Strategies in the synthesis of dibenzo[b,f]heteropines

Introduction

The dibenzo[b,f]azepine (1a) scaffold (Figure 1) is featured in commercial pharmaceuticals and other lead compounds , ligands and in materials science with possible applications in organic light emitting diodes (OLEDs) and dye-sensitized solar cells (DSSCs) .

Figure 1: Dibenzo[b,f]azepine (1a), -oxepine (1b) and -thiepine (1c) as examples of dibenzo[b,f]heteropines (1) with the corresponding 10,11-dihydro derivatives (2).

Commercial pharmaceutical agents based on dibenzo[b,f]azepine (1a), or the 10,11-dihydro derivative thereof (2a), include imipramine (3) and clomipramine (4) (tricyclic antidepressants) , opipramol (5) (generalized anxiety disorder) and carbamazepine (6) (seizure disorders) (Figure 2).

Figure 2: Selected pharmaceuticals with the dibenzo[b,f]azepine skeleton.

10,11-Dihydrodibenzo[b,f]azepine-based ligand 7 and a methyl analogue thereof are known to form pincer complexes with Pd, Ir, Rh and Ln , whereas a copper(II) wagon wheel complex of 8 was reported in a molecular organic framework (MOF) (Figure 3) .

Figure 3: Examples of 10,11-dihydrodibenzo[b,f]azepine-based ligands.

4,4'-(5-(Pyridin-2-yl)-10,11-dihydro-5H-dibenzo[b,f]azepine-2,8-diyl)bis(N,N-diphenylaniline) (9) exhibits properties suitable for the use in organic light emitting diodes whereas dyes 10–12 were found suitable for the use in dye-sensitised solar cells (Figure 4) .

Figure 4: The dibenzo[b,f]azepine moiety in dyes with properties suitable for the use in organic light emitting diodes and dye-sensitised solar cells.

Though analogous dibenzo[b,f]oxepines 1b, with an oxygen in the heterocyclic ring as opposed to nitrogen in azepines, are known from natural sources (compounds 13–18 as examples) , the application thereof in a clinical setting is limited (Figure 5). Novartis CGP 3466 (19), a propargylamine derivative, showed excellent neuroprotective properties for the treatment of Parkinson’s disease in rat models (Figure 5) . Unfortunately, the promising preclinical studies of 19 could not be replicated in human trials .

Figure 5: Selective bioactive natural products (13–18) containing the dibenzo[b,f]oxepine scaffold and Novartis CGP 3466 (19).

For more details on the early synthesis of dibenzo[b,f]azepine (1a), the extensive review prepared by Kricka and Ledwith in 1974, is recommended. While the review is lacking modern metal catalysis, it is still an excellent work covering early syntheses and properties. An analogous review published by Olivera et al. covers the topic of dibenzo[b,f]oxepines (1b) up to 2002.

Other heteroatoms (e.g., O, N, S, P, B and Si) in the heterocyclic ring result in analogues of dibenzo[b,f]azepines and -oxepines. This group of compounds will thus be broadly referred to as dibenzo[b,f]heteropines (1).

The first section of this review will cover the synthesis of dibenzo[b,f]heteropines (1) and 10,11-dihydrodibenzo[b,f]heteropines (2). The following section will briefly touch on functionalisation of the scaffold.

While some reports are limited to the introduction of a single heteroatom, e.g., nitrogen in the case of azepines 1a or oxygen in the case of oxepines 1b, some approaches allow for the incorporation of a diverse scope of heteroatoms (e.g., O, N, S, P, B and Si) and may give access to a range of dibenzo[b,f]heteropines 1 using common intermediates . Therefore, this section will be broadly organised by reaction type responsible for ring closure.

Review 1 Industrial route to 5H-dibenzo[b,f]azepine (1a)

10,11-Dihydro-5H-dibenzo[b,f]azepine (2a), also known as iminobibenzyl (2a), is used as precursor for several compounds, including 5H-dibenzo[b,f]azepine (iminostilbene) (1a), and therefore will be discussed in a section on large scale industrial synthesis.

While extensive patent literature documenting methods exists, it is difficult to find accurate, up to date information regarding the industrial synthesis of 5H-dibenzo[b,f]azepine (1a) and derivatives. The following strategy (Scheme 1) was noted by chemists at Novartis as standard in 2005 .

Oxidative coupling of o-nitrotoluene (22) Reduction to 2,2'-diaminobibenzyl (20) Ring-closing via amine condensation Catalytic dehydrogenation

Scheme 1: Retrosynthetic approach to 5H-dibenzo[b,f]azepine (1a) from nitrotoluene (22).

1.1 Oxidative coupling of o-nitrotoluene (22) and reduction to 2,2'-diaminobibenzyl (20)

The preparation of dinitrobibenzyl (21) can be achieved by the oxidative coupling of nitrotoluene (22) under alkaline conditions (e.g., O2, KOt-Bu; O2, KOH, MeOH, ethylenediamine, etc.), as reported by Stansbury and Proops . Aerobic oxidation of 22 in alkaline methanol with added ethylenediamine, gave 21 in 36% yield (Scheme 2), which is poor compared to that reported for the p-nitro derivative (75%). Moormann, Langbehn and Herges recently optimized the method by the introduction of Br2 as oxidizing agent (t-BuOK, Br2, THF) to give the desired 2,2'-dinitrobibenzyl (21) in 95% yield.

Scheme 2: Oxidative coupling of o-nitrotoluene (22) and reduction of 2,2'-dinitrobibenzyl (21) to form 2,2'-diaminobibenzyl (20).

In a method patented in 1987 , 22 is coupled oxidatively in the presence of a variety of transition metal (Ni, Fe, V) porphyrin catalysts and oxygen. Catalytic reduction (H2, Pd/C) affords 2,2'-diaminobibenzyl (20) in the subsequent step .

1.2 Ring-closing via amine condensation

The initial synthesis of 10,11-dihydro-5H-dibenzo[b,f]azepine (2a) was reported in 1899 by Thiele and Holzinger via the polyphosphoric acid (PPA) catalysed cyclisation of 2,2'-diaminobibenzyl (20) at elevated temperatures (Scheme 3) .

Scheme 3: Synthesis of 10,11-dihydro-5H-dibenzo[b,f]azepine (2a) via amine condensation.

1.3 Catalytic dehydrogenation

An early synthesis of 5H-dibenzo[b,f]azepine (1a) involved the gas phase dehydrogenation of 10,11-dihydro-5H-dibenzo[b,f]azepine (2a) to 1a in poor yield (20–50%) . The starting material 2a was distilled through a heated (≈150 °C) column packed with Pd/C and glass wool. Crude 1a was collected as a solid and purified. Further research has been conducted on the effect of catalyst choice and composition for large scale synthesis. Knell et al. reported a comparison of several catalysts, which included potassium-promoted iron, cobalt and manganese oxide catalysts, for the synthesis of 1a. Industrially, 1a is produced by the vapour phase dehydration of 2a over an iron/potassium/chromium catalyst system (Scheme 4) .

Scheme 4: Catalytic reduction of 10,11-dihydro-5H-dibenzo[b,f]azepine (2a).

2 Ring expansion through rearrangement

Several methods utilise ring expansion to prepare the required 7-membered azepine and oxepine rings of 1a and 1b.

2.1 Ring expansion of acridin-9-ylmethanols

In 1960, Bergmann and Rabinovitz reported a simple ring expansion of acridin-9-ylmethanol (23) to 1a in good yield (80%) by heating 23 in polyphosphoric acid (Scheme 5).

Scheme 5: The Wagner–Meerwein rearrangement of acridin-9-ylmethanol (23) into 5H-dibenzo[b,f]azepine (1a).

Independently, in an effort to synthesise phenothiazine isosteres, Craig et al. prepared 1a via a Wagner–Meerwein rearrangement of 23 with P2O5 (Scheme 5) the following year. The method was used to successfully synthesise unsubstituted as well as chloro-substituted derivatives of 1a. Storz et al. have reported on an analogous method to prepare dibenzo[b,f]oxepines 1b through the rearrangement of 9-(α-hydroxyalkyl)xanthenes.

2.2 Ring expansion of 2-(9-xanthenyl)malonates

Oxidative ring expansion of 2-(9-xanthenyl)malonates 24 was reported by Cong et al. as a method for the synthesis of substituted dibenzo[b,f]oxepines 25 (Scheme 6). Treatment of the malonate derivative 24 with Mn(OAc)3 in 90% acetic acid gave C-10 carboxylate derivatives of dibenzo[b,f]oxepine 25. The authors proposed a one-electron oxidation of the enol carboxylate and subsequent 1,2 radical rearrangement and decarboxylation. Moderate to good yields of dibenzo[b,f]oxepine carboxylates 25 were achieved (63–85%).

Scheme 6: Oxidative ring expansion of 2-(9-xanthenyl)malonates 24.

Stopka et al. reported on tandem C–H functionalisation and ring expansion as an alternative to the Wagner–Meerwin rearrangement (Scheme 7). Several azepine 30 and oxepine 31 examples were prepared in good yield from the respective acridane (26) and xanthene (27) derivatives. As an alternative to the thermal Wagner–Meerwin rearrangement (Scheme 5 and Scheme 6), which requires elevated temperatures, Stopka et al. used mild copper-catalysed oxidative conditions to effect the transformation to 30 and 31.

Scheme 7: Ring expansion via C–H functionalisation.

2.3 Ring expansion from N-arylisatins

Elliott et al. reported the four-step synthesis of fluorinated 5H-dibenzo[b,f]azepine 38 from N-arylisatin 34 via Wagner–Meerwein rearrangement of 9-acridinemethanol 37 (Scheme 8).

Scheme 8: The synthesis of fluorinated 5H-dibenzo[b,f]azepine 38 from isatin (32).

2.4 Ring expansion of N-arylindoles (41)

The polyphosphoric acid (PPA)-catalysed rearrangement of N-arylindoles 41 was first reported by Tokmakov and Grandberg . The reaction provided moderate yields with a simple 2 step linear sequence from indole 39. The reaction requires heating at elevated temperatures and reaction times of up to 150 hours. The electronic properties of the rings have a significant influence, with the strongly electron-withdrawing groups, p-NO2 and m-CF3, preventing the rearrangement and electron-donating groups (e.g., p-OMe, -CH3) promoting the rearrangement. The authors postulated an intramolecular electrophilic substitution via a carbocation intermediate 42 (Scheme 9).

Scheme 9: The synthesis of substituted dibenzo[b,f]azepines 43 from indoles 39.

Elliott et al. investigated several methods to synthesise substituted dibenzo[b,f]azepines, which included the ring expansion of N-arylindoles 41 to synthesise 43 and the rearrangements of 9-acridine methanol 37 (Scheme 8) and N-arylindoles 41 (Scheme 9). The authors reported an excellent two-step synthesis of substituted dibenzo[b,f]azepines 43 via commercially available substituted indole 39 precursors based on the method of Tokmakov and Grandberg . N-Arylindoles 41 were successfully synthesised via a copper-catalysed Ullman-type coupling or a palladium-catalysed Buchwald–Hartwig amination (Scheme 9). Performing the rearrangement at high temperatures resulted in the undesirable formation of acridine byproducts 44. Cleaner reaction profiles could be obtained at a lower temperature (100 °C). In contrast to the effect reported for NO2 and CF3 substituents by Tokmakov and Grandberg , electron-withdrawing halogen substituents on the aryl ring did not prevent rearrangement to dibenzo[b,f]azepine 43 . The isolated yield of unsubstituted 43 was good (67%), however, substitution resulted in a decreased yield. While fluoro groups were well tolerated, a major drawback of the method is the acid-catalysed dehalogenation of chloro- and bromo-substituted dibenzo[b,f]azepines. The brominated analogue was only isolated in 5% yield, compared to 67% for the unsubstituted 43. In addition, several methods of carboxamidation were tested, thus allowing the authors to synthesize carbamazepine (CBZ) derivatives of 43.

3 Metal-catalysed cyclisation

Diverse metal-catalysed coupling methods exist for the preparation of the dibenzo[b,f]heteropine ring system. The following approaches are broadly categorised according to the major or final catalytic step employed to form the 7-membered heterocycle as several synthetic methods use multiple catalytic steps.

3.1 Buchwald–Hartwig amination, etherification and thioetherification

The Buchwald–Hartwig reaction gives access to arylamines, -ethers and thioethers from aryl halides and triflates through palladium catalysis . Scheme 10 provides a retrosynthesis of amination in the synthesis of dibenzo[b,f]azepine 45 as an example.

Scheme 10: Retrosynthetic pathways to dibenzo[b,f]azepines via Buchwald–Hartwig amination.

Arnold et al. reported an excellent method for the convergent synthesis of variable sized dibenzo-fused heterocycles. Among these, Heck reaction conditions allowed for the coupling of aryl acrylates 50 to aryl halides 48 and 49, followed by intramolecular Pd-catalysed amination or etherification to give C-10 carboxylates of dibenzo[b,f]azepine 55 and dibenz[b,f]oxepine 54 in good yield (Scheme 11). However, no ring-substituted derivatives were reported. The authors used alpha-substituted acrylates to reduce the effect of poor endo/exo regioselectivity in the intramolecular Heck reaction (cf. Scheme 19).

Scheme 11: Synthesis of dibenzo[b,f]oxepine 54 and -azepine 55 derivatives via (i) Heck reaction and (ii) Buchwald–Hartwig-type etherification or amination.

Božinović et al. reported the synthesis of symmetrical 5H-dipyridoazepines 60a and unsymmetrical 5H-pyridobenzazepines 60b via cyclisation of 2,2'-dihalostilbene analogue 58 through a Pd-catalysed double Buchwald–Hartwig amination. The stilbene analogues 58 were prepared by a Wittig reaction with reported yields of the desired Z-isomer around 55%. The amination step was performed on a series of primary alkylamines (RNH2) with moderate to good yields (47–87%). The strategy was also successfully applied to the synthesis of thiepines 59 with moderate yield (49–51%, Scheme 12).

Scheme 12: Double Buchwald–Hartwig amination and thioetherification in the synthesis of tricyclic azepines 60 and thiepines 52.

Zhang et al. applied a Buchwald–Hartwig amination in 2012 to assemble substituted dibenzo[b,f]azepines 62. The reaction pathway includes the synthesis of intermediate stilbenes 61 by Wittig coupling. The authors elected to use a Pd2dba3/DPEphos (L4)/Cs2CO3 system (dba = dibenzylideneacetone; DPEphos = bis[(2-diphenylphosphino)phenyl] ether) in toluene after catalyst and ligand screening. Cyclisation of several substituted 2,2'-dibromostilbenes 61 by means of a double Buchwald–Hartwig amination gave yields between 62% and 96% using aniline as the amine reactant (Scheme 13). The reaction proved to be compatible with both aromatic and aliphatic amines and the reaction time varied between 11 and 24 hours. Fluoro, chloro, nitrile, alkyl, and methyl ether aromatic substituents were tolerated.

Scheme 13: Double Buchwald–Hartwig amination towards substituted dibenzoazepines 62.

Unsymmetrical 10,11-dihydro-5H-dibenzo[b,f]azepine derivatives 71 have been synthesised by ortho-bromination of functionalised dihydrostilbenes 67, followed by intramolecular cyclisation using Buchwald–Hartwig amination (Scheme 14) . The pathway relies on a double Sonogashira coupling [(i) and (iii)], reduction (iv), and bromination (v), followed by Buchwald–Hartwig amination (viii) (Scheme 14). While interesting, the reaction has limited substrate scope due to the reliance on a late-stage bromination. To achieve the correct ortho-bromo substitution pattern, it requires a para-substituted ester as a directing group. The strategy furthermore cannot access 5H-dibenzo[b,f]azepines 1a as the ethylene bridge would cross react with the brominating agent .

Scheme 14: Double Buchwald–Hartwig amination towards 10,11-dihydro-5H-dibenzo[b,f]azepine derivatives 71.

N-Aryl and N-alkyldihydropyridobenzazepines 75 and 76 were synthesised by Tsoung et al. through a multicomponent reaction system . The authors provided a series of substituted derivatives through Pd/Rh-catalysed domino coupling. The reaction proceeded via a Suzuki coupling, followed by an in situ Buchwald–Hartwig amination. The authors reported moderate to good yields in a series with electron-donating and electron-withdrawing groups, as well as N-aryl and N-alkylamines (Scheme 15).

Scheme 15: One-pot Suzuki coupling–Buchwald–Hartwig amination.

Lam et al. expanded on the multicomponent method to form substituted dihydropyridobenzazepines 80–82 wherein vinylpyridines 77 are coupled with boronate ester anilines 78 in a Suzuki reaction, whereafter Buchwald–Hartwig amination afford the various diarylazepines. A three-component one-pot process allowed for a second in situ Buchwald–Hartwig amination of the diarylazepine with aryl or benzyl halides to give the respective N-aryl and N-benzylazepine derivatives 83 and 84 (Scheme 16).

Scheme 16: One-pot Rh/Pd-catalysed synthesis of dihydropyridobenzazepines.

3.2 Mizoroki–Heck coupling

Whereas Arnold et al. reported the preparation of dibenzo[b,f]heteropines via consecutive Heck and Buchwald–Hartwig reactions (Scheme 11), amination may also precede the introduction of the double bond (Scheme 17). The formation of the dibenzo[b,f]heteropine skeleton by means of a final Mizoroki–Heck reaction will be discussed in the following section.

Scheme 17: A retrosynthetic pathway to dibenzo[b,f]azepines via Mizoroki–Heck reaction.

The Buchwald group reported a ligand-controlled divergent synthesis involving intramolecular cyclisation, allowing for the formation of several heterocycles, including dibenzo[b,f]azepines 89, in two steps. Screening of reaction conditions during the investigation resulted in the synthesis of dibenzo[b,f]azepine 89 directly from 2-bromostyrene 86 and 2-chloroaniline 87 in up to 99% yield (Scheme 18). Several substituted dibenzo[b,f]azepines 89 and heteroaryl analogues were reported with excellent yields and regioselectivity. A later correction to the article revised the yield from 99% to 70% and with overall poorer selectivity . The correction is in line with reports of poor selectivity when performing intramolecular Heck reactions (cf. Jepsen et al. ).

Scheme 18: One-pot domino Pd-catalyzed Mizoroki–Heck–Buchwald–Hartwig synthesis of dibenzo[b,f]azepines.

An analogous reaction pathway by Jepsen et al. was used to synthesise dibenzo[b,f]thiapines 1c and dibenzo[b,f]oxepines 1b in three steps through a styrene (95 and 96) intermediate (Scheme 19). While the reported conversion was excellent, the yield was low due to moderate selectivity, resulting in a mixture of 7-endo (1c and 1b) and 6-exo (97 and 98) cyclised products.

Scheme 19: Dibenzo[b,f]thiapine and -oxepine synthesis via SNAr (thio)etherification, Wittig methylenation and Mizoroki–Heck reaction.

3.3 Ullmann-type coupling

Copper-catalysed Ullmann etherification (Scheme 20) offers an alternative to SNAr and Buchwald–Hartwig etherification.

Scheme 20: A retrosynthetic pathway to dibenzo[b,f]oxepines via Ullmann coupling.

Olivera et al. reported a copper-catalysed Ullmann-type etherification as a key step in the synthesis of their pyrazole-fused dibenzo[b,f]oxepine derivatives 101 (Scheme 21).

Scheme 21: Ullmann-type coupling in dibenzo[b,f]oxepine synthesis.

Lin et al. used copper-catalysed coupling in their total synthesis of bulbophylol-B (105), a substituted dihydrobenzo[b,f]oxepine. The authors synthesised an intermediate stilbene via Wittig reaction, followed by hydrogenation to give dihydrostilbene 104, which underwent intramolecular Ullmann-type coupling catalysed by CuBr·DMS to form the fused dihydro[b,f]oxepine ring system in 89% yield, whereafter hydrogenation afforded 105 in almost quantitative yield (Scheme 22). The method is a sequence of 12 steps, the majority of which are to prepare Wittig reagent precursor 102 and the complementary aldehyde 103.

Scheme 22: Wittig reaction and Ullmann coupling as key steps in dihydrobenz[b,f]oxepine synthesis.

3.4 Catellani-type reaction

The Catellani reaction involves palladium-norbornene cooperative catalysis to functionalise the ortho- and ipso-positions of aryl halides by alkylation, arylation, amination, acylation, thiolation, etc. .

Della Ca' et al. reported the synthesis of substituted dibenzo[b,f]azepines 110 as unexpected products during their investigation of the Catellani reaction. The Pd-catalysed reaction of an aryl iodide 106, bromoaniline 107, norbornadiene (108) and base resulted in the norbornene-azepine intermediate 109. Heating to 130 °C induces a retro-Diels–Alder reaction, giving dibenzo[b,f]azepine 110 in good yield (Scheme 23). The authors synthesised a series of derivatives, with substituents including -OMe, -Me, -Cl and –F, with good yield (50–78%) in one step.

Scheme 23: Pd-catalysed dibenzo[b,f]azepine synthesis via norbornene azepine intermediate 109.

In the follow-up reported in 2018 , the method was extended to aryl bromides and electron-withdrawing groups. The authors found that the addition of potassium iodide, and thus in situ palladium-catalysed halogen exchange, improved the yield of dibenzo[b,f]azepine 110. Unsymmetrical derivatives of 110 containing -CO2Me, -CF3, -NO2 and -CN substituents were synthesised in moderate to good yield (35–82%).

3.5 Ring-closing metathesis

Olefin metathesis is a metal-catalysed reaction wherein carbon–carbon double bonds are cleaved and formed through an intermediate cyclometallacarbene 114, thus allowing for transalkylidenation and the formation of mixed alkenes 115 (Scheme 24) . Variations of this reaction include alkyne metathesis and carbonyl metathesis .

Scheme 24: A simple representation of olefin metathesis resulting in transalkylidenation.

Ring-closing metathesis (RCM) gave access to a series of dibenzo[b,f]heteropines, as reported by Matsuda and Sato (Scheme 25). The authors synthesised a series of Si-, Sn-, Ge- and B-tethered dienes 118 from 2-bromostyrene (116) via halogen–lithium exchange and quenching with the appropriate heteroatom source (SiR2Cl2, SnMe2Cl2, GeR2Cl2, BBr3). P-Tethered dienes were synthesised via quenching of a 2-vinylphenyl Grignard reagent with phenylphosphonic dichloride (PhPOCl2). O-Tethered dienes were prepared by Wittig methylenation of commercially available bis(2-formylphenyl) ether (119), whereas a formylation–Wittig methylenation sequence of commercial diphenylsulfone (120) and protected bis(2-bromophenyl)amine 121 afforded the S- and N-tethered diene, respectively. Ruthenium (2nd generation Hoveyda–Grubbs catalyst) catalysed ring-closing metathesis gave dibenzo[b,f]heteropines 122 in excellent yields (>80%). Unfortunately, the metathesis reaction required elevated temperatures (>100 °C) and dilute solutions to reduce unwanted self-metathesis competing with RCM. While excellent yields for synthesising the tethers and RCM products are reported, the method does not currently allow for the synthesis of unsymmetrical compounds.

Scheme 25: Ring-closing metathesis as key step in the synthesis of dibenzo[b,f]heteropines.

3.6 Alkyne–aldehyde metathesis

Bera et al. reported on the synthesis of a series of 10-acyldibenzo[b,f]oxepines 125 by alkyne–aldehyde metathesis catalysed by iron(III) chloride (Scheme 26). Alkyne–carbonyl metathesis is proposed to proceed via [2 + 2] cycloaddition and –reversion steps, catalysed by a Brønsted or Lewis acid, with the catalyst proposed to form a σ-complex with the carbonyl group and/or a π-complex with the alkyne .

Scheme 26: Alkyne–aldehyde metathesis in the synthesis of dibenzo[b,f]heteropines.

3.7 Hydroarylation

The construction of an N-triarylated dibenzo[b,f]azepine scaffold 129 by means of Au(I)-catalysed hydroarylation was reported by Ito et al. . While the attempted synthesis of an N-phenyldibenzazepine derivative 127 was unsuccessful, the authors were able to prepare a fused carbazole-dibenzo[b,f]azepine 129 in 90% yield via a gold/silver catalyst system (Scheme 27).

Scheme 27: Hydroarylation of 9-(2-alkynylphenyl)-9H-carbazole derivatives.

4 Oxidative C–C coupling

Whereas oxidative C–C coupling precedes amination in the industrial route to 5H-dibenzo[b,f]azepine, oxidative C–C coupling may also be the final step in the construction of the dibenzo[b,f]heteropine skeleton.

Comber and Sargent synthesised pacharin (

View original article

BEILSTEIN JOURNAL OF ORGANIC CHEMISTRY

Like

分享书签

0 0 0 0 0 0 0

留言 (0)

登入或註冊以發表你的留言

More from this channel

Strategies in the synthesis of dibenzo[b,f]heteropines

留言 (0)