|Home | About | Journals | Submit | Contact Us | Français|
The gold(I)-catalyzed cycloisomerization 1,5-enynes and 1,4-allylallenes to tetracyclododecane and tetracyclotridecane derivatives is reported. Complexation of the cationic gold(I) complex to either the alkyne or the allene moiety induces an intramolecular addition of the alkene leading to gold(I)-stabilized carbenoid intermediate. This intermediate undergoes a formal sp3-C-H insertion to generate the tertacyclic adduct. A series of deuterium labeling experiments show that the C-H functionalization step proceeds with an inverse kinetic isotope effect.
Cationic gold(I) complexes have emerged as powerful π-acids capable of promoting a diverse range of enyne cycloisomerization reactions.1 In these reactions, the gold catalyst not only serves to activate the alkyne towards nucleophilic addition,2 but also plays a role in stabilizing the cationic intermediates produced in the cyclization event.3 In many of these cases, the stabilization arises from interaction of the gold catalysts with the directly bonded carbocation.4 Thus, transformations in which these intermediates display reactivity that is reminiscent of electrophilic metal-stabilized carbenes have been developed.5,6 However, in contrast to other electrophilic metal-carbenoid species, insertion of the gold-stabilized cationic intermediates into sp3-C-H bonds is rare.7,8 We envisioned that access to such a transformation might provide insight into the importance of the gold complex, not only in stabilizing of the cationic intermediates, but also the subsequent reactivity of these species.9
With the aim of examining the potential of cationic gold(I)-carbenoid intermediates to participate in this class of reactions, we revisited the gold-catalyzed cycloisomerization of 1,5-enynes A.10 While we had previously observed that gold-catalyzed cycloisomerizations of cyclobutane and -pentane derived enynes (A, n=1,2) were terminated by a ring expansion to afford C,10b we postulated that larger more flexible rings might allow for an intramolecular C-H insertion via gold-carbenoid intermediate B (eq 1).
In order to explore this hypothesis, 1,5-enyne 1a possessing cycloheptyl skeleton at the C4 position was allowed to react under our standard conditions (2% (Ph3P)AuCl/AgSbF6) for cycloisomerization of 1,5-enynes (eq 2). We were pleased to find that the reaction proceeded smoothly to provide tetracyclic compound 2a in 75% yield after one hour. On the basis the observation that electron donating ligands are preferred for reactions involving trapping of the gold(I)-carbenoid intermediates, we examined N,N-bis(2,6-diisopropylphenyl)imidazol-2-ylidene (IPr) and tri(tert-butyl)phosphinegold(I)chloride as catalysts. Gratifyingly, both of these complexes showed improved reactivity (10–15 minutes) in the cycloisomerization reaction and produced 2a in better yield. Notably, in all cases, neither competing deprotonation from gold-carbenoid intermediate B to afford spiro[5.6]dodeca-1,4-diene nor 1,2-alkyl shift to provide ring expansion product (C) were observed.
Under these optimized reactions conditions, a number of cycloheptyl or cyclooctyl-substituted 1,5-enynes 1 undergo the gold(I)-catalyzed cycloisomerization/C-H insertion reaction (Table 1). For example, in addition to terminal alkynes, substrates containing alkyl and aryl alkynes participate in the gold(I)-catalyzed reaction to afford the desired products in excellent yield (entries 2–4). Furthermore, the C-H insertion reaction occurs independent of the electronic nature of substituent at the acetylenic position (entries 5–9). Moreover, gold(I)-catalyzed cycloisomerization of 1k and 1l afforded bis-cyclopropanes 2k and 2l in excellent yield without competing functionalization of the aromatic C-H bonds (entries 10–12). The structure of 2l was unambiguously confirmed by X-ray diffraction analysis (Figure 1).
We considered that intermediate B might also be accessed through a gold(I)-catalyzed cyclization of a 1,4-enallenes.11 While the desired reaction did not occur at room temperature, we were pleased to find that 1,4-enallenes were converted into the anticipated cyclopropane adducts at slightly elevated temperatures (Table 2).
A proposed mechanism for these transformations is outlined in Scheme 1. Complexation of the cationic gold(I) complex to either the alkyne of the allene moiety induces an intramolecular addition of the alkene leading to gold(I)-stabilized cationic intermediate 4. This intermediate then undergoes a formal sp3-C-H insertion to generate the tetracyclic adduct.
To gain insight into the mechanism of the sp3-C-H functionalization, a double-labeling crossover experiment was carried out (eq. 3). Gold(I)-catalyzed cycloisomerization of a 1 : 1 mixture of d2-1l and 1k afforded d2-2l and 2k with complete deuterium incorporation exclusively in d2-2l. This experiment strongly suggests an intramolecular transfer of the sp3-C-H bond to gold(I)-stabilized cationic intermediate 4.
The fact that no H/D scrambling occurred during the cycloisomerization (eq 3), allowed us to directly measure a kinetic isotope effect by monitoring the reaction of a 1:1 mixture of deuterium and hydrogen labeled compounds. Thus, a small normal intermolecular kinetic isotope effect (kH/D) of 1.08 was measured for the gold-catalyzed cycloisomerization of 1l (eq 4). In order to interpret this result, we considered two possibilities: (1) the coordination/cyclization steps in the proposed mechanism (Scheme 1) are reversible and therefore the isotope effect is due (at least partially) to the C-H functionalization event, and (2) one of the first two steps is irreversible and therefore the small measured isotope effect is derived from the coordination/cyclization step. Moreover, the fact that a change in the C-H bond containing fragment from a 7-membered ring in 1l to an 8-membered ring in 1b (eq 5)12 had only a minor impact on the observed kinetic isotope lead us to conjecture that that these values are largely due to the coordination and/or cyclization steps. The observation that essentially no kinetic isotope effect was measured in the gold-catalyzed cycloisomerization of diastereomers of d1-1l (eq 6) effect is consistent with this possibility.
In order to probe this hypothesis, we examined the gold-catalyzed cyclization of d2-1d. In this case, the coordination/cyclization of d2-1d produces cationic intermediate 5 that can insert into either a C-H or C-D bond; this allows for direct measurement of the isotope effect of the C-H insertion step, irrespective of the reversibility of the first steps (eq 7). Given that this step involves cleavage of a C-H bond, we were surprised to find that an inverse kinetic isotope effect was measured for the tri(tert-butyl)phosphinegold(I)-catalyzed cycloisomerization of d2-1l. Moreover, similar values were obtained using cationic (IPr)gold(I) as the catalyst, independent of counterion.
To further verify this observation, the gold-catalyzed cycloisomerization of allene d2-3d was examined. Given that the identical gold-carbenoid 5 is postulated as an intermediate cycloisomerization of d2-3d, a similar kinetic isotope effect would be anticipated. Indeed, an inverse isotope effects were also measured for the phosphinegold(I)- and N-heterocycliccarbenegold( I)-catalyzed rearrangement of allenes d2-3d and d2-3e (eq 8).
Intramolecular C-H insertion reaction of metal-carbenoid complexes typically exhibit primary kinetic isotope effects (kH/D =1.1-3.1).13 Given that inverse primary kinetic isotope effects (kH/D = 0.93-0.89) were measured for the C-H insertion of the gold(I)-stabilized cationic intermediate, a mechanism analogous to those proposed for metal-carbenoid insertion14 into C-H bonds does not adequately account for the measured isotope effects. Significant normal kinetic isotope effects are also generally observed for the C-H bond undergoing hydride transfer to a carbocation.15 Thus, our measured inverse kinetic isotope effects for the C-H insertion suggest that a mechanism involving a simple hydride to a carbocation-like intermediate is unlikely; the experimental inverse isotope effects requires that transition state for hydride transfer have higher force constants for the C-H bond than for cationic intermediate 5.16 Alternatively, and in analogy to other metal-catalyzed C-H activation reactions,17 formation of a σ-complex between the hydrogen atom and cationic gold(I) preceding the hydrogen transfer event may account for the observed inverse isotope effect.
In conclusion, we have developed a gold(I)-catalyzed sequential cycloisomerization/sp3-C-H bond functionalization of 1,5-enynes and 1,4-allylallenes that provides tetracyclododecane and tetracyclotridecane derivatives. These transformations represent rare examples of sp3-C-H bond insertion by a cationic gold(I)-carbenoid intermediate. It is difficult to draw conclusions on the nature of the bonding and stabilization of the cationic intermediate by the gold catalyst from these experiments; however, the studies presented herein further support the hypothesis that the transition metal plays an integral role in the subsequent transformations of these cationic intermediates. Further study on the mechanism and scope of the reactivity of cationic gold(I)-carbenoid intermediates as well as application of this present strategy are ongoing and will be reported in due course.
Typical experimental procedures and spectral data for all new compounds. This material is available free of charge via the Internet at http://pubs.acs.org,
FDT thanks NIHGMS (RO1 GM073932), Merck Research Laboratories, Bristol-Myers Squibb, and Novartis for funding. YH gratefully acknowledges a Grant-in-Aid for Young Scientists (B) 19750071 for partial support. We thank Prof. R. G. Bergman (UC Berkeley) for helpful suggestions and discussions.