2022 年 98 巻 4 号 p. 190-205
Biaryl synthesis continues to occupy a central role in chemical synthesis. From blockbuster drug molecules to organic electronics, biaryls present numerous possibilities and new applications continue to emerge. Transition-metal-catalyzed coupling reactions represent the gold standard for biaryl synthesis and the mechanistic steps, such as reductive elimination, are well established. Developing routes that exploit alternative mechanistic scenarios could give unprecedented biaryl structures and expand the portfolio of biaryl applications. We have developed metal-free C–H/C–H couplings of aryl sulfoxides with phenols to afford 2-hydroxy-2′-sulfanylbiaryls. This cascade strategy consists of an interrupted Pummerer reaction and [3,3] sigmatropic rearrangement. Our method enables the synthesis of intriguing aromatic molecules, including oligoarenes, enantioenriched dihetero[8]helicenes, and polyfluorobiaryls. From our successes in aryl sulfoxide/phenol couplings and a deeper understanding of sigmatropic rearrangements for biaryl synthesis, we have established related methods, such as aryl sulfoxide/aniline and aryl iodane/phenol couplings. Overall, our fundamental interests in underexplored reaction mechanisms have led to various methods for accessing important biaryl architectures.
As the biaryl skeleton is structurally rigid, yet electronically flexible and peripherally modifiable, it occupies a central position in the design and synthesis of electronically and structurally tunable molecules. Needless to say, transition-metal-catalyzed cross-coupling reactions have been widely used to reliably prepare biaryls (Scheme 1).1),2) Recently, catalytic direct C–H arylation reactions have been established as a more atom- and step-economical alternative.3)–5) In both methods, the key intermediate is a transition metal complex bearing two aryl groups on the metal center, from which reductive elimination provides the desired biaryl.
Transition-metal-catalyzed coupling through reductive elimination.
Despite extensive studies, these catalytic methods for the synthesis of biaryls are not universal. For example, there remain several classes of substituted biaryls that are not easily accessible by current catalytic methods. Such limitations mostly originate from the difficulty in generating the key diaryl transition metal complex, for example due to steric hindrance around the metal center or poor regioselectivity of metalation. Thus, developing new methods for the construction of biaryls based on different reaction mechanisms without recourse to the biaryl-forming reductive elimination is highly sought after.6)–8)
We would like to draw your attention to the benzidine rearrangement (Scheme 2a).9),10) This 150-year-old reaction represents the isomerization of N,N′-diphenylhydrazine (hydrazobenzene) under acidic conditions into 4,4′-diaminobiphenyl (benzidine) via a protonation-induced sigmatropic rearrangement. Its N–O analogues, N,O-diphenylhydroxylamine derivatives, also undergo similar rearrangement (Scheme 2b).11),12) By looking at these classical reactions from the viewpoint of modern organic chemistry, we become aware of the following interesting features. (1) This unique C–C bond-forming event accompanies dearomatization of both the benzene rings, which is made possible through cleavage of the weak N–N or N–O bond. (2) The transformations are regarded as a formal dehydrogenative C–H/C–H coupling, which proceeds with high regioselectivity controlled by [5,5] sigmatropic rearrangement. (3) The products naturally have two heteroatom substituents that stem from the tether. Similarly, [3,3] sigmatropic rearrangement gives 2,2′-disubstituted biaryls that are often useful synthetic intermediates. As an elegant recent example, Kürti13) and List14) independently reported in 2013 enantioselective benzidine rearrangement of hydrazonaphthalene by using an Akiyama-Terada catalyst, which affords 2,2′-diaminobinaphthyl, a privileged motif for axially chiral catalysts (Scheme 3a). Similarly, Kürti/Gao15)–17) and Li/Tan18) reported the coupling reaction of N-naphthylhydroxylamine with naphthyliodonium salts to synthesize racemic 2-amino-2′-hydroxybinaphthyl via sigmatropic rearrangement of in situ generated N,O-dinaphthylhydroxylamine (Scheme 3b). Currently, the construction of biaryl skeletons via sigmatropic rearrangement is attracting increased attention.
(a) Benzidine rearrangement and (b) its N,O variant.
Such couplings to prepare biaryls via sigmatropic rearrangement is distinct from transition-metal-catalyzed coupling and hence intriguing. However, significant progress is needed to establish it as a truly useful strategy for biaryl synthesis. For example, these reports suffer from several limitations. (1) There are only a few methods to prepare the precursors for the rearrangement. (2) The available tethers between the two aromatic rings are limited to an N–N or N–O bond.
Since 2005, we have been interested in chemical transformations that utilize the unique reactivity of organosulfur compounds. During our study,19),20) we developed a new method for the synthesis of biaryls via sigmatropic rearrangement of sulfur-tethered intermediates. Here we summarize our progress regarding this new sulfur-based method,21) together with related work by others, after briefly detailing how our idea for sulfonium-aided biaryl synthesis was evolved in our laboratory in the next section.
It was when we had been exploring the use of ketene dithioacetal monoxides (KDMs) for extended Pummerer reactions22)–32) that we accidentally realized the sigmatropic rearrangement of sulfur-tethered intermediates. As described in Scheme 4, treatment of KDM 1 with trifluoromethanesulfonic anhydride (Tf2O, Tf = CF3SO2) results in activation of the sulfoxide moiety of 1. Typically, a nucleophile (Nu−) would then attack at the distant exocyclic vinylic carbon of the sulfonium cation 2 (an additive Pummerer reaction) to yield intermediate 3 and eventually ketene dithioacetal product 4, a useful disubstituted ketene equivalent.
Typically conceived mechanism of functionalization of KDM 1 at the β position.
By considering the importance of allylic groups in organic synthesis, we tested allylic silane 5 as the nucleophile (Scheme 5)33) and encountered a totally unexpected result: The reaction exclusively afforded the linear product 6 via a formal α-attack, without forming the branched product 7, the regioselectivity of which is opposite to the γ-attack that the Hosomi–Sakurai reactions always undertake. We considered that the following rationalizes the abnormal regioselectivity: The activated sulfonium cation 2 reacts with allylic silane 5 in a normal Hosomi–Sakurai fashion. However, the nucleophilic attack does not occur at the exocyclic vinylic carbon of 2 but at the cationic sulfur atom to yield allyl vinyl sulfonium intermediate 8 (an interrupted Pummerer reaction). The intermediate 8 then undergoes charge-accelerated sigmatropic rearrangement34) at room temperature very smoothly to deliver the linear product 6 after the final deprotonation. We realized that such a cascade of interrupted Pummerer reaction and [3,3] sigmatropic rearrangement would provide unusual reactivity/selectivity and thus great possibilities in organic synthesis and started pursuing new transformations utilizing this cascade. It is worth noting that our work on the allylation of ketene dithioacetal monoxides inspired others to develop related interesting allylation and propargylation processes of aryl sulfoxides.35)–38)
Discovery of a cascade of interrupted Pummerer reaction and [3,3] sigmatropic rearrangement.
For example, instead of allylic silanes, ketones are suitable nucleophiles for the cascade reaction (Scheme 6).39) Sulfonium cation 2 undergoes nucleophilic attack of the carbonyl oxygen at the cationic sulfur atom to yield 9, which rearranges to make a carbon–carbon bond between the α position of the ketone and the exocyclic vinylic carbon. The initial product 10 is a 2-trifluoromethyl-1,4-dicarbonyl equivalent that can be easily converted to the corresponding 3-trifluoromethylheteroles by Paal–Knorr-type transformations. This report was followed by interesting reports by others using aryl sulfoxides instead of KDMs to develop transition metal-free α-arylation of carbonyls.40)–42)
α-Functionalization of ketones for the synthesis of 3-trifluoromethylheteroles.
Another notable example is the reaction of phenols with alkenyl sulfoxides such as KDM 11 to yield benzofuran 12 (Scheme 7).43)–47) As phenol is a tautomer of 2,4-cyclohexadienone, similarly to Scheme 6, phenol interrupts the sulfonium intermediate from 11 to yield 13. Interestingly, subsequent [3,3] sigmatropic rearrangement occurs with concomitant dearomatization at or below room temperature to form a C–C bond at the ortho position of phenol. Cyclization and global aromatization of intermediate 14 provided 12. This sulfonium-aided annulation reaction is operationally very simple, proceeding by an addition of anhydride to a mixture of phenol and alkenyl sulfoxide. The reaction has a wide substrate scope, even N-sulfonylanilines such as 15 can be used to prepare indoles (Scheme 8).48) This beautiful cascade process reminds us of the Fischer indole synthesis.
Annulation between phenol and alkenyl sulfoxide into benzofuran.
Annulation between N-sulfonylaniline and KDM into indole.
We then envisioned using aromatic sulfoxides as unsaturated sulfoxides instead of alkenyl sulfoxides. To our delight, treatment of phenol and 2-benzothienyl sulfoxide 16 with trifluoroacetic anhydride provided biaryl 17a in high yield with exclusive regioselectivity (Scheme 9).49) The reaction should proceed in a similar fashion to Scheme 7. After formation of the phenoxy-substituted sulfonium intermediate by an interrupted Pummerer reaction, [3,3] sigmatropic rearrangement takes place with concomitant dearomatization of both the aromatic rings to form a C–C bond between the ortho position of phenol and the 3 position of 16. Subsequent rearomatization of both the rings furnishes 17a. Other regioisomers were not observed, which suggests the reaction is not a simple Friedel–Crafts-type electrophilic aromatic substitution of phenol.
Formation of biaryl via a cascade of interrupted Pummerer reaction and [3,3] sigmatropic rearrangement.
A series of aromatic sulfoxides were found to participate in the synthesis of biaryls (Scheme 10). Other five-membered heteroaromatic sulfoxides participated to yield 17b–f. 2-Benzothienyl 4-tolyl sulfoxide reacted selectively on the less aromatic and more electron-rich benzothiophene ring to yield 17c exclusively. The reaction of 2-naphthyl sulfoxide with 2-naphthol afforded 1,1′-binaphthyl 17g in high yield. However, the reaction of phenyl sulfoxide with 2-naphthol afforded the expected product 17h in low yield in a complex product mixture. In contrast, electron-rich 3,5-dimethoxyphenyl sulfoxide is a good substrate to yield 17i in excellent yield. These electronic effects provide important information on the detailed nature of the sigmatropic rearrangement (vide infra).
Scope of the biaryl synthesis. Compound 17u was obtained as a 2.6:1 mixture of regioisomers and the asterisk (*) in 17u indicates the position where the C–C bond formation occurred in the minor isomer.
The scope with respect to phenols is wide as shown in Scheme 10. Functional groups such as iodo in 17o and 17q and pinacolatoboryl in 17p survived under the conditions, which is in sharp contrast to transition-metal-catalyzed coupling reactions. Arylation of 3- or 2-substituted phenols occurred at the less hindered 6 positions to afford 17q–u with exclusive or high regioselectivity. These regioselectivities strongly suggest again that the reactions proceeded not via Friedel–Crafts-type electrophilic aromatic substitution of phenols but via the sigmatropic rearrangement of sulfonium-tethered intermediates.
Interestingly, the reaction of 2,6-dimethylphenol resulted in the preferential arylation at the 3 position of 2,6-dimethylphenol to yield biaryl 18 with high regioselectivity (Scheme 11). By following a similar mechanism to that shown in Scheme 9, the sterically encumbered interrupted intermediate 19 is formed and undergoes [3,3] sigmatropic rearrangement to yield 20, despite the presence of the methyl group. The benzothiophene skeleton should be easily regenerated, but the cyclohexadienone would fail to aromatize. Instead, the trifluoroacetic acid that is generated in situ would protonate cyclohexadienone 21 to induce 1,2-shift of the benzothienyl group, which followed by deprotonation furnishes 18.
Arylation of 2,6-dimethylphenol at the 3 position.
It is worth noting that Procter reported a similar method detailing the reaction of in situ generated benzothiophene S-oxides with phenols and trifluoroacetic anhydride to give the phenol and benzothiophene coupled products linked at their ortho and 3 positions, respectively (Scheme 12).30),50)–55)
Arylation of benzothiophene at the 3 position by Procter.50)
This new sulfonium-aided coupling of two arenes is mechanistically totally different from the conventional transition-metal-catalyzed coupling reactions. Although similar benzidine rearrangements have been extensively investigated, they remain mechanistically obscure. In light of the importance of such coupling reactions via sigmatropic rearrangement, we investigated the mechanism of our sulfonium-aided coupling experimentally and computationally.56) For clarity, here we mainly discuss computationally obtained reaction pathways (Fig. 1) without detailing experimental results that coincide with the computational results.
Energy diagram of the whole process of the reaction of 16 with phenol. Structures of transition states are shown by ball and stick model, with representative bond lengths in angstrom (Å).
As we expected, the reaction was found to proceed via an interrupted Pummerer reaction followed by sigmatropic rearrangement. Interestingly, activation of sulfoxide 16 and the subsequent interrupted Pummerer reaction of the resulting sulfurane INT1 with phenol to yield INT2 were calculated to be reversible. The transition state of the interrupted Pummerer reaction TS1 was found to be six-membered and cyclic, wherein the S–OPh-bond formation and deprotonation by the CF3CO2 unit occurs in a concerted way. From INT2, the rate-determining sigmatropic rearrangement proceeds via a boat-like transition state TS2 with an activation free energy of 20.3 kcal/mol. The rate-determining step is supported by the fact that INT2 was observable by 1H NMR experiments. The subsequent two tautomerization processes are both kinetically and thermodynamically facile.
Detailed analysis of the intrinsic reaction coordinate (IRC) of the sigmatropic rearrangement revealed the apparent asynchronous nature of the concerted rearrangement. As shown in Fig. 2a, the sigmatropic rearrangement shows a nearly flat region in energy (deep blue solid line) before TS2. The root mean squared (RMS) gradient (red dotted line) clearly shows a trough around the flat region, which indicates the existence of a “hidden intermediate”. This asynchronous nature is also observed in the bond lengths of the cleaving S–O bond and the developing C–C bond (Fig. 2b). The S–O bond cleavage takes place at the early stage of the sigmatropic rearrangement with the C–C bond underdeveloped, and the C–C bond formation follows to reach TS2 and then INT3.
(a) IRC pathway from TS2. (b) Potential energy surface around TS2.
In the reaction of an electron-deficient phenol, the “hidden intermediate” was not observed and the sigmatropic rearrangement is synchronous. In contrast, in the reaction of an electron-rich phenol, the “hidden intermediate” is not hidden any more and was computationally obtained as an energy minimum (Fig. 3). The Wiberg bond indexes (WBI) of the cleaving S–O bond and the developing C–C bond in the emerging intermediate INTπ are 0.07 and 0.06, respectively, which suggests the benzothiophene and phenoxy units are not connected through apparent σ bonds and that INTπ can be regarded as a π-complex. The changes of the natural population analysis (NPA) charge distribution during the sigmatropic rearrangement indicates the following: Initially, the methylsulfanylbenzothiophene part and the phenoxy part have positive (+1.27) and negative (−0.39) NPA charges, respectively. As the reaction proceeds, the electron density moves to the benzothienyl part (reaching +0.72 at INTπ) from the phenoxy part (reaching +0.26 at INTπ). From INTπ, nucleophilic attack by the benzothienyl unit onto the electron-deficient phenoxy moiety (phenoxonium) occurs to form the C–C bond of INT3′ and yield the cationic thionium (+0.92) and almost neutral cyclohexadienone (+0.06). Thus, when comparing the formation of products 17h and 17i, the positive effect of methoxy substituents in 17i can be explained as an increase in electron density on the sulfoxide unit, which facilitates smoother nucleophilic attack. This detailed analysis of the sigmatropic rearrangement is highly informative for the design and realization of new efficient coupling reactions of two arenes.
Stepwise rearrangement in the reaction of 16 with electron-rich p-cresol.
With a new unique method for regioselectively coupling aryl sulfoxides and phenols in hand, we explored applications that highlight the advantages over transition-metal-catalyzed coupling reactions.
4.1. Synthesis of oligoaryls via repetitive couplings.Oligoaryls represent an important class of organic frameworks and are used as functional core structures for biological and materials applications. In order to synthesize well-designed oligoaryls, stepwise elongation via iterative cross-coupling has been regarded as a powerful and reliable methodology.57) In particular, the utilization of on/off-switchable difunctional coupling partners has occupied a central position to this end, as exemplified in Fig. 4 by the pioneering work using an OH/OTf switch by Hamilton58) and Manabe59) and by elaborate work concerning on/off organometallic switches by Suginome (Bdan),60) Burke (BMIDA),61) and Nakao/Hiyama (HOMSi).62)
On/off-switchable units for iterative cross-coupling.
With this in mind, we envisioned that our coupling of aryl sulfoxide with phenols is applicable to metal-free iterative synthesis of oligoaryls.63) Two-fold arylation of aryl sulfoxide 22 with two different naphthols is exemplified in Scheme 13. Of note, concomitant with the coupling, the sulfinyl moiety is reduced to the sulfanyl group which is “off” and no longer able to promote the dehydrogenative coupling with phenols, thus preventing unwanted polymerizations. The resulting aryl sulfides 23 can be readily switched “on” through oxidation to the corresponding sulfoxide 24, which consists of two diastereomers that have axial chirality in the biaryl motif and central chirality at the sulfur atom. Another dehydrogenative coupling is possible for the synthesis of teraryl 25, which consists of two diastereomers that have two stereogenic axes in the skeleton. It is worth stating that this switching ‘on’ and ‘off’ of reactivity by sulfur oxidation was also recognized and exploited by Procter and his team.64)
Sulfoxide-based iterative couplings.
Our method always provides biaryls bearing an ortho-hydroxy group on one aromatic ring and an ortho-methylsulfanyl group on the other aromatic ring. Utilizing these two functional groups in proximity, one can envision synthesizing π-extended heteroles via intramolecular cyclization. We have indeed achieved systematic syntheses of a series of enantiomerically pure dihetero[8]helicenes.65) This presents a unique route to heterohelicenes, which have attracted attention because of their intriguing helical π-conjugated systems and resulting chiral optical properties.66),67)
The synthetic scheme starts with assembling a common synthetic intermediate 27 from bissulfinylnaphthalene 26 and 2 equivalents of 2-naphthol through our method (Scheme 14a). It is worth noting that the reaction proceeds with high efficiency and anti selectivity even though 27 has a highly crowded ternaphthyl motif, which highlights a clear advantage of our method. Optical resolution of 27 was facile to obtain pure material of both enantiomers of 27 on multi-gram scales.
Synthesis of a series of (P)-dihetero[8]helicenes. (a) Synthesis of enantiopure common intermediate (Sa,Sa)-27. (b) Construction of helicenes from (Sa,Sa)-27. The synthesis of the (M) isomer is omitted for clarity. Values of ee after recrystallization are in parentheses.
The situation is set for systematic syntheses of enantiomerically pure dihetero[8]helicenes (Scheme 14b). After oxidation of (Sa,Sa)-27, disulfone (Sa,Sa)-28 underwent two-fold SNAr cyclization using the sulfur functionality as a leaving group to yield dioxa[8]helicene (P)-29O. Alternatively, E2 elimination of (Sa,Sa)-27 to yield bisthiol 30 and subsequent ring-closing condensation with triflic acid afforded (P)-29S. These cyclization reactions took place with excellent axial-to-helical chirality conversion, especially in the room temperature synthesis of (P)-29S. Dithia[8]helicene (P)-29S was oxidized to the corresponding tetraoxide 29SO2 quantitatively, which was further transformed into the nitrogen and the carbon analogs by replacing the two endocyclic SO2 units via SNAr-based aromatic metamorphosis.68),69) Diaza[8]helicene (P)-29N was synthesized by treatment of (P)-29SO2 with 2-phenylethylamine under strongly basic conditions. This reaction proceeded via an SNAr-based replacement of the SO2 units to give intermediate 31 followed by E2 elimination to furnish (P)-29N.70),71) A slight decrease in the optical purity of 29N to 77%ee was observed during the substitution because of the high reaction temperature (110 °C). In a similar fashion, xanthene and fluorene served as carbon nucleophiles for the endocyclic substitution reaction72) to yield (P)-29CX and (P)-29CF. For all the heterohelicenes, recrystallization provided the enantiomerically pure (P) isomers. Notably, the enantiomers of all the heterohelicenes were obtained from a common precursor (Sa,Sa)-27. Analogously, the (M)-heterohelicene isomers can be obtained in similar yields and enantioselectivities from (Ra,Ra)-27 by following the same synthetic pathways.
We thus succeeded in the syntheses of a series of enantiomerically pure dihetero[8]helicenes with extremely high efficiency. The efficient construction of this library allowed us to comprehensively and systematically evaluate the chemical and photophysical properties of the series of dihetero[8]helicenes. Several comments are worth noting. 1) The kinetic studies for racemization dynamics revealed that the dihetero[8]helicene skeleton is fixed due to the rigid 5-membered rings and exhibits unexpectedly higher thermal conformational stability in comparison with carbo[8]helicene that shows larger overlap between the terminal edges. This observation offers a seemingly contradictory viewpoint in designing helicenes: introduction of a smaller 5-membered heterole ring to stabilize helicity. 2) Spirocyclic 29CX and 29CF were found to exhibit by far the highest fluorescence quantum yields of 62% and 66%, respectively. In addition, 29CF was found to emit blue fluorescence in the solid state (ΦF = 40%). 3) 29O and 29N show both high fluorescence efficiency (ΦF = 39% and 13%) and dissymmetric factors of fluorescence |glum| (6.1 × 10−3 for 29O and 9.5 × 10−3 for 29N) among reported helicenes.73) Thus, the potential utility of our systematic synthetic strategy will be applicable to efficient generations of chemical libraries of π-extended compounds to find ‘hit’ molecules as well as to accelerate the progress of the chemistry of heterohelicenes.
4.3. Synthesis of fluorinated arenes via defluorinative strategy.Organofluorine compounds are widely used in pharmaceutical, agrochemical, and material science and vastly related industrial applications. Development of methods for the synthesis of fluorinated compounds has been an important topic in organic synthesis. While introduction of a fluorine atom to a molecule is a straightforward and conventional method, the recent increase in the availability of polyfluorinated compounds pose questions about developing selective and controlled defluorinative transformations.74),75) However, this new contrasting methodology remains underdeveloped in many aspects such as regioselectivity and mono/di/tri-defluorination selectivity.
In analogy to the reaction of 2,6-dimethylphenol in Scheme 11, we investigated the use of 2,6-difluorophenol for our Pummerer chemistry (Scheme 15).76),77) The reaction with 16 should provide dearomatized intermediate 33, which then undergoes reductive defluorination by means of zinc in the same pot to recover aromaticity and to afford monofluorobiaryl 32a. This defluorinative transformation accommodates a broad range of polyfluorophenols including heptafluoro-2-naphthol and heteroaryl sulfoxides (Scheme 16). The biaryl bond formations took place exclusively at the ortho position of the phenols and selectively removed only one fluorine atom. On a negative note, simple phenyl sulfoxide is not a suitable substrate because of the stronger aromaticity of the phenyl ring. The chemistry is applicable to the synthesis of a fluorinated analog of Maxipost (BMS-204352), a competent potassium channel modulator (Scheme 17). Indolyl sulfoxide 35, prepared from commercially available 34, reacted with 4-chloro-2,6-difluorophenol to afford 36 in high yield. Methylation of the phenolic hydroxy group, Boc deprotection, and electrophilic fluorination with Selectfluor culminates in the synthesis of the fluorinated analog.
Defluorinative synthesis of biaryls from 2,6-difluorophenol and benzothienyl sulfoxide with the aid of zinc as a reducing agent.
Defluorinative synthesis of biaryls from polyfluorophenols and heteroaryl sulfoxide.
Synthesis of a fluorinated analog of Maxipost.
We also focused on the usefulness of dearomatized intermediates such as 33 as building blocks for the synthesis of multi-substituted phenols.77) In the absence of zinc powder, the reaction of 2,6-difluorophenol produces 33 in the reaction flask. The subsequent addition of BF3·OEt2 induces acid-mediated 1,2-shift of the benzothienyl unit as shown in Scheme 11 to result in the meta-substituted 2,6-difluorophenol 37 without defluorination (Scheme 18). Instead of treatment with acid, treatment of 33 with a nucleophile triggers 1,4-addition onto the cyclohexadienone skeleton. Then, elimination of HF and concomitant aromatization yields monofluorophenol 38 bearing the nucleophilic unit at the meta position (Scheme 19).78)
meta-Arylation of 2,6-difluorophenol without defluorination.
Nucleophilic meta-amination and thiolation of 2,6-difluorophenol with ortho-defluorinative arylation.
In the previous sections, the reactions of phenols via the transient formation of S–O bonds were discussed. If one can use different combinations of heteroatom linkers, biaryl synthesis based on sigmatropic rearrangement becomes a more popular synthetic strategy.
5.1. Synthesis of 2-amino-2′-sulfanylbiphenyls.We used a simple aniline instead of phenols and tried reactions with trifluoroacetic anhydride as an activator under similar conditions. However, trifluoroacetylation of highly nucleophilic aniline took place exclusively and activation of sulfoxide did not occur. Interestingly, triflic anhydride Tf2O was found to react preferentially with sulfoxide leading to an interrupted Pummerer reaction (Scheme 20).79) Unfortunately, another problem arose from the high stability of sulfyliminium intermediate 39; no sigmatropic rearrangement took place because the cationic charge on the sulfur is shared with the neighboring nitrogen. We thus envisioned that protonation of 39 would form dicationic intermediate 40, which was expected to undergo sigmatropic rearrangement. This was indeed the case, and treatment of 39 with TfOH yielded 2-amino-2′-sulfanylbiaryl 41 in good yield.
Biaryl synthesis from aryl sulfoxide and aniline via dicationic intermediate.
One drawback of the reaction of unprotected anilines is its limited scope. To achieve a more general method, we investigated the effect of protecting groups on the aniline substrates. Electron-withdrawing protecting groups on the nitrogen should impart several favorable features, for example, reduce the nucleophilicity of the nitrogen to avoid direct reaction with anhydride, increase the acidity of the N–H unit to construct an S–N bond by concerted deprotonation (Fig. 1, from INT1 to INT2), and weaken the S–N bond in the precursor to charge-accelerated sigmatropic rearrangement. After the screening of electron-withdrawing groups, sulfonyl groups were found to participate in the biaryl formation efficiently (similar to the indole synthesis in Scheme 8), and p-methoxybenzenesulfonyl group showed the best performance among those tested (Scheme 21).80) While the scope of the reactions is not so wide as in the case of phenols, a range of heteroaryl sulfoxides and naphthyl sulfoxides reacted to yield sulfonylaminobiaryls 42.
Biaryl synthesis from aryl sulfoxides and sulfoanilides.
We also sought sulfur-free surrogates of aryl sulfoxides for biaryl synthesis. Considering the similarity in structure and reactivity,26),81) we focused on hypervalent iodine compounds. However, because trivalent iodine compounds are strong oxidants, we naturally suffered from oxidative side reactions of phenols. After extensive screening, we found that 2-naphthols are suitable for the coupling and that treatment with 2-iodonaphthalene diacetate (43) in acetic acid yielded 2-hydroxy-2′-iodobinaphthyls 44 (Scheme 22).82) Furthermore, iodoarene diacetate can be generated in situ and used without isolation; iodobenzene 45 was treated sequentially with mCPBA and 2-naphthol to afford biaryl 47 via 46 (Scheme 23).
Binaphthyl synthesis from 2-iodonaphthanlene diacetate and 2-naphthols.
Biaryl synthesis from iodoarene diacetate generated in situ.
Transition metal-catalyzed cross-coupling is currently considered the first choice in the synthesis of biaryls. The key step in this process is the reductive elimination of two aryl ligands on the transition metal center. In contrast, our biaryl synthesis totally differs from catalytic cross-coupling in terms of the reaction mechanism and reaction scope. By transiently connecting two aryl groups with a heteroatom–heteroatom linker, sigmatropic rearrangements can be harnessed to construct biaryls. Remarkably, the aromaticity of both aryl rings is lost during the sigmatropic rearrangement, though this aromaticity is regained upon formation of the final product. Our method has provided access to biaryls that are difficult to synthesize via conventional methods. We have demonstrated the power of our methods by using them to access important compound classes, such as oligoarenes, drug candidates, atropisomers and heterohelicenes. Considering the future of biaryl skeletons, we hope that new methods to prepare biaryls by underexplored mechanisms will be discovered.83)
H.Y. wishes to thank his co-workers, whose names appear on the cited articles, for making this chemistry possible. H.Y. thanks JSPS KAKENHI Grants (JP19H00895, JP18H04409, JP16H01149), JST CREST grant JPMJCR19R4, the Naito Foundation, and the Mitsubishi Foundation for financial support. G.J.P.P. is the recipient of a JSPS Postdoctoral Fellowship for Research in Japan (KAKENHI Grant Number JP21F21039).
Edited by Keisuke SUZUKI, M.J.A.
Correspondence should be addressed to: H. Yorimitsu, Department of Chemistry, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto 606-8502, Japan (e-mail: yori@kuchem.kyoto-u.ac.jp).
p-methoxybenzenesulfonyl
Boct-butoxycarbonyl
Bpinpinacolatoboryl
HMDShexamethyldisilazide
KDMketene dithioacetal monoxide
LDAlithium diisopropylamide
mCPBAm-chloroperbenzoic acid
Tftrifluoromethanesulfonyl
Tsp-toluenesulfonyl
Hideki Yorimitsu was born in Kochi Prefecture, Japan, in 1975 and graduated from the Faculty of Engineering, Kyoto University in 1997. He obtained his Ph.D. in 2002 from Kyoto University under the tutelage of Professor Koichiro Oshima by engaging in research on synthetic radical reactions in aqueous media. He then served as a JSPS postdoctoral fellow with Professor Eiichi Nakamura at the Department of Chemistry, the University of Tokyo, where he worked on the synthesis of 15O-labeled 2-deoxyglucose of 2-min half-life for PET imaging as well as chemical modifications of nanocarbons. Subsequently, he became Assistant Professor (2003) and Associate Professor (2008) in the Graduate School of Engineering, Kyoto University. In 2009, he moved to the Graduate School of Science, Kyoto University, where he was promoted to Full Professor in 2015. He was Project Leader of ACT-C (2012–2018) and is currently Project Leader of CREST (since 2019), supported by Japan Science and Technology Agency. He received the Chemical Society of Japan Award for Young Chemists in 2009, the Young Scientists’ Prize from MEXT in 2011, the Mukaiyama Award in 2016, the Negishi Award in 2018, the JSPS Prize in 2020, and the Japan Academy Medal in 2020. His research has always focused on the development of new organic transformations in order to create new molecules, phenomena, and concepts. Currently, he is interested in electron injection into unsaturated molecules to generate and use the resulting carbanion species for modern organic synthesis, in addition to sulfur-aided organic synthesis.
Gregory J. P. Perry received his Master of Chemistry from the University of Liverpool (U.K.) in 2012. During his undergraduate studies, he carried out research under the supervision of Professor P. Andrew Evans. He then joined the group of Professor Igor Larrosa and was awarded his Ph.D. from the University of Manchester (U.K.) in 2016. His doctoral studies focused on the development of decarboxylative and C–H transformations for organic synthesis. In 2017, he moved to Professor Kenichiro Itami’s group at the Institute of Transformative Bio-Molecules, Nagoya University (Japan) where he worked on the preparation of small molecules for chemical biology. Greg then moved back to the University of Manchester (U.K.) in 2018 as a Lecturer in Organic Chemistry. During this time, he worked within the group of Professor David J. Procter whilst developing his own research themes. In 2021, Greg was awarded a JSPS Postdoctoral Fellowship to continue his research within the group of Professor Hideki Yorimitsu at Kyoto University (Japan). His current research interests include the application of sulfur (VI) compounds (e.g., sulfonamides, sulfones) in synthesis and the carboxylation/decarboxylation of organic molecules.