Selectivity in the Elaboration of Bicyclic Borazarenes

Abstract

Among aromatic compounds, borazarenes represent a significant class of isosteres in which carbon-carbon bonds have been replaced by B-N bonds. Described herein is a summary of the selective reactions that have been developed for known systems, as well as a summary of computationally-based predictions of selectivities that might be anticipated in reactions of yet unrealized substructures.

Graphical Abstract

1. Introduction

Isosteres provide novel and versatile molecular scaffolds to explore and optimize chemical space for disparate applications. By providing a diverse array of alternative, but electronically and sterically related, chemical motifs, isosteres allow novel molecular platforms to be fine-tuned for specific needs.

Two of the isosteric relationships between carbon-carbon bonds and boron-nitrogen bonds are depicted in Figure 1. Among BN/CC congeners, the sp² version has been exploited the most because of their incorporation into aromatic and heteroaromatic systems, the most abundant and significant class of materials. Perhaps the most predominant subset of such isosteres is exemplified by the 1,2-azaborines.^[1]

Figure 1.

Isosteric relationship between various B-N bonds and C-C bonds.

Although more polar than a carbon-carbon double bond, the B-N motif isosteric to a carbon-carbon double bond is significantly less polar than the B-N bond that is isosteric to a carbon-carbon single bond because the electronegativity of nitrogen is offset to some extent by lone pair-donation to the vacant p orbital on boron, leading to boron and nitrogen bearing similar partial charges (Figure 2). Theoretical calculations, as well as experimental data, have shown that 1,2-azaborines are stable aromatic molecules, whose aromaticity and thermal stability are only slightly lower than that of the corresponding C=C aromatic system (e.g., ~16.1 kcal/mol of resonance stabilization energy in 1,2-azaborine versus 36.6 kcal/mole in benzene).^[2] These substructures thus provide structural mimicry endowed with unique electronic properties and hence distinctive chemical reactivity.

Figure 2.

Electronics of BN/CC isosterism.

Azaborines have fascinated organic chemists for more than sixty years. Although historically studied as a means to understand aromaticity,^[3] boron-based aromatic systems have more recently been recognized as unique substructures for expanding molecular diversity,^[4] and the azaborine isosteric motif has become a powerful tool to increase chemical/structural space in many fields, including the pharmaceutical and agrochemical industry^[5] where there is increasing interest in the use of 1,2-azaborine substructures for such purposes.^[6] In addition to their medicinal chemistry/agrochemical applications, 1,2-azaborines have made a significant impact in materials science owing to their photophysical properties.^[7] Replacing one or more of the carbon atoms in a polycyclic aromatic hydrocarbon (PAH) with boron decreases the HOMO-LUMO gap of the material, providing access to potentially useful molecules with chemiluminescent properties. Azaborines have thus been shown to function as organic light emitting diodes (OLEDs)^{[7a, 7b]} and organic field-effect transistors (OFETs).^[7c–e]

The beauty of 1,2-azaborines lies in their stunning synthetic versatility as a result of the electronic desymmetrization of the heterocyclic ring, providing several modes of highly selective functionalization not accessible to their all-carbon homologues as recently reviewed by McConnell and Liu.^[8] Capitalizing on the ability of the B-N bond to mimic a C=C bond in some ways, yet at the same time alter electron density dramatically throughout the aromatic pi system, 1,2-azaborine cores have been successfully manipulated for diversification in ways not available in the parent system.^[9] For example, in an aromatic hydrocarbon such as naphthalene, a standard electrophilic aromatic substitution can be complex, but generally provides quite reasonable 1-substituted derivatives. No matter how polar this functional group might be, however, it is exceedingly challenging to direct further functionalization selectively at other positions,^[10] and site selectivity gets more and more complex as further substituents are added. By contrast, as outlined herein, the ease with which selective functionalization can be achieved in the borazarenes relative to the all-carbon aromatics allows practitioners to build substantial, highly substituted systems exceedingly efficiently in an exploration of new chemical space. Not only are such reactions more selective, but site selectivity is often complementary to that obtained in the all-carbon systems, with the unique electronic distribution created by the B-N bond overriding the natural electron-donating/electron-withdrawing effects of embedded substituents in subsequent transformations. Thus, even taking into account the commercial availability of the all-carbon aromatic cores, the facility with which borazine systems can be rapidly and selectively functionalized is unmatched in terms of efficiency and diversity.

This review highlights the progress that has been made in exploring the synthesis and selective functionalization of several bicyclic 1,2-azaborines. An emphasis is placed on the use of computational methods as a means to predict site-selectivity in highly functionalized systems, permitting precise substitution at nearly every unique site about the rings. Finally, these same computational tools are used in a predictive manner to envisage those substructures that might provide the most fertile and productive targets for further exploration.

2. 2,1-Borazaronaphthalene

The isosteric replacement of a carbon-carbon double bond with a B-N bond in a naphthalene system affords 2,1-borazaronaphthalenes, which were first prepared by Dewar in 1959 through the reaction of 2-aminostyrene and boron trichloride.^[11] This type of cycloaddition reaction is at the core of almost all reported routes to the 2,1-borazaronaphthalenes. Paetzold later isolated the unstable 2-chloro-2,1-borazaronaphthalene and analogous unstable 2-bromo-2,1-borazaronaphthalene.^[12] After forming the azaborine core, exposure to air results in formation of the corresponding anhydride (Scheme 1).

Scheme 1.

Initial Synthesis of the 2,1-Borazaronaphthalene Core.

The combination of an organodihaloborane and 2-aminostyrene to access these cores was an intriguing disconnection. 2-Aminostyrene can be easily prepared in one-step; however, the lack of commercially available organodihaloboranes limited the ease of preparing these azaborines. Building upon examples in the literature, where air- and moisture-stable potassium organotrifluoroborates were liberated of KF using fluorophiles to yield organodihaloboranes, we envisioned the assembly of a library of 2,1-borazaronaphthalenes from relatively simple starting materials. By employing SiCl₄ as the fluorophile, the azaborine cores were easily constructed with a wide variety of alkyl-, alkenyl-, (hetero)aryl-, and alkynyltrifluoroborates (Scheme 2).^[9a] Borazaroquinoline and borazaroisoquinoline derivatives were likewise prepared using the same method, albeit requiring harsher reaction conditions to promote product formation.^{[9k, 13]}

Scheme 2.

Accessing the 2,1-Borazaronaphthalene Core from Potassium Organotrifluoroborates.

The cycloaddition presumably proceeds via N-B coordination followed by subsequent cyclization with loss of HCl. Paetzold (1968)^[14] and later Pei (2017)^[9n] realized similar cyclizations from anilines, alkynes, and organodihaloboranes (Schemes 3 and 4, respectively). Paetzold reported the intermolecular reaction of the three reagents. Because phenylacetylene is required for this transformation, there will be an aryl substituent at the C4 position of the azaborine. Pei later demonstrated that incorporation of the alkyne on the aniline allows the cyclization to proceed and affords the corresponding 4-chloro-2,1-borazaronaphthalene core.

Scheme 3.

Accessing the 2,1-Borazaronaphthalene Core from Potassium Organotrifluoroborates.

Scheme 4.

Accessing the 2,1-Borazaronaphthalene Core from 2-Alkynylanilines and Dihaloboranes.

Utilizing an alternative approach that does not utilize a cycloaddition reaction to build the azaborine core, Yorimitsu more recently employed a ring expansion of indoles with Li powder followed by reaction with an alkyl- or arylboronate pinacol ester to access 2,1-borazaronaphthalenes (Scheme 5).^[15]

Scheme 5.

Accessing the 2,1-Borazaronaphthalene Core from Indoles.

A comparison of the different published routes to this azaborine core in the context of substitution around the azaborine ring shows that substitution is possible at only some of the core positions of the 2,1-borazaronaphthalene (Table 1). Although substitution may be possible at other positions based on functionalizing precursors for these routes, this table is limited to published examples.

Table 1.

Comparison of the Substitution Patterns in Routes to 2,1-Borazaronaphthalenes

	Scheme	N(1)	B(2)	C3	C4	C5	C6	C7	C8
Molander	2	Yes	Yes	No	No	Yes	Yes	Yes	No
Paetzold	3	No	Yes	No	Yes	No	Yes	No	No
Pei	4	No	Yes	Yes	Yes	No	No	No	No
Yorimitsu	5	Yes	Yes	No	No	No	Yes	No	No

Calculations of selectivity in reactions of 2,1-borazaronaphthalene:

The 2,1-borazaronaphthalene core was examined computationally in parallel to experimental studies to establish the basis for expansion of these methods to alternative azaborine cores. DFT calculations were performed using Gaussian 16,^[16] and WebMO^[17] was used to visualize the structures.

Nuclear independent chemical shift (NICS) calculations were used to gauge the aromaticity of the 2,1-borazaronaphthalenes relative to the C=C isostere. Imaginary (ghost) atoms were placed at distances of 0.0 Å [A(0) and B(0)] and 1 Å [A(1) and B(1)] from the center point of each ring in the system and the absolute middle of the bicyclic system, in a perpendicular direction, prior to performing NMR chemical shift calculations on the system (Figure 3). NICS(1) values are emphasized because of their better predictive power. The chemical shifts of the imaginary atoms were used to probe ring current and hence aromaticity. Ring current can be correlated to aromaticity - greater aromatic character in a system results in greater ring current, which in turn results in shielding being observed in NMR calculations for the ghost atoms that are placed above the ring system.^[18] Therefore, a more negative calculated chemical shift for the ghost atom signals greater aromatic character. Comparison against the corresponding carbon isostere is needed to understand the relative aromaticity, as the absolute value for the chemical shift is not of any significance. Structure optimizations and NMR chemical shift calculations for the ensuing NICS measurement were performed in the gas phase at the B3LYP/6–311+G(2d,p) level of theory.

Figure 3.

NICS Ghost Atom Placement in 2-Phenyl-2,1-borazaronaphthalene.

Probing the aromaticity of 2,1-borazaronaphthalene and its comparison to naphthalene revealed that NICS values in ring A are comparable to that of naphthalene. However, ring B, which contains the B-N bond, showed compromised aromaticity (Figure 4). A less negative NICS(1) value indicates decreased ring current and therefore decreased π-delocalization, indicative of less aromatic character. This trend is observed in all the examined azaborine systems.

Figure 4.

NICS Calculations for 2-Phenyl-2,1-borazaronaphthalene and 2-Phenylnaphthalene.

Because of the desymmetrization of the 2,1-borazaronaphthalene as a result of the B-N bond, site specific electrophilic aromatic substitution is possible. To understand the regioselectivity at a deeper level, we investigated bromination in CH₂Cl₂ as the model reaction, both experimentally and computationally.^[19] Therefore, the energy of the intermediate arenium ion resulting from addition of bromine to different positions was calculated and compared in CH₂Cl₂ using the polarizable continuum model (PCM)^[20] at the B3LYP/6–311+G(2d,p) level of theory to reveal the most accessible site for reaction. Site-selectivity in electrophilic aromatic substitution can be attributed to relative rates of reaction at different positions. Because the formation of the arenium intermediate is an uphill step and is the rate-determining step of the reaction, the relative rate of formation of different arenium intermediates correlates with their stability. It is safe to conclude that the more stable arenium intermediate forms faster, and as a result this will determine the site-selectivity of the reaction.

To probe the regioselectivity of electrophilic aromatic substitution, the energies of various arenium intermediates arising from addition of bromine around the 2,1-borazaronaphthalene core were computed and compared (Figure 5). The relative energies of arenium ions are reported with respect to the most stable intermediate for ease of comparison (see SI for raw computational data). As can be seen, the arenium ion resulting from bromination at C3 is predicted as the lowest energy intermediate, followed by the arenium ion afforded from bromination at C6. Notably, attempts to optimize the structure of the arenium ion resulting from bromination at C4 were unsuccessful, presumably because of the instability of the arenium intermediate. These results are in accord with the experimental data, wherein monobromination of 2,1-borazaronaphthalene was observed selectively at position 3, and upon treatment with excess bromine, 3,6-dibromination was accomplished.^[9b] Furthermore, bromination of 3-substituted-2,1-borazaronaphthalenes yielded site-selective bromination at C6.

Figure 5.

Energies of Arenium Intermediates for Bromination on 2,1-Borazaronaphthalene Core (kcal/mol).

Having an understanding of the regiochemical outcome for electrophilic aromatic substitution, we sought to investigate the anion stability at different positions around the ring system to help predict the site of a C-H borylation.^[21] Given that the transition state for iridium catalyzed C-H activation is suggested to have a proton transfer archetype,^[22] computational studies of all-carbon analogs exhibit significant partial negative charge residing on the carbon undergoing borylation as part of the rate determining step. However, several other factors have been demonstrated to play a role in site-selectivity of Ir-catalyzed borylation, namely borylation site accessibility based on steric constraints.^{[9b, 21c]} Nonetheless, by comparing the energy of deprotonated derivatives in various positions and taking into account the steric factors, the C-H borylation site could be narrowed down to a few plausible positions. The energy of anions resulting from deprotonation at different positions were calculated and compared in the gas phase using self-consistent reaction field model (SCRF) at the B3LYP/6–311+G(2d,p) level of theory.

To probe the predicted site-selectivity of Ir-catalyzed C-H borylation, the structure of anions arising from deprotonation at various positions around 2-phenyl-2,1-borazaronaphthalene ring were calculated to determine which best stabilize negative charge (Table 2). The relative energy of anions is reported with respect to the most stable anionic intermediate for ease of comparison. C8 was demonstrated to be significantly more capable than other positions in anion stabilization. As minimal steric constraints were also predicted for C8, it was anticipated to be the initial site for C-H borylation, which was confirmed with the development of a site specific C-H borylation.^[9m]

Table 2.

Anionic Stability at Carbons in 2-Phenyl-2,1-borazaronaphthalene.

Site	ΔG (kcal/mol)	Site	ΔG (kcal/mol)
C3	0.000	C8	−11.831
C4	−3.506	o-C	−7.236
C5	−5.533	m-C	−0.486
C6	−2.718	p-C	−1.757
C7	−5.407

As we furthered our studies on the azaborines, we sought to realize our goal of preparing highly decorated 2,1-borazaronaphthalenes. Previously, functionalization at nitrogen or other positions on the azaborine ring required substitution before the cycloaddition, limiting the ability to access more diverse structures. To date, the cycloaddition by Pei is the only other method that directly affords a functional handle (halide) on the azaborine core.^[9n] All other functional groups were required to be installed prior to cyclization. As detailed in the following sections, methods now expand to a wide swath of chemical space stemming from the 2,1-borazaronaphthalene core.

A summary of the different methods developed for elaboration of the 2,1-borazaronaphthalene core is shown in Figure 6. For each position, methods that rely on the synthesis of the core are colored red, whereas methods that functionalize the core are colored green. To date, there are no general methods to functionalize the C5 and C7 positions of the azaborine. Similarly, site-selective functionalization of the 1,2-azaborine has been well studied and conforms to the structural and electron isosterism for BN ring systems.^{[1a, 8, 22b, 23]}

Figure 6.

Functionalization Around the 2,1-Borazaronaphthalene Core. (Red – diversity installed prior to core synthesis, Green – diversification of the borazaronaphthalene core directly)

Functionalizing the N1 position of the 2,1-borazaronaphthalene:

In relation to the naphthalene isostere, the reaction of the N-H subunit of the borazine by deprotonation and reaction with an electrophile would be analogous to an electrophilic C-H functionalization. The earliest reports on the synthesis and reactivity of 2,1-borazaronaphthalene demonstrated the ability to deprotonate the N-H bond with methyllithium and then substitute with an electrophile.^[24] We later revisited this substitution reaction by utilizing KHMDS as a base, but the reaction was still limited in substrate scope and required a second addition of KHMDS and electrophile to reach complete conversion (Scheme 6 shows both N substitution reactions).^[25]

Scheme 6.

N-Alkylation of the 2,1-Borazaronaphthalenes.

Substitution can also occur on a pendant alkyl chloride in the case of various nitrogen-based nucleophiles reacting with a 3-chloropropyl-2,1-borazaronaphthalene (Scheme 7).^[24]

Scheme 7.

Pendant Alkylation of 2,1-Borazaronaphthalenes.

The substitution with iodomethylboronate pinacol ester was realized, followed by conversion to the corresponding trifluoroborate, providing an ideal handle to functionalize the N-position through a nickel-catalyzed photoredox coupling with aryl halides (Scheme 8). The generated azaborinyl radical was also subjected to other single electron chemistry to extend substitution past aryl/heteroaryl groups (Scheme 9).^[25]

Scheme 8.

Preparation and Photoredox Coupling of N-Substituted 2,1-Borazaronaphthalenes.

Scheme 9.

Single-Electron Transformations of N-Substituted 2,1-Borazaronaphthalenes.

Functionalizing the B2 position of the 2,1-borazaronaphthalene:

Substitution at boron through substitution of 2-chloro-2,1-borazaronaphthalene with an organometallic reagent was the first reported example of substitution at boron. Dewar^[11] and Paetzold^[12] reported substitution with various aryl Grignard and -lithium reagents, and there is one report of substitution with pyridinyllithium^[7g] (Scheme 10).

Scheme 10.

Aryl Substitution on Boron by Arylmetallic Reagents.

We sought to access a building block by preparing the azaborine with a 2-chloromethyl substituent at boron. This compound was equivalent to a benzyl halide and was subjected to many of the same types of transformations enjoyed by such activated electrophiles. By serving as an electrophile in substitution reactions, the corresponding amines, ethers, and carbonates were prepared. When sodium azide was used, the product was subjected to a subsequent click-coupling to elaborate the structure of the azaborine (Scheme 11).^[9e] Many of these products could not be easily incorporated directly through the cycloaddition reaction used to prepare the azaborine core.

Scheme 11.

Nucleophilic Substitution of 2,1-Chloromethyl-2,1-Borazaronaphthalene.

This building block also provided access to aryl, alkynyl, and alkenyl substitution by serving as the electrophilic partner in palladium-catalyzed cross-coupling reactions (Scheme 12).^[9f] This disconnection obviates the need to prepare the corresponding benzyl-, allyl-, or propargyltrifluoroborate by employing the commercially available potassium aryl-, heteroaryl-, and alkenyltrifluoroborates as well as terminal alkynes in the coupling reaction. Alternatively, the 2-chloromethyl group can be converted to the corresponding trifluoroborate and cross coupled to access the same products.

Scheme 12.

Cross-Coupling Reactions Starting From 2,1-Chloromethyl-2,1-Borazaronaphthalene.

Functionalizing the C3 position of the 2,1-borazaronaphthalene:

The majority of transformations to elaborate the 2,1-borazaronaphthalene core has occurred at the C3 position. As first reported by Dewar^[26] and subsequently by us using milder conditions, site-selective bromination with Br₂ occurs at the C3 position.^[9b] Dewar also reported the site selective chlorination with Cl₂ (Scheme 13).

Scheme 13.

Halogenation of the 2,1-Borazaronaphthalene.

We demonstrated the ability of this azaborinyl bromide to serve as an electrophilic handle for several different transition metal-catalyzed reactions. The preliminary report showcased the ability of this electrophile to couple potassium aryl- and heteroaryltrifluoroborates in high yield with low catalyst loading.^[9b] Interestingly, in a subsequent report, we realized that N-substituted, B-aryl, brominated 2,1-borazaronaphthalenes undergo a self-arylation reaction in which the azaborine serves as both the nucleophile and the electrophile in the cross-coupling to provide access to 2,1-borazaronaphthols, which had previously only been observed in solution (Scheme 14).^[9d]

Scheme 14.

Cross-Coupling of 3-Bromo-2,1-Borazaronaphthalenes.

The 3-bromo-2,1-borazaronaphthalenes were cross-coupled with potassium alkenyltrifluoroborates to provide access to a library of 3-alkenyl-substituted azaborines (Scheme 15).^[9g]

Scheme 15.

Cross-Coupling of alkenyltrifluoroborates with 3-Bromo-2,1-Borazaronaphthalenes.

Alkyl substitution at the C3 position was the next goal. Instead of developing conditions for an alkyl Suzuki coupling with potassium alkyltrifluoroborates, we wanted to demonstrate the stability of this azaborine core to a different type of catalysis, and thus developed a nickel-catalyzed reductive coupling with alkyl iodides, with the main focus of installing non-aromatic, nitrogen-containing heterocycles.^[9c] To expand further the alkyl substitution at the C3 position and develop a milder, more sustainable method, we subjected the 3-bromo-2,1-borazaronaphthalenes to nickel-catalyzed photoredox with ammonium bis(catecholato)silicates to afford the C3-alkylated azaborines in high yield (Scheme 16).^[9i]

Scheme 16.

Alkylation of 3-Bromo-2,1-Borazaronaphthalenes.

Having demonstrated the ability of this azaborinyl bromide to be converted to aryl-, alkenyl-, and alkyl- groups under an array of different reaction conditions, we sought to demonstrate the umpolung reactivity in a cross-coupling reaction. The 3-bromo-2,1-borazaronaphthalenes were easily converted into their corresponding trifluoroborates through a palladium-catalyzed Miyaura borylation with bis(pinacolato)diboron. Both the subsequent azaborinylBpin and azaborinyltrifluoroborate served as nucleophiles in palladium-catalyzed cross couplings with aryl bromides. The azaborinyltrifluoroborates were also converted to the corresponding chloride with tetrachloroisocyanuric acid (TCICA) (Scheme 17).^[27]

Scheme 17.

Borylation and Reactivity of 2,1-Borazaronaphthalenes.

Functionalizing the C4 position of the 2,1-borazaronaphthalene:

To access a functional handle at the C4 position of the azaborine, one must utilize the cycloaddition reported by Pei, which installed a bromide or chloride at the C4 position. Pei reported the palladium-catalyzed cross-coupling of the 4-bromo-2,1-borazaronaphthalene with arylboronic acids (Scheme 18).^[9n]

Scheme 18.

Cross-Coupling of 4-Bromo-2,1-Borazaronaphthalenes.

Functionalizing the C6 position of the 2,1-borazaronaphthalene:

Addition of excess Br₂ during the bromination of 2,1-borazaronaphthalenes led to a second, site-selective bromination at the C6 position. Alternatively, bromination of a 3-substituted-2,1-borazaronaphthalene led to selective bromination at the C6 position. We demonstrated that the 3,6-dibromo- and the 3-substituted-6-bromo-2,1-borazaronaphthalenes could be cross-coupled with potassium aryltrifluoroborates (Scheme 19)^[9b]

Scheme 19.

Cross-Coupling of 6-Bromo-2,1-Borazaronaphthalenes.

Interestingly, the C6 position is also the site for a second selective C-H borylation using iridium and B₂pin₂ as the borylating agent (Scheme 20).^[9m]

Scheme 20.

Dual C-H Borylation of 2,1-Borazaronaphthalenes.

Functionalizing the C8 position of the 2,1-borazaronaphthalene:

Having elaborated the structures of 2,1-borazaronaphthalenes at the C3 and C6 positions through halogenation and subsequent transformation, we developed a method to access C8-substituted azaborines. The 2,1-borazaronaphthalenes undergo a site selective borylation, driven by a confluence of steric and electronic effects. The functionalized azaborines effectively serve as nucleophiles in the palladium-catalyzed cross coupling with aryl bromides (Scheme 21).^[9m]

Scheme 21.

C-H Borylation and Cross-Coupling of 2,1-Borazaronaphthalenes.

3. Borazaronaphthalene Isomers

Other borazaronaphthalene isomers have been explored (Figure 7), though none as extensively as the 2,1-borazaronaphthalene system. Each provides different strategies for synthesis and alternative positions for diversification.

Figure 7.

Isomers of Borazaronaphthalene

1,2-Borazaronaphthalene:

1,2-Borazaronaphthalene affords the regioinverted structure to the 2,1-borazaronaphthalene, originally reported by Cui and coworkers in 2015.^[28] Synthetically, these can be accessed by reacting phenethyl imines, which can isomerize under basic conditions, with haloboranes (Scheme 22). Upon heating with base, the intermediate thus formed undergoes electrophilic cyclization to afford the 1,2-borazaronaphthalene core in a telescoped procedure. Through this route, the core products maintain a boron-halogen bond that can be further derivatized by nucleophilic substitutions. Because derivatization at both positions of the imine precursors can be achieved, along with nucleophilic substitution at boron, this currently affords three points for limited diversification of the 1,2-borazaronaphthalene core. Representative examples of carbonyl olefination with an appended α-amine substituent have been reported,^[29] with substrates cyclizing to provide the 1,2-borazaronaphthalene core in a complementary, albeit less developed, strategy.

Scheme 22.

Preparation and Boron Functionalization of 1,2-Borazaronaphthalene.

1,4-Borazaronaphthalene:

Extending the B-N isosterism through conjugation affords the 1,4-borazaronaphthalene, where the nitrogen delocalizes electron density through the conjugated ring system to the empty orbital on boron. Accessible by a five-step synthesis from 2-bromoanilines, allylation and alkene isomerization is followed by borylation with an appended vinyl substituent to set up a final ring-closing metathesis (Scheme 23).^[30] Although cumbersome for diversification, this route does leave a replaceable group on boron that has been exchanged for various leaving groups and can ultimately be transformed into arylboron substituents with the addition of Grignard- or lithium nucleophiles. This method has afforded a pathway to azaborine-containing ligands for transition metal catalysis.^[31] Subsequently, a single-step transformation has been reported to afford the 1,4-borazaronaphthalene scaffold by condensation of various 2-aminophenylboronic acids and alkyne dicarboxylates (Scheme 24).^[32] Although this enables rapid access to the 1,4-borazaronaphthalene core, it is limited by the remaining dicarbonyl substituents required for the reaction to occur.

Scheme 23.

Ring Closing Metathesis Strategy toward 1,4-Borazaronaphthalene and Subsequent Boron Derivatization.

Scheme 24.

Condensation Strategy toward 1,4-Borazaraonaphthalene.

1,9-Borazaronaphthalene:

1,9-Borazaronaphthalene is one of the least stable borazaronaphthalene isomers as there is minimal delocalization across the two ring systems, as indicated by a high HOMO energy.^[7f] Accessible via a ring-closing metathesis strategy, 2-vinylpyridine can form a coordination complex with allyldichloroborane, which undergoes cyclization and final elimination to afford the parent 1,9-borazaronaphthalene (Scheme 25). Though a few analogues have been reported by changing substituents on the starting vinylpyridine, no functionalization strategies have been reported to date.

Scheme 25.

Synthetic Access to 1,9-Borazaronaphthalene.

9,1-Borazaronaphthalene:

Similarly, 9,1-borazaronaphthalene is also accessible by a route applying a key ring-closing metathesis (Scheme 26).^[23b] With the B-N bond preinstalled in vinyl 1,2-azaborine, accessible from established routes developed for the borazine isostere of benzene, boron substitution with 3-butenyl Grignard reagent sets up the aforementioned ring-closing metathesis. A penultimate dehydrogenation, followed by nitrogen protecting group removal affords the core borazaronaphthalene structure. This core and the homologous 1,9-borazaronaphthalene contains one of the heteroatoms of the azaborine motif at the juncture of the rings. This limits one of the typical sites of functionalization either through substitution of the nitrogen or facile methods of nucleophilic displacement of labile groups at boron.

Scheme 26.

Synthetic Access to 9,1-Borazaronaphthalene.

Fused 9,10-Borazaronaphthalene:

The fused 9,10-borazaronaphthalene is the sole isomer where all the peripheral sites on the ring maintain a C-H bond, so the effect of the B-N isosterism solely functions through electronic perturbation. Strategies to access this core focus around ring-closing metathesis. In the initial reported synthesis,^[33] the divinyl, diallyl B-N complex undergoes an initial metathesis to close one of the rings to a pentacycle (Scheme 27). Upon deprotonation to the 6π intermediate, a methyl homologation affords the 1,2-azaborine intermediate. The same strategy was employed to close the second ring, which affords the fused-borazaronaphthalene core. A shortened route began with diallylamine-dichloroborane, which was alkylated with allylmagnesium bromide to afford the tetraallyl-aminoborane complex (Scheme 28). The bicyclic core was accessed via double ring-closing metathesis, followed by aromatization by either oxidation or dehydrogenation strategies.^[34]

Scheme 27.

Homologation Strategy toward Fused 9.10-Borazaronaphthalene.

Scheme 28.

Double Ring Closing Metathesis Strategy toward Fused 9,10-Borazaronaphthalene.

Functionalization of this core has been reasonably well studied. 9,10-Borazaronaphthalene readily undergoes electrophilic halogenation at the carbon alpha to boron (Scheme 29).^[34a] When the halogenated product was resubjected to the same conditions, site selective halogenation at the other alpha carbon occurred. 1-Iodo-9,10-borazaronaphthalene has been shown to be a suitable electrophile for an array of metal-catalyzed transformations, in particular couplings with alkynes in Sonogashira reactions, electron-rich and electron-poor arylboronic acids in Suzuki reactions, and is a viable substrate in Heck reactions. Additionally, copper-catalyzed phosphorylation of 1-iodo-9,10-borazaronaphthalene has been applied to develop competent ligands for Suzuki reactions.^[35]

Scheme 29.

Halogenation and Subsequent Functionalization of 9,10-Borazaronaphthalene.

Friedel-Crafts acylation is also a viable transformation, with the same selectivity as halogenation (Scheme 30).^[36] Interestingly, when applied to amidation in the presence of silver tetrafluoroborate, the Friedel-Crafts selectivity changes to the C3 position. Copper-mediated nitration of 9,10-borazaronaphthalene also occurs at the C3 position.^[37] This selectivity is suggested to arise from a radical mechanism, and is substantiated by evidence that Lewis acids are inconsequential to reaction progress and that the reaction is shut down with the addition of a radical trapping agent. 9,10-Borazaronaphthalenes have also been demonstrated to undergo site-selective deprotection with an alkyllithium reagent, which can then be trapped by an aldehyde electrophile.^[38]

Scheme 30.

Substitution of Fused 9,10-Borazaronaphthalene.

Fused BN-Indole:

A fused azaborine isostere of indole is also synthetically accessible.^[39] Starting from allyl-diamine precursors, coordination of allylboron dichloride to form a five-membered ring sets up a ring closing metathesis to form the 6–5 bicyclic scaffold (Scheme 31). A penultimate palladium/carbon dehydrogenation, followed by protecting group removal, affords the BN-indole azaborine structure. The N-H is labile under basic conditions, with an experimentally derived pK_a roughly around 30 in DMSO, and can be alkylated under deprotonating conditions.^[40] This core has been demonstrated to undergo electrophilic substitution at the C3 position, matching the reactivity of indole.^[41]

Scheme 31.

Synthetic Pathway to the Fused BN Indole Azaborine.

4. Computationally-Based Predictions of Selectivity in Azaborines

As elucidated earlier, computational strategies can be developed to probe and make predictions concerning the aromaticity and relative reactivity of the azaborines.^[42] In the 2,1-borazaronaphthalene system, the experimental evidence was corroborated with the results of the computational study in explaining the site selectivity of the electrophilic aromatic substitution and C-H borylation. We therefore sought to apply these methods to assess the aromatic character and potential reactivity of novel azaborine cores. Certain azaborine cores were selected based on the prevalence of their carbon isosteres in biologically active compounds.^[43] Several 6–6 and 6–5 bicyclic systems are herein explored, as well as one tricyclic system, carbazole.

Aromaticity:

NICS calculations were used to quantify the aromatic character of novel azaborine systems computationally. NICS(1) values are emphasized, owing to their greater ability to show ring current from π-delocalization.^[18b] Consistent patterns arise in comparing NICS values for azaborines and their carbon isosteres. Each B-N isostere explored demonstrated decreased overall ring current and π-delocalization compared to its corresponding C-C analogue. Specifically, rings containing the B-N bond consistently have weaker ring current, and thereby less π-delocalization, based on the proclivity of boron and nitrogen to hold some part of their sp³ character.

Previously synthesized azaborines isosteric to naphthalene were explored. 9,10- and 9,1-Borazaronaphthalene are both isosteric to unsubstituted naphthalene (Figure 8). Because the parent naphthalene is symmetric, a consistent ring current is observed in both A and B rings. Both 9,10- and 9,1-borazaronaphthalene have relatively less negative NICS(1) values for rings A and B, as well as the center of the system, which indicates decreased aromatic character. 9,10-Borazaronaphthalene is symmetric and shows similarly weaker ring currents in both A and B rings. By contrast, 9,1-borazaronaphthalene has a weaker ring current in its B ring compared to its A ring.

Figure 8.

NICS Calculations for 9,10- and 9,1-Borazaronaphthalene, Isosteres of Naphthalene.

Other azaborine systems were explored with phenyl substituents on boron, owing to the increased stability of such systems. NICS calculations of 1,2-, 1,4-, and 1,9-borazaronaphthalene, which are isosteric to 1-phenylnaphthalene, are shown in Figure 9. The 1,2-borazaronaphthalene is overall the most delocalized as indicated by its most negative center value. The A ring of 1,2-borazaronaphthalene matches the parent naphthalene almost exactly in NICS value, indicating that it has similar aromatic character. In contrast, the 1,4- and 1,9-systems show the A ring contains less aromatic character than the parent naphthalene. 1,4-Borazaronaphthalene demonstrates similar aromatic character to 1,2-borazaronaphthalene, whereas 1,9-borazaronaphthalene is significantly less aromatic overall, as indicated by its center NICS value.

Figure 9.

NICS Calculations for 1,2-, 1,4-, and 1,9-borazaronaphthalene, Isosteres of 1-Phenylnaphthalene.

We extended our computational study to include novel azaborines. The isostere of 7-phenylquinoline, 2,1,8-borazaroquinoline, was next explored. 7-Phenylquinoline shows less aromatic character than 2-phenylnaphthalene in both rings and overall NICS(1) values. The 2,1,8-borazaroquinoline isostere shows less aromatic character overall and in both rings compared to its carbon analogue. The B ring is markedly less π-delocalized, while the A ring shows a slight reduction in delocalization (Figure 10).

Figure 10.

NICS Calculations for 2,1,8-borazaroquinoline and 7-phenylquinoline.

Four 6–5 bicyclic systems of biological importance were evaluated.^[43–44] 2,1-Borazarobenzimidazole is shown by NICS(1) calculations to have less aromatic character in the A ring (now the ring containing a B-N bond) than the analogous ring in benzimidazole. Interestingly, the imidazole ring (B) of 2,1-borazarobenzimidazole showed slightly greater delocalization than that of benzimidazole. Overall, the 2,1-borazarobenzimidazole shows weaker ring current, demonstrated by a lower central NICS(1) value (Figure 11).

Figure 11.

NICS Calculations for 2,1-Borazarobenzimidazole and its Isostere 6-Phenylbenzimidazole.

Three isosteres of indole were explored: a 2,1,5-borazaroindole, 2,1,7-borazaroindole, and the “fused” BN indole prepared by Liu and co-workers.^{[39, 41]} Each was compared to a carbon analogue for aromatic character (Figure 12). The results are consistent with trends in the 6–6 systems; the azaborine isosteres show decreased aromatic character relative to the indole analogues. The A ring is affected to a greater degree by the presence of the B-N bond in the 2,1,5- and 2,1,7-borazaroindoles. The “fused” BN-indole shows both rings significantly lowered in aromatic character relative to indole, akin to the 9,10-borazaronaphthalene.

Figure 12.

NICS Calculations of Isosteres of Indoles.

Isosteres of 5- and 6-phenylbenzofuran, 2,1,5-borazarobenzofuran and 2,1,7-borazarobenzofuran, respectively, were probed. Both isosteres show less aromatic character than their carbon analogues. The 2,1,5- isomer shows slightly greater aromatic character than its 2,1,7-counterpart, in both A and B rings as well as overall (Figure 13). Similarly, isosteres of 5- and 6-phenylbenzothiophene, 2,1,5-borazarobenzothiophene, and 2,1,7-borazarobenzothiophene, proved to yield similar results. In general, benzothiophenes and borazarobenzothiophenes show more aromatic character than the benzofuran counterparts (Figure 14).

Figure 13.

NICS Calculations of Isosteric Benzofurans.

Figure 14.

NICS Calculations of Isosteric Benzothiophenes.

One tricyclic system, carbazole, was probed for aromatic character in an azaborine isostere. NICS calculations were carried out for ghost atoms in the center of each ring, as well as the centers of rings A/B and B/C. The results are reported in Figure 15. Ring A is consistent between carbazole and its isostere, although 2,1-borazarocarbazole shows slightly less aromatic character there. Ring B, however, shows greater aromatic character in 2,1-borazarocarbazole than the parent carbazole itself. Ring C follows the consistent trend observed in other azaborines: the isosteric ring is significantly less aromatic in character than the carbon analogue.

Figure 15.

NICS Calculations of 4-Phenylcarbazole and 2,1-Borazarocarbazole.

Electrophilic Aromatic Substitution:

As detailed above, to understand the site selectivity in electrophilic aromatic substitutions, we probed the 2,1-borazaronaphthalene system computationally and observed that the free energies of the brominated arenium intermediates provided a quantitative means to compare potential reaction pathways. Computed relative free energies (kcal/mol) of arenium intermediates were taken at the sites with the greatest electron density for the HOMO. The difference in energy, “ΔΔG”, between the two lowest-energy intermediates could be used to gauge selectivity. In most of our computationally explored cases, borazarenes are predicted to have greater selectivity for monobromination than their corresponding carbon isosteres. However, certain isomeric azaborines are predicted to be markedly less regioselective for bromination. Furthermore, the placement of the B-N bond greatly influences the predicted position of bromination and selectivity, which is indicative of the potential for isomeric isosteres to enable access to different positions on any core. In some cases, the differences in free energy indicate that a second iterative bromination could be selective as typified in the 2,1-borazaronaphthalene.^[9b] Data are reported as ΔG in kcal/mol relative to the highest energy arenium ion intermediate for a given ring system.

Calculations of different isomeric borazaronaphthalenes show that regioselectivity is strongly influenced by the particular class of azaborine (Figure 16). Each isomer is predicted to be regioselective for monobromination. In every case, the lowest energy Wheland intermediate^[45] is favored by more than 3.0 kcal/mol. The 1,4-borazaronaphthalene has the greatest difference between the two lowest-energy intermediates, while 1,2-borazaronaphthalene has the least difference. Notably, the second and third-lowest energy intermediates of 1,2-borazaronaphthalene are within 0.7 kcal/mol of each other, so a second selective bromination may not be feasible. It is noteworthy that almost every position on the borazaronaphthalene core is selectively accessible by electrophilic bromination, depending on the isomer of choice.

Figure 16.

Predicted Reactivity toward Electrophilic Bromination for Isomeric Borazaronaphthalenes.

The novel 2,1,8-borazaroquinoline is predicted to show complementary reactivity to quinoline (Figure 17). The C3 position of 2,1,8-borazaroquinoline is calculated to be favored over the C6 position by 6.99 kcal/mol, similar to that seen in 2,1-borazaronaphthalene.

Figure 17.

Predicted Reactivity toward Electrophilic Bromination for Quinoline and 2,1,8-Borazaroquinoline.

2,1-Borazarobenzimidazole is also shown to be potentially selective for the C3 position. This is the equivalent position to where benzimidazole has its most stable arenium intermediate (Figure 18). However, 2,1-borazarobenzimidazole is predicted to be more selective based on the difference between the two lowest-energy intermediates.

Figure 18.

Predicted Reactivity toward Electrophilic Bromination for Benzimidazole and 2,1-Borazarobenzimidazole.

Two novel borazaroindoles and the “fused” BN indole were evaluated similarly (Figure 19). Computations of the “fused” BN indole match the findings of Liu and coworkers^[41a] in their investigation: monobromination is most favorable at the C3 position. The 2,1,5-borazaroindole is predicted to be highly selective for the 3-position, followed by the C6 position. However, calculations show that arenium intermediates at the 3-position and 5-position are within 1.5 kcal/mol of each other, indicating that monobromination at 2,1,7-borazaroindole may not be selective.

Figure 19.

Predicted Reactivity toward Electrophilic Bromination for Indole, the “Fused” BN Indole, and Two Novel Azaborine Isosteres.

This result is mirrored in isomers of borazarobenzofuran (Figure 20) and borazarobenzothiophene (Figure 21). In both cases, the 2,1,5-isomer is predicted to be selective, whereas the 2,1,7-isomer is calculated to be less so. The 2,1,7-borazarobenzofuran is predicted to be selective at the C6 position by 4.48 kcal/mol, but the 2,1,7-borazarobenzothiophene yielded results similar to the 2,1,7-borazaroindole.

Figure 20.

Predicted Reactivity toward Electrophilic Bromination for Benzofuran and Two Novel Azaborine Isosteres.

Figure 21.

Predicted Reactivity toward Electrophilic Bromination for Benzothiophene and Two Novel Azaborine Isosteres.

The tricyclic 2,1-borazarocarbazole (Figure 22) is predicted to be very selective for monobromination, but of poor selectivity for a second bromination.

Figure 22.

Predicted Reactivity toward Electrophilic Bromination for Carbazole and 2,1-Borazarocarbazole.

C-H Functionalization:

As in the case of the C-H borylation of 2,1-borazaronaphthalene, we can calculate the site of greatest anionic stability as a potential site of reaction selectivity, given that steric constraints are satisfied.^{[9m, 21a, 21c]} As expected, the position most favored for borylation is influenced strongly by proximity to heteroatoms, and can also be swayed by steric factors and substitution patterns. Energies are reported as ΔG in kcal/mol relative to the least stable carbanion around the ring. The size of differences (“ΔΔG”) between positions correlate with predicted selectivity. Many novel cores were explored with a phenyl substituent at boron, and energies for anions at the ortho, meta, and para positions of the phenyl substituent were computed. In most cases, these positions were assessed to be electronically unfavorable, in addition to the known steric factors preventing borylation at the ortho and meta sites.

Exploration of isomeric borazaronaphthalenes shows that most are expected to be electronically selective for reaction at the carbon closest to the nitrogen heteroatom (Figure 23). This is consistent with computations and experimental findings for 2,1-borazaronaphthalene (vide supra). The 9,1-, 9,10-, and 1,2-borazaronaphthalenes are calculated to have highly differentiated anionic stability as indicated by the gap in ΔG of the most stable anions (in green). With minimal steric interference at those predicted sites, selective C-H borylation seems likely. The 1,9- and 1,4- isomers are not expected to show much selectivity. The C8 position is calculated to be favored on 1,9-borazaronaphthalene. However, standard substitution on boron will sterically impose the substitution, suggesting borylation may be nonselective; the C3 position is minimally favored on 1,4-borazaronaphthalene. In addition, the 2,1,8-borazaroquinoline was explored; the anion at the C5 position is projected to be favored by a computed ~4.3 kcal/mol (Figure 24). The introduction of another heteroatom into the system complicates the hierarchy of anionic stability, but still yields a potentially selective point of reaction.

Figure 23.

Most Stable Carbanions at Various Positions on Isomeric Borazaronaphthalenes. Reported as ΔG (kcal/mol) Relative to the Highest Energy Carbanion.

Figure 24.

Most Stable Carbanions at Various Positions on 2,1,8-Borazaroquinoline. Reported as ΔG (kcal/mol) Relative to the Highest Energy Carbanion.

Results from the 6–5 heterocycles illustrate the complex effects of multiple heteroatoms in governing anionic stability. The most stable anion of 2,1-borazarobenzimidazole is predicted to be at the C6 position on the imidazole ring (Figure 25). It is calculated to be favored by ~11.8 kcal/mol over the C4 position.

Figure 25.

Most Stable Carbanions at Various Positions on 2,1-Borazarobenzimidazole. Reported as ΔG (kcal/mol) Relative to the Highest Energy Carbanion.

Three isomers of indole were evaluated. The “fused” BN-indole presents three likely sites of borylation with similar anionic stabilities (Figure 26). Interestingly, the novel 2,1,5-borazaroindole is predicted to be less selective than the 2,1,7-borazaroindole, demonstrated by the gap in ΔG. However, both are projected to be selective for the C6 position.

Figure 26.

Most Stable Carbanions at Various Positions for Isosteres of Indole. Reported as ΔG (kcal/mol) Relative to the Highest Energy Carbanion.

Isosteres of benzofuran (Figure 27) and benzothiophene (Figure 28) show a similar pattern: the 2,1,5-isomer is projected to be less selective than the 2,1,7-isomer. Furthermore, most stable anions are calculated to be closer in energy than that of indole, more so for borazarobenzofuran than borazarobenzothiophene. The placement of the B-N bond seems to influence the electronic factors governing selectivity for Ir-catalyzed borylation in 6–5 heterocycles.

Figure 27.

Most Stable Carbanions at Various Positions for Isosteres of Benzofuran. Reported as ΔG (kcal/mol) Relative to the Highest Energy Carbanion.

Figure 28.

Most Stable Carbanions at Various Positions for Isosteres of Benzothiophene. Reported as ΔG (kcal/mol) Relative to the Highest Energy Carbanion.

Lastly, 2,1-borazarocarbazole was evaluated. Many carbon centers on the ring were calculated to provide similar anionic stabilities (Figure 29). Steric factors may influence borylation to be site-selective, but electronic factors point to the likelihood of non-selective borylation.

Figure 29.

Most Stable Carbanions at Various Positions on 2,1-Borazarocarbazole. Reported as ΔG (kcal/mol) Relative to the Highest Energy Carbanion.

5. Conclusion

The azaborines represent a readily accessible group of isosteres possessing extraordinary potential for highly selective elaboration. Although detailed synthetic studies have been carried out for a few systems, the vast majority of the more important substructures remain completely unexplored. The research carried out to date indicates that not only can more highly selective reactions be carried out on these systems than on their carbon-based isosteres, but often complementary reactivity patterns are observed owing to the unique electronic desymmetrization of the B-N ring platforms. Elaboration of the borazarenes thus serves as a highly useful and convenient way to access new chemical space that would be highly challenging or tedious to achieve in the carbon-carbon congeners.

The data presented herein demonstrates many effective means to functionalize select borazarenes, and also provides a road-map for future research on novel substructures. Thus, computational methods that have proven effective in predicting reactivity and selectivity patterns for known systems point to the most promising leads for future endeavours, providing practitioners with a template for success.

An area that has not been addressed extensively to date is that of the stability and suitability of the borazarenes for applications in biological systems. Thus, the absorption, distribution, metabolism, and excretion (ADME) properties of these core structures are vastly understudied. Studies that have been conducted to date indicate that some substitution patterns render the borazarenes highly stable and thus suitable for evaluation in vitro and in vivo, but other systems are highly susceptible to decomposition even under mild, aqueous conditions. In such instances, the borazarenes may afford entry to initial hits because of their easy accessibility, but more advanced studies may require a switch to the carbon-carbon congeners for lead compounds and beyond.

In summary, what began as an interest in fundamental theoretical/physical properties such as aromaticity has become a field that is increasingly viewed as being of interest in a variety of domains in the pharmaceutical, agrochemical, and electronics industries. It is also evident that the surface has barely been scratched in terms of the understanding of the reactivity and stability of the vast majority of possible substructures, but the future appears quite bright for potential applications.

Biographies

Ayan Bhattacharjee is an undergraduate at the University of Pennsylvania in the College of Arts & Sciences, studying Chemistry. He is currently a research assistant in the group of Professor Gary Molander.

Geraint Davies completed his undergraduate degree at Emory University before performing his graduate studies at the University of Pennsylvania under the tutelage of Gary Molander. Upon earning his PhD in 2017, Geraint joined Celgene as a medicinal chemist. Upon their merger with Bristol Myers Squibb, he is currently a researcher within Small Molecule Drug Development.

Borna Saeednia graduated from Sharif University of Technology in 2017 where he received a bachelor’s degree in chemistry under the supervision of Prof. Matloubi-Moghaddam. He then started his graduate studies at the University of Pennsylvania where he is currently a PhD candidate in the research laboratory of Prof. Molander.

Steve Wisniewski completed his undergraduate studies at The College of New Jersey and his graduate studies at the University of Pennsylvania under the direction of Gary Molander. After earning his PhD in 2014, Steve joined the process group within Chemical Process Development at Bristol Myers Squibb, where he is currently a Senior Research Investigator.

Gary Molander completed his undergraduate studies in chemistry at Iowa State University under the tutelage of Professor Richard C. Larock. He earned his Ph.D. at Purdue University with Professor Herbert C. Brown and undertook postdoctoral training with Professor Barry M. Trost at the University of Wisconsin, Madison. He began his academic career at the University of Colorado, Boulder, moving to the University of Pennsylvania in 1999, where he is currently the Hirschmann–Makineni Professor of Chemistry.