luminescence quenching experiments unambiguously confirmed a single-electron transfer (SET) through a reductive quenching of the excited photocatalyst and concomitant oxidation of the (benzo)thiophene. To further support this hypothesis, cyclic voltammetry (CV) was conducted for both substrates and no oxidation peak was observed for the model bicyclobutane; thus, excluding a SET between the strained reactant and the photocatalyst. The quantum yield (QY) of the reaction was determined to be slightly inferior to unity. This value indicates that a stepwise mechanism is more likely in this reaction. As anticipated by the authors, stoichiometric radical scavengers completely inhibited the reaction. Supported by their calculations, the mechanism involves the formation of a (benzo)thiophene radical cation resulting from the first SET with the photocatalyst. Addition of the radical and strain-release of the bicyclobutane takes place with an inverted regioselectivity for thiophenes and benzothiophenes. In fact, the highest spin density of the thiophene radical cation is located on C2, whereas it is equally distributed on the heterocycle for the (benzo)thiophene radical cation. Subsequent cyclization followed by reduction of the intermediates through SET from the reduced photocatalyst triggers irreversible C–S bond cleavage. Overall, the rearrangement occurs with a different mechanism for both substrates, as seen by comparing the free-energy barriers. In the case of thiophenes, the rate-limiting step is the addition on the bicyclobutane. On the other hand, the cyclization step becomes the rate-limiting step for benzothiophenes.

Zoom Image
Scheme 8 (Benzo)thiophenes to bridged 8-membered heterocycles: postulated mechanism. RDS = rate-determining step.

# 2.2

Pyrazoles, Isoxazoles, Isothiazoles and Their Fused Analogues

In this section, the presented methodologies exclusively rely on the relative weakness of the N–X bond, where X is either a nitrogen, oxygen or sulfur atom, to incorporate an additional carbon atom. In fact, the key to success consists in selectively and irreversibly breaking the bond between the two heteroatoms. Two main approaches can be distinguished: a base-mediated rearrangement of azolium ylides and carbenoid internalization reactions. The mechanistic features will be discussed for each strategy.

The first example of a base-mediated dearomative enlargement on 2-vinylpyrazolium salts was reported in 1985 by Pardo (Scheme [9]).[16] Upon treatment with aqueous potassium carbonate at room temperature, the authors were able to isolate the desired product as an inseparable mixture of diastereoisomers. In this scenario, the hydroxide ion adds onto the exocyclic double bond to generate the zwitterionic intermediate. A concerted ring enlargement was proposed even though the report lacks evidence to support this hypothesis. Many crucial details are missing; however, this pioneering observation paved the way for recent developments.

Zoom Image
Scheme 9 A vinylpyrazolium salt to a dihydropyrimidine by Pardo

Inspired by the work of Pardo,[16] two different groups reported analogous transformations in 2017 and 2024. First, Chen and co-workers developed the thermal rearrangement of N-benzyl or N-acetylpyrazolium salts in the presence of potassium carbonate (Scheme [10]A).[17] To evaluate the regioselectivity of the rearrangement for unsymmetrical substrates, they performed deuterium-labelling experiments. When only one pending benzylic system is decorated with either strong electron-donating (4-MeO) or electron-withdrawing (4-O2N) groups, only one product is observed. As expected, the most acidic proton is readily abstracted by the base. On the other hand, in the event one of the benzyl arms is substituted with a weak electron-donating group such as 4-methyl, a 4:1 mixture of regioisomers is obtained. Regarding the mechanism, after deprotonation of the most acidic site, the resulting pyrazolium ylide ring opens to afford the π-conjugated acyclic intermediate. According to the authors, the cyclization occurs either via a nucleophilic addition (6-endo-trig), or via a 6π electrocyclization.

More recently, the group of Derksen evaluated the ability of pyrazolium salts to rearrange in the presence of isocyanates (Scheme [10]B).[18] Optimization of the reaction conditions enabled the authors to efficiently convert pyrazolium salts into dihydropyrimidines under mild conditions. Noteworthy, using an isothiocyanate does not permit the ring enlargement. The authors were able to isolate a stable ylide that remains inert, even upon heating or when subjected to basic conditions. At least two equivalents of potassium carbonate are mandatory to achieve this transformation. The first deprotonation causes the addition of the pyrazolium ylide on the isocyanate to reveal the corresponding amide moiety. A second deprotonation forms a second ylide, the stability of which is increased by the amide group. Following an analogous mechanism similar to that proposed by Chen, the latter undergo ring opening followed by cyclization to yield the 6-membered heterocycle.

Zoom Image
Scheme 10 Pyrazolium salts to dihydropyrimidines by Chen (A) and Derksen (B)

As noted in Section 2.1 with pyrroles and indoles, the reactivity of metallic carbenoids is particularly coveted in dearomative enlargement reactions. These highly reactive species can be easily made in situ from various carbene precursors. In this regard, diazo compounds are readily activated by rhodium- or silver-based metallic complexes (Scheme [11]). These electrophilic organometallic species can be intercepted by nucleophilic addition of the most basic nitrogen atom of a 5-membered heteroaromatic ring to generate a metallic zwitterion. Ring opening of the latter forms the acyclic intermediate, the behavior of which resembles the one disclosed for the base-mediated ring enlargement of pyrazolium salts.

Zoom Image
Scheme 11 Dearomative enlargement of pyrazoles, isoxazoles and isothiazoles: unified mechanism

A first example of this approach on isothiazolones was described in 1983 by Gosney and collaborators (Scheme [12]).[19] In their report, diazomalonate starting materials are readily activated with a catalytic amount of dirhodium tetraacetate to form the corresponding carbenoid. Only 7 examples were detailed in this seminal work. Nonetheless, it paved the way for future developments.

Zoom Image
Scheme 12 Isothiazolones to thiazinones by Gosney

Later, in 2008, the ring expansion of isoxa(thia)zole derivatives was reported by Davies and co-workers (Scheme [13]A).[20] Initially, the authors were attempting to complete the total synthesis of elisabethin C. Their retrosynthetic design involved a C–H activation step on a fused isoxazole ring using a rhodium-stabilized vinylcarbenoid reactant. Instead of the expected product, the authors serendipitously recovered a rearranged product comprising a dearomatized oxazine scaffold. Interested by their discovery, they optimized their conditions to selectively promote the ring-enlargement reaction. Thus, a combination of diazoacetates and commercially available rhodium acetate dimer in catalytic amounts enabled the authors to prepare a series of (benzo)oxa(thia)zines in very good yields. They further capitalized on this reaction to prepare highly functionalized pyridines from isoxazoles and vinyl diazoacetates (Scheme [13]B).[21] In this scenario, the rhodium-catalyzed dearomative enlargement takes place under the previously developed conditions. Heating the crude oxazines in boiling toluene triggers a subsequent retrocyclization/6π-cyclization to afford the corresponding dihydropyridines. Final oxidation of the latter furnished a large panel of substituted pyridines in good yields after 3 sequential steps.

Zoom Image
Scheme 13 (Benzo)isoxazoles and benzoisothiazoles to (benzo)oxazines and benzothiazines by Davies

Further developments of this strategy were conducted by Khlebnikov in 2014, and Jurberg and Davies in 2017 (Scheme [14]). In the former, the scope was extended to highly functionalized isoxazoles while preserving a similar efficiency (Scheme [14]A).[22] Moreover, Khlebnikov and co-workers demonstrated that carbenes could directly react with the substrates in a catalyst-free approach at 103 °C, albeit in poor yields. Furthermore, a thorough theoretical study further supports the original mechanistic hypothesis formulated by Davies back in 2008.

Zoom Image
Scheme 14 Isoxazoles and isoxazolones to oxazines and oxazinones by Khlebnikov (A) and Jurberg and Davies (B)

Jurberg and Davies adapted their protocol to the ring expansion of isoxazolones en route to the preparation of the interesting oxazinone core (Scheme [14]B).[23] Aryl diazoacetates were successfully engaged under the classical rhodium-catalyzed conditions to afford the corresponding products in good yields. In addition, the authors developed an alternative metal-free protocol using p-toluenesulfonic acid (TsOH·H2O) as a Brønsted acid. The latter first reacts with the diazo compound to form a transient tosylate intermediate. In the presence of triethylamine, representative isoxazolones were readily converted into their 6-membered counterparts, although in lower yields compared to the carbenoid approach.

Cyclic diazo compounds were also engaged under transition-metal-catalyzed protocols to generate spiro(benz)oxazines upon insertion of the carbenoid into the N–O bond of (benz)isoxazoles (Scheme [15]). In 2019, Chen and Pu successfully employ diazo indolinimines under rhodium-catalyzed conditions to prepare a panel of representative spirobenzoxazines (Scheme [15]A).[24] Screening of catalysts revealed that copper(I) salts were completely inactive for achieving the desired transformation. Almost 20 structures were isolated by the authors, with diverse substitution patterns on both reaction partners. Unfortunately, their protocol failed to achieve the ring expansion of monocyclic isoxazoles.

Zoom Image
Scheme 15 (Benz)isoxazoles to spiro(benz)oxazines by Chen and Pu (A) and Lu and Wang (B). esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate.

Similarly, Lu and Wang recently explored the reactivity of diazo dihydroisoquinolines for the dearomative enlargement of isoxazoles (Scheme [15]B).[25] Contrasting with the observations made by Chen and Pu,[24] tetrakis(acetonitrile) copper(I) hexafluorophosphate outperformed both rhodium(II) and palladium(II) catalysts. Moreover, monocyclic isoxazoles were amenable under the developed conditions, thus providing 27 densely functionalized spiro-structures in good yields.

Zoom Image
Scheme 16 (Benzo)isoxa(thia)zoles to (benzo)oxa(thia)zines by Bi. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate. DIPEA = diisopropylethylamine.

Finally, fluoroalkyl N-triftosyl hydrazones were proven efficient to permit the dearomative enlargement on 5-membered heteroaromatic rings bearing 2 heteroatoms, as reported by Bi in 2024 (Scheme [16]).[26] A large scope of more than 70 examples is detailed on both benzo-fused and monocyclic substrates. Two sets of conditions were developed depending on the nature of the chalcogen atom. It is noteworthy that a pending N-methylated indole was fully preserved in compound 68, probably owing to the steric hindrance generated by the trifluoromethyl moiety.

Another illustration of the carbenoid approach on pyrazoles can be found in a report by Novikov and co-workers from 2018 (Scheme [17]A).[27] Using a commercial source of rhodium with a low loading (1 mol%), the authors prepared more than 30 dihydropyrimidines in good to excellent yields. Many interesting moieties are tolerated under their conditions; for instance, an isoxazole ring is preserved in the structure of 77 and does not undergo the dearomative enlargement despite a significant excess of the diazo compound. Most of the substitution patterns posed no difficulties; nonetheless, blocking the C3-position of the pyrazole was deleterious as the starting material remained inactive.

Zoom Image
Scheme 17 Pyrazoles and indazoles to dihydropyrimidines and dihydroquinazolines by Novikov (A) and Bi (B and C). esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate. DIPEA = diisopropylethylamine.

As for pyrroles, indoles and (benzo)isoxa(thia)zoles (see Schemes 2, 3 and 16), the group of Bi published two different reports on the rhodium- and silver-catalyzed enlargements of pyrazoles and indazoles using their N-triftosyl hydrazones as carbene precursors. In their first article, they focused on the insertion of fluoroalkylated carbenoids (Scheme [17]B).[26] More than 80 different structures could be synthesized, emphasizing the impressive compatibility of their system towards many different functional groups. Among others, they demonstrated that an indazole reacted faster than a pyrazole heterocycle, as evidenced by product 83 that was obtained in 40% yield. Contrasting with the limitation encountered by the group of Novikov,[27] a 3-substituted pyrazole readily reacted to afford compound 84. Finally, biorelevant drug molecules were also successfully transformed, leaving the peripheral functional groups intact.

Their second report described a similar reaction on indazoles using donor-type carbene precursors (Scheme [17]C).[28] Using N-triftosyl hydrazones prepared from readily available diarylketones, aryl/alkylketones or even aldehydes, the authors could internalize diversely substituted methylenes into the N–N bond. Sensitive functional groups were again well-tolerated, for instance, a terminal alkyne was preserved in both 88 and 92. Late-stage modification of bioactive substrates was also undertaken to provide new analogues in nearly quantitative yields. It is worth mentioning that this operationally simple protocol could be conducted on gram scale without dramatic erosion of the yield.

Zoom Image
Scheme 18 Divergent preparation of benzoxazines or benzoxazepines from benzoisoxazoles and ynamides by R.-S. Liu. L = P(tBu)2(o-biphenyl) (JohnPhos). IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene.

Beyond base-mediated ring enlargement or carbenoid internalization, ynamides and propargyl esters in conjunction with a gold catalyst proved to be efficient in the ring expansion of benzisoxazoles. Depending on the nature of the ligand, two different outcomes can be obtained from ynamides, as demonstrated by R.-S. Liu and co-workers in 2018 (Scheme [18]).[29] In their report, the in situ generated cationic gold(I) complex promotes either the formation of the 6-membered Z-benzoxazines using a hindered phosphine ligand (JohnPhos), or alternatively, the 7-membered benzoxazepine with an NHC ligand (IPr). Both methods were applied to a wide variety of starting materials to yield the desired products in very good to excellent yields. Regarding the benzoxazepines, incorporation of the ynamide occurs in a regioselective manner. In addition, the 6-membered products were not observed under these conditions. On the contrary, the phosphine/gold(I) catalytic system often led to a mixture of Z/E isomers, as well as the 7-membered compound, although in marginal amounts (6–18%). Interestingly, a paper published several weeks later by Y. Liu reported the same transformations under similar conditions.[30] If the synthesis of benzoxazepines remains the same with AuBr3 instead of IPrAuCl/AgNTf2, the E/Z selectivity of the benzoxazine preparation contrasts with the previous study by R.-S. Liu.[29] Indeed, 3-substituted benzisoxazoles with either an aryl or an alkyl group afforded the E-benzoxazine isomer as the major product in most of the cases. However, when a methoxy substituent was installed on the same position, the Z-isomer was predominant.

Further experimental details provided by Gandon and Sahoo in 2019 shed some light on the mechanism. An in-depth theoretical study by DFT allowed the authors to rationalize the different selectivities observed depending on the substitution of the benzisoxazoles.[31]

Regarding the proposed mechanism, both paths start with nucleophilic addition of the benzisoxazole on the ynamide promoted by a gold(I) catalyst (Scheme [19]). Consecutive ring opening of the 5-membered ring affords the common Au-carbenoid. With JohnPhos as the ligand, a 6π electrocyclization proceeds preferentially (path a). Next, 1,2-migration of the amino group followed by a deauration step delivers the benzoxazine structure. In the case of an NHC ligand, direct O-addition on the carbenoid followed by a deauration regioselectively affords the benzoxazepine (path b). Hence, this divergence in the observed outcome only results from the nature of the ligand associated with the metallic center.

Zoom Image
Scheme 19 Ligand-controlled chemoselective dearomative enlargement of benzisoxazoles. L = P(tBu)2(o-biphenyl) (JohnPhos). IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene.
Zoom Image
Scheme 20 Benzisoxazoles to benzoxazines by Y. Liu. IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene.

The group of Y. Liu also reported that propargyl esters could serve as a C1 source in the gold-catalyzed dearomative enlargement of benzisoxazoles (Scheme [20]).[30] The mechanism is initiated by the formation of a gold-carbenoid via a 1,2-acyloxy migration, which undergoes nucleophilic attack of the heterocycle. Deaurative ring opening of the obtained adduct followed by a 6π electrocyclization furnished the expected products. This protocol was applied for the synthesis of 9 products in good average yields.


#
# 3

Dearomative Enlargement of Six-Membered Rings

Regarding the dearomative enlargement of 6-membered rings, an historical milestone was set by Buchner and Curtius in 1885 (Scheme [21]A).[32] In their original article, the authors reported the formation of ethyl cycloheptatrienate resulting from the thermal treatment of ethyl diazoacetate and benzene. Advances in organometallic catalysis as well as the diversity of carbene precursors contributed to improve the efficiency and the selectivity of the Buchner reaction, rendering this strategy compelling for several applications.[9]

Zoom Image
Scheme 21 A brief overview of dearomative enlargement reactions on benzene rings. MTAD = 4-methyl-1,2,4-triazoline-3,5-dione. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate.

Interestingly, the dearomative enlargement of benzene rings is not restricted to carbon insertions. In fact, research groups have successfully incorporated heteroatoms to furnish the corresponding 7-membered congeners. For instance, the group of Sarlah reported a two-step procedure enabling the monooxygenation of fused arenes for the synthesis of fused oxepines (Scheme [21]B).[33] A visible-light-promoted [4+2] cycloaddition between the polycyclic arenes and MTAD results in dearomatization of the substrates, enabling the formal manganese-based epoxidation of the adduct. Basic treatment of the latter followed by a mild oxidation with CuCl2 led to the 7-membered analogues. Intramolecular internalization of nitrenes was also recently explored by two research groups. In 2023, Xu and Wei prepared a wide panel of polycyclic azepines resulting from the insertion of a rhodium nitrenoid into a pendant benzene ring (Scheme [21]C, top).[34] On the other hand, the group of Ruffoni and Leonori described a multi-step protocol to access highly substituted azepanes (Scheme [21]C, bottom).[35] The sequence starts with the photogeneration of singlet aryl nitrene upon deoxygenation of a triplet nitroarene. Internalization of the nitrogen atom followed by the addition of a secondary amine nucleophile affords the azepine derivative. Saturation of the latter is made possible via catalytic hydrogenation with heterogeneous catalysts.

Although the dearomative enlargement of benzenes is beyond the scope of this short review, these key precedents were undoubtedly a source of inspiration for the development of analogous reactions on heterocyclic scaffolds.

Pyridine and its benzo-fused congeners are probably among the most represented heteroaromatic rings in bioactive molecules.[36] [37] [38] [39] In fact, this simple 6-membered ring can be easily produced in many efficient ways.[40] Furthermore, a plethora of pyridine-containing building blocks are commercially available. Therefore, in the context of dearomative enlargement protocols, several groups envisaged pyridines as abundant feedstocks for the development of innovative reactions aiming at inserting carbon, nitrogen or even oxygen atoms within the backbone of pyridines and (iso)quinolines.

A first example illustrating this strategy was reported in 2004 by the group of Yadav (Scheme [22]).[41] After a preliminary step involving activation of the nitrogen atom with ethyl chloroformate, the in situ generated (iso)quinolinium salts are successfully enlarged with a combination of diazocarbonyl compounds and a catalytic amount of copper(II) as the catalyst at 75 °C. The corresponding benzoazepines were obtained in very good to quantitative yields on millimolar scale. Even though the mechanism was not detailed in their report, one can anticipate the transient formation of a copper carbenoid intermediate.

Zoom Image
Scheme 22 (Iso)quinolines to benzoazepines by Yadav

This strategy was further explored in two subsequent reports by the group of Hari[42] in 2010 and García Mancheño[43] in 2019. In the former, the authors engaged a methylated quinolinium salt with freshly prepared diazo(trimethylsilyl)methyllithium or -magnesium bromide at room temperature (Scheme [23]A). Fine-tuning of both the organometallic species and temperature demonstrated that magnesium quantitatively promotes nucleophilic attack at the 2-position of the quinolinium salt. Finally, heating the resulting intermediate in the presence of a catalytic amount of copper(I) generates the carbenoid upon dediazotization, as suggested by the authors. The latter rearranges via the transient formation of a bicyclic azanorcaradiene intermediate. A prompt 6π electrocyclic ring opening furnishes the corresponding benzoazepines. Noteworthy, depending on the electronic nature of the pyridinium salt (N-CO2Et for Yadav, N-Me for Hari), the bond cleavage resulting in the ring expansion does not occur on the same bond.

Zoom Image
Scheme 23 Pyridine derivatives with diazo compounds to (benzo)azepines by Hari (A) and García Mancheño (B)

On the other hand, the group of García Mancheño were able to improve the synthetic applicability of this method by circumventing the use of harsh organometallic reactants, and therefore tackle the incompatibility with sensitive functional groups (Scheme [23]B). In fact, they demonstrated that diazo(trimethylsilyl)methane was nucleophilic enough to react with the in situ generated quinolinium salts, obtained by oxidizing the starting dihydroquinolines with tritylium perchlorate. Following a similar pathway to the one proposed by Hari and co-workers,[42] more than 20 examples bearing electrophilic functional moieties were successfully prepared. Interestingly, using the same conditions, the authors could extend their strategy to a dihydroisoquinoline and a dihydropyridine, yielding 115 and 116. Interestingly, the transformation developed by García Mancheño does not require any transition metal.

In 2015, the group of Liu and Li developed a method to efficiently promote the ring expansion of 2-alkynyldihydropyridines and their fused analogues under gold catalysis (Scheme [24]).[44] The starting materials are easily prepared in one step from the parent pyridines. In the presence of pyridine N-oxide as mild stoichiometric oxidant, the authors could prepare monocyclic as well as benzo-fused 3-acylazepines in very good yields. Regarding the mechanism, the authors postulated a nucleophilic attack of the oxygen atom of the pyridine N-oxide onto the alkynyl moiety promoted by the Au(I) cation. Subsequent rearrangement of the latter releases the pyridine and reveals the gold-carbenoid intermediate. Finally, a cyclization step affords the cationic azanorcaradiene, which undergoes ring expansion to furnish the corresponding azepine and regenerate the catalyst. In addition to their mechanistic experiments, the authors supported each elementary step with DFT calculations. Four years later, Cai and co-workers reported the same transformation using a heterogeneous gold catalyst. The latter is easily recovered by filtration and can be re-engaged up to eight times without noticeably altering the yield.[45]

Zoom Image
Scheme 24 Dihydropyridines and dihydroquinolines to benzoazepines by Liu and Li

More recently, Yoo and Kim developed a silver-catalyzed sequence involving quinolinium zwitterions and diazoesters (Scheme [25]).[46] Silver is not the most common catalyst reported to generate carbenoids; however, canonical transition metals such as rhodium, gold or copper failed to yield the expected benzoazepine. According to their mechanistic experiments, the authors concluded that the nucleophilic addition of the diazo compounds occurs on the C4 atom, thus affording a partially dearomatized intermediate. This regioselectivity contrasts with that observed in Scheme [23]; the latter can be explained by the moderate steric hindrance of the pendant vinyl group at the 2-position. An azanorcaradiene intermediate is also suggested by the authors to explain the ring enlargement. The third ring, namely the dihydroimidazole, is formed after an intramolecular hydroamination step.

Zoom Image
Scheme 25 Quinolinium zwitterions to fused azepines by Yoo and Kim

Rather than adding an external diazo compound to their system, Beeler and co-workers capitalized on the visible-light-induced rearrangement of pyridinium (and their fused analogues) ylides to azepines (Scheme [26]).[47] These photoactive intermediates are generated in situ upon deprotonation with DBU (1,8-diazabicyclo(5.4.0)undec-7-ene) or TMG (1,1,3,3-tetramethylguanidine). Noteworthy, monocyclic azepines are prepared using a flow photoreactor whereas the fused congeners are simply synthesized in batch. Upon irradiation, the pyridinium ylide reaches a singlet excited state. Radical recombination of the latter forms the azanorcaradiene intermediate, which ring-opens to deliver the azepine analogue.

Zoom Image
Scheme 26 Pyridinium or quinolinium salts to monocyclic or fused azepines by Beeler

Besides carbon atom insertion reactions, Moreau and Ghiazza recently reported the photochemical enlargement of pyridines, through the insertion of an additional nitrogen atom, to provide 1,2-diazepine-containing scaffolds (Scheme [27]).[48] This sequential one-pot protocol enables the synthesis of more than 40 products, including bioactive substances with relevant functional groups, which were previously complicated to prepare via classical disconnections. Moreover, two electronically different pyridine moieties could be easily discriminated leading to product 134 selectively. Their reaction design starts with electrophilic amination of the pyridine derivative, followed by an in situ functionalization of the exocyclic atom to deliver the corresponding pyridinium ylide. According to the seminal work of Streith from the late 1960s and early 1970s,[49] [50] the latter are known to rearrange under UV light irradiation and form the 7-membered analogues.

Mechanistic investigations were conducted on the aforementioned system and revealed the involvement of a singlet excited state as a productive key intermediate (Scheme [28]). On the other hand, photoinduced formation of the triplet excited state via energy transfer from the thioxanthone led to cleavage of the N–N bond, and therefore to the generation of a nitrene. The latter could be trapped with an olefin which resulted in the isolation of the aziridine derivative. In addition, an inverse secondary kinetic isotope effect (KIE) was observed for the photochemical rearrangement, thus supporting the formation of a transient (Csp3)-containing diazanorcaradiene intermediate during the rate-determining step (RDS).

Zoom Image
Scheme 27 Pyridines to 1,2-diazepines by Moreau and Ghiazza. Mes = mesityl.
Zoom Image
Scheme 28 Pyridines to 1,2-diazepines: postulated mechanism

Finally, a proof-of-concept was recently validated by the group of Sarlah (Scheme [29]).[51] Capitalizing on their previously established MTAD-based dearomative approach (see Scheme [21]),[33] [52] they were able to achieve the oxidative dearomatization of pyridines in two steps to form pyridine oxides, which are in equilibrium with their 1,4-oxazepine tautomers. Visible-light-activation of a mixture of both MTAD and pyridine at –78 °C triggers the formation of a bridged adduct resulting from a [4+2] cycloaddition. Epoxidation of the latter yields an isolable tetracyclic intermediate. Base-promoted retrocyclization forms the corresponding pyridine oxide.

Zoom Image
Scheme 29 Pyridines to 1,4-oxazepines by Sarlah. MTAD (X) = N-methyl-1,2,4-triazoline-3,5-dione.

Overall, the yields are synthetically useful, although the scope remains restricted to 2-substituted or 2,4-disubstituted pyridines with only a handful of functional groups evaluated. Nonetheless, the synthesis of this rather unusual heterocycle is far from trivial and could be telescoped to downstream functionalizations to achieve the preparation of complex architectures.


# 4

Conclusion and Perspectives

In brief, recent decades have witnessed the emergence of modern strategies aiming both at dearomatizing and enlarging the skeleton of (hetero)aromatic rings in a single-flask operation. The newly synthesized structures are often intricate to prepare via tedious multi-step syntheses following canonical disconnections. Additionally, the three-dimensionality conferred by the dearomatization increases the complexity of the analogues, and therefore their likelihood to selectively interact with biological receptors. In other words, the dearomative enlargement strategy merges the benefits of both skeletal editing, as a powerful tool to generate new potential hits from available substances from our chemical libraries, and the saturation of aromatic compounds, as the insertion of additional atoms slightly alters the planarity and the overall geometry of the molecule.

Overall, these examples mostly rely on the insertion of metallic carbenoids within heteroaromatic rings; only a handful of methods report the internalization of other heteroatoms. As a result, the heteroatom to carbon ratio decreases during the dearomative ring enlargement process. In our humble opinion, future developments covering the insertion of chalcogen or pnictogen atoms would be remarkably beneficial in expanding our chemical space. This does not relegate the insertion of carbenoids to the past, rather the opposite. As these methods, in their vast majority, rely on the use of transition metals, the enantioselective incorporation of chiral centers is of utmost interest.


#
#

Conflict of Interest

The authors declare no conflict of interest.


Corresponding Author

Clément Ghiazza
Université Paris-Saclay, UVSQ, CNRS, UMR 8180 Institut Lavoisier de Versailles
78035 Versailles Cedex
France   

Publication History

Received: 02 May 2024

Accepted after revision: 29 May 2024

Accepted Manuscript online:
29 May 2024

Article published online:
17 June 2024

© 2024. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


   Permissions and Reprints All articles of this category
Zoom Image
Clément Ghiazza obtained his Ph.D. in 2019 under the supervision of Dr. Anis Tlili and Dr. Thierry Billard during which he was involved in the development of new fluoroalkylselenolation methods at the University of Lyon, France. He subsequently earned a postdoctoral fellowship from the Alexander von Humboldt foundation to work in the group of Dr. Josep Cornella at the Max Planck Institut für Kohlenforschung in Mülheim an der Ruhr, Germany. During this period, his research topics were mainly focused on the late-stage modification of highly functionalized heterocycles, the synthesis of hybrid fluorinated isosteres and the development of catalytically active bismuth complexes. In October 2022, Clément was recruited as a CNRS researcher at the Institut Lavoisier de Versailles, France. In a group led by Prof. Xavier Moreau, he is interested in innovative heterocyclic synthesis and asymmetric (organo)catalysis.
Zoom Image
Scheme 1 Context of modern dearomative enlargement reactions
Zoom Image
Scheme 2 Indoles to dihydroquinolines by Liu and Bi. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate. PMP = p-methoxyphenyl.
Zoom Image
Scheme 3 Pyrroles to dihydropyridines by Liu and Bi. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate.
Zoom Image
Scheme 4 Indoles to benzoazepines by Bakthadoss
Zoom Image
Scheme 5 Benzofurans to benzoxaborins by Yorimitsu. IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene. OA = oxidative addition. TM = transmetalation. RE = reductive elimination. TBS = tert-butyldimethylsilyl.
Zoom Image
Scheme 6 Benzofurans to 6-membered oxaheterocycles by Yorimitsu. TMEDA = tetramethylethylenediamine.
Zoom Image
Scheme 7 (Benzo)thiophenes to bridged 8-membered heterocycles by Houk and Glorius
Zoom Image
Scheme 8 (Benzo)thiophenes to bridged 8-membered heterocycles: postulated mechanism. RDS = rate-determining step.
Zoom Image
Scheme 9 A vinylpyrazolium salt to a dihydropyrimidine by Pardo
Zoom Image
Scheme 10 Pyrazolium salts to dihydropyrimidines by Chen (A) and Derksen (B)
Zoom Image
Scheme 11 Dearomative enlargement of pyrazoles, isoxazoles and isothiazoles: unified mechanism
Zoom Image
Scheme 12 Isothiazolones to thiazinones by Gosney
Zoom Image
Scheme 13 (Benzo)isoxazoles and benzoisothiazoles to (benzo)oxazines and benzothiazines by Davies
Zoom Image
Scheme 14 Isoxazoles and isoxazolones to oxazines and oxazinones by Khlebnikov (A) and Jurberg and Davies (B)
Zoom Image
Scheme 15 (Benz)isoxazoles to spiro(benz)oxazines by Chen and Pu (A) and Lu and Wang (B). esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate.
Zoom Image
Scheme 16 (Benzo)isoxa(thia)zoles to (benzo)oxa(thia)zines by Bi. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate. DIPEA = diisopropylethylamine.
Zoom Image
Scheme 17 Pyrazoles and indazoles to dihydropyrimidines and dihydroquinazolines by Novikov (A) and Bi (B and C). esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate. DIPEA = diisopropylethylamine.
Zoom Image
Scheme 18 Divergent preparation of benzoxazines or benzoxazepines from benzoisoxazoles and ynamides by R.-S. Liu. L = P(tBu)2(o-biphenyl) (JohnPhos). IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene.
Zoom Image
Scheme 19 Ligand-controlled chemoselective dearomative enlargement of benzisoxazoles. L = P(tBu)2(o-biphenyl) (JohnPhos). IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene.
Zoom Image
Scheme 20 Benzisoxazoles to benzoxazines by Y. Liu. IPr = 1,3-bis(diisopropylphenyl)imidazol-2-ylidene.
Zoom Image
Scheme 21 A brief overview of dearomative enlargement reactions on benzene rings. MTAD = 4-methyl-1,2,4-triazoline-3,5-dione. esp = α,α,α′,α′-tetramethyl-1,3-benzenedipropionate.
Zoom Image
Scheme 22 (Iso)quinolines to benzoazepines by Yadav
Zoom Image
Scheme 23 Pyridine derivatives with diazo compounds to (benzo)azepines by Hari (A) and García Mancheño (B)
Zoom Image
Scheme 24 Dihydropyridines and dihydroquinolines to benzoazepines by Liu and Li
Zoom Image
Scheme 25 Quinolinium zwitterions to fused azepines by Yoo and Kim
Zoom Image
Scheme 26 Pyridinium or quinolinium salts to monocyclic or fused azepines by Beeler
Zoom Image
Scheme 27 Pyridines to 1,2-diazepines by Moreau and Ghiazza. Mes = mesityl.
Zoom Image
Scheme 28 Pyridines to 1,2-diazepines: postulated mechanism
Zoom Image
Scheme 29 Pyridines to 1,4-oxazepines by Sarlah. MTAD (X) = N-methyl-1,2,4-triazoline-3,5-dione.