Download PDF
Synlett 2010(5): 692-706  
DOI: 10.1055/s-0029-1219527
ACCOUNT
© Georg Thieme Verlag Stuttgart ˙ New York

Gold-Catalyzed Reaction of Propargylic Carboxylates via an Initial 3,3-Rearrangement

Shaozhong Wang, Guozhu Zhang, Liming Zhang*
Department of Chemistry and Biochemistry, University of California, Santa Barbara, CA 93106
Fax: +1(805)8934120; e-Mail: zhang@chem.ucsb.edu;

Further Information

Publication History

Received 16 September 2009
Publication Date:
18 February 2010 (online)

 PDF Download Permissions and Reprints All articles of this category

Biographical Sketches

Shaozhong Wang was born in P. R. of China, in 1975. He received his BSc degree in Chemistry from Nanjing University in 1996 and his PhD in 2002 from the same university working with Prof. Yuefei Hu. He started his independent academic career at Nanjing University and was promoted to an associate professor in 2005. From September 2005 to August 2006, he carried out post-doctoral research with Prof. Liming Zhang at the University of Nevada, Reno. His research interests include methodology development and alkaloid synthesis.
Guozhu Zhang was born in Chongqing, P. R. of China, in 1977. He received his BSc and Master’s degrees from Nankai University in Organic Chemistry with Professor Jinpei Cheng in 2000 and 2003, respectively. He spent the next three years in Wuxi Pharmatec company, Shanghai, working as a synthetic chemist. In 2006, he moved to the University of Nevada, Reno, and joined Professor Liming Zhang’s group. In June 2009, he moved with the Zhang group to UC Santa Barbara. His research interests include transition-metal catalysis and natural product synthesis.
Liming Zhang was born in Pingxiang, P. R. of China, in 1972. He received his BSc degree in Chemistry from Nanchang University in 1993, his first Master’s degree in Organometallic Chemistry from Nankai University (P. R. of China) in 1996, and his second Master’s degree in Organic Chemistry from the University of Alabama with Professor Michael P. Cava in 1998. He obtained his PhD degree with Professor Masato Koreeda from the medicinal chemistry program at the University of Michigan in 2003 and then carried out a post-doctoral study with professor Sergey A. Kozmin at the University of Chicago. He started his independent academic career as assistant professor at the Department of Chemistry of the University of Nevada, Reno in July 2005 and continued at the University of California Santa Barbara in July 2009. His research ­interests include late-transition-metal-catalyzed reactions, natural product synthesis and medicinal chemistry.

Abstract:

Gold-catalyzed 3,3-rearrangement of propargylic carboxylates offers a versatile entry into a range of fascinating subsequent transformations, leading to various functional structures.

1 Introduction

1.1 Two Competing Mechanistic Pathways

1.2 Experimental Evidence for 3,3-Rearrangement

2 Synthetic Application via Gold-Containing Oxocarbenium Intermediates

2.1 General Considerations/Design

2.2 Using Indole-3-acetyl as the Acyl Group

2.3 Varying the R³ Group

2.3.1 Using Enyne Substrates (R³ = Alkenyl)

2.3.2 Using the TMSCH2 Group as R³

2.4 Using the Aryl Group for R¹ (R² = H)

2.5 Hydrolytic Treatment

2.5.1 Formation of Enones

2.5.2 Formation of α-Iodo/Bromoenones

2.6 Intramolecular Acyl Migration

2.7 Incorporating Gold(I)/Gold(III) Catalysis

2.7.1 Gold-Catalyzed Oxidative Homo-Coupling

2.7.2 Gold-Catalyzed Oxidative Cross-Coupling Reaction

2.7.3 Gold-Catalyzed Oxidative C-O Bond Formation

3 Attack of Au-Activated Carboxyallene at the β- or
γ-Position

3.1 Nucleophilic Attack at the γ-Position

3.2 Nucleophilic Attack at the β-Position

4 Carboxyallenes as Nucleophiles

5 Summary

Key words

rearrangements - carbocations - cyclizations - oxidations - allenes

1 Introduction

Propargylic carboxylate is a versatile structural motif that has been studied extensively in transition-metal catalysis, allowing access to a variety of functional structures. [³] Due to their easy and modular preparation, various additional functional groups can be readily incorporated to decorate the core structure and, importantly, be rendered to participate in reactions in an intramolecular manner, therefore offering highly functionalized products and increasingly diverse reaction manifolds.

Homogeneous gold catalysis has recently experienced dramatic development, delivering a large array of useful synthetic methods. [4] With a few exceptions, [5] homogeneous gold chemistry hinges on the exceptional capacity of gold complexes to activate alkynes and allenes toward nucleophilic attack, which is often the entry point of catalytic cycles. It comes with no surprise that propargylic carboxylates have been studied extensively in gold catalysis in the past few years.

This account attempts to survey in an inclusive manner recent developments in gold-catalyzed transformations of propargylic carboxylates via an initial 3,3-rearrangement. The competing 1,2-acyloxy migration will only be briefly discussed for comparison purpose.

1.1 Two Competing Mechanistic Pathways

With increasing experimental evidence, two divergent initial transformations in gold-catalyzed reactions of propargylic carboxylates have been proposed, namely 1,2-acyloxy migration [6] and 3,3-rearrangement [7-9] (Scheme  [¹] ). In the case of 1,2-acyloxy migration, an initial 5-exo-dig cyclization of the ester carbonyl group to the gold-activated alkyne leads to the formation of alkenyl gold carbene A [¹0] upon heterolytic fragmentation of the original ester bond; alternatively, a 6-endo-dig cyclization of the ester carbonyl group results in the formation of a six-membered, gold-containing oxocarbenium intermediate (i.e., B), which can undergo ring opening to form carboxyallene C.

In literature, alternative terms including ‘1,3-acyl migration’ and ‘3,3-sigmatropic rearrangement’ have been used to describe the 3,3-rearrangement. We prefer ‘3,3-rearrangement’, as: a) mechanistically this rearrangement is most likely to involve metal-containing intermediate B and not a concerted pericyclic reaction; b) the migration group is acyloxy instead of just the acyl group. Even the term ‘1,3-acyloxy migration’ does not convey mechanistic information that has been substantiated by calculations [¹¹] and experimental data. [¹²] Similarly, for the 1,2-acyloxy migration, it is incorrect to describe it as 1,2-acyl migration.

Scheme 1 Two competing mechanistic pathways for gold-catalyzed reactions of propargylic carboxylates

Recent calculations by Cavallo [¹³] suggested that these two pathways interconvert and a net 3,3-rearrangement may be the result of two consecutive 1,2-acyloxy migrations (Scheme  [¹] ), which however was not supported by an O¹8-labeling study by Toste. [¹²a]

Substitution patterns on the propargyl moiety play essential roles in dictating reaction pathways: in general, sterically and/or electronically unbiased substrates undergo reactions via an initial 3,3-rearrangement, while biased substrates prefer 1,2-acyl migrations. In the biased cases, R¹ is either H6 or an EWG, and R²/R³ are either aryl/H or both non-hydrogen groups. [6] For example, tertiary propargyl pivalate 1 with a terminal alkyne undergoes intermolecular cyclopropanation with alkenes in the presence of Ph3PAu+ (generated from Ph3PAuCl/AgSbF6, Equation  [¹] a). [6c] In this chemistry, the reactive intermediate, gold carbene D, [¹0] generated via 1,2-acyloxy migration, is formed selectively due to the drastic steric difference between the two ends of the propargyl moiety (i.e., H vs Me/Me). In a related platinum catalysis, the electron-withdrawing carboxylate promotes 1,2-acetoxy migration due to polarization of the C-C triple bond via resonance, eventually leading to pentannulation (Equation  [¹] b). [¹4]

Equation 1 Selective initial 1,2-acyloxy migrations in biased substrates

1.2 Experimental Evidence for 3,3-Rearrangement

The initial evidence for gold-catalyzed 3,3-rearrangement of propargylic carboxylates was reported as a side reaction by Miki, Ohe and Uemura during their pioneer discovery of gold(III)-catalyzed 1,2-acyloxy migration. [6a] In contrast to Equation  [¹] a, the use of gold(III) chloride (AuCl3) allowed the observation of acetoxyallene 2 as a minor product (Equation  [²] a). In our first study on gold-catalyzed reaction of propargylic indole-3-acetates, with a sterically relatively unbiased substrate (i.e., 3), carboxyallene 4 was isolated in 52% yield in the presence of catalytic AuCl3 (Equation  [²] b). [7a] Notably, no competing 1,2-acyloxy migration was observed in this case, offering evidence to the importance of the substrate substitution pattern on the reaction outcome. Similarly, using the Ph3Au+ cation, generated from Ph3PAuCl/AgSbF6, again only 3,3-rearrangement occurred. The formation of tetracyclic lactones like 5 is discussed in Section 2.2.

Equation 2 Evidence for gold-catalyzed 3,3-rearrangement of propargylic carboxylates

The ability of gold complexes to catalyze 3,3-rearrangements comes with no surprise, however. It has been well established in the literature that salts/complexes of other coinage metals (i.e., Ag [¹5] and Cu [¹6] ) as well as Pt [¹7] can catalyze this transformation. Surprisingly, however, only a few synthetic applications [¹5d] [e] [¹6] [¹7a] of this transformation had been reported before our study although carboxyallenes are densely functionalized and should be synthetically versatile. It is most likely due to the reversible nature of this rearrangement (Scheme  [¹] ) and in general only partial conversion of propargylic carboxylates into carboxyallenes can be achieved; in addition, their separation from propargylic carboxylates is challenging.

2 Synthetic Application via Gold-Containing Oxocarbenium Intermediates

2.1 General Considerations/Design

At the onset of our research in gold catalysis at the University of Nevada, Reno, we were interested in using allenes [¹8] as substrates. While gold complexes were shown to activate the cumulative double bonds efficiently, the field was in its infancy and only limited substrate scopes were under investigation. [¹9] We were particularly attracted to allenes with heteroatom(s) directly bonded to the 1,2-diene unit. Several advantages were anticipated: a) more functionalized products due to a higher level of substrate functionalization; b) offer potential regioselectivity in double-bond activation; c) added dimension of reactivities due to closely positioned heteroatom-containing groups. We expected that a rich array of reactivities and transformations could be discovered. However, the challenge was to gain rapid access to these types of functionalized allenes so that succinct synthetic methods could be developed.

Scheme 2 Gold-catalyzed tandem reaction of propargylic carboxylates via an initial 3,3-rearrangement: design

Carboxyallenes, where a carboxyl group is directly bonded to the 1,2-diene unit, are one class of functionalized allenes we were particularly interested in. Although their preparations are often complicated by equilibration with propargylic carboxylates, we anticipated that such equilibrium may be non-consequential as long as an irreversible downstream transformation can be incorporated in a competitive manner. In this approach, carboxyallenes do not need to be isolated and propargylic carboxylates become suitable substrates. Moreover, the gold catalyst used for the 3,3-rearrangement of propargylic carboxylates into carboxyallenes could be employed to catalyze further transformations of the in-situ formed allenes. As shown in Scheme  [²] , we anticipated that activation of carboxyallenes by the same homogeneous gold complex would generate oxocarbenium E. This intermediate contains multiple reaction sites including an activated acyl group, a nucleophilic Au-C(sp²) bond, a C-C double bond, and an electrophilic oxocarbenium center and could undergo various useful transformations. To sum up, this design would possess the following features:

- Propargylic carboxylates can be prepared modularly under mild reaction conditions via three-component coupling in one or two steps.

- Carboxyallene C is generated and used in situ. No isolation is needed.

- The same gold complex catalyzes multiple transformations.

- Substituents containing additional functionalities can be easily incorporated into already densely functionalized intermediate E.

2.2. Using Indole-3-acetyl as the Acyl Group

The first study to implement this general design was to append a carbon nucleophile in the acyl group to engage intramolecular nucleophilic attack of the oxocarbenium moiety in E, leading to the formation of C-C bonds (Figure  [¹] ).

Figure 1

An indole ring appeared to be ideal to us, and the corresponding indole-3-acetates can be readily prepared from commercially available indole-3-acetic acid. When indole-3-acetate 6 was treated with Ph3PAu+ SbF6 - (1 mol%), to our great surprise, highly functionalized tetracyclic cyclobutane 7 was formed in excellent yield and with excellent diastereoselectivity (Equation  [³] ). [7a]

Equation 3

Notable is that the exocyclic double bond of 7 has exclusive E-geometry. A scope study rendered additional examples of this chemistry and offered X-ray crystallographic data to support the structural assignment. The mechanism of this chemistry is proposed in Scheme  [³] : an initial gold-catalyzed 3,3-rearrangement should result in the formation of carboxyallene 8 in an equilibrium; subsequent activation of the allene by the very same cationic gold(I) catalyst would yield two oxocarbenium intermediates, i.e., the Z-isomer G and the E-isomer F, which differ in the geometric relationship between Ph3PAu and the R¹ group. We speculated that intermediates F and G should be in equilibrium via carboxyallene 8. As we had designed, an intramolecular nucleophilic attack by the electron-rich indole ring could readily explain the formation of the γ-lactone ring; the final cyclization to form the strained cyclobutane ring would afford the observed products. Several insights with this proposal mechanism are worth commenting on and offer general guidelines for further exploring this design:

- The exclusive formation of the E-exocyclic double bond indicates that oxocarbenium G is most likely the predominant, if not exclusively formed, intermediate. This can be rationalized by evoking detrimental A¹,³-strain in F while the cis arrangement of Ph3PAu and R¹ is sterically tolerable due to the long Au-C(sp²) bond. [²0] This preferred formation of Z-alkenylgold is supported by various later studies. [7c-g]

- For the first time, the nucleophilicity of the Au-C(sp²) bond leading to C-C-bond formation is suggested. This spurs the development of functionalization of such organometallic bonds instead of simple protodeauration.

This study offers strong support for the validity of the general design, and one of the examples (Equation  [²] b) offered support for the intermediacy of carboxyallene 8. An alternative route for the formation of G/F via electrocyclic ring opening of intermediate H, however, cannot be ruled out.

Scheme 3 Proposed mechanism for gold-catalyzed formation of tetracyclic cyclobutanes

A complete reactivity divergence was observed when platinum(II) chloride was used as catalyst under CO atmosphere (Equation  [4] ). [¹7b]

Equation 4

Scheme 4 Enynyl carboxylates as substrates

Equation 5

Instead of cyclobutanes, tetracyclic cyclopentenes were formed in excellent yields. Mechanistically, this reaction diverges at lactone intermediate I : instead of 4-exo-trig cyclization, cyclization to the iminium moiety using the distal end of the platinoalkene in a 5-exo-trig manner forms a 5-membered ring containing a platinum carbene, which is transformed into cyclopentene smoothly. This chemistry highlights the substantial difference in catalysis with platinum and gold complexes.

Equation 6

Equation 7

2.3. Varying the R ³ Group

2.3.1 Using Enyne Substrates (R³ = Alkenyl)

It was anticipated that intermediate E would become pentadienyl cation J (Scheme  [4] ) if R³ is an alkenyl group. Such an intermediate should undergo electrocyclic ring closure to afford cyclopentenylidene gold intermediate K, which could undergo 1,2-C-H insertion and hydrolysis to afford cyclopentenones (Nazarov-type reaction). Cyclopentenones are indispensable building blocks in organic synthesis and were indeed formed in this chemistry in mostly good yields using substrates containing an alkyl group at the propargylic position (e.g., Equation  [5] a). [7b]

Interestingly, when an aryl group was present (e.g., Ph in Equation  [5] b), in situ enol acetate hydrolysis did not occur and cyclopentadienyl acetate 9 was isolated in excellent yield. Beside the very mild reaction conditions (room temperature, open flask), a salient feature of this chemistry is the excellent regioselectivity, and only one isomer was detected, in contrast to conventional Nazarov reaction. The reaction mechanism was studied computationally, and a novel role of water in facilitating the reaction was revealed. [¹¹]

With an additional C-C double bond appropriately tethered, Fensterbank and Malacria showed that the gold carbene intermediate K could cyclopropanate the tethered double bond rapidly, outcompeting 1,2-C-H insertion in some cases, yielding polycyclic structures (Equation  [6] ). [9d] [²¹] Calculations suggested that the competing 1,2-C-H insertion requires higher activation energy.

Recently, Liu reported an interesting hydride migration and cyclization involving gold carbene K, constituting a formal 1,3-addition of a C(sp³)-H bond to the alkenyl gold carbene moiety (Equation  [7] ). [9h]

2.3.2 Using the TMSCH 2 Group as R ³

Alternatively, a TMSCH2 group could be incorporated into oxocarbenium E as R³, thus affording intermediate L (Figure  [²] ). We anticipated a facile desilylation would occur to yield dienyl carboxylate products. Not surprisingly, in the presence of isopropanol, dienyl esters were indeed obtained smoothly (Equation  [8] a). [7d] The role of isopropanol was to convert TMS+ into H+, which facilitates catalyst turnover via protodeauration. Interestingly, carbonate derivative 10 underwent the same reaction, yielding dienyl carbonate 11 in 78% yield (Equation  [8] b).

Equation 8

Figure 2

2.4 Using the Aryl Group for R ¹

Scheme 5 Nolan’s efficient synthesis of substituted indenes

Nolan and coworker reported in early 2006 an efficient synthesis of substituted indenes from propargylic acetates derived from aromatic aldehydes (i.e., R¹ = aryl, Scheme  [5] ). [9a] The authors proposed two consecutive 1,2-acetoxy shifts to account for the allene formation and did not invoke intermediate M for the final cyclization. In the context of this account, a 3,3-rearrangement could count for the two 1,2-shifts and the final step could be viewed as an intramolecular electrophilic aromatic substitution by the oxocarbenium moiety in M.

2.5. Hydrolytic Treatment

2.5.1 Formation of Εnones

α,β-Unsaturated ketones are fundamental building blocks in organic synthesis. A simple hydrolysis of intermediate E followed by protodeauration could provide access to this versatile structural motif as a useful alternative to the HWE reaction (Scheme  [6] ). Hence, under exceptionally mild reaction conditions, propargyl acetates were converted into enones in regular 2-butanone or acetonitrile-water (80:1) at ambient temperature (Equation  [9] ). [7e] Notably, the reaction was highly E-selective for substrates derived from secondary alcohols.

Scheme 6 Hydrolytic transformations

Equation 9

Nolan at the same time reported enone/enal formation from propargylic acetates. [²²] With the use of [(ItBu)AuCl]/AgSbF6 [ItBu: di(tert-butyl)imidazol-2-yliden] as catalyst and tetrahydrofuran-water as solvent, a broader scope was realized and an interesting mechanism involving an SN2′-like attack by the gold-bound OH group was proposed following DFT calculations (Scheme  [7] ).

Notably, Akai recently reported a highly efficient preparation of enones and enals from propargylic alcohols using a combination of gold(I) and molybdenum catalysts. [²³]

2.5.2 Formation of α-Iodo/Bromoenones

We envisioned that the nucleophilic Au-C(sp²) bond in intermediate N (Scheme  [6] ) could react with electrophilic halogens besides protonation, affording α-haloenones. To our delight, the chemistry indeed worked beautifully using either NIS [7f] or NBS. [7g] Excellent Z-selectivity was achieved with aliphatic substrates (Equation  [¹0] a), while low stereoselectivity was observed with aromatic substrates (Equation  [¹0] b). Compared to known procedures for preparing linear α-iodo-/bromoenones, [²4] this method is concise, flexible, and highly efficient. Recently, we reported a much improved version of this chemistry using propargylic alcohol directly as substrate. With a combination of gold and molybdenum complexes, the reaction scope was expanded to include α-iodo-/bromoenal products (Equation  [¹0] c). [²5]

Scheme 7 Nolan’s enone synthesis and the proposed mechanism

Equation 10

2.6 Intramolecular Acyl Migration

The coexistence of an activated acyl group and a nucleophilic Au-C(sp²) bond in oxocarbenium E invites speculation that these two may react with each other, perhaps in an intramolecular manner (Figure  [³] ).

Figure 3

Upon heating, such reaction indeed occurred in the presence of 5 mol% of dichloro(pyridine-2-carboxylato)gold(III) (12), and α-ylidene-β-diketones were formed in excellent yields and in some cases high diastereoselectivities (Equation  [¹¹] [] ). Crossover experiments established that the acyl migration is intramolecular or within a tight solvent cage. α-Ylidene-β-diketones are versatile intermediates. Although they can be prepared in general via Knoevenagel reaction, this reaction offers an efficient alternative that tolerates sensitive functional groups (e.g., Equation  [¹¹b] ).

Equation 11

With alkynyl ether substrates, Barluenga reported a similar acyl migration reaction. [²6] Interestingly, both platinum(II) chloride and [Cu(MeCN)4]+ BF4 - were capable of catalyzing the acyl migration but with contrasting diastereoselectivities (Equation  [¹²] ). Different mechanisms were proposed to explain the outcomes.

Equation 12

Scheme 8 Proposed mechanism for the gold-catalyzed oxidative couplings

2.7 Incorporating Gold(I)/Gold(III) Catalysis

Cross-coupling reactions catalyzed by late transition metals such as palladium and nickel are indispensable tools in modern organic synthesis. The power of these reactions lies in the capacity of the metal to undergo oxidation and reduction readily and efficiently in the same catalytic cycle, thus promoting novel and mild bond-forming processes unattainable previously.

Equation 13

Gold-catalyzed homogeneous reactions [²7] involving an Au(I)/Au(III) catalytic cycle are novel and represent a new area for further expanding gold catalysis. It is foreseeable that the potential of this new reactivity manifold could be significant if coupled with a plethora of gold-­catalyzed transformations of alkyne or allene substrates.

Lately there were a few studies suggesting that such a gold(I)/gold(III) catalysis in solution is feasable. [²8] [²9] However, those reactions were mostly homo-couplings, and no cross-coupling reaction involving alkyne/allene substrates was previously implemented.

2.7.1 Gold-Catalyzed Oxidative Homo-Coupling

In our study with propargylic substrates using Selectfluor as electrophilic fluorine source, we discovered that enone dimers were formed efficiently even with hindered substrates under the optimized reaction conditions (Equation  [¹³] ). [8c] As shown in Scheme  [8] , a Au(I)/Au(III) catalytic cycle involving Au(I)-catalyzed tandem transformation of propargyl acetates (see Scheme  [²] ), oxidation of enonyl Au(I) complex N to Au(III) intermediate O by Selectfluor, transmetallation from another molecule of N to the formed O, and subsequent reductive elimination was proposed to account for the homo-coupling. Notably in this mechanism, the transmetallation occurs between gold intermediates, Selectfluor is acting as oxidant, and reductive elimination allows regeneration of the gold(I) catalyst and completion of the catalytic cycle.

2.7.2 Gold-Catalyzed Oxidative Cross-Coupling Reaction

We speculated that an external organometallic reagent could be introduced to compete with the enonyl Au(I) complex N during transmetallation. As a result, upon reductive elimination, the first example of cross-coupling reaction utilizing Au(I)/Au(III) catalysis could be realized (Scheme  [8] ). Arylboronic acids were chosen as the organometallic reagents due to their commercial availability and stability. After extensive catalyst and reaction condition optimization, cross-coupling products, i.e., α-arylen­ones, were indeed obtained in mostly good yields (Equation  [¹4] ). [8a] The use of water in the reaction was the key to minimize the competing homo-coupling as it facilitated aryl transfer from arylboronic acid likely through the formation of tetra-coordinated borate species.

Equation 14

Scheme 9 Proposed mechanism for gold-catalyzed oxidative C-O bond formation

2.7.3 Gold-Catalyzed Oxidative C-O Bond Formation

This novel reaction manifold could lead to C-O bond formation as well. It was found that with substrates without substitution at the propargylic position α-benzoxyenones were formed efficiently instead of enone dimers (Equation  [¹5] ); moreover, dienones with both electron-­deficient and capto-dative double bonds could be easily accessed. Mechanistic studies including crossover experiments in the presence of external carboxylic acids all strongly suggested that the benzoxy group migrated intramolecularly during the reaction, and a corresponding mechanism was proposed. As shown in Scheme  [9] , intermediate P, instead of undergoing hydrolysis to form an intermediate similar to N, may undergo oxidation by Selectfluor first, forming Au(III) complex Q. The highly electrophilic gold center in Q invites intramolecular nucleophilic attack by the benzoxy group, thus initiating a two-step migration process. Hydrolysis of the bridge intermediate R completes the intramolecular benzoxy migration and forms complex S, which could undergo reductive elimination to afford the product. This unique mechanism was further supported by the observation that substrates with substitution at the propargylic position only led to enone dimers, which could be explained readily by invoking detrimental steric congestion in bridge intermediate R (Figure  [4] ).

Equation 15

Figure 4

3 Attack of Au-Activated Carboxyallene at the β- or γ-Position

Scheme 10 Alternative transformations of the carboxyallene intermediate

Equation 16

Equation 17

Following the initial gold-catalyzed 3,3-rearrangement, the thus-formed carboxyallene can be activated by the same gold catalyst to proceed via alternative reaction pathways instead of via oxocarbenium E . As shown in Scheme  [¹0] , instead of activating the double bond proximal to the carboxy group (i.e., the α,β-double bond) yielding E , activation of the distal double bond of the carboxyallene intermediate (i.e., the β,γ-double bond) would possibly lead to nucleophilic attack at either end of the double bond; alternatively, the cationic intermediate S may be formed. In this case, it is more likely that intermediate S is accessed via resonance structure E formed via activation of the more electron-rich enol acetate moiety (i.e., the α,β-double bond). In other words, the participation of E cannot be ruled out in nucleophilic attacks at the γ-position of gold-activated carboxyallenes (vide infra).

3.1 Nucleophilic Attack at the γ-Position

Gagosz reported an efficient synthesis of bicyclo- [3.1.0]hexenes and cycloalkenones from propargylic acetates with a tethered C-C double bond (Equation  [¹6] ). [9b] In the proposed mechanism, the cyclization step occurs via the alkene attacking the γ-position of the activated carboxyallene.

De Brabander proposed cycloetherification as the cyclization step in their work on gold-catalyzed reactions of ω-hydroxy propargylic esters (Equation  [¹7] ). [9i]

Echavarren showed that soft carbon nucleophiles could readily attack the carboxyallene intermediate at the γ-position via intermediate S and in one case at the α-position as well (Equation  [¹8] ). [9f]

Equation 18

A mechanistically interesting and facile formation of naphthalenes by Gevorgyan involves tandem 6-endo-dig cyclization, 1,2-aryl/alkyl/H migration, cyclization and aromatization (Equation  [¹9] ). [9g] [j] The 1,2-migration can be viewed as the migrating group as a nucleophile attacks an equivalent of cation S. This reaction is limited to esters with an aryl-containing tertiary substituent at the propargylic position. Notably, diethyl phosphates could replace carboxylates as substrates. For starting materials without aryl group or the tertiary substituent, dienyl esters were obtained. [9j]

Equation 19

Wang, Nevado and Goeke reported the formation of various carbocycles from cyclopropyl alkynyl acetates. When the alkyne terminus was substituted, enol acetate U was formed efficiently (Equation  [²0] ). [9k] Mechanistically, it was proposed to proceed via cyclopropane ring opening in an intermediate of type S.

Equation 20

Equation 21

3.2 Nucleophilic Attack at the β-Position

An interesting regioselectivity reversal was realized by Ohfune using substrates with silyl substitution at the propargyl position. [9l] Although gold activation of allenes generally leads to nucleophilic attacks at either 1- or 3-position, [³0] interestingly the β-silicon effect (silicon stabilization of β-carbocation) effectively reverted this selectivity, and α-acyloxy-α′-silyl ketones were formed in generally good yields (Equation  [²¹] ).

4 Carboxyallenes as Nucleophiles

Carboxyallenes, once formed, can alternatively react with electrophiles instead of being activated by the existing gold catalyst. The key for the success of this approach is that the intended reaction has to outcompete other reactions promoted by gold activation of carboxyallenes. Consequently, intramolecular reactions are the ones reported in literature.

Equation 22

Equation 23

Toste reported the synthesis of aromatic ketones using gold- or silver-activated alkyne as the internal electrophile (Equation  [²²] ). [9c] Interestingly, AgSbF6 in the presence of Ph3P and MgO worked well as catalyst and showed somewhat different substrate scope in comparison to a gold catalyst. Oh later reported their independent work detailing the same transformation using NaAuCl4 as catalyst. [9m]

Equation 24

A rapid construction of complex α-pyrones was developed by Schreiber using gold-activated propiolates as the internal electrophile (Equation  [²³] ). [9e] Notable in this chemistry is that the propiolate underwent 3,3-rearrangement with no incident and the cationic intermediate was readily trapped by electron-rich arenes.

A cycloisomerization reaction of 1,6-enyne was employed to generate an electrophilic center as the internal electrophile, leading to aryl ketone products. [9n] As shown in Equation  [²4] , Liang showed that PtCl2 was a better catalyst than either gold(I) or AuCl3, which led to mostly substrate decomposition.

In a study extending their earlier work, [9m] Oh reported the formation of 2,3-bis(alkylidene)cycloalkanones using gold-activated alkyne as the internal electrophile (Equation  [²5] ). [9o] With a tethered alkyne, tricyclic arene-fused alkanones were formed in acceptable yields. Interestingly, substrates with a one-carbon longer tether resulted in [2+2] cycloaddition. [³¹]

Equation 25

5 Summary

Via an initial 3,3-rearrangement, propargylic carboxylates in the presence of gold complexes can be transformed into carboxyallenes albeit in an equilibrium. These functionalized allenes, difficult to generate using other approaches, do not need to be purified and are rich in chemistry. They can either be further activated by the very same gold catalyst to engage in electrophilic reactions or act as nucleophiles, leading to various versatile synthetic methods. The irreversible nature of downstream transformations makes the equilibrating 3,3-rearrangement inconsequential.

More synthetic reactions based on this initial transformation of propargylic carboxylates are anticipated, and it is foreseeable that carboxyallenes will be regularly accessed via this approach for organic synthesis.

Acknowledgment

The authors thank the University of California, Santa Barbara, the University of Nevada, Reno and NSF (CAREER CHE 0969157) for generous financial support. Most of our work was done in the University of Nevada, Reno. LZ thanks all the students and postdoctoral associates who worked in his group for their contribution.

1

Current address: Department of Chemistry, Nanjing University, Nanjing, P. R. of China

2

Previous address: Department of Chemistry, University of Nevada, Reno, NV 89557

20

ConQuest® searching of the Cambridge structural database for structures containing C(sp²)-Au(I)(PPh3) bonds gave an average bond length of ˜2.04 Å, while the bond length of C(sp²)-C(sp³) is 1.50 Å.

1

Current address: Department of Chemistry, Nanjing University, Nanjing, P. R. of China

2

Previous address: Department of Chemistry, University of Nevada, Reno, NV 89557

20

ConQuest® searching of the Cambridge structural database for structures containing C(sp²)-Au(I)(PPh3) bonds gave an average bond length of ˜2.04 Å, while the bond length of C(sp²)-C(sp³) is 1.50 Å.

   Permissions and Reprints All articles of this category

Scheme 1 Two competing mechanistic pathways for gold-catalyzed reactions of propargylic carboxylates

Scheme 2 Gold-catalyzed tandem reaction of propargylic carboxylates via an initial 3,3-rearrangement: design

Figure 1

Scheme 3 Proposed mechanism for gold-catalyzed formation of tetracyclic cyclobutanes

Scheme 4 Enynyl carboxylates as substrates

Figure 2

Scheme 5 Nolan’s efficient synthesis of substituted indenes

Scheme 6 Hydrolytic transformations

Scheme 7 Nolan’s enone synthesis and the proposed mechanism

Figure 3

Scheme 8 Proposed mechanism for the gold-catalyzed oxidative couplings

Scheme 9 Proposed mechanism for gold-catalyzed oxidative C-O bond formation

Figure 4

Scheme 10 Alternative transformations of the carboxyallene intermediate