Xiaoqiang
Huang
a,
Xinyao
Li
a and
Ning
Jiao
*ab
aState Key Laboratory of Natural and Biomimetic Drugs, Peking University, School of Pharmaceutical Sciences, Peking University, Xue Yuan Rd. 38, Beijing 100191, China. E-mail: jiaoning@pku.edu.cn
bState Key Laboratory of Organometallic Chemistry, Chinese Academy of Sciences, Shanghai 200032, China
First published on 20th July 2015
A novel direct transformation of aliphatic terminal alkynes to alkenyl nitriles through the incorporation of a nitrogen atom into the simple hydrocarbons has been reported. The usage of inexpensive copper catalyst, O2 as the sole oxidant, broad substrate scope as well as feasibility for “late-stage modification” make this protocol very promising. Mechanistic studies including DFT calculation demonstrate a novel 1,2-hydride shift process for this novel nitrogenation reaction.
Recently, novel transformation of simple alkynes has been disclosed through the assistance of transition metals. Yamamoto's group significantly developed palladium/acid catalyzed alkylation and hydroamination reaction of internal alkynes with nucleophiles (Scheme 1, A).10 By using a Rh(I)/phosphine ligand/benzoic acid catalyst system, Breit and co-workers pioneeringly achieved the intermolecular coupling of aliphatic terminal alkynes with carboxylic acids and sulfonyl hydrazides under redox-neutral conditions (Scheme 1, B).11 Despite these breakthroughs, new catalytic systems and new strategies are highly desirable to disclose novel transformations of aliphatic terminal alkynes.
![]() | ||
Scheme 1 Direct transformation of simple alkynes involving the cleavage of a propargylic C(sp3)–H bond. |
Herein, we report a novel Cu-catalyzed aerobic oxidative transformation of simple terminal alkynes to alkenyl nitriles (Scheme 1, C). In this present chemistry: (1) a very simple hydrocarbon is successfully converted into an N-containing compound through the incorporation of a nitrogen atom into the substrate; (2) inexpensive Cu-catalyst, the green molecular oxygen oxidant, as well as the broad substrate scope make this protocol very attractive and low-cost; (3) a novel propargylic C(sp3)–H bond cleavage through 1,2-H shift mechanism is proved. (4) DFT calculation reasonably explains the mechanism and the stereoselectivity of products.
Entry | Variation from standard conditions | Yieldb (%) |
Z![]() ![]() |
---|---|---|---|
a Standard conditions: 1a (0.40 mmol), TMSN3 (0.80 mmol), CuBr (0.08 mmol), pyridine (0.80 mmol) and NaOAc (0.40 mmol) in PhCl (2.0 mL) under O2 (balloon) was stirred at 90 °C for 48 h. b Isolated yields. c Determined by 1H NMR measurement of the crude mixture. DMEDA = N,N′-dimethyl-1,2-ethanediamine. | |||
1 | None | 78 |
63![]() ![]() |
2 | Ar instead of O2 | Trace | — |
3 | Without CuBr | 0 | — |
4 | Cu(OAc)2 instead of CuBr | 49 | 69![]() ![]() |
5 | CuBr2 instead of CuBr | 62 | 67![]() ![]() |
6 | 10 mol% CuBr was employed | 45 | 65![]() ![]() |
7 | Without pyridine | 0 | — |
8 | DMEDA instead of pyridine | 0 | — |
9 | L-proline instead of pyridine | 0 | — |
10 | 0.4 equiv. pyridine was employed | 26 | 69![]() ![]() |
11 | Without NaOAc | 48 | 64![]() ![]() |
12 | LiOAc instead of NaOAc | 48 | 64![]() ![]() |
13 | NaOMe instead of NaOAc | 47 | 67![]() ![]() |
With the optimal conditions in hand, we next investigated the substrate scope of this transformation. This reaction exhibited a good functional group compatibility (Table 2). Long-chain-alkyl substituted alkynes were successfully transformed to the corresponding alkenyl nitriles in good yields (2a–2e). Notably, propargylic 3° C–H of 1f could be cleaved, giving 2f in 61% yield. To our satisfaction, terminal alkyne 1g, bearing a TBDMS protected hydroxyl group, worked well (2g, 68%). Remarkably, linkages, including ether bonds (2h–2n) and ester bonds (2o–2q), did not reduce effectiveness. Several functional groups (trifluoromethyl, chlorine, vinyl and thienyl) were well tolerated in the present catalytic system. Furthermore, reasonable yields were obtained for alkynes containing phthalimide and sulfonamide group, respectively (2r–2s). Interestingly, C–H bond adjacent to internal ethynyl group was inactive, which leads to the high regioselectivity.
Entry | 2 | Yield of 2b (%) |
Z![]() ![]() |
|
---|---|---|---|---|
a Standard conditions: see entry 1, Table 1. b Isolated yields. c Determined by 1H NMR measurement of the crude mixture. d Two portions of TMSN3 (0.60 mmol) were added every 24 h. | ||||
1 |
![]() |
n = 3 | 78 (2a) | 65![]() ![]() |
2 | n = 2 | 44 (2b) | 67![]() ![]() |
|
3 |
![]() |
m = 7 | 76 (2c) | 61![]() ![]() |
4 | m = 6 | 62 (2d) | 66![]() ![]() |
|
5 |
![]() |
63 (2e) | 60![]() ![]() |
|
6 |
![]() |
61 (2f) | — | |
7 |
![]() |
R1 = TBDMS | 68 (2g) | 69![]() ![]() |
8 | R1 = n-C9H19 | 59 (2h) | 69![]() ![]() |
|
9 |
![]() |
R2 = C6H5 | 60 (2i) | 68![]() ![]() |
10 | R2 = 4-MeOC6H4 | 50 (2j) | 69![]() ![]() |
|
11 | R2 = 4-CF3C6H4 | 66 (2k) | 63![]() ![]() |
|
12 | R2 = 2-ClC6H4 | 71 (2l) | 66![]() ![]() |
|
13 | R2 = 1-naphth | 60 (2m) | 64![]() ![]() |
|
14 | R2 = 2-naphth | 57 (2n) | 64![]() ![]() |
|
15 |
![]() |
R3 = Me | 69 (2o) | 69![]() ![]() |
16d | R3 = ![]() |
40 (2p) | 64![]() ![]() |
|
17 |
![]() |
73 (2q) | 66![]() ![]() |
|
18 |
![]() |
65 (2r) | 69![]() ![]() |
|
19 |
![]() |
61 (2s) | 70![]() ![]() |
|
20 |
![]() |
46 (2t) | 66![]() ![]() |
Alkenyl nitriles are not only useful building blocks in synthetic chemistry but also important structure motifs commonly found in drugs.13 Moreover, late-stage modification is a highly valuable strategy for medicinal chemistry research.14 Therefore, several complex bioactive molecule derivatives were submitted to the optimal conditions (Table 3). Natural alcohol derivatives containing ester or ether linkages, such as menthol, borneol, nopol and cholesterol, worked well in the current transformation, generating the corresponding alkenyl nitriles in 49–73% yield (4a, 4d–f), respectively. Alkyne 3b that was prepared from antibacterial metronidazole afforded alkyl alkenyl nitrile 4b in 60% yield. Besides, terminal alkyne with a protected sugar moiety selectively underwent aerobic oxidation, giving nitrogenation product 4c in 67% yield. These results demonstrate that the present protocol could be applied in late-stage bioactive compound modification.
To gain mechanistic insight into this transformation, some control experiments were conducted under the standard conditions. Allene 5, which could be generated from alkyne 1a, failed to afford nitriles under the present conditions (eqn (1)), indicating a novel mechanism different from Breit's works.11 In addition, propargylic azide 6 or allylic azide 7 could not furnish alkenyl nitriles either (eqn (2) and (3)). These results ruled out the possibility of 6 and 7 as intermediates of the transformation.15
![]() | (1) |
![]() | (2) |
![]() | (3) |
Considering that the current nitrogenation reaction could only be catalyzed by copper salt, and the Glaser–Hay homocoupling product3a could be detected in some cases, we postulated that copper acetylide might be an intermediate of this reaction. Although, no product was formed employing copper(I)-acetylide 8 as a substrate (eqn (4)), which might due to the aggregation of 8,162c could be obtained in comparative yield when 8 was used as a catalyst (eqn (5)). When C(sp)–H bond deuterated alkynes 1a-1-d1 was subjected to the reaction, no deuterium was detected in the product, which is in accordance with the existence of copper acetylide species (eqn (6)).
![]() | (4) |
![]() | (5) |
Furthermore, labeling experiment with propargylic C–H bond deuterated 1a-3,3-d2 was performed. To our surprise, nearly 100% incorporation of deuterium at the both α and β positions of the nitrile was observed (eqn (7)). Hence, the cleavage of propargylic C–H bond might proceed via a 1,2-hydride shift.17 Then, an intermolecular kinetic isotopic experiment was conducted giving the result of kH/kD = 2.2 (eqn (8)).18
![]() | (6) |
![]() | (7) |
![]() | (8) |
On the basis of all these results and previous reports, a proposed mechanism is depicted in Scheme 2. The reasonable first step is the formation of copper(I)-acetylide intermediate A.16 Then, copper triazolide B formed via Cu-catalyzed azide–alkyne cycloaddition (CuAAC)19 undergoes ring-opening reaction affording cuprated diazoimine C.20,21 The oxidation of C under aerobic conditions with assistance of pyridine gives α-diazonitrile D and regenerates the copper(I) catalyst.22 Subsequently, upon loss of dinitrogen D would afford carbene E or copper carbene F.23 Finally, 1,2-hydride shift of the carbene species generates the alkenyl nitrile.17 Alternatively, a mechanism with ethynyl azide could also be possible. Cu-catalyzed aerobic oxidative cross-coupling of terminal alkynes with TMSN3 might generate ethynyl azide G,3b,4 which is known to liberate dinitrogen leading to the formation of cyanocarbene species E.24
To further explore the stereoselectivity of the reaction, density functional theory (DFT) calculation investigation was carried out (Fig. 1).25 After the sequential CuAAC19 and ring-opening process,20,21 the α-diazonitrile INT1 is generated (see ESI† for details). The pyrolysis of INT1 has two pathways. In pathway A, the thermal induced release of N2 through TS1 requires an activation free energy of 22.3 kcal mol−1 to give cyanocarbene carbene INT2. Alternatively, INT2 generated from ethynyl azide could not be excluded.24 The subsequent 1,2-hydride shift process17viaZ-TS2 and E-TS2 almost barrierlessly delivers Z-2 and E-2, respectively. It is noteworthy that the energy barrier gap between Z-TS2 and E-TS2 is insignificant (only 0.3 kcal mol−1), which might be due to the similar steric hindrance between hydrogen and cyano group. The calculated Z:
E ratio of 2via pathway A is predicted to be 64
:
36, which is qualitatively consistent with the experimentally observed 66
:
34 Z
:
E ratio for this reaction.
In alternative pathway B, Cu(I) catalyst can induce the Cu–C bond formation on INT3 with the release of N2 through TS3 in a stepwise manner, which requires an activation free energy of 22.3 kcal mol−1 to form Cu-carbene INT4.23 The subsequent 1,2-H shift process17viaZ-TS4 and E-TS4 also barrierlessly furnishes Z-2 and E-2, respectively. Notably, Z-TS4 is also only 0.7 kcal mol−1 lower in energy than E-TS4, which is corresponding to a 76:
24 Z
:
E ratio of 2, in good agreement with the experimental observation.
Moreover, E-2 is only 0.3 kcal mol−1 lower in energy than Z-2, indicating that the Z to E isomerization of alkenyl nitriles 2 is short of driving force thermodynamically. These results could explain why the E:
Z ratio of the products is so difficult to optimize whether by dynamic or thermodynamic means.
Footnote |
† Electronic supplementary information (ESI) available: Characterization data and experimental procedures. See DOI: 10.1039/c5sc02126j |
This journal is © The Royal Society of Chemistry 2015 |