A Pd-catalyzed highly selective three-component protocol for trisubstituted allenes

Abstract

Herein we report the first example of a Pd-catalyzed highly selective three-component reaction of alkynyl-1,4-diol dicarbonates, organoboronic acids, and malonate anions for the efficient synthesis of trisubstituted 2,3-allenyl malonates not readily available by the known protocols. The reaction demonstrates an excellent regio- and chemo-selectivity for both the oxidative addition referring to the two C–O bonds and the subsequent coupling with the nucleophile with a remarkable functional group compatibility. A series of control experiments confirm a unique mechanism involving β-O elimination forming alka-1,2,3-triene and the subsequent insertion of its terminal C C bond into the Ar–Pd bond.

A novel three-component protocol to synthesize trisubstituted allenes from 2-alkynyl-1,4-diol dicarbonate with organoboronic acid and malonate has been developed.

Introduction

Allenes have attracted significant attention¹ in the fields of natural products,² modern organic synthesis,^3,4 medicinal chemistry,⁵ and materials science,⁶ and thus, much attention has been paid to the development of new methodologies for the syntheses of allenes from readily available starting chemicals.⁷ Among them, a two-component reaction between readily available alkynes with an appropriate propargylic leaving group and organometallic reagents has been developed as a straightforward method for the syntheses of allenes⁸ (Scheme 1a). We envisioned a new concept of Pd-catalyzed reaction of readily available 2-alkynyl-1,4-diol dicarbonate with an organometallic reagent and a nucleophile for allene synthesis (Scheme 1b). As we know the Pd-catalyzed reaction of 2-alkynyl-1,4-diol dicarbonate with organometallic reagents would afford the double coupling products, 1,3-dienes C.⁹ Thus, the challenges of this strategy are (1) the regioselectivity of the oxidative additions of the two different C–O bonds in the dicarbonate; (2) the selectivity issue for the formation of different allenes A-1 or A-2, alkynes B-1 or B-2, or 1,3-conjugated dienes C-1 or C-2 (Scheme 1b). Herein, we report our realization of the first example of a Pd-catalyzed three-component reaction of 2-alkynyl-1,4-diol dicarbonates with organoboronic acids and malonates, affording trisubstituted allenes 4 (ref. 10 and 11) exclusively, enjoying an excellent regio- and chemo-selectivity with an unprecedented mechanism (Scheme 1c). Such A-1-type allenyl malonates have been demonstrated as highly versatile building blocks for natural allene product synthesis¹² and are traditionally prepared via the transition metal catalyzed reactions of 2,3-allenol derivatives with malonates.^13,14 Although the precursors for 2,4-disubstituted 2,3-allenol derivatives, 2,3-allenols, are conveniently available via allenation of the propargyl alcohol (ATA) with aldehydes^15,16 (Scheme 1d), the current method is highly efficient and diverse due to the readily availability of the three starting materials and irreplaceable due to the inaccessibility of D-type allenols via the ATA reaction.

Scheme 1. Approaches to trisubstituted allenes.

Results and discussion

At the outset, we examined the reaction of 2-alkynyl-1,4-diol dicarbonate 1a, phenylboronic acid 2a, and dimethyl malonate 3a in the presence of [Pd(allyl)Cl]₂, SPhos (L1),^9a and K₂CO₃ (Table 1, entry 1) at room temperature in THF for 22 h. To our delight, the allene product 4aaa was obtained albeit in only 36% yield with 56% yield of an unexpected by-product, 2,3-allenyl methyl ether 5aa. Other allenes A-2, alkynes B-1 and B-2, or conjugated dienes C-1 and C-2 were not detected, indicating exclusive chemo- and regio-selectivity. It is quite surprising to note that the oxidative addition occurred with the sterically more crowded secondary C–O bond. The study on the ligand effect indicated that although L2 or L3 led to a better yield for allene product 4aaa, a small amount of dicarbonate 1a was also recovered (Table 1, entries 2 and 3). No reaction occurred with bulkier ligand L4 (Table 1, entry 4). Since Gorlos-Phos (L5·HBF₄) and LB-Phos (L6·HBF₄)¹⁷ were fully ineffective for this reaction (Table 1, entries 5 and 6), L1 was applied for further screening. The base was proven to be essential (Table 1, entry 7) and slightly increasing the loading of K₂CO₃ would improve the yield of 4aaa to 41% (Table 1, entry 8). By running the reaction at 50 °C, the yield of 4aaa could be improved to 77% together with 15% of 2,3-allenyl methyl ether 5aa (Table 1, entries 9–11). Solvent screening led to the observation that DCE was the best furnishing product 4aaa in 88% yield (Table 1, entries 12–16). A reaction with 1.0 mmol furnished a slightly better result with 4aaa being isolated in 91% yield (Table 1, entry 17).

Optimization of the reaction conditions^a.

Entry	Ligand	Solvent	Temp.	Yield of 4aaa/1a recovered/yield of 5aab %
1	L1	THF	r.t.	36/0/56
2	L2	THF	r.t.	49/10/24
3	L3	THF	r.t.	48/23/12
4	L4	THF	r.t.	0/71/0
5	L5·HBF4	THF	r.t.	0/90/0
6	L5·HBF4	THF	r.t.	0/84/0
7c	L1	THF	r.t.	0/99/0
8d	L1	THF	r.t.	41/0/53
9d	L1	THF	30 °C	46/0/49
10d	L1	THF	40 °C	64/0/26
11d	L1	THF	50 °C	77/0/15
12d,e	L1	DCE	50 °C	88/0/0
13d,e	L1	Toluene	50 °C	32/0/53
14d,e	L1	EtOAc	50 °C	83/0/4
15d,e	L1	CH3CN	50 °C	69/0/0
16d,e	L1	1,4-Dioxane	50 °C	80/0/0
17d,e,f	L1	DCE	50 °C	94 (91)g/0/0

^aReaction conditions: 1a (0.2 mmol), 2a (1.0 equiv.), 3a (1.5 equiv.), [Pd(allyl)Cl]₂ (2.5 mol%), ligand (12 mol%), and base in solvent unless otherwise noted.

^bDetermined by ¹H NMR analysis of the crude product using CH₂Br₂ as the internal standard.

^cThe reaction was run in the absence of K₂CO₃.

^d2.0 equiv. of K₂CO₃ were used.

^eThe reaction was run for 24 h.

^fThe reaction was carried out on a 1 mmol scale.

^gIsolated yield.

After establishing the optimal conditions, we turned to investigate the scope of 2-alkynyl-1,4-diol dicarbonates 1 (Table 2). The reaction could be applied to a wide range of substrates 1 bearing either electron-poor (1c–1f) or electron-rich (1g–1h) aryl groups (R¹) generating the corresponding allene products 4 in 67–88% yields. Interestingly, ortho-fluoro (1i) and meta-fluoro (1j) substituted substrates were accommodated to afford the multi-functionalized allene products 4iaa and 4jaa. Delightedly, commercial drug adapalene derived 2-alkynyl-1,4-diol dicarbonate 1k could also afford the desired allene 4kaa in an excellent yield. Moreover, R¹ may also be an alkyl group (4laa–4naa). Next, the substrate scope of the organoboronic acid 2 was studied. Obviously, the electronic nature of the organoboronic acids has a limited impact on the reactivity affording the allene products 4aba–4aka in up to 90% yields. Bioactive molecules, such as ibuprofen and estrone, modified organoboronic acids could also be applied, delivering products 4ala and 4ama in moderate yields. The practicality was demonstrated by the gram scale reaction of 1a with 2a and 3a affording 3.5 g of product 4aaa in 90% yield under 50 °C for 24 h. The structure of 4aaa was further confirmed by single-crystal X-ray diffraction.¹⁸

The scope studies^a.

^aUnless otherwise indicated, the reaction was performed with 1 mmol of 1, 1.0 equiv. of 2, 1.5 equiv. of 3, 6 or 8, 2.5 mol% of [Pd(allyl)Cl]₂, 12 mol% of SPhos, and 2.0 equiv. of K₂CO₃ in DCE (0.1 M) at 50 °C for 24 h. Yields of isolated products are given.

^b5 mol% of [Pd(allyl)Cl]₂ and 24 mol% of SPhos were used. The reaction was run for 48 h.

^c3.5 mol% of [Pd(allyl)Cl]₂ and 15 mol% of SPhos were used. The reaction was run for 36 h.

^d3.5 mol% of [Pd(allyl)Cl]₂, 15 mol% of SPhos, and 2.0 equiv. of 3a were used. The reaction was run for 36 h.

^eThe reaction was run for 36 h.

^f3.5 mol% of [Pd(allyl)Cl]₂ and 15 mol% of SPhos were used. The reaction was run for 48 h.

^g2.0 equiv. of Cs₂CO₃ were used instead of K₂CO₃. The reaction was run for 48 h.

^hThe reaction was performed with 1 mmol of 1, 1.0 equiv. of 2a, 2 equiv. of 3, 3.5 mol% of [Pd(allyl)Cl]₂, 15 mol% of SPhos, and 2.0 equiv. of Cs₂CO₃ in DCE (0.1 M) at 50 °C for 48 h.

With 2-substituted malonates 3, K₂CO₃ should be replaced with Cs₂CO₃ for complete transformation obviously due to the steric effect. Methyl, benzyl, allyl and 2,3-allenyl groups may be introduced into the 2-position of malonates to prepare allenes 4aab–4aac, 1,6-vinylallene 4aad, and bisallene 4aae in moderate to excellent yields. The protocol also works for 2-acetamidomalonate 3f and 2-fluoromalonate 3g. Besides, bis(phenylsulfonyl)methane 6a and dibenzylamine 8a¹⁹ could also serve as suitable nucleophiles to deliver the desired products 7 and 9 in moderate to high yields (42–73%). Diallylamine 8b could provide the desired product 9aab in 16% yield with 44% recovery of 1a. Other amines such as tetrahydroisoquinoline, morpholine, pyrrolidine, N-allylmethylamine and benzylamine were also studied; however, no corresponding allenyl amine products were obtained.

To demonstrate the synthetic potential of this methodology, synthetic transformations of 4caa have been demonstrated (Scheme 2): firstly, Krapcho decarboxylation was achieved, affording 4,5-allenoate 10 in 73% yield;²⁰ reduction of 10 with LiAlH₄ formed 4,5-allenol 11 in 74% yield; 4,5-allenoic acid 12 could be yielded by its treatment with LiOH; next, Weinreb amide 13 could be obtained in high yield by using N,O-dimethylhydroxylamine hydrochloride; gold(i)-catalyzed butyrolactone (E)-14 in 96% yield with an exclusive E-selectivity.²¹

Scheme 2. Synthetic applications.

To gain insight into the mechanism, a couple of control experiments were performed (Scheme 3): at first, the reaction of 2-alkynyl-1,4-diol dicarbonate 1a with 1 equiv. of phenylboronic acid 2a under the standard conditions for 24 h afforded 2,3-allenyl methyl ether 5aa in 81% yield (Scheme 3(a)); by increasing the loading of 2a to 2 equiv., the conjugated diene 16 was obtained in 36% yield (E/Z = 1.25 : 1) together with 60% yield of allenylic ether 5aa (Scheme 3(b)); a control experiment showed that only 4% yield (E/Z = 1 : 1) of conjugated diene 16 was formed when 2 equiv. of 2a were applied under the standard reaction conditions together with allenylic malonate 4aaa being formed in 90% yield (Scheme 3(c)); interestingly, when 1a was treated with phenylboronic acid 2a (1 equiv.) for 3 h, ether 5aa was already formed in 79% NMR yield. The subsequent addition of malonate 3a (1.5 equiv.) led to the formation of ether 5aa in 73% yield, exclusively (Scheme 3(d)); when the reaction of 2-alkynyl-1,4-diol dicarbonate 1a and malonate 3a was conducted in the absence of phenylboronic acid 2a, no products were observed with complete recovery of 1a, indicating the importance of phenylboronic acid for the transformation (Scheme 3(e)). Furthermore, the reaction of ether 5aa with 1.5 equiv. of 3a under the standard conditions failed to afford the product 4aaa, indicating that 2,3-allenyl methyl ether 5aa is certainly not the intermediate for the formation of 4aaa (Scheme 3(f)). We then deliberately prepared the envisioned intermediate, methyl 2,3-allenyl carbonate 15. Its reaction under the standard conditions afforded 6% yield of 1,3-enyne 17 with 81% recovery (Scheme 3(g)); the reaction of allenylic carbonate 15 with malonate 3a under the standard conditions in 24 h afforded 4aaa in a lower yield of 62% without the formation of 1,3-enyne 17 or allenylic ether 5aa, as compared with the three-component reaction (Scheme 3(h)). The reaction of 1a in the presence of phenylboronic acid 2a and malonate 3a was also monitored and the formation of 15 was not observed during the whole reaction time (Scheme 3(i)). These experiments (Scheme 3(g)–(i)) indicated that 15 is not the intermediate product for the formation of either 4aaa or ether 5aa under the current reaction conditions.

Scheme 3. Mechanistic studies.

Based on these experimental mechanistic studies, we proposed a plausible mechanism as shown in Scheme 4: initially, the oxidative addition occurred unexpectedly with the sterically more crowded secondary C–O bond exclusively to afford the allenyl palladium species Int-I rather than its regioisomer Int-I′, most probably due to the steric effect in these two intermediates. Int-I would undergo transmetalation exclusively with boronic acid 2a to generate the intermediate Int-II. Int-II would prefer β-O elimination over reductive elimination, affording 1,2,3-butatriene-coordinated palladium intermediate Int-III. Reductive elimination could be ruled out by the monitoring experiment showing that the formation of 15 was not observed during the whole reaction time. Subsequently, highly regioselective insertion of the terminal C C bond of the 1,2,3-butatriene into Ph-PdL_nOMe formed methylene-π-allyl Pd species Int-IV. After deprotonation of malonate 3a and nucleophilic attack, product allene 4aaa was afforded and the catalytically active Pd(0) species was regenerated, finishing the catalytic cycle. In the absence of malonate, reductive elimination would afford the methyl ether 5aa.

Scheme 4. Proposed mechanism.

Conclusions

In conclusion, a novel protocol to synthesize trisubstituted allenes from 2-alkynyl-1,4-diol dicarbonate with organoboronic acid and malonate has been developed. The method utilizes readily available coupling partners, affording trisubstituted double functionalized allenes in an excellent regio- and chemo-selectivity while enjoying a broad substrate scope and step-economy. It should be noted that such 2,2,4-trisubstituted allenyl malonates require lengthy steps via known protocols (see Scheme S1†). A unique mechanism involving β-O elimination forming 1,2,3-alkatrienes has been proposed. This reaction provides a new entity for the syntheses of multiple substituted allenes. Further studies, including the asymmetric version of this reaction, are being conducted in our laboratory.

Data availability

All experimental data and detailed procedures are available in the ESI.†

Author contributions

S. M. directed the research and developed the concept of the reaction with C. L., who also performed most of the experiments and prepared the ESI. Z. Z. also performed some of the experiments. C. L. and S. M. checked the experimental data. C. L. and S. M. wrote the manuscript with contributions from the other authors.

Conflicts of interest

There are no conflicts to declare.