Xi-Sha
Zhang‡
a,
Zhao-Wei
Li‡
a and
Zhang-Jie
Shi
*ab
aBeijing National Laboratory of Molecular Sciences (BNLMS) and Key Laboratory of Bioorganic Chemistry and Molecular Engineering of Ministry of Education, College of Chemistry and Molecular Engineering and Green Chemistry Center, Peking University, Beijing, 100871, China. E-mail: zshi@pku.edu.cn
bState Key Laboratory of Organometallic Chemistry, Chinese Academy of Sciences, Shanghai 200032, China
First published on 20th December 2013
A palladium-catalyzed base-accelerated ortho-selective C–H alkenylation of phenols to synthesize bioactive coumarin derivatives was developed. The reaction condition was mild and the substrate scope was broad with both electron-neutral and electron-deficient phenols, which is complementary to the previous methods to synthesize electron-rich coumarins. Several bioactive molecules were functionalized and several coumarins with bioactive properties were synthesized. Mechanistic studies showed that this reaction underwent C–H bond activation via direct metallation rather than the Friedel–Crafts pathway.
Coumarin derivatives17 are one kind of important natural products in plants and display good biological and pharmacological activity, including antitumor,18 anti-HIV,19 antioxidant,20 antibacterial,21 and anti-inflammatory22 activity. In addition, they also have good optical properties and are widely used in laser devices, light-emitting diodes, as fluorescent probes and in other areas23 (Scheme 1).
Traditional synthesis of coumarin derivatives is based on condensation reactions, such as Pechmann,20 Perkin, Knoevenagel, Wittig,24 Ponndorf, Horner–Wittig, Houben–Hoesch, and so on.25 Although these methods are widely used and efficient, harsh conditions and stoichiometric quantities of strong Lewis or Brønsted acids are usually required, and the regioselectivity is not ideal and leads to mixtures. Transition metal-catalyzed intermolecular or intramolecular coupling reactions (Heck reaction, carbonylation, etc.) partially solved this problem.26 However, prefunctionalization of the starting materials (halides and organometallic reagents) made these reactions non-economic. Transition metal-catalyzed direct C–H bond addition to alkynes developed by Trost and Fujiwara is more atom economic and has seen wide application in recent years14d,e,27 (Scheme 2, (2)). However, this reaction is generally limited to electron-rich phenols and a Friedel–Crafts pathway rather than a C–H activation pathway cannot be ruled out clearly. Direct C–H bond alkenylation provide another strategy for the synthesis of coumarins. This reaction was first realized in 2005 by Pd catalysis with TFA (trifluoroacetic acid) as solvent, although only electron-rich phenols can be adopted28 (Scheme 2, (3)). Very recently, a tandem reaction starting from cyclohexanone with phenol as intermediate was reported.29 Again, the reaction was conducted in acid and the Friedel–Crafts pathway rather than C–H activation cannot be ruled out clearly. As we were preparing our manuscript, a Pd-catalyzed alkenylation of phenols to synthesis benzofurans and coumarins was reported by Maiti.30 Here, we report a base-accelerated alkenylation reaction of phenol to synthesize coumarins under mild conditions (without acid as solvent) and good functional group tolerance (Scheme 2).
Initially, we selected 4-(tert-butyl)phenol (1a) and methyl acrylate (2a) as model substrates to screen the conditions (Table 1). When Pd(OAc)2 was used as catalyst and Cu(OAc)2 as oxidant, the alkenylation/cyclization product 6-(tert-butyl)-2H-chromen-2-one (3a) was obtained in 11% NMR yield (Table 1, entry 1). Screening of other palladium catalysts showed that Pd(CH3CN)4(BF4)2 gave the best yield (Table 1, entry 3). We proposed that addition of proper base may promote the formation of phenolic anion, which will coordinate with the palladium catalyst more readily and thus promote the ortho-C–H bond activation. Screening of a series of bases showed that many of the bases could accelerate the reaction to different degrees, and NaOPiv and KOPiv showed the highest activity (Table 1, entries 6 to 12). Changing the solvent from DCE to mesitylene further promoted the yield to 67% (Table 1, entry 14). Increasing or decreasing the temperature showed inferior results (Table 1, entries 15 and 16). No product was obtained in the absence of the catalyst (Table 1, entry 17). Replacement of the palladium catalyst by Lewis acids, Cu(OTf)2 or Sc(OTf)3, all showed no reaction, ruling out the Friedel–Crafts reaction pathway (Table 1, entries 18 and 19).
Entry | Catalyst (10 mol%) | Solvent | Additive | Yieldb (%) |
---|---|---|---|---|
a Reaction conditions: 1a (0.20 mmol, 1.0 equiv.), methyl acrylate 2a (0.40 mmol, 2.0 equiv.), Pd catalyst (0.020 mmol, 10 mol%), Cu(OAc)2 (0.40 mmol, 2.0 equiv.) and additive were reacted in solvent (2.0 mL) at 120 °C for 20 h under N2 atmosphere. DCE (1,2-dichloroethane). b NMR yield with CH2Br2 as internal standard. c Isolated yield. d 140 °C. e 90 °C. | ||||
1 | Pd(OAc)2 | DCE | None | 11 |
2 | Pd(OTFA)2 | DCE | None | <10 |
3 | Pd(CH3CN)4(BF4)2 | DCE | None | 14 |
4 | Pd(CH3CN)2(OTs)2 | DCE | None | 8 |
5 | Pd(dba)2 | DCE | None | <5 |
6 | Pd(CH3CN)4(BF4)2 | DCE | Na2CO3 (100 mol%) | 19 |
7 | Pd(CH3CN)4(BF4)2 | DCE | K2CO3 (100 mol%) | 19 |
8 | Pd(CH3CN)4(BF4)2 | DCE | Cs2CO3 (100 mol%) | 15 |
9 | Pd(CH3CN)4(BF4)2 | DCE | K3PO4 (100 mol%) | 9 |
10 | Pd(CH3CN)4(BF4)2 | DCE | NaOPiv (100 mol%) | 30 |
11 | Pd(CH3CN)4(BF4)2 | DCE | KOPiv (100 mol%) | 30 |
12 | Pd(CH3CN)4(BF4)2 | DCE | NaOPiv (80 mol%) | 33 |
13 | Pd(CH3CN)4(BF4)2 | Decalin | NaOPiv (80 mol%) | 53 |
14 | Pd(CH3CN)4(BF4)2 | Mesitylene | NaOPiv (80 mol%) | 67 (70)c |
15d | Pd(CH3CN)4(BF4)2 | Mesitylene | NaOPiv (80 mol%) | 48 |
16e | Pd(CH3CN)4(BF4)2 | Mesitylene | NaOPiv (80 mol%) | 38 |
17 | None | Mesitylene | NaOPiv (80 mol%) | 0 |
18 | Cu(OTf)2 | Mesitylene | NaOPiv (80 mol%) | 0 |
19 | Sc(OTf)3 | Mesitylene | NaOPiv (80 mol%) | 0 |
With the reaction conditions in hand, we first investigated the substrate scope of phenol derivatives (Table 2). Both electron-rich and electron-deficient phenols could be tolerated, and electron-deficient ones (3ea, 3ga, 3ka, 3ja, 3la, 3oa, 3qa) are better than electron-rich ones (3fa, 3na). This kind of electronic effect complements the developed methods, which makes this transformation synthetically useful.14e,27b,31 Apart from alkyl (3aa, 3ba, 3ma) and aryl (3da) substituents, functional groups like nitro (3ea), halide (3ga to 3ia), acetyl (3ja), formyl (3ka), and ester (3la) could all be well tolerated, providing access to synthesize more complex molecules by further transformation of these functional groups.32Meta-substituted phenols underwent this reaction selectively at the less sterically hindered position (3ma to 3oa), avoiding the difficult separation of regioisomers.27b,e Di-substituted phenols also reacted smoothly to give moderate to good yields (3ra to 3ua). The tolerance of the OTs group (3va) also provided the possibility for further cross-coupling reactions.26a,33 It should be noted that this reaction can even be conducted on a 1.0 mmol scale without a decrease in yield (3aa), further suggesting its synthetic applications.
a Reaction conditions: 1 (0.20 mmol, 1.0 equiv.), methyl acrylate (0.40 mmol, 2.0 equiv.), [Pd(CH3CN)4](BF4)2 (0.020 mmol, 10 mol%), Cu(OAc)2 (0.40 mmol, 2.0 equiv.) and NaOPiv (0.16 mmol, 0.80 equiv.) were reacted in mesitylene (2.0 mL) at 120 °C for 20 h under N2 atmosphere. Isolated yields were reported. b 1.0 mmol scale in 10.0 mL of mesitylene. |
---|
![]() |
Next, we began to investigate the substrate scope of the alkenes (Table 3). Variation of the protecting groups of the ester showed a reduced activity of more sterically hindered acrylate (Table 3, entries 1 to 6). Apart from mono-substituted alkenes, 1,2-di-substituted alkenes can also show acceptable to moderate yields (Table 3, entries 7 to 9). The reactivity of methyl cinnamate (2g) provided a method to synthesize 4-aryl coumarins with bioactivity.34 However, the 1,1-di-substitued alkenes showed no reactivity (Table 3, entry 10).
Entry | 2 | 3 | Yield |
---|---|---|---|
a Reaction conditions: 1 (0.20 mmol, 1.0 equiv.), alkene (0.40 mmol, 2.0 equiv.), [Pd(CH3CN)4](BF4)2 (0.020 mmol, 10 mol%), Cu(OAc)2 (0.40 mmol, 2.0 equiv.) and NaOPiv (0.16 mmol, 0.80 equiv.) were reacted in mesitylene (2.0 mL) at 120 °C for 20 h under N2 atmosphere. Isolated yields were reported. b NMR yield. | |||
1–6 |
![]() |
![]() |
R = Me, 70%b; Bu, 52%b |
R = Et, 32%b; Ph, 5%b | |||
R = tBu, 20%b; Bn, 36%b | |||
7 |
![]() |
![]() |
R1 = Ph, 3sg, 57% |
8 |
![]() |
R1 = Me, 3sh, 37% | |
9 |
![]() |
R1 = nPr, 3si, 21%b | |
10 |
![]() |
![]() |
3sj, Trace |
When [1,1′-biphenyl]-4,4′-diol (4) was selected as the substrate, di-functionalization occurred and the product 5 with more complexity was obtained in 34% yield (Scheme 3, Eq. 1). Interestingly, for 2-phenylphenol (6) as substrate, the compound 8 was produced via an oxidative alkenylation/oxidative cyclization sequence rather than the alkenylation/Michael addition product reported in the literature (Scheme 3, Eq. 2).35
To verify whether this reaction was initiated by alkenylation or esterification, we conducted the reaction with phenyl acrylate (9) or (E)-methyl 3-(2-hydroxyphenyl)acrylate (10) as starting materials in parallel with the standard reaction (Scheme 4, Eq. 3–5). (E)-Methyl-3-(2-hydroxyphenyl)acrylate (10) can be converted to the product in comparable yield with the standard reaction while only a trace amount of product can be obtained from phenyl acrylate, indicating the C–H bond alkenylation occurs first. A tandem process of C–H activation, olefin insertion, proto-demetalation, condensation and oxidation was also ruled out by using 11 (chroman-2-one) as the starting material under the standard conditions (Scheme 4, Eq. 6).36
An intramolecular deuterium labeling experiment showed a kinetic isotopic effect (KIE) value of 1.9 (Scheme 5, Eq. 7), showing that the C–H activation was involved in the rate-determining step. When 4-methylphenol (1b) was conducted under the conditions in the presence of 4.0 equivalents of D2O but in the absence of alkene, no deuterium exchange we observed in the recovered starting material (Scheme 5, Eq. 8).
Competing experiments between phenols with different electronic factors showed that electron-rich phenols reacted faster than electron-deficient ones (Scheme 6).
Based on these results and the literature reports,13,14c,37 we proposed the mechanism as shown in Scheme 7. In the presence of base, phenol was transformed to its sodium salt I, which coordinated with the palladium catalyst to form intermediate II. Palladium 1,3-migration viaIII generated the C–H bond metalation intermediate IV. Alkene insertion and β-H elimination released the alkenylated product and the PdH species VI, which was oxidized to the Pd(II) species to furnish the catalytic cycle. Intramolecular esterification produced the final product.
With our method, several bioactive molecules can also be functionalized to synthesize more complex structures (Scheme 8). Estrone (1,3,5(10)-estratrien-3-ol-17-one)38 and Phth-Me-D-tyrosine39 all reacted smoothly under the standard conditions to afford moderate yields.40 In addition, with this methodology, several bio-active molecules can also be constructed (Scheme 9). Apart from the herniarin (7-methoxy-2H-chromen-2-one, 3na) with anti-inflammatory activity,41 deprotection of 2-oxo-2H-chromen-7-yl-4-methylbenzenesulfonate (3va) can also generate the natural umbelliferone molecule with anti-oxidant activity.17m,42
In summary, a palladium-catalyzed base-accelerated ortho-selective phenol C–H alkenylation reaction was developed without acid as solvent. The substrate scope for the phenol was very wide and both electron neutral and electron-deficient phenols were well tolerated, which is complementary to the previous methods to synthesize electron-rich coumarins. Several bioactive molecules were functionalized and several coumarins with bioactive or fluorescent properties were synthesized. Mechanistic studies showed that this reaction underwent C–H bond activation via metallation but not the Friedel–Crafts pathway. Expansion of this chemistry to synthesize other useful molecules is undergoing in our laboratory.
Footnotes |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3qo00010a |
‡ These authors contributed equally. |
This journal is © the Partner Organisations 2014 |