Gold(I)-Catalyzed Allene–Diene–Alkyne Coupling Reaction to Polycycles

In general, the design of complex polycycles from simple building blocks is challenging and requires many reaction steps. Herein, we present that complex polycycles can be synthesized starting from two alkyne building blocks in a onepot reaction using gold(I) as catalyst, in yields up to 77 %. Mechanistically, the process can be described as cascade reaction in two steps: first, the allene is formed by gold(I)-catalyzed dearomatization, second, the desired product is then obtained In organic chemistry the development of efficient and new C–C coupling reactions is highly important for the synthesis of complex molecules. Especially, catalysis by transition metals has become an important tool to design such molecules.[1] One of these transition metal-catalyzed C–C coupling reactions is the catalytic asymmetric dearomatization (CADA).[2–6] By using cheap, functionalizable and ubiquitous aromatic systems for this C–C coupling reaction it is possible to synthesize more complex polycycles.[2–6] In search of efficient transition metal catalysts to build up new C–C bonds, homogeneous gold catalysis has come to the fore.[7–11] Currently, it is possible to form up to four new C–C bonds[12] during a reaction cascade by means of homogeneous gold catalysis if the combination of identical building blocks (e.g. oligomerization) is not taken into account.[13] However, gold has been rarely used in CADA reactions compared to other transition metals.[14–17] An example of gold-catalyzed dearomatization involving one alkyne unit was shown by Bandini et al.[18] They converted naphthyl propargyl ethers 1 to naphthalene-2-one derivatives 2 (Scheme 1a). This reaction can be considered as a Claisen rearrangement with subsequent intramolecular hydroxy group addition to the newly formed allene. Our previous work in the field of gold catalysis was focused on the conversion of haloacetylenes (Scheme 1b). The advantage of these compounds is their easy accessibility[19] and their thermal stability at standard conditions.[20–22] During the last [a] N. Semleit, M. Kreuzahler, Prof. Dr. G. Haberhauer Institut für Organische Chemie, Universität Duisburg-Essen Universitätsstraße 7, 45117 Essen Germany E-mail: gebhard.haberhauer@uni-due.de Supporting information and ORCID(s) from the author(s) for this article are available on the WWW under https://doi.org/10.1002/ejoc.202001190. © 2020 The Authors. European Journal of Organic Chemistry published by Wiley-VCH GmbH. · This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Eur. J. Org. Chem. 2020, 6629–6634 © 2020 The Authors. European Journal of Organic Chemistry published by Wiley-VCH GmbH 6629 by allene–diene–alkyne coupling. Quantum chemical investigations confirm the assumed mechanism. In sum, five new C–C bonds are built in one sweep, which is extraordinary since nonidentical building blocks are combined in this reaction. The introduced new allene–diene–alkyne coupling reaction paves the way for several new syntheses of cycles and polycycles considering potential intermolecular reactions. Scheme 1. Gold(I)-catalyzed dearomatization (a), haloalkynylation (b), cyclization (c) of alkynes and allene–diene–alkyne coupling reaction (d). few years we and other research groups presented haloalkynylation reactions of different alkynes[23–25] and alkenes.[26–29] Mechanistic investigations proved that the formation of products obtained by addition has to take place via rearrangement.[30] However, the restriction of previous systems is that only conjugated alkynes can be used. While attempting to transfer the concept of haloalkynylation to non-conjugated haloacetylenes, we discovered a gold(I)-catalyzed allene–diene– alkyne coupling reaction to polycycles. Herein, we describe the application range as well as mechanistical investigations of this novel type of cascade reaction. In the course of studies relating to gold(I)-catalyzed reactions of non-conjugated haloacetylenes with alkynes, compound Communication doi.org/10.1002/ejoc.202001190 EurJOC European Journal of Organic Chemistry classes 12 and 13 got into our focus of interest (Scheme 1d). These molecules can be prepared in a few steps from readily accessible compounds. The use of two methyl groups in ortho position is necessary to avoid gold(I)-catalyzed cyclization to the corresponding 2H-chromene; this type of reaction has already been observed for terminal alkynes (Scheme 1c).[31–34] A similar concept was used for propargylic ethers 1 (Scheme 1a).[18] The undesired cyclization could actually be prevented and 2,5-dihydrofurans 2 emerged from gold(I)-catalyzed asymmetric dearomatization. Regarding haloacetylenes 12 and 13 adequate cyclization reactions are not expected as halogen atoms are monovalent. As test reaction the conversion of chloroacetylene 12 with alkyne 5a and [JohnPhosAu(NCMe)]SbF6 as gold catalyst was investigated via NMR experiments in deuterated chloroform. Within only two hours at room temperature the starting material was completely consumed. Fortunately, one single main product (yield: 47 %) was obtained (see Figure S8 and Table 1). The yields of all side products amounted to less than 5 %. However, the NMR spectra showed that not the expected enyne dimer but a polycycle was formed. Structural investigations via 2D NMR spectroscopy revealed that the main product can be ascribed to tetracycle 14a (Figures S9–S11). Analysis of the connectivity in tetracycle 14a shows that five C–C bonds were formed in the course of the reaction. Additionally, one C–O bond was broken, and one hydrogen atom was rearranged. Until now, we are not aware of any case in which so many C–C bonds were formed by gold catalysis in one reaction cascade, if the combination of two identical building blocks (e.g. oligomerization) is not taken into account.[13] Table 1. Optimization of the reaction conditions for the gold(I)-catalyzed coupling of 12 with alkyne 5a.[a] Entry Catalyst Conditions Yield [%] Conversion cycle allene [%] 1 [JohnPhosAu(NCMe)]SbF6 (5 mol-%)[35] RT, 2 h, CDCl3 47 – 100 2 JohnPhosAuCl (5 mol-%)/AgNTf2 (10 mol-%)[37] RT, 2 h, CDCl3 37 – 97 3 JohnPhosAuCl (5 mol-%)/NaBArF24 (10 mol-%) RT, 2 h, CDCl3 9 42 100 4 tBuXPhosAuNTf2 (5 mol-%)[37] RT, 2 h, CDCl3 30 – 92 5 CyJohnPhosAuCl (5 mol-%)/AgSbF6 (10 mol-%) RT, 2 h, CDCl3 38 – 100 6 Me3PAuCl (5 mol-%)/AgSbF6 (10 mol-%) RT, 2 h, CDCl3 – – 0 7 IPrAuCl (5 mol-%)/AgNTf2 (10 mol-%) RT, 2 h, CDCl3 24 13 97 8 Dichloro(2-picolinato)gold(III) (5 mol-%)[36] RT, 2 h, CDCl3 – 98 100 9 [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) RT, 2 h, CDCl3 52 – 100 10 [JohnPhosAu(NCMe)]SbF6 (1 mol-%) RT, 2 h, CDCl3 – 19 21 11 [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) RT, 2 h, C6D6 43 – 100 12 [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) RT, 2 h, DCM 40 – 100 13 [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) RT, 2 h, DCE 26 – 100 14 [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) –10 °C, 2 h, CDCl3 37 23 100 15 [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) 0 °C, 2 h, CDCl3 44 9 100 16[b] [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) RT, 2 h, CDCl3 49 – 100 17[c] [JohnPhosAu(NCMe)]SbF6 (2.5 mol-%) RT, 2 h, CDCl3 34 – 100 [a] Yields were determined by 1H NMR using cyclooctane as internal standard. If not stated otherwise, the concentration of alkyne 13 was 0.15 M and alkyne 5a was added in 2 molar equivalents. [b] 3 molar equivalents of alkyne 5a were added. [c] 1 molar equivalent of alkyne 5a was added. Eur. J. Org. Chem. 2020, 6629–6634 www.eurjoc.org © 2020 The Authors. European Journal of Organic Chemistry published by Wiley-VCH GmbH 6630 We wanted to examine if the yield obtained in the cascade reaction can be increased. Therefore, we tested different gold catalysts in the first step (entries 1–8 in Table 1). Nevertheless, we were not able to improve the yield (47 %) that was obtained by using [JohnPhosAu(NCMe)]SbF6. The usage of sterically demanding ligands like XPhos led to distinctly lower yields for polycycle 14a (30 %; entry 4). If another type of phosphine ligand like trimethylphosphine was employed, no conversion was observed (entry 6). If JohnPhosAuCl was used with NaBArF24 as counterion, the main product was no longer polycycle 14a but the multiple unsaturated system 15a (entry 3). This system was identified as allene by NMR spectroscopy. The use of gold(III) complex dichloro(2-picolinato)gold(III)[36] led exclusively to allene 15a (entry 8). The use of an NHC ligand[37] ended up in a product mixture in which polycycle 14a and allene 15a were formed in a 2:1 ratio (entry 7). Thus, the initially applied catalyst [JohnPhosAu(NCMe)]SbF6 turned out to be the best choice. In the next step, the equivalents of both starting materials and catalyst as well as temperatures and solvents were varied (entries 9–17). The highest yield of polycycle 14a (52 %) was obtained, when chloroacetylene 12 was converted with 2.5 mol-% of the catalyst and two equivalents of alkyne 5a at room temperature in chloroform. After optimization of the reaction conditions, we evaluated the substrate scope of this reaction on a preparative scale. Therefore, we used different arylacetylenes (3–7; see Scheme 2) in which both triple bond substituents were varied. At first, we had a closer look at alkyl arylacetylenes 5. The length of the alkyl chain plays no important role, if the second substituent is a phenyl group (14a–c). In all cases the yields amount to ca. Communication doi.org/10.1002/ejoc.202001190 EurJOC European Journal of Organic Chemistry 40 %. The yield was increased by a methoxy group in para position (57 %, 14d), if a propyl group is attached to the triple bond as second substituent. A lower yield was found if the second substituent is a methyl group (26 %, 14e). Scheme 2. Evaluation of the substrate scope of the gold(I)-catalyzed coupling of haloacetylenes 12 and 13 with alkynes 3–7. A methyl group in ortho position lowers the yield to 23 % (14f ). If the alkyl group in 5 is substituted by chlorine (3a), no significant change was observed (40 %, 14h). The corresponding bromo derivative 14g was isolated in 13 % yield. Terminal alkynes (6) are also tolerated; the corresponding products were obtained in yields of ca. 25 % (14i, 14j and 14l) and 10 % (14k), respectively. Diarylacetylenes (7) form as well polycycles with 12. The resulting products were isolated in yields up to 77 %. In these products the more electron-rich aromatic ring is bound to the position which is closer to the carbonyl group. If bromo derivative 13 was used instead of chloroacetylene 12, the conEur. J. Org. Chem. 2020, 6629–6634 www.eurjoc.org © 2020 The Authors. European Journal of Organic Chemistry published by Wiley-VCH GmbH 6631 version to polycycle 14q was not complete. Even after one week at room temperature or after a few days at 40 °C, the allene and only small amounts of polycycle 14q (11 %) were found. Next, we wanted to elucidate the mechanism of this remarkable coupling reaction. In the first step, the role of the allene was investigated; we wanted to find out whether this allene embodies an intermediate or a side product. Therefore, we converted chloroacetylene 12 without alkyne 5a using gold(III) as catalyst. The allene 15a was isolated in 57 % yield (Scheme 3a). Afterwards, we converted 15a under the optimized reaction conditions with [JohnPhosAu(NCMe)]SbF6 as catalyst and 5a by NMR experiments. Polycycle 14a was obtained in 55 % yield. Thus, the formation of 15a via gold-catalyzed Claisen rearrangement can be described as the first sequence of the cascade reaction. Scheme 3. Gold(I)-catalyzed reaction of chloroacetylenes 12 and 16 in presence and absence of alkyne 5a. Interestingly, the NMR yield of 14a after conversion of 15a with alkyne 5a (55 %) was not much higher than the NMR yield of 14a obtained in the reaction of 12 and 5a (52 %, Scheme 3a). This means that the one-pot cascade reaction is almost as efficient as the stepwise process. Please note, the reaction of 12 with [JohnPhosAu(NCMe)]SbF6 as catalyst without alkyne 5a led to a complete conversion (Scheme 3a), but only a variety of oligomeric compounds was formed, which could not be separated by column chromatography. Therefore, we assume that – utilizing JohnPhosAu+ as catalytic species – allene 15a (which is formed in the first step) reacts with the diene unit of the cycle. This leads to the formation of oligomeric structures due to the absence of intercepting alkynes. Another preparative mechanistic indication was achieved via conversion of 16 under the optimized reaction conditions. Bicycle 17 was obtained in 67 % yield whether alkyne 5a was present or not. The formation of the product underlines the assumption that the allene unit is cyclized with the diene unit building a five-membered ring. The formed cation can be stabilized by deprotonation. Thus, the second sequence of the cascade reaction probably starts with the formation of a five-membered ring. Communication doi.org/10.1002/ejoc.202001190 EurJOC European Journal of Organic Chemistry The above-mentioned experimental indications were used to calculate the cascade reaction mechanism via quantum chemical methods. Arylacetylene 12 and alkyne 6a were utilized as model substances. In order to optimize the geometrical parameters of all stationary points, the density functional B3LYP[38–40] together with the dispersion correction via Becke– Johnson damping[41] (D3BJ) was employed. As basis sets 6-31G(d) was applied for the elements C, H, O, P, and Cl, whereas def2-TZVP was used for Au. Furthermore, single-point calculations on the optimized structures were performed using B3LYP-D3BJ with basis sets 6-311++G(d,p) (for C, H, O, P, and Cl) and def2-TZVP (for Au). To take solvent effects into account, chloroform was considered as reaction solvent by using the SMD[42] model. The first sequence of the cascade reaction is the formation of the allene starting with chloroacetylene 12 and JohnPhosAu+ as catalytic-active species. The coordination of the gold(I) comFigure 1. Reaction pathway of the gold(I)-catalyzed dearomatization of chloroacetylene 12. The indicated free-energy values (ΔG in kcal/mol) were calculated using B3LYP-D3BJ(SMD) and are relative to gold complex 18a. [Au]+ = JohnPhosAu+. Figure 2. Reaction pathway of the addition of acetylene 6a to cation 24a. The indicated free-energy values (ΔG in kcal/mol) were calculated using B3LYPD3BJ(SMD) and are relative to acetylene 6a and cation 24a. [Au]+ = JohnPhosAu+. Eur. J. Org. Chem. 2020, 6629–6634 www.eurjoc.org © 2020 The Authors. European Journal of Organic Chemistry published by Wiley-VCH GmbH 6632 plex is accompanied by free-energy reduction amounting to 15.4 kcal/mol (Figure 1). Please note that the coordination of JohnPhosAu+ to acetylene 6a (ΔG = –16.2 kcal/mol) is energetically favored compared to the complexation of 12. However, the reactions of the thusly formed complex with acetylenes 6a and 12, respectively, require high activation energies, making the complexation of 6a a dead end (Figures S8–9). Complex 18a, which is formed by coordination of JohnPhosAu+ to acetylene 12, was used as reference for all subsequent reaction steps. The rate-determining step of the first cascade reaction sequence (allene formation) is the cyclization to oxonium ion 20a (10.0 kcal/mol). This ion converts into gold(I) allene complex 22a in the second step. Therefore, a barrier of 7.4 kcal/mol needs to be overcome. After that, two reaction paths are possible: on the one hand, decomplexation can take place leading to the catalytic species and allene 15a (this allene formation corresponds to the end Communication doi.org/10.1002/ejoc.202001190 EurJOC European Journal of Organic Chemistry Figure 3. Possible reaction mechanism for the gold(I)-catalyzed dearomatization (black and green arrows) leading to allene 15a and the allene–diene–alkyne coupling (blue arrows) leading to polycycle 14i. of the first cascade reaction sequence). On the other hand, the formation of bicyclic cation 24a can occur initiating the second cascade reaction sequence (allene–diene–alkyne coupling). According to calculations, the second sequence (intramolecular cyclization) is more favored (12.5 kcal/mol) than the formation of free allenes (15.3 kcal/mol). This agrees well with the results showing a polymerization in case of converting 12 in absence of 5a with JohnPhosAu+ as catalytic species. Let us now consider the second cascade reaction sequence (allene–diene–alkyne coupling): the first step is again ratedetermining and corresponds to the addition of alkyne 6a to cation 24a with formation of vinyl cation 26a (Figure 2). The activation energy for this step amounts to 9.1 kcal/mol. The vinyl cation is stabilized by the phenyl ring, which also explains the regioselectivity of the attack of 6a to 24a. A possible competing reaction to the addition of alkyne 6a to cation 24a is the addition of alkyne 12 to cation 24a. However, the required activation energy for the latter reaction path is distinctly higher (15.5 kcal/mol; Figure S10). In the next step, two intramolecular C–C bonds are simultaneously formed leading to gold(I)-stabilized cation 28a. Proceeding from this σ complex, π complex 30a is formed by a 1,2-hydride shift. Decomplexation leads to the final product 14i. In Figure 3 the most important intermediates are illustrated in a catalyst cycle. The whole cycle consists of two sequences; in both gold(I) complexed allene 22a is passed. Allene formation takes place via two-step gold-catalyzed Claisen-like rearrangement (black in Figure 3). Allene 22a functions as origin of two branches: on the one hand, the formation of allene 15a is enabled by decomplexation (green in Figure 3), on the other hand, the intramolecular 5-endo-trig-cyclization to 24a is possible (blue in Figure 3). Alkyne 6a can be added to intermediate 24a ending up in the formation of gold complexed polycycle 30a via two further steps. Decomplexation leads to final product 14i and finishes the catalyst cycle. Eur. J. Org. Chem. 2020, 6629–6634 www.eurjoc.org © 2020 The Authors. European Journal of Organic Chemistry published by Wiley-VCH GmbH 6633 In sum, we were able to show that polycycles can be designed from haloacetylenes and alkynes by gold(I) catalysis. In our case, the haloacetylene was an o,o-dimethyl-substituted propargyl ether. The reaction tolerates a wide range of alkyne substrates like alkylaryl, diaryl, haloand terminal alkynes. The achieved preparative yields amount up to 77 %. Experimental and quantum chemical investigations showed that the mechanism can be described as cascade reaction with two sequences. The first sequence is a gold(I)-catalyzed dearomatization in which the corresponding allene is built from propargyl ether by formal Claisen rearrangement (isolation of the allene is possible). The second sequence is a new reaction type: four new C–C bonds are formed by an allene–diene–alkyne coupling leading to the corresponding polycycle. Altogether, even five C–C bonds are newly formed during the whole cascade. The described allene–diene–alkyne coupling reaction has the potential to exploit new ways of polycycle formation starting from simple building blocks. Especially, the option of expanding the reaction to obtain intermolecular products is promising and desires further exploration. Supporting Information (see footnote on the first page of this article): Figures and Tables syntheses of the new compounds, NMR experiments, computational details, cartesian coordinates and absolute energies for all calculated compounds, 1H NMR and 13C NMR spectra of the new compounds. Conflict of Interest There is no conflict of interest to declare.


Abstract:
In general, the design of complex polycycles from simple building blocks is challenging and requires many reaction steps. Herein, we present that complex polycycles can be synthesized starting from two alkyne building blocks in a onepot reaction using gold(I) as catalyst, in yields up to 77 %. Mechanistically, the process can be described as cascade reaction in two steps: first, the allene is formed by gold(I)-catalyzed dearomatization, second, the desired product is then obtained In organic chemistry the development of efficient and new C-C coupling reactions is highly important for the synthesis of complex molecules. Especially, catalysis by transition metals has become an important tool to design such molecules. [1] One of these transition metal-catalyzed C-C coupling reactions is the catalytic asymmetric dearomatization (CADA). [2][3][4][5][6] By using cheap, functionalizable and ubiquitous aromatic systems for this C-C coupling reaction it is possible to synthesize more complex polycycles. [2][3][4][5][6] In search of efficient transition metal catalysts to build up new C-C bonds, homogeneous gold catalysis has come to the fore. [7][8][9][10][11] Currently, it is possible to form up to four new C-C bonds [12] during a reaction cascade by means of homogeneous gold catalysis if the combination of identical building blocks (e.g. oligomerization) is not taken into account. [13] However, gold has been rarely used in CADA reactions compared to other transition metals. [14][15][16][17] An example of gold-catalyzed dearomatization involving one alkyne unit was shown by Bandini et al. [18] They converted naphthyl propargyl ethers 1 to naphthalene-2-one derivatives 2 (Scheme 1a). This reaction can be considered as a Claisen rearrangement with subsequent intramolecular hydroxy group addition to the newly formed allene.
Our previous work in the field of gold catalysis was focused on the conversion of haloacetylenes (Scheme 1b). The advantage of these compounds is their easy accessibility [19] and their thermal stability at standard conditions. [20][21][22] During the last by allene-diene-alkyne coupling. Quantum chemical investigations confirm the assumed mechanism. In sum, five new C-C bonds are built in one sweep, which is extraordinary since nonidentical building blocks are combined in this reaction. The introduced new allene-diene-alkyne coupling reaction paves the way for several new syntheses of cycles and polycycles considering potential intermolecular reactions. few years we and other research groups presented haloalkynylation reactions of different alkynes [23][24][25] and alkenes. [26][27][28][29] Mechanistic investigations proved that the formation of products obtained by addition has to take place via rearrangement. [30] However, the restriction of previous systems is that only conjugated alkynes can be used. While attempting to transfer the concept of haloalkynylation to non-conjugated haloacetylenes, we discovered a gold(I)-catalyzed allene-dienealkyne coupling reaction to polycycles. Herein, we describe the application range as well as mechanistical investigations of this novel type of cascade reaction.
In the course of studies relating to gold(I)-catalyzed reactions of non-conjugated haloacetylenes with alkynes, compound classes 12 and 13 got into our focus of interest (Scheme 1d). These molecules can be prepared in a few steps from readily accessible compounds. The use of two methyl groups in ortho position is necessary to avoid gold(I)-catalyzed cyclization to the corresponding 2H-chromene; this type of reaction has already been observed for terminal alkynes (Scheme 1c). [31][32][33][34] A similar concept was used for propargylic ethers 1 (Scheme 1a). [18] The undesired cyclization could actually be prevented and 2,5-dihydrofurans 2 emerged from gold(I)-catalyzed asymmetric dearomatization. Regarding haloacetylenes 12 and 13 adequate cyclization reactions are not expected as halogen atoms are monovalent.
After optimization of the reaction conditions, we evaluated the substrate scope of this reaction on a preparative scale. Therefore, we used different arylacetylenes (3-7; see Scheme 2) in which both triple bond substituents were varied. At first, we had a closer look at alkyl arylacetylenes 5. The length of the alkyl chain plays no important role, if the second substituent is a phenyl group (14a-c). In all cases the yields amount to ca.

EurJOC
European Journal of Organic Chemistry 40 %. The yield was increased by a methoxy group in para position (57 %, 14d), if a propyl group is attached to the triple bond as second substituent. A lower yield was found if the second substituent is a methyl group (26 %, 14e). A methyl group in ortho position lowers the yield to 23 % (14f ). If the alkyl group in 5 is substituted by chlorine (3a), no significant change was observed (40 %, 14h). The corresponding bromo derivative 14g was isolated in 13 % yield. Terminal alkynes (6) are also tolerated; the corresponding products were obtained in yields of ca. 25 % (14i, 14j and 14l) and 10 % (14k), respectively. Diarylacetylenes (7) form as well polycycles with 12. The resulting products were isolated in yields up to 77 %. In these products the more electron-rich aromatic ring is bound to the position which is closer to the carbonyl group. If bromo derivative 13 was used instead of chloroacetylene 12, the con- version to polycycle 14q was not complete. Even after one week at room temperature or after a few days at 40°C, the allene and only small amounts of polycycle 14q (11 %) were found. Next, we wanted to elucidate the mechanism of this remarkable coupling reaction. In the first step, the role of the allene was investigated; we wanted to find out whether this allene embodies an intermediate or a side product. Therefore, we converted chloroacetylene 12 without alkyne 5a using gold(III) as catalyst. The allene 15a was isolated in 57 % yield (Scheme 3a). Afterwards, we converted 15a under the optimized reaction conditions with [JohnPhosAu(NCMe)]SbF 6 as catalyst and 5a by NMR experiments. Polycycle 14a was obtained in 55 % yield. Thus, the formation of 15a via gold-catalyzed Claisen rearrangement can be described as the first sequence of the cascade reaction. Interestingly, the NMR yield of 14a after conversion of 15a with alkyne 5a (55 %) was not much higher than the NMR yield of 14a obtained in the reaction of 12 and 5a (52 %, Scheme 3a). This means that the one-pot cascade reaction is almost as efficient as the stepwise process. Please note, the reaction of 12 with [JohnPhosAu(NCMe)]SbF 6 as catalyst without alkyne 5a led to a complete conversion (Scheme 3a), but only a variety of oligomeric compounds was formed, which could not be separated by column chromatography. Therefore, we assume thatutilizing JohnPhosAu + as catalytic species -allene 15a (which is formed in the first step) reacts with the diene unit of the cycle. This leads to the formation of oligomeric structures due to the absence of intercepting alkynes. Another preparative mechanistic indication was achieved via conversion of 16 under the optimized reaction conditions. Bicycle 17 was obtained in 67 % yield whether alkyne 5a was present or not. The formation of the product underlines the assumption that the allene unit is cyclized with the diene unit building a five-membered ring. The formed cation can be stabilized by deprotonation. Thus, the second sequence of the cascade reaction probably starts with the formation of a five-membered ring.
The above-mentioned experimental indications were used to calculate the cascade reaction mechanism via quantum chemical methods. Arylacetylene 12 and alkyne 6a were utilized as model substances. In order to optimize the geometrical parameters of all stationary points, the density functional B3LYP [38][39][40] together with the dispersion correction via Becke-Johnson damping [41] (D3BJ) was employed. As basis sets 6-31G(d) was applied for the elements C, H, O, P, and Cl, whereas def2-TZVP was used for Au. Furthermore, single-point calculations on the optimized structures were performed using B3LYP-D3BJ with basis sets 6-311++G(d,p) (for C, H, O, P, and Cl) and def2-TZVP (for Au). To take solvent effects into account, chloroform was considered as reaction solvent by using the SMD [42] model.
The first sequence of the cascade reaction is the formation of the allene starting with chloroacetylene 12 and JohnPhosAu + as catalytic-active species. The coordination of the gold(I) com-  plex is accompanied by free-energy reduction amounting to 15.4 kcal/mol (Figure 1). Please note that the coordination of JohnPhosAu + to acetylene 6a (ΔG = -16.2 kcal/mol) is energetically favored compared to the complexation of 12. However, the reactions of the thusly formed complex with acetylenes 6a and 12, respectively, require high activation energies, making the complexation of 6a a dead end (Figures S8-9). Complex 18a, which is formed by coordination of JohnPhosAu + to acetylene 12, was used as reference for all subsequent reaction steps. The rate-determining step of the first cascade reaction sequence (allene formation) is the cyclization to oxonium ion 20a (10.0 kcal/mol). This ion converts into gold(I) allene complex 22a in the second step. Therefore, a barrier of 7.4 kcal/mol needs to be overcome. After that, two reaction paths are possible: on the one hand, decomplexation can take place leading to the catalytic species and allene 15a (this allene formation corresponds to the end of the first cascade reaction sequence). On the other hand, the formation of bicyclic cation 24a can occur initiating the second cascade reaction sequence (allene-diene-alkyne coupling). According to calculations, the second sequence (intramolecular cyclization) is more favored (12.5 kcal/mol) than the formation of free allenes (15.3 kcal/mol). This agrees well with the results showing a polymerization in case of converting 12 in absence of 5a with JohnPhosAu + as catalytic species.
Let us now consider the second cascade reaction sequence (allene-diene-alkyne coupling): the first step is again ratedetermining and corresponds to the addition of alkyne 6a to cation 24a with formation of vinyl cation 26a (Figure 2). The activation energy for this step amounts to 9.1 kcal/mol. The vinyl cation is stabilized by the phenyl ring, which also explains the regioselectivity of the attack of 6a to 24a. A possible competing reaction to the addition of alkyne 6a to cation 24a is the addition of alkyne 12 to cation 24a. However, the required activation energy for the latter reaction path is distinctly higher (15.5 kcal/mol; Figure S10). In the next step, two intramolecular C-C bonds are simultaneously formed leading to gold(I)-stabilized cation 28a. Proceeding from this σ complex, π complex 30a is formed by a 1,2-hydride shift. Decomplexation leads to the final product 14i.
In Figure 3 the most important intermediates are illustrated in a catalyst cycle. The whole cycle consists of two sequences; in both gold(I) complexed allene 22a is passed. Allene formation takes place via two-step gold-catalyzed Claisen-like rearrangement (black in Figure 3). Allene 22a functions as origin of two branches: on the one hand, the formation of allene 15a is enabled by decomplexation (green in Figure 3), on the other hand, the intramolecular 5-endo-trig-cyclization to 24a is possible (blue in Figure 3). Alkyne 6a can be added to intermediate 24a ending up in the formation of gold complexed polycycle 30a via two further steps. Decomplexation leads to final product 14i and finishes the catalyst cycle.
In sum, we were able to show that polycycles can be designed from haloacetylenes and alkynes by gold(I) catalysis. In our case, the haloacetylene was an o,o-dimethyl-substituted propargyl ether. The reaction tolerates a wide range of alkyne substrates like alkylaryl, diaryl, halo-and terminal alkynes. The achieved preparative yields amount up to 77 %. Experimental and quantum chemical investigations showed that the mechanism can be described as cascade reaction with two sequences. The first sequence is a gold(I)-catalyzed dearomatization in which the corresponding allene is built from propargyl ether by formal Claisen rearrangement (isolation of the allene is possible). The second sequence is a new reaction type: four new C-C bonds are formed by an allene-diene-alkyne coupling leading to the corresponding polycycle. Altogether, even five C-C bonds are newly formed during the whole cascade. The described allene-diene-alkyne coupling reaction has the potential to exploit new ways of polycycle formation starting from simple building blocks. Especially, the option of expanding the reaction to obtain intermolecular products is promising and desires further exploration.

Conflict of Interest
There is no conflict of interest to declare.