Leveraging Fleeting Strained Intermediates to Access Complex Scaffolds

Abstract

Arynes, strained cyclic alkynes, and strained cyclic allenes were validated as plausible intermediates in the 1950s and 1960s. Despite initially being considered mere scientific curiosities, these transient and highly reactive species have now become valuable synthetic building blocks. This Perspective highlights recent advances in the field that have allowed access to structural and stereochemical complexity, including recent breakthroughs in asymmetric catalysis.

Introduction

In 1902, a provocative proposal was put forth by Stoermer and Kahlert, who contemplated the intermediacy of benzofuranyne 1 (Figure 1).¹ Although the existence of 1 would ultimately be called into question,²⁻⁴ the proposal led chemists to consider if triple bonds could exist in small rings. Roughly 50 years later, benzyne (2) and cyclohexyne (3) were validated experimentally, thanks to pioneering efforts by Roberts,^5,6 Wittig,⁷ and Huisgen,⁸ in particular. Soon thereafter, in 1966, Wittig showed that 1,2-cyclohexadiene (4), an unusual-looking counterpart to benzyne (2) and cyclohexyne (3), could be generated and intercepted in cycloaddition processes.⁹ These results were striking at the time, as 2–4 possess functional groups (i.e., alkynes and allenes) that would ordinarily be linear. The bent nature of these species renders them transient and highly reactive intermediates that initially received little use in chemical synthesis. However, over the past few decades, the field of arynes and related strained intermediates has undergone a substantial period of growth, as shown in Figure 1.¹⁰

Figure 1.

Historical perspective, growth of “aryne” or “benzyne” chemistry, and select synthetic applications.

The synthetic utility of arynes and related strained intermediates is often underappreciated, perhaps owing to common misconceptions regarding safety or high reactivity (and consequently, misconceptions regarding low selectivity in aryne reactions). Conversely, we highlight the impact of arynes and related intermediates in the syntheses of 5–9. Popular ligands, such as XPhos (5), are made using aryne chemistry.¹¹ The medicinal chemistry route toward Chantix (6) was enabled by an aryne cyclization.¹² Syngenta has prepared isopyrazam (7), an important fungicide, using an aryne Diels–Alder reaction on multi-kilogram scale.¹³ Lastly, there are numerous examples of arynes and related intermediates in total synthesis,¹⁴⁻¹⁶ such as the use of an aryne insertion in Sarpong’s synthesis of cossonidine (8) and a cyclohexyne insertion in Carreira’s synthesis of guanacastepene N (9).^17,18 As these examples highlight, arynes and related intermediates can be used to build two new bonds in a single transformation. Moreover, it should be noted that arynes can be generated under exceedingly mild, safe, and operationally simple fluoride-based reaction conditions, thanks to the advent of Kobayashi silyl triflates.¹⁹⁻²² In turn, chemists have sought to leverage the synthetic utility of these once controversial species, in addition to deepening our fundamental understanding through electrophilicity studies²³ and regioselectivity studies.²⁴⁻²⁶

Rather than providing a comprehensive overview of the field, several of which are available elsewhere,^{14−16,20,27−37} this Perspective examines some of the quickly emerging areas of aryne chemistry and related strained intermediates. We assess the following recent advances: (a) the use of heterocyclic strained alkynes to construct scaffolds of value to medicinal and materials chemistry, (b) access to quaternary stereocenters using noncatalytic reactions of arynes and cyclic alkynes, (c) catalytic asymmetric reactions of arynes, (d) the assembly of intricate heterocyclic scaffolds via strained cyclic allene intermediates, and (e) stereospecific and catalytic asymmetric transformations of strained cyclic allenes. We hope this discussion underscores the current excitement in the field and, moreover, helps to draw researchers into an area that continues to rapidly evolve with opportunities for discovery.

Use of Heterocyclic Alkynes in the Synthesis of Medicinal and Materials Scaffolds

New strategies to access heterocyclic compounds, especially those with a high degree of sp³-rich character, remain highly sought after due to the numerous applications of heterocycles in drugs, agrochemicals, natural products, and materials.³⁸⁻⁴³ Just within the pharmaceutical industry, the vast majority of small-molecule drugs approved by the U.S. Food and Drug Administration contain a nitrogen or oxygen-containing heterocycle.^44,45 A few examples are Plavix (10), an antiplatelet, rhoeadine (11), a sedative and antitussive, and the antimalarial drug artemisinin (12) (Figure 2).

Figure 2.

Representative natural products and pharmaceuticals containing heterocycles and representative nitrogen- and oxygen-containing strained cyclic alkynes.

Strained cyclic alkynes have emerged as valuable building blocks to access sp³-rich³⁹ medicinally relevant heterocycles (Figure 2).⁴⁶⁻⁴⁹ Heteroatom-containing cyclohexyne derivatives, such as 2,3-piperidynes 13, 3,4-piperidynes 14, and oxacyclohexyne 15 have been far less studied than related aryne counterparts, (e.g., 3,4-pyridyne (16)).⁵⁰⁻⁵³ For comparison, the 3,4-pyridyne (16) was first disclosed in 1955,⁵⁴ but studies involving 3,4-piperidynes were only reported recently. Specifically, the Danheiser group reported access to 13 in 2014⁴⁷ and our group described generation and interception of 14 and 15 in 2015 and 2016, respectively.^48,49 These studies demonstrate the synthetic value of transient intermediates 13–15.

The initial breakthrough in this area by Danheiser and co-workers involved piperidyne precursor 17 (Figure 3). Upon treatment of 17 with CsF or KF, in the presence of a cycloaddition partner, a variety of (3 + 2) and (2 + 2) cycloadducts 20 were obtained. Presumably, these reactions proceed via the regioselective trapping of cyclic alkyne 19. Three examples are shown to highlight the diverse products (i.e., 22, 24, and 26) that can be obtained using this methodology. Moreover, each of the cycloadducts shown were isolated as one major constitutional isomer, with the significant regioselectivity of the reaction being attributed to the electronic effect of the nitrogen atom.

Figure 3.

Selected trapping experiments involving piperidyne 19.

Our laboratory had simultaneously engaged in complementary studies, which were geared toward accessing cyclic alkynes 28 (Figure 4). After developing synthetic routes to the requisite silyl triflates 27, we performed experiments to generate and trap the corresponding heterocyclic alkynes. As summarized, strained intermediate trapping via (4 + 2) and (3 + 2) cycloadditions led to an array of interesting heterocyclic products, such as aza- and oxa-bicycles 30 and 32, respectively, isooxazolines 34 and 35, pyrazole 36, and triazole 37. Where applicable, reactions occurred regioselectively, indicative of initial bond formation occurring at C4 of the reactive strained intermediate. It should be emphasized that by strategically varying the trapping partner, one can utilize common silyl triflate precursors (i.e., 27) to rapidly arrive at a variety of diverse heterocyclic scaffolds, including previously unknown scaffolds and analogs of medicinally relevant compounds.⁵⁵⁻⁵⁸

Figure 4.

Selected trapping experiments of heterocyclic alkynes using silyl triflate precursors.

An attractive feature related to strained cyclic alkyne methodologies is the ability to make reliable predictions regarding regioselectivities using the aryne distortion model. This model, proposed by Houk and co-workers via our earlier collaborative studies,^25,33 shows that nonsymmetrical cyclic alkynes are geometrically distorted in their ground state. Nucleophilic addition occurs more favorably at the alkyne terminus with a larger internal angle (i.e., more distorted toward linearity), which correlates to a lower calculated distortion energy seen in the corresponding transition state. Moreover, larger differences between the internal angles typically correlate with regioselectivities in trapping experiments. Thus, by analyzing the ground state structure of a strained cyclic alkyne, one can typically make meaningful regioselectivity predictions. In the cases of piperidyne 14 and oxacyclohexyne 15, both react with a preference for C4 addition (Figure 5). Consistent with the distortion model, the ground state geometries of 14 and 15 shown by DFT calculations demonstrate that the C4 alkyne terminus is more distorted toward linearity. Furthermore, the difference between the internal angles of 14 and 15 are 12° and 15°, respectively. The greater angle difference of 15° seen in 15 is consistent with observed regioselectivities. For example, as shown in Figure 4, the nitrone trapping proceeds with higher selectivity when using 15 compared to 14.

Figure 5.

Regioselectivity of piperidyne 14 and oxacyclohexyne 15 trappings.

These methods demonstrate the value of strained heterocyclic alkynes for the synthesis of medicinally relevant scaffolds. Using common building blocks, rapid structural diversification can be achieved, with regioselectivity being predicted using simple calculations. Moreover, this chemistry provides a new and unusual strategy for accessing decorated heterocycles (i.e., using new strained intermediates). We expect these reports to influence future retrosynthetic analyses for the synthesis of heterocycles seen commonly in pharmaceuticals.

In addition to their value for the synthesis of medicinally relevant scaffolds, strained heterocyclic alkynes have seen recent utility in the synthesis of aromatic structures relevant to materials chemistry. Highly conjugated small molecules and polycyclic hydrocarbon frameworks have many potential applications in the field of organic electronics and material science.^59,60 For example, these conjugated small molecules have been widely used in light-emitting diodes (OLEDs),⁶¹⁻⁶⁴ field-effect transistors (OFETs),⁶⁵⁻⁶⁸ and photovoltaics (OPVs).⁶⁹⁻⁷¹ Two common conjugated small molecules are 9,10-diphenylanthracene (38) and triphenylene (41) (Figure 6).⁷²⁻⁷⁵ Heteroatom-containing derivatives of these compounds, such as 39, 40, 42, and 43, have been gaining interest as the presence of heteroatoms can modulate electronic properties.⁷⁶⁻⁷⁸ New synthetic methods to access novel heterocyclic conjugated materials are highly desirable. As such, in 2017, our laboratory synthesized heterohelicene⁷⁹43 (and derivatives) using an indolyne cyclotrimerization reaction.^80,81 Indole trimers have been used in various materials applications.^78,82−87 Building on these studies, we sought to access heteroatom-containing 9,10-diarylanthracene scaffolds using arynes and strained alkyne building blocks.

Figure 6.

Notable polycyclic aromatic hydrocarbons.

In 2019, we disclosed a modular strategy toward N-containing 9,10-diphenylanthracene derivatives, wherein all four aryl rings could be easily modified.⁸⁸ The sequence was inspired by the collective seminal studies by Steglich, Nuckolls, and Wudl pertaining to double-aryne annulations of oxadiazinones.⁸⁹⁻⁹¹ As shown in Figure 7, our approach involved two steps. First, 3,4-piperidyne (14), generated in situ from the corresponding silyl triflate, undergoes reaction with oxadiazinones 44 to give pyrones 45 through a Diels–Alder/retro-Diels–Alder sequence.⁹² The oxadiazinones are easily prepared with two different aryl substituents from readily available starting materials.⁹³ In a second Diels–Alder/retro-Diels–Alder sequence, pyrones 45, which are bench stable and isolable, react with an aryne or nonaromatic strained cyclic alkyne (i.e., 46) to deliver polycyclic hydrocarbon frameworks 47. 48–50 are representative products. As needed, subsequent oxidation can be performed to access more highly conjugated derivatives.

Figure 7.

Modular strategy to access 9,10-diphenylanthracene derivatives and representative products 48–50.

Figure 8 highlights two additional aspects of this chemistry. In the first, 51–53 were shown to react at room temperature in the presence of CsF. This three-component coupling obviates the need to isolate a pyrone intermediate and delivers 54 in 56% yield. Additionally, the methodology could be used to access 55, a novel heterocyclic PAH scaffold that fluoresces, displaying a blue emission. In the presence of acid, pyridinium salt 56 forms, which displays an orange emission. Stimuli responsive materials are useful in a host of applications such as pH fluorescence sensors^94,95 and solid-state fluorescent switches.⁹⁶

Figure 8.

Utilizing arynes to access 9,10-diphenylanthracene derivatives.

The chemistry described above pertaining to PAHs showcases an exciting new direction of modern strained intermediate chemistry. Notably, Hosoya and co-workers published a complementary method in 2019 that utilizes a variety of strained intermediates, including thienobenzyne, to access PAH skeletons.⁹⁷ Our group has published a follow-up study as well, which provides additional experimental results, as well as a computational mechanistic investigation.⁹⁸ These efforts demonstrate that heterocyclic strained intermediates can be strategically leveraged through a cycloaddition cascade sequence to rapidly generate PAH scaffolds. The transformation enables access to structurally diverse products through the formation of four carbon–carbon bonds. It is expected that this chemistry and variants thereof will prompt the development of related methods that rely on strained intermediates to access compounds of value to materials chemistry.

Use of Arynes and Cyclic Alkynes to Access Quaternary Stereocenters (Noncatalytic)

The majority of reported methodologies and synthetic applications that utilize strained cyclic alkynes are intermolecular reactions that generate achiral or racemic products. However, efforts have been put forth to generate enantioenriched products, including those that rely on the use of chiral auxiliaries or reagents. For example, reports by Lautens and co-workers^99,100 and Barrett et al.^101,102 demonstrate stereocontrolled intermolecular reactions of arynes for the introduction of tertiary stereocenters using Oppolzer or Schöllkopf reagents, respectively. As accessing quaternary stereocenters is especially valuable,^103−106 we were interested in utilizing arynes and strained cyclic alkynes to access stereodefined quaternary centers in an intermolecular fashion.

Our efforts, which were reported in 2018,¹⁰⁷ concerned the reaction between α-ketoester 57 and strained cyclic alkyne 46 to yield α-substituted product 58, as shown in Figure 9. Prior efforts to achieve this transformation in a racemic sense were accompanied by concomitant C–C bond fragmentation,¹⁰⁸ so we considered an alternative, two-step approach. First, α-ketoester 57 would be treated with amine 59 to afford the corresponding enamine 60. Enamine 60 could then be used to trap aryne 46, affording α-substituted product 58 after hydrolysis. We envisioned that employing a chiral amine (i.e., 59) would render the reaction diastereoselective and yield 58 in enantioenriched form. It should be noted that enantioselective α-arylation of β-ketoesters, one of the net transformations we hoped to develop, had remained a challenging synthetic problem.^109−116

Figure 9.

Intercepting arynes and cyclic alkynes for the installation of stereodefined quaternary stereocenters.

At the time of our study, the use of enamines and strained cyclic alkynes to construct quaternary stereocenters was unknown; thus, the racemic arylation^117−119 was first developed (Figure 10). Benzylamine was condensed onto ketoester 61 to afford enamine 62 (R = Bn). Next, enamine 62 was used to trap benzyne, which was generated in situ from silyl triflate 53 using CsF in DME. Following hydrolysis with 1 M HCl, α-arylated product 63 was obtained in 92% yield. To effect the desired stereoselective variant, we ultimately arrived at the use of an anthracenyl derivative (i.e., R = 67). Several products that were obtained are shown in Figure 10 (i.e., 64–66), which highlight the use of heterocyclic arynes and cyclic alkynes as trapping agents, as well as a 7-membered β-ketoester.

Figure 10.

Selected substrate scope of α-arylation methodology.

Another attractive aspect of this reaction methodology is the one-pot variant shown in Figure 11. Treatment of ketoester 68 with chiral amine 67 yielded the corresponding enamine, which was not isolated. This intermediate was then subjected to silyl triflate 53 and CsF, followed by the addition of aqueous 1 M HCl. The product, ketoester 68, was isolated in 68% yield and 92% ee. Of note, amine 67 was recovered in 67% yield. Overall, this methodology provided a means to access stereodefined quaternary stereocenters by the interception of strained intermediates in intermolecular processes. As will be discussed later in this review, an elegant catalytic asymmetric variant of this methodology has subsequently been developed.¹²⁰

Figure 11.

α-Arylation reaction demonstrated in one-pot.

The synthesis of natural products and their derivatives also provides an opportunity to use arynes in complex settings, as has been well demonstrated in the literature.^14,16,33,34 With regard to several recent efforts, the Hoye group has used arynes generated from hexadehydro-Diels–Alder reactions with readily abundant natural products to access complex derivatives.¹²¹ Additionally, our laboratory has performed late-stage intermolecular aryne cycloadditions to access derivatives of strictosidine, the last common biosynthetic precursor to all monoterpene indole alkaloids.¹²² More commonly, natural products have been accessed through diastereoselective intramolecular aryne trappings of enantioenriched substrates. Early on, our laboratory used an “indolyne” cyclization to build the complex bridged bicyclic core of coveted welwitindolinone natural products.^123−127 Subsequently, we targeted the tubingensin alkaloids, wherein we envisioned using an aryne cyclization to access a quaternary center.

Tubingensins A and B (70 and 71, respectively) are complex indole diterpenoids first isolated from Aspergillus tubingensis in 1989 (Figure 12).^128,129 Both natural products feature disubstituted carbazole moieties. In tubingensin A (70), the carbazole is fused to a cis-decalin containing four contiguous stereocenters, two of which are vicinal quaternary stereocenters. Tubingensin B (71), however, contains a carbazole fused to a [3.2.2]-bridged bicycle containing five stereocenters, four of which are contiguous. As an additional layer of complexity, three of the stereocenters are quaternary, two of which are vicinal. Moreover, tubingensins A and B (70 and 71, respectively) possess antiviral activity against herpes simplex virus type 1 (HSV-1) along with pesticidal activity.^128,129 At the time we began our efforts, there were no total syntheses of tubingensins A or B, although Li and Nicolaou reported an elegant route to (±)-70 in 2012.¹³⁰

Figure 12.

Tubingensin alkaloids and natural products with quaternary stereocenters introduced using aryne chemistry.

Given the intriguing structural features of tubingensins A and B, our group proposed synthetic routes to these compounds during a “Molecule of the Month” (MOM) exercise in a 2009 group meeting. Several routes that would employ aryne intermediates were proposed, some of which were tested in the laboratory. Although our initial strategies proved unsuccessful in the laboratory, they helped us appreciate the challenge of establishing quaternary stereocenters in complex frameworks using arynes. Prior studies had largely focused on the use of arynes to access tertiary stereocenters.^{122,131−134} With regard to setting quaternary centers of natural products and drug candidates using arynes, notable precedent was available in the synthesis of crinine (72)¹³⁵ and ibutamoren mesylate (73)¹³⁶ (Figure 12). There were no such examples where scaffolds bearing vicinal quaternary stereocenters, like those seen in the tubingensin alkaloids, were accessed using aryne cyclizations.

Following our initial total synthesis of (+)-tubingensin A (70),¹³⁷ we were well-poised to address the more complex family member, (−)-tubingensin B (71). As summarized in Figure 13, (+)-dihydrocarvone (74), an abundant chiral terpene building block, and 2-hydroxycarbazole (75) were used as starting materials and elaborated to give carbazolyne precursor 76. Of note, 76 possesses all of the carbons that would be needed to complete the total synthesis. This key intermediate was subjected to sodium amide and tert-butanol in THF^138,139 at 23 °C to effect carbazolyne cyclization. Although we were expecting to form a single C–C bond, we instead observed the formation of two new C–C bonds as evident by product cyclobutenol 78. Although this outcome was undesired, the formation of 78 highlights the structural complexity accessible using aryne chemistry. The reaction likely proceeds via a formal [2 + 2] cycloaddition (e.g., 77) and occurs with excellent diastereoselectivity. Nonetheless, this transformation served to establish the necessary vicinal quaternary stereocenter framework, construct the natural product’s seven-membered ring, and preserve the phenylselenide moiety necessary for a late-stage radical cyclization. To complete the synthesis, cyclobutenol 78 was subjected to Murakami’s Rh-catalyzed fragmentation conditions,¹⁴⁰ which delivered the ketone 79 in 53% yield via fragmentation of the cyclobutenol ring. In turn, ketone 79 could be elaborated to (−)-tubingensin B (71).¹⁴¹ This effort demonstrates the utility of highly reactive aryne intermediates for the assembly of sterically congested carbon–carbon bonds in natural products.

Figure 13.

Overview of (−)-tubingensin B total synthesis.

Arynes in Asymmetric Catalysis

The aforementioned section highlights two key tactics for accessing enantioenriched products using aryne intermediates. More specifically, chiral auxiliaries, chiral reagents, or enantioenriched substrates have been the most common and successfully used to build intricate scaffolds and quaternary stereocenters. In turn, chemists have questioned: is it possible to leverage asymmetric catalysis in aryne trapping experiments? Such processes may be very challenging, as they could require two transiently generated intermediates,^29,32 including a highly reactive aryne, to come together and undergo a productive reaction with stereocontrol. In addition to the fundamental scientific interest of using a fleeting aryne intermediate in a catalytic reaction, there are also practical advantages of asymmetric catalysis.^142−144 We highlight three key studies in this emerging field involving the synthesis of helicenes (axial chirality)^145,146 and the introduction of quaternary stereocenters (point chirality).¹²⁰

A breakthrough involving the use of arynes in asymmetric catalysis was reported by Guitián and co-workers in 2006.¹⁴⁵ More specifically, the authors studied [2 + 2 + 2] cycloadditions of arynes and alkynes to access enantioenriched helicenes. Accessing helicenes in an enantioselective manner has remained an important area of research^79,147,148 and prior aryne-based approaches had given rise to racemic products.^149,150 As shown in Figure 14, silyl triflate 80 was treated with CsF, alkyne 81, catalytic Pd₂(dba)₃, and BINAP to afford helicene 82. The reaction likely proceeds through a palladium-catalyzed cyclotrimerization of two in situ generated aryne intermediates (from silyl triflate 80) and alkyne 81. Although the yield was modest, helicene 82 was obtained in 67% ee. This study demonstrated an important proof-of-concept with regard to arynes and their use in asymmetric catalysis.

Figure 14.

Enantioselective synthesis of helicenes using arynes and asymmetric catalysis.

Building upon this pioneering study and their own prior work in this area,¹⁵¹ the Kamikawa group published an enantioselective synthesis of helicenes in 2020.¹⁴⁶ For example, as shown in Figure 15, a cyclotrimerization was effected between aryne 83 and alkyne 84 using Pd₂dba₃·CHCl₃, CsF, and (S)-QUINAP (86).¹⁵² This transformation features a design similar to the aforementioned example by Guitián and co-workers, and affords triple helicene (M,P,M)-85 in 49% yield and 96% ee. Notably, despite using racemic aryne precursor 83, both terminal helicenes exclusively formed the (M)-[5]-helicenyl moiety, which suggests a kinetic resolution or dynamic kinetic resolution occurs in the reaction. Overall, this study achieved the first enantioselective synthesis of a triple helicene, while also providing the first practical catalytic asymmetric aryne trapping reaction.

Figure 15.

Enantioselective synthesis of triple helicenes via the catalytic asymmetric trapping of arynes.

Also in 2020, the Luo group reported a pioneering study in the catalytic asymmetric trapping of arynes and related strained intermediates.¹²⁰ Their methodology strategically blends organocatalysis and electrochemistry to achieve the α-functionalization of 1,3-dicarbonyl substrates and establish point chiral stereochemistry. For example, treatment of ketoester 61 with 1-aminobenzotriazole (87) and amine catalyst 88, under their optimal electrochemical oxidation conditions, afforded α-arylated ketoester 63 in 71% yield and 94% ee (Figure 16). Mechanistically, the reaction is thought to proceed by way of an intermediate enamine species, which reacts with the aryne generated by electrochemical oxidation of 87. The cobalt catalyst is proposed to bind the aryne triple bond and stabilize excess benzyne intermediate that is generated.^153−155 Products 89–92 provide further examples of the products that were generated using this methodology. These studies demonstrate an electrochemical oxidation approach to access benzyne and related strained intermediates, which nicely complements prior studies involving the use of nonelectrochemical oxidation conditions for aryne generation.^156−160 Moreover, this study demonstrates that arynes can be engaged in asymmetric catalysis to provide access to point chiral products in synthetical useful enantioselectivities. The ability to access stereodefined quaternary stereocenters is especially notable.

Figure 16.

Catalytic enantioselective α-arylation of 1,3-dicarbonyls.

Use of Strained Cyclic Allenes to Access Polycyclic Scaffolds

Whereas arynes and strained cyclic alkynes have been studied extensively, a related counterpart, strained cyclic allenes (i.e., 4, 93–94, Figure 17), have received significantly less attention. Of note, the discovery of 1,2-cyclohexadiene (4) by Wittig was reported in 1966,⁹ only 9 years after the validation of cyclohexyne (and 13 years after the validation of benzyne in 1953).^14,27,31 Theoretical investigations of these species have been reported^161−169 as well as synthetic advances, most notably by Christl.^170−176 In a key finding, Guitián and co-workers developed Kobayashi precursors to strained cyclic allenes and demonstrated their utility in several transformations, such as [4 + 2] cycloadditions.^177,178 Subsequently, our laboratory and West’s laboratory expanded the synthetic utility of Kobayashi cyclic allene precursors to encompass other intermolecular cycloadditions, especially (3 + 2) cycloadditions.^179−182

Figure 17.

Strained cyclic allenes and representative trapping reactions.

Here, we feature selected recent efforts aimed at harnessing strained cyclic allenes to build complex polycyclic scaffolds bearing heteroatoms and sp³ centers. Generally speaking, these reactions proceed by treating silyl triflate precursors 95 with a fluoride source to generate cyclic allenes 96. In situ trapping with cycloaddition partners 18 provides adducts 97 (Figure 17). As will be discussed in the remaining sections of this Perspective, strained cyclic allene chemistry provides exciting opportunities for study related to regiochemistry, pertaining to which olefin of the allene undergoes reaction, and stereochemistry, due to the inherent chirality of cyclic allenes and generation of a new sp³ center.

A striking example of cyclic allene chemistry was reported by West and co-workers in 2019, where in situ generated 1,2-cyclohexadienes underwent (4 + 2) cycloaddition with tethered furans to access complex tetracyclic products (Figure 18).¹⁸³ Substrates 98 were subjected to TBAF to afford cycloadducts 100, presumably by way of cyclic allene intermediate 99. The products contain three stereocenters, set in a relative sense, including a quaternary center. Representative products 101 and 102 highlight that the linker length could be varied (n = 1 or 2). In addition, heteroatoms were tolerated in the linker, as shown by cycloadducts 103 and 104. Of note, the transformations proceeded diastereoselectively in favor of the endo product and regioselectively, with the latter being governed by the tether. This methodology provides the first examples of an intramolecular cycloaddition with a six-membered cyclic allene and showcases the structural complexity that is accessible using cyclic allene chemistry.

Figure 18.

Intramolecular Diels–Alder reactions afford tetracyclic products.

Our laboratory was particularly interested in heterocyclic allenes and the possibility of accessing them using Kobayashi-type precursors. Many heterocyclic allenes had been accessed previously, but never using a silyl triflate as the precursor.^184−190 After developing syntheses of the appropriate precursors to azacyclic and oxacyclic allenes (i.e., 105 and 106, respectively), we demonstrated their use in cycloaddition reactions, with select results shown in Figure 19.^191,192 Operationally, reactions are performed by treating the appropriate silyl triflate precursor (105, X = NCbz; or 106, X = O) with a trapping agent 18 in the presence of CsF at 23 °C, ultimately leading to cycloadducts 108. Three cycloadducts arising from azacyclic allene precursor 105 are shown reflective of (4 + 2), (3 + 2), and (2 + 2) cycloadditions (entries 1–3). The cycloadducts, 109, 111, and 112, are formed in good yield and useful diastereoselectivities (when applicable). Similarly, oxacyclic allene precursor 106 could be employed in (4 + 2), (3 + 2), and (2 + 2) cycloadditions, thus giving rise to cycloadducts 114, 116, and 118, respectively (entries 4–6). Overall, by varying the heterocyclic allene precursor and the trapping agent, one can now use strained cyclic allene chemistry as a strategy to synthesize structurally complex heterocyclic compounds.

Figure 19.

Representative azacyclic and oxacyclic allene trapping experiments.

Whereas the results shown in Figure 19 showcase reactions of unsubstituted strained cyclic allenes, another important facet of these compounds is observing and understanding what happens when substituents are present on the allene carbons in intermolecular trapping experiments. Our laboratory has probed this issue in the context of azacyclic allenes,¹⁹¹ with key results provided in Figure 20. Silyl triflate 119, 122, and 125 were accessed as precursors to the corresponding cyclic allenes 120, 123, and 126, in order to probe the influence of electron-donating and electron-withdrawing substituents (i.e., alkyl and ester groups, respectively). As shown by the formation of 121, cycloaddition occurs on the olefin of 120 distal to the methyl group, whereas the cycloaddition takes place on the olefin proximal to the ester of 123. When both methyl and ester substituents were present, cycloaddition occurs proximal to the ester and distal to the methyl of cyclic allene 126, as one might expect in this matched scenario (125 + 113 → 127). All three reactions proceeded in good yields and excellent diastereoselectivities to give structurally complex products, including two with quaternary centers (i.e., 124 and 127), further highlighting the synthetic utility of this methodology. In a collaboration with the Houk group at UCLA, a distortion-interaction analysis demonstrated that interaction energies, related to electron-factors, likely govern regioselectivity.¹⁹¹ With regard to diastereoselectivity, reactions proceed with endo selectivity with respect to the unreactive cyclic allene double bond.^191,193 Given the synthetic utility of this methodology and the ability to understand the regiochemical and stereochemical outcomes, we expect this methodology will see increased usage in chemical synthesis. For example, Schreiber and co-workers have recently harnessed heterocyclic allene chemistry for the synthesis of DNA-encoded libraries.¹⁹⁴

Figure 20.

Regioselectivity investigation of substituted azacyclic allenes.

Strained Cyclic Allenes in Enantioselective and Stereospecific Reactions

The aforementioned studies demonstrate that cyclic allenes can undergo regio- and diastereoselective trapping to give polycyclic heterocyclic products. In an emerging area, efforts have also been put forth to control absolute stereochemistry in reactions of strained cyclic allenes. Seminal studies performed by Christl and co-workers demonstrated that enantioenriched cyclic allenes can be generated and intercepted using the Skattebøl rearrangement.^171,172 Our laboratory devised the alternative approach shown in Figure 21.^191,192 Silyl triflates 128 would be prepared with control of absolute stereochemistry and subjected to mild fluoride-based conditions. This could lead to the transmission of stereochemical information to cyclic allenes 129, which in turn could undergo stereospecific trapping with 18 to yield cycloadducts 130. Of note, this overall process would proceed with the transfer of point chirality from 128 to 130, via the intermediacy of axially chiral intermediates (i.e., 129), with potential ablation of the sole stereocenter present in the substrate. This approach involving enantioenriched silyl triflate precursors had not been examined previously.

Figure 21.

Transfer of stereochemical information in allene cycloadditions.

As shown in Figure 22, the success of the aforementioned approach was validated in the context of both aza- and oxacyclic allenes. In the first example, silyl triflate (+)-119 was accessed in >99% ee by performing preparative chiral supercritical fluid chromatography (SFC) on a synthetic precursor. Under standard cyclic allene generation and trapping conditions, with diene 31 as the trapping agent, a Diels–Alder adduct was obtained in 98% ee (98% stereoretention).¹⁹¹ In the second example, silyl triflate (−)-132 was obtained in 81% ee. The absolute stereochemistry was introduced by performing asymmetric allylic alkylation^195−199 of a ketone precursor in collaboration with Stoltz and co-workers, thus obviating the need for preparative chiral chromatography. Nonetheless, cyclic allene generation and trapping provided (+)-134 in 81% ee, reflective of >99% stereoretention.¹⁹² Collectively, these results demonstrate the feasibility of transferring stereochemical information from the silyl triflate to the cycloadduct through an axially chiral cyclic allene, while providing access to enantioenriched polycyclic products.

Figure 22.

Stereospecific cycloadditions of heterocyclic allenes using enantioenriched silyl triflates.

An exciting new approach to control the absolute stereochemistry in reactions of cyclic allenes was recently demonstrated. This strategy was motivated by an interest in cyclic allene racemization, in contrast to the aforementioned approach involving transfer of stereochemical information. A key result that prompted this study is shown in Figure 23. Silyl triflate 122, accessible in 87% ee (via chiral separation of a precursor), was subjected to cyclic allene generation and trapping.¹⁹¹ However, cycloadduct 135 was obtained racemically. This result suggested the importance of substituent effects and that racemization could plausibly outcompete cyclic allene trapping. Interestingly, computational studies suggest that ester-substituted cyclic allene 122 undergoes racemization readily, with an estimated barrier of only 14.1 kcal/mol.¹⁹¹ Thus, efforts were put forth toward developing an asymmetric trapping of cyclic allene intermediates using transition metal catalysis.

Figure 23.

Cyclic allene racemization.

Only one prior example of a transition metal-mediated strained cyclic allene reaction was known in the literature, which involved a cyclotrimerization process.¹⁷⁷ Our laboratory has since developed two variants of asymmetric transition metal-catalyzed annulations of cyclic allenes, as shown in Figure 24. It was found that, under nickel-catalyzed annulation conditions,²⁰⁰ cyclic allene precursor 136 and benzotriazinone 137 afforded tricycle 139 in 85% yield and 94% ee.²⁰¹ More recently, we demonstrated that reaction of silyl triflate 140 and iodopyridine 141 under palladium-catalyzed annulation conditions afforded tricycle 143 in 64% yield and 90% ee.²⁰² These transformations are thought to proceed via the reaction of an in situ generated, racemic cyclic allene and a transient metallocycle formed by oxidative addition. Mechanistic details, including how absolute stereochemistry is introduced, have been proposed for the nickel-catalyzed annulation based on computations.²⁰¹ As such, these methodologies not only provide an advance in strained cyclic allene chemistry but also are some of the few catalytic asymmetric reactions across the larger family of transiently generated cyclic intermediates (e.g., arynes, cyclic alkynes, cyclic allenes).

Figure 24.

Catalytic asymmetric reactions of cyclic allenes.

Outlook and Future Directions

The study of arynes, strained cyclic alkynes, and strained cyclic allenes has seen tremendous growth in recent years. As highlighted in this Perspective, such strained intermediates were contemplated as early as 1902 and validated 50+ years later. In the modern era, these species can be strategically employed in a host of impressive synthetic applications. The increase in practical use can be attributed to a combination of factors, including the mild conditions used for their generation, the commercial availability of silyl triflate precursors,²⁰³ and advances in our ability to understand and predict how strained cyclic intermediates will react in complexity-generating transformations.²⁴⁻²⁶

What does the future hold for arynes, cyclic alkynes, and cyclic allenes? We have shown in this Perspective how these transient species can be strategically leveraged to build complexity in the form of polycyclic scaffolds. Notably, many of newly developed methods discussed herein have potential applications in drug discovery and materials as well as in the synthesis of heterocycles, natural products, and sp³-rich compounds. Moreover, key advances to control absolute stereochemistry using asymmetric catalysis recently emerged and represent a particularly exciting area for future reaction development. As such, we curiously and enthusiastically await further advances from the synthetic community in the chemistry of arynes, strained cyclic alkynes, and strained cyclic allenes.

Author Contributions

^† L.G.W. and M.S.M. contributed equally.

The authors declare no competing financial interest.