Intermolecular oxidative dehydrogenative 3,3 ’ -coupling of benzo[ b ]furans and benzo[ b ]thiophenes promoted by DDQ/H + : total synthesis of shandougenine B †‡

With an excess of a strong acid, 2,3-dichloro-5,6-dicyano-1,4-quinone (DDQ) is shown to promote metal-free intermolecular oxidative dehydrogenative (ODH) 3,3 ’ -coupling of 2-aryl-benzo[ b ]furans and 2-aryl-benzo[ b ]thiophenes up to 92% yield as demonstrated with 9 substrates. Based on the analysis of oxidation potentials and molecular orbitals combined with EPR, NMR and UV-Vis observations, the studied reaction is initiated by a DDQ-substrate charge transfer complex and presumably proceeds via oxidation of the substrate into an electrophilic radical cation that further reacts with another molecule of a neutral substrate. The coupling reactivity can easily be predicted from the oxidation potential of the substrate and the morphology of its frontier molecular orbitals. The intermolecular ODH coupling reaction allowed a concise total synthesis of the natural product shandougenine B.


Introduction
Quinones are important oxidizers in both biological and synthetic chemical conversions. 1For example, the high oxidative power of 2,3-dichloro-5,6-dicyano-1,4-quinone (DDQ) makes it a frequently used oxidant in various chemical conversions.Broadly categorizing, there are three commonly accepted mechanistic pathways associated with oxidation: hydride abstraction, hydrogen atom transfer and single electron transfer (SET). 2 It has previously been shown that quinone promoted heteroatomic oxidations associated with α-deprotonation (Scheme 1a) or direct hydride transfer (Scheme 1b) lead to the formation of good electrophiles (e.g.iminium, oxonium or allylic cations) that can be coupled with various nucleophiles (Nu). 3,46a-c Common to these reactions is that quinones are used in the presence of a strong acid promotor (e.g.TFA, MsOH, H 2 SO 4 , TfOH, HClO 4 or HFSO 3 ), which has been shown to increase the oxidative power of the quinone considerably. 6heme 1 Quinone mediated metal-free approaches for intermolecular oxidative dehydrogenative (ODH) couplings.† This paper is dedicated to Professor Ari M. P. Koskinen on the occasion of his 60 th birthday.‡ Electronic supplementary information (ESI) available: Crystallographic data, xyz-coordinates, cyclic voltammetry data and EPR, 1 H and 13 C NMR spectra (PDF).And X-ray crystallographic data for DDQ : 1c (1 : 2) complex (CIF).CCDC 1477482 (1c_DDQ).][9][10][11][12][13] This structural motif is found in natural products, 14 diphosphine metal ligands, 7a,8a,b diphosphine oxide organocatalysts 15 and in organic chiral electrodes. 16Nevertheless, many of the modern high yielding coupling methods require the use of prefunctionalized starting materials, stoichiometric or substoichiometric amounts of transition metals or air sensitive or pyrophoric reagents (Scheme 2a).However, benzo [b]furans and benzo [b]thiophenes can be 3,3′-coupled from unfunctionalized monomers directly by the Brønsted acid mediated Friedel-Crafts reaction to dihydro-3,3′-benzofused intermediates in variable yields (Scheme 2b). 10,17The McMurry coupling of 3-oxobenzofurans has also been shown to offer a viable route to dihydrogenated intermediates (Scheme 2b). 7These intermediates can then be dehydrogenated either thermally (Friedel-Crafts) or with a stoichiometric amount of an oxidant such as DDQ (McMurry).
In our previous studies, we have observed that carbonyl rich heterogeneous carbon materials can catalyze the oxidative coupling of various unfunctionalized benzo-fused heterocycles. 18In that paper, we also observed that DDQ is capable of mediating a similar 3,3′-homocoupling of 2-phenyl-benzo [b]  furan in the presence of MsOH.18b Herein we report the optimization of the synthetic procedure, the study of the reaction scope and an improved understanding of the reaction mechan-ism via experimental and computational studies (Scheme 2c).We also demonstrate the usefulness of the devised method in the first total synthesis of the 3,3′-bibenzo [b]furan natural product, shandougenine B. 14b

Results and discussion
We began the development of our method with non-optimized reaction conditions, 1 eq. of DDQ, 5 eq. of MsOH and 1 eq. of 1a in DCM (0.05 M) at RT, which yielded 2,2′-bis( phenyl)-3,3′bibenzo [b]furan 2a in 61% yield in 1 h.18b In preliminary studies, we found that an acid additive was necessary for the progress of the reaction.Therefore, we tested the effect of several acids on the reaction and found that, out of these, only MsOH worked satisfactorily (Table 1, entries 1-3, see the ESI ‡ for kinetic data).Later we noticed that the behaviour of TFA was both concentration and substrate dependent.At higher concentrations TFA could promote the reaction of 1a to 2a, whereas in more dilute solutions it promoted the coupling of the more electron rich substrate 1c to 2c. 19 Next, we screened solvents and found that DCE worked better than DCM or CHCl 3 , and that MeNO 2 gave a low yield (Table 1, entries 3-6).We then employed a two level full factorial design for three factors to optimize the reaction conditions. 20Entry 3 was chosen as the centre point and molarity, temperature and the amount of DDQ were set as variables (Table 1, entries 3 and 7-14).One corner of the screened reaction space (+1, −1, −1) gave a very high (99%) yield (entry 11) and no further optimization was deemed necessary. 21fter determining the optimized reaction conditions, the scope of the reaction was studied next.In the case of 2-arylbenzo [b]furans, we were able to obtain the corresponding 2,2′bis(aryl)-3,3′-bibenzo [b]furans in good to excellent isolated yields (Scheme 3, 2a-c).A slightly lower yield (68%) of the coupling product 2d was obtained when the 2-aryl substituent was functionalized with an amide group at the para-position.However, the electron poor 2-(4-nitrophenyl)benzo [b]furan 1i could not be coupled even in the presence of a stronger acid TfOH or with different oxidants such as p-tetrachloro-oquinone or ammonium persulfate (see below).To our pleasure, we could, however, obtain the corresponding 2,2′-bis(aryl)-3,3′bibenzo [b]thiophenes, 2e-g, in equally high yields as the related bibenzo [b]furans 2a-c.Undesirably, the 3,3′-homocoupling product 2h, 2,2′-bis(2-thienyl)-3,3′-bibenzo [b]thiophene was obtained only in 14% yield.We believe that the reason for such a low yield is in the structure of 1h that enables reactivity at two sites, benzo [b]thiophene 3-and thiophene 5-positions, leading to oligomerization.The oxidative 3,3′-homocoupling of the electron rich substrate 1j would have given a direct route to the natural product kynapcin-24, 14a but we could not obtain any product under the reaction conditions employed (see below).
Based on previous studies, 6 it seemed reasonable that the reactions are initiated by a SET between the substrate and the DDQ/H + and the oxidation potential of a particular substrate would be directly related to its reactivity.Consequently, we measured the relative oxidation potentials of compounds 1a-j vs. FcH/FcH + couple, and compared the results with data from theoretical calculations (Table 2).It is clearly evident that both oxidation potentials normalized to 1c have strong systematic correlation (the absolute value is underestimated by ca.0.31 V). 23 Furthermore, both the calculated and the experi-mental oxidation potentials correlate well with the observed reactivity and provide a rationale as for why the homocoupling of 1i was unsuccessful.With an oxidation potential of 1.310 V, 1i is not oxidized by DDQ.
The oxidation potentials do not, however, explain why 1j fails to homocouple even though it has an oxidation potential comparable to 1a and 1f.If we consider that dimerization involves the energetically closest MOs, 24 we can then pay our attention to the morphology of the SOMO of the radical cation and the HOMO of the neutral substrate.For a successful 3,3′homocoupling, a high contribution from the C3-carbon to the nucleophile's HOMO and the electrophile's SOMO is imperative.As shown in Fig. 1, this is certainly true for 1a but not for 1j.Thus, it is clear that, in order to predict reactivity, both    the oxidation potential of the substrate and the morphology of its frontier orbitals need to be considered.Finally, to demonstrate the applicability of our approach, we endeavoured to synthesise shandougenine B. Prior to the synthesis, we predicted the oxidative homocoupling to be favourable based on its low oxidation potential and a high contribution from the C3-carbon to the frontier MOs (see the ESI ‡).Substrate 1k was prepared with the reported literature procedure 25 and the homocoupling reaction promoted with 6 eq. of TFA and 0.5 eq. of DDQ (Scheme 4).To our delight, the dimerization of 1k proceeded smoothly and the corresponding dimer could be isolated in 80% yield.After subsequent removal of the benzyl groups, the antioxidant shandougenine B was obtained in 73% yield (NMR) under standard conditions.Interestingly, the biological origin of shandougenine B has been hypothesised to be oxidative coupling of the monomer Sophora as it is also present in the roots of Sophora tonkinensis.14b Our successful synthesis of shandougenine B via an oxidative route certainly supports this hypothesis.
In the proposed mechanism (Scheme 5), DDQ initially forms a charge transfer (CT) complex with the substrate (Int-1).The formation of a CT complex is seen visually by a colour change and confirmed by UV-vis spectroscopy for 1a-c (Fig. 2 top).Furthermore, the structure of a 2 : 1 complex of 1c and DDQ was determined by single crystal X-ray diffraction (Fig. 2 bottom, see the ESI ‡ for details). 26The layered structure of the complex between 1c and DDQ shows a short (3.14 Å) distance between the electron donor and the acceptor, which is a prerequisite for efficient electron transfer. 27n our mechanistic scenario, the actual coupling reaction is initiated by a SET from 1 to DDQ, which is, in turn, triggered and promoted by addition of an acid that increases the oxidative power of DDQ. 6 The formed electrophilic radical cation 1 +• can then react with a nucleophilic neutral substrate 1 to form a radical cation intermediate Int-2, from which product 2 is obtained after elimination of two protons and a second SET.Alternatively, 1 +• could deprotonate and couple with neutral 1 to form a neutral radical intermediate, leading to 2 via elimination of one proton and a second SET.At this point, the radical cation mechanism seems more likely as it is analogous to the one reported for MoCl 5 -mediated dehydrogenative coupling of arenes.28 Monitoring the reaction with NMR and EPR spectroscopy implies that the purported mechanism is conceivable as the disappearance of the NMR signals is accompanied by an appearance of an EPR signal upon  addition of the acid (see the ESI ‡).Moreover, our preliminary calculations show that the Int-2 structure is a local minimum on the potential energy surface when 1 +• and 1 are coupled together via their C3-carbon atoms (see the ESI ‡).Detailed computational investigations of the reaction mechanism are currently underway.

Conclusions
We have developed a metal free DDQ/H + mediated 3,3′-homocoupling reaction of 2-aryl-benzo [b]furans and 2-aryl-benzo [b]  thiophenes.The feasibility of the reaction was found to correlate with the oxidation potential of the substrate and the morphology of its frontier molecular orbitals.The usefulness of this approach was demonstrated in a priori examination of the monomer Sophora that was successfully dimerized with the developed synthetic method, resulting in the first total synthesis of a natural product shandougenine B.

Computational methods
All calculations were performed with the Turbomole 7.0 program package. 29The solvent environment was taken into account with the COSMO approach with a dielectric constant of 10.125 for 1,2-dichloroethane. 30The geometries were optimized using the TPSS functional 31 augmented with D3 dispersion correction. 32Ahlrichs' def2-TZVP basis sets were used for all atoms as implemented in Turbomole. 33A modified integration grid M4 was employed along with the MARI-J approximation augmented with a suitable auxiliary basis set. 34,35ibrational frequencies were calculated numerically for all stationary points and were found to verify their correspondence to true minima (zero imaginary frequencies) and to obtain Gibbs free energies.Absolute oxidation potentials were calculated using a standard protocol (E°= ΔG 298.15K /nF; where n = 1 and F = 96.48kcal mol −1 V −1 ) and were further normalized to the FcH/FcH + couple in DCE by subtracting 4.927 eV from all calculated values. 23The use of FcH/FcH + couple as a reference for data employing nonaqueous solvents is recommended by IUPAC. 36Molecular orbitals were visualized with the program VMD. 24b

aScheme 3
Scheme 3 Reaction conditions and isolated yields for DDQ/H + mediated homocoupling of benzofurans and benzothiofurans 1a-h to 2a-h.Substrates 1i and 1j failed to yield a product (see the text for details).

a
n.) b /V Exp.− Calc.(Exp.n.− Calc.n.) Measurements were carried out in DCM with 0.05 M NBu 4 PF 6 as the supporting electrolyte.b Values in parenthesis have been normalized to 1c to see relative potentials.

Fig. 2 (
Fig. 2 (Top) UV-vis spectra of DCE solutions of 1a-c and DDQ in a 2 : 1 ratio.(Bottom) Side view of the stacked packing of DDQ and 1c in the crystal structure of a 1 : 2 adduct between DDQ and 1c.

Table 1
Optimization of reaction conditions

Table 2
22perimental and calculated oxidation potentials (vs.FcH/ FcH + )22 13C{ 1 H} NMR spectra were recorded using a Varian Mercury 300 or Varian Inova 500 spectrometer at 27 °C.1Hspectrawerereferenced to tetramethylsilane (TMS, 0.0 ppm) or to residual solvent peaks (CDCl 3 7.26 ppm; d 6 -DMSO 2.50 ppm, CD 3 OD 3.31 ppm and CD 2 Cl 2 5.32 ppm).13C{ 1 H} NMR spectra were referenced to residual solvent peaks (CDCl 3 77.16ppm; d 6 -DMSO 39.52 ppm and CD 3 OD 49.00 ppm).HRMS data were acquired with a JEOL JMS-700 instrument using EI ionization mode.UV-vis spectra were recorded on a Varian Cary 50 spectrometer at 27 °C.Cyclic voltammetry was performed with a Gamry Reference 600 potentiostat using a scan rate of 150 mV s −1 .Solid samples of 1a-j (1-2 mg) were dissolved into a DCM solution containing 0.05 M of NBu 4 PF 6 as the supporting electrolyte.Potentials were scanned at RT in a single compartment cell fitted with Pt electrodes and referenced to FcH/FcH + couple.Details of single crystal X-ray structure determinations are given in the ESI.‡