Pilar Calleja1, Óscar Pablo1, Beatrice Ranieri1, Morgane Gaydou1, Anthony Pitaval1, María Moreno1, Mihai Raducan1, Antonio M Echavarren2,3. 1. Institute of Chemical Research of Catalonia (ICIQ), Barcelona Institut of Science and Technology (BIST), Av. Països Catalans 16, 43007, Tarragona, Spain. 2. Institute of Chemical Research of Catalonia (ICIQ), Barcelona Institut of Science and Technology (BIST), Av. Països Catalans 16, 43007, Tarragona, Spain. aechavarren@iciq.es. 3. Departament de Química Analítica i Química Orgánica, Universitat Rovira i Virgili, C/ Marcel⋅li Domingo s/n, 43007, Tarragona, Spain. aechavarren@iciq.es.
Abstract
1,6-Enynes bearing OR groups at the propargyl position generate α,β-unsaturated gold(I)-carbenes/ gold(I) stabilized allyl cations that can be trapped by alkenes to form cyclopropanes or 1,3-diketones to give products of α-alkylation. The best migrating group is p-nitrophenyl ether, which leads to the corresponding products without racemization. Thus, an improved formal synthesis of (+)-schisanwilsonene A has been accomplished. The different competitive reaction pathways have been delineated computationally.
1,6-Enynes bearing OR groups at the propargyl position generate α,β-unsaturated pan class="Chemical">gold(I)-carbenes/ gold(I) stabilized allyl cations that can be trapped by alkenes to form cyclopropanes or 1,3-diketones to give products of α-alkylation. The best migrating group is p-nitrophenyl ether, which leads to the corresponding products without racemization. Thus, an improved formal synthesis of (+)-schisanwilsonene A has been accomplished. The different competitive reaction pathways have been delineated computationally.
The study of gold(I)‐catalyzed reactions of 1,n‐pan class="Chemical">enynes has led to the discovery of a wealth of cyclization modes, including mechanistically intriguing skeletal rearrangements and nucleophilic addition reactions to cycloaddition processes.1 In this context, we recently found that 1,6‐enynes such as 1 bearing propargyl alcohols, ethers, or silyl ethers react with gold(I) catalysts through the usual type of highly delocalized cyclopropyl gold(I) intermediates 2,2 which then undergo a new type of 1,5‐migration of the OR groups to generate species 3
3 that we postulated as intermediate between α,β‐unsaturated gold(I) carbenes and gold(I)‐stabilized allyl cations.2, 4 In the presence of carbonnucleophiles such as indole or furans, products 4
3 or trienes such as 5
5 were obtained by Friedel–Crafts‐type reactions (Scheme 1). Intermediates 3 can also react with electron‐rich alkenes to form the corresponding cyclopropanes,3 a reaction which is also characteristic of gold(I) carbenes 2.6, 7, 8 We demonstrated the potential of this tandem cyclization/1,5‐OR migration/cyclopropanation to form products 6 that were key intermediates in the first total synthesis of the natural sesquiterpene (+)‐schisanwilsonene A (7).9
1
Scheme 1
Intermolecular trapping of the intermediates of the gold‐catalyzed cyclization/1,5‐OR migration of enynes 1.
Intermolecular trapping of the intermediates of the gold‐catalyzed cyclization/1,5‐OR migration of enynes 1.This type of gold(I)‐catalyzed 1,5‐migration has been found to compete10 with 1,2‐ and 1,3‐migrations of propargylic carboxylate groups.10 Related processes have been found in the pan class="Chemical">gold(I)‐catalyzed reactions of dienynes 8, which undergo cyclization/1,5‐OR migration/intramolecular cyclopropanation through intermediates 9 to form stereoselectively hexahydroazulenes 10,3 which were the key intermediates in our total synthesis of the sesquiterpernes (−)‐epiglobulol (11) and (−)‐4β,7α‐aromadendranediol (12) (Scheme 2).11 Alternatively, when the gold(I)‐catalyzed reaction was performed in the presence of allyl alcohol, this external nucleophile reacted to give 9′, which underwent intramolecular cyclopropanation to give rise to the sesquiterpene (−)‐4α,7α‐aromadendranediol (13).11, 2
Scheme 2
Intramolecular trapping of the intermediates of the gold‐catalyzed cyclization/1,5‐OR migration of dienynes 8.
Intramolecular trapping of the intermediates of the gold‐catalyzed cyclization/1,5‐OR migration of dienynes 8.The proposed mechanism for these inter‐ and intramolecular reactions was based on the isolation of diverse products but not on a rigorous study of this intriguing process and its several possible competitive cycloisomerization pathways. Furthermore, although the chirality transfer in the intramolecular processes was satisfactory, partial racemization was observed in the formation of intermediates 2 when RO=AcO, which led to (+)‐schisanwilsonene A (7) with 9:1 e.r. Here we report that 1,3‐dipan class="Chemical">carbonyl compounds can also be used as the C‐nucleophiles to trap the putative α,β‐unsaturated gold(I) carbenes 3 leading to products of formal alkylation. This and additional studies on the intermolecular trapping of intermediates 3 with alkenes have allowed selecting p‐nitrophenyl ether as the protecting group of choice in these reactions. This led us to develop an improved formal synthesis of (+)‐schisanwilsonene A (7) in which the key step proceeds with total retention of configuration. We have also found cases in which a skeletal rearrangement takes place preferentially to form six‐membered ring compounds. A detailed computational study has been performed to understand the mechanisms of these complex transformations.
Results and Discussion
Selection of the best OR migration group
1,3‐Dicarbonyl compounds react as C‐pan class="Chemical">nucleophiles with 1,6‐enynes in the presence of gold(I) catalysts by formal attack at the alkene via opening of the cyclopropane of intermediates of type 2.1h We therefore decided to examine whether enynes 1 would react with this type of nucleophiles at the alkene, through intermediates 2, or at the terminal alkynecarbon through intermediates 3. In the event, reaction of enynes 1 with 1,3‐diphenyl‐1,3‐propandione in the presence of gold(I) complexes gave products of α‐alkylation of the dicarbonyl compounds 14 by trapping of intermediates 3 at C‐1 (Table 1). Using catalyst [(JohnPhos)Au(NCMe)]SbF6 (A), the best migrating group proved to be p‐nitrophenyl (PNP) ether in substrate 1 a, which led to adduct 14 a in 58 % yield after 30 min at 24 °C (Table 1, entry 1). p‐Anisyl ether derivative 1 e gave 14 e in a moderate 44 % yield (Table 1, entry 5), whereas lower yields (14–18 %) were obtained with the free alcohol 1 b, methyl ether 1 c, and acetate 1 d, and benzyl ethers 1 f, g failed to give any of the expected products 14 f, g (Table 1, entries 2–7). Poor results were obtained with catalysts B and C bearing tBuXphos as the ligand (Table 1, entries 8 and 9). The best results were obtained using the less electrophilic catalysts D–G with more donating NHC ligands (Table 1, entries 10–13), which have been proposed to enhance the carbene‐like character of the intermediates in gold(I)‐catalyzed reactions.1g,1h, 1
Table 1
Gold(I)‐catalyzed reaction of enynes 1 a–g with 1,3‐diphenyl‐1,3‐propandione.
Entry
[AuLL′]X
1 a–g
14 a–g (yield) [%]
1
A
1 a
14 a (58)
2
A
1 b
14 b (14)
3
A
1 c
14 c (18)
4
A
1 d
14 d (14)
5
A
1 e
14 e (44)
6
A
1 f
14 f (−)
7
A
1 g
14 g (−)
8[a]
B
1 a
14 a (8)
9[a]
C
1 a
14 a (8)
10
D
1 a
14 a (67)
11
E
1 a
14 a (67)
12
F
1 a
14 a (62)
13
G
1 a
14 a (71)
[a]=70–75 min
.
Gold(I)‐catalyzed reaction of pan class="Chemical">enynes 1 a–g with 1,3‐diphenyl‐1,3‐propandione.
[a]=70–75 min
.1,6‐Enyne 1 a reacted smoothly with other 1,3‐dicarbonyl compounds and β‐ketoesters as the C‐pan class="Chemical">nucleophiles using the optimal IPr gold(I) complex G to form adducts 14 h–n in 59–82 % yield in 30 min at 24 °C (Table 2).2
Table 2
Products of the reaction of 1,6‐enyne 1 a with 1,3‐dicarbonyl compounds and β‐ketoesters in the presence of catalyst G (24 °C, 30 min).
Products of the reaction of 1,6‐enyne 1 a with 1,3‐dicarbonyl compounds and β‐ketoesters in the presence of catalyst G (24 °C, 30 min).p‐Nitrophenyl ether 1 a was also the substrate of choice for the cyclization/1,5‐OR migration/intermolecular cyclopropanation with cyclohexene and pan class="Chemical">norbornene to form cyclopropanes 15 a, b in satisfactory yields as mixtures of exo and endo diastereomers by using catalyst A under very mild conditions (Scheme 3). Dihydropyrane reacted similarly to give 14 c, although in this case catalyst D gave better results. Indene, benzofuran, 5‐bromobenzofuran, and benzothiophene gave the corresponding adducts 15 d–g more stereoselectively. The configuration of the mayor product 15 g in the reaction with benzothiophene was assigned by X‐ray diffraction.12 This product was formed in lower yield most likely as a consequence of the inhibition of the catalytic reaction by coordination of gold(I) to the thioether 15 g. Interestingly, in the last three cases, the electron‐rich heterocycles undergo cyclopropanation, which is in contrast to what we observed previously for indole, which gave product 4 after formal alkylation at C‐3 (Scheme 1).3, 3
Scheme 3
Gold(I)‐catalyzed cyclization/1,5‐OR migration/intermolecular cyclopropanation of 1 a.
Gold(I)‐catalyzed cyclization/1,5‐OR migration/intermolecular cyclopropanation of 1 a.
A second‐generation formal synthesis of (+)‐schisanwilsonene A
The gold(I)‐catalyzed reaction of racemic 1 a with the bis‐TBS ether of 2‐methylenepropane‐1,3‐diol gave 6 a in good yield as a single diasteromer (Scheme 4).9 Enantioenriched pan class="Chemical">acetate 1 d (96:4 e.r.) reacted in slightly lower yield to afford 6 b in 91:9 e.r. This partial racemization was explained as a result of a 1,2‐acyloxy rearrangement competing as a minor pathway.9, 4
Scheme 4
First‐generation synthesis of (+)‐schisanwilsonene A (7) via cyclization/1,5‐OR migration/cyclopropanation.9
First‐generation synthesis of (+)‐schisanwilsonene A (7) via cyclization/1,5‐OR migration/cyclopropanation.9Although the formation of enantioenriched 6 b allowed us to complete the first total synthesis of (+)‐schisanwilsonene A (7) via hexahydroazulene diol 16 a (90:10 e.r.) (Scheme 4), a better solution has now been found by using enantioenriched 1 a as the substrate for the cyclization/1,5‐OR migration/cyclopropanation cascade. Thus, pan class="Chemical">alcohol 1 b was protected as the PNP ether by nucleophilic aromatic substitution of its potassium salt with p‐fluoronitrobenzene and 18‐crown‐6 to give 1 a with full retention of the configuration (94.5:5.5 e.r.) (Scheme 5). The key gold(I)‐catalyzed reaction of 1 a with the bis‐TBS ether of 2‐methylenepropane‐1,3‐diol, followed by desilylation with TBAF, provided diol 6 c in 68 % yield without detectable racemization, within experimental error (94:6 e.r.). The configuration of 6 c was determined by X‐ray diffraction.12 This result demonstrates that the enyne cyclization/1,5‐OR migration/cyclopropanation cascade takes place without any racemization, which excludes the involvement of a propargyl carbocation as an intermediate in the process. In the racemic series, 6 c has been converted into hexahydroazulene diol 16 a according to the original sequence followed by a two‐setp deprotection of the PNP group (76 % yield).13, 5
Scheme 5
Enantio‐retentive synthesis of key intermediate 6 c in the synthesis of (+)‐schisanwilsonene A (7).
Enantio‐retentive synthesis of key intermediate 6 c in the synthesis of (+)‐schisanwilsonene A (7).
Exceptions to the rule: no migration of the propargyl group
We also tried the reaction of 1,6‐enynes 17 a, b with a pan class="Chemical">propargylic amine (Scheme 6). In the case of 17 a, the electron‐rich p‐methoxyphenyl amine undergoes a gold(I)‐catalyzed intramolecular hydroarylation with the terminal alkyne14 to give dihydroquinoline 18, whereas PNP derivative 17 b afforded 7‐methylene‐2‐azabicyclo[2.2.1]heptane 19 as the major product, whose structure was determined by X‐ray diffraction.12, 6
Scheme 6
Reaction of propargyl amines 17 a, b with gold(I) catalyst A.
Reaction of propargyl amines 17 a, b with pan class="Chemical">gold(I) catalyst A.
We had reported that 1,6‐enyne 20 a reacts with methanol in the presence of catalyst A to give stereoselectively adduct 21 in which the migration of the benzyloxy group has not taken place15 (Scheme 7). Interestingly, pan class="Chemical">alcohol 20 b also reacted in the presence of catalyst A without migration of the OH group to give 22, the product of an endo‐type single cleavage rearrangement.16 This reaction was better performed in the presence of 4 Å molecular sieves. Surprisingly, in the absence of molecular sieves, diene 23 was obtained as the major product of the reaction, whose structure was determined by X‐ray diffraction. A speculative mechanism for the formation of this unexpected product could involve a reaction of 22 with allyl cation 24 to form a new allyl cation 25, followed by aromatization by proton loss and dehydration (Scheme 7).7
Scheme 7
Cyclizations of enynes 20 a, b and 22 with gold(I) catalyst A without OR migration and the proposed mechanism for the formation of 23.
Cyclizations of enynes 20 a, b and 22 with pan class="Chemical">gold(I) catalyst A without OR migration and the proposed mechanism for the formation of 23.
1,6‐Enynes 26 a–h with tertiary pan class="Chemical">propargyl hydroxyl or trimethylsilyloxy groups and different aryl groups at the alkene also react with catalyst A to give endo‐type single cleavage rearrangement products 27 under mild conditions (Table 3). The reaction proceeds satisfactorily with substrates bearing phenyl or aryl groups with electron‐withdrawing substituents (Table 3, entries 1–5 and 8; Figure 1), whereas enynes 26 f, g with a more electron‐rich anisyl group gave complex reaction mixtures (Table 3, entries 6 and 3, 1). Somewhat surprisingly, the reaction of the corresponding methyl ethers led only to decomposition.
Table 3
Gold(I)‐catalyzed reaction of 1,6‐enynes 26 a–h to give endo‐type single‐cleavage rearrangement products 27 a–h.
Entry
Enyne
R1
R2
Product (yield) [%]
1
26 a
H
H
27 a (52)
2
26 b
TMS
H
27 b (95)
3
26 c
H
p‐NO2
27 c (88)
4
26 d
TMS
p‐NO2
27 d (72)
5
26 e
H
o‐NO2
27 e (98)
6
26 f
H
p‐OMe
27 f (−)
7
26 g
TMS
p‐OMe
27 g (−)
8
26 h
H
o‐Br
27 h (84)
Figure 1
X‐ray diffraction structure of product 27 h (Table 3, entry 8).
Gold(I)‐catalyzed reaction of 1,6‐pan class="Chemical">enynes 26 a–h to give endo‐type single‐cleavage rearrangement products 27 a–h.
X‐ray diffraction structure of product 27 h (Table 3, entry 8).
Mechanistic discussion
We examined computationally the evolution of enynes Ia–c coordinated with pan class="Chemical">AuL+ (L=PMe3). In all cases, in agreement with calculations, Ia–c reacted by 5‐exo‐dig pathways to form preferentially IIa–c (Schemes 8 and 9), which correspond to intermediates of type 2 in Scheme 1.17 The alternative 6‐endo‐dig pathway leading to IIIa–c was less favorable in all cases. Whereas in the case of PNP‐protected intermediate IIa, the migration to form VIa proceeds in a direct manner through TS (Scheme 8), the migration of the OR group in complexes IIb and IIc proceeds via bicyclic intermediates IVb and IVc, which then lead to products of 1,5‐migration VIb and VIc through very low barrier transition states TS and TS, respectively (Scheme 9). The alternative evolution of IIa–c to products of single‐cleavage rearrangement Va–c was found to be thermodynamically the most favorable pathway,18 although kinetically the 1,5‐migration is more favorable as TS
– have higher energies than TS
– (Schemes 8 and 9).8, 9
Scheme 8
Alternative reaction pathways for the evolution of gold(I) complexes Ia into the product of 1,5‐OR migration VIa (L=PMe3). DFT calculations (B3LYP, 6‐31G(d,p) (C, H, P, O, N) and SDD (Au), CH2Cl2). Values for free energies in kcal mol−1.
Scheme 9
Alternative reaction pathways for the evolution of gold(I) complexes Ib and Ic into the products of 1,5‐OR migration VIb and VIc. DFT calculations (B3LYP, 6‐31G(d,p) (C, H, P, O) and SDD (Au), CH2Cl2). Values for free energies in kcal mol−1.
Alternative reaction pathways for the evolution of gold(I) complexes Ia into the product of 1,5‐OR migration VIa (L=pan class="Chemical">PMe3). DFT calculations (B3LYP, 6‐31G(d,p) (C, H, P, O, N) and SDD (Au), CH2Cl2). Values for free energies in kcal mol−1.
Alternative reaction pathways for the evolution of gold(I) complexes Ib and Ic into the products of 1,5‐OR migration VIb and VIc. DFT calculations (B3LYP, 6‐31G(d,p) (C, H, P, O) and SDD (pan class="Chemical">Au), CH2Cl2). Values for free energies in kcal mol−1.
The calculated structure for minimum VIb shows very similar C−C bond lengths of 1.41 and 1.39 Å (Figure 2). The Au−C bond length is 2.04 Å, which might correspond to a single pan class="Chemical">metal–carbon bond, and is similar to that found in well‐characterized heteroatom‐stabilized gold(I) carbenes.3, 4 Overall, the calculated structure fits better with a gold(I)‐stabilized allylic cation.2
Figure 2
Computed structure for minimum VIb (Scheme 9). Bond lengths are in Å.
Computed structure for minimum VIb (Scheme 9). Bond lengths are in Å.The observed reactivity trends when L=PMe3 are reproduced in cases where the ligand on pan class="Chemical">gold(I) is changed to bulkier PPh3phosphine or the model NHC ligand 1,3‐dimethylimidazol‐2‐ylidene (Table 4).4
Table 4
Evaluation on the ligand effect for the computed trends of gold(I) complex Ib.[a]
PMe3
PPh3
NHC[b]
Ib
0.0
0.0
0.0
TS1b
3.6
5.8
9.5
IIb
−1.1
−1.5
1.6
TS2b
7.4
9.3
11.3
IIIb
−3.1
−2.5
0.1
TS3b
2.1
4.4
7.9
IVb
−6.6
−4.3
−2.3
TS5b
−6.6
−5.1
−0.7
VIb
−15.7
−13.9
−13.3
TS4b
6.7
7.0
11.0
Vb
−37.5
−35.4
−33.1
[a] ΔG energies are given in kcal mol−1. [b] 1,3‐Dimethylimidazol‐2‐ylidene.
Evaluation on the ligand effect for the computed trends of pan class="Chemical">gold(I) complex Ib.[a]
[a] ΔG energies are given in kcal mol−1. [b] 1,3‐Dimethylimidazol‐2‐ylidene.DFT calculations for the reaction of gold(I) complex Id led to much less clear‐cut results (Scheme 10). Although the 5‐exo‐pan class="Chemical">dig pathway leading to IId was again more favorable than the 6‐endo‐dig cyclization to form IIId, the 1,5‐migration of IId through IVd to form VId was found to be the kinetically most favorable pathway, which is not what was observed experimentally for 20 b and 26 a–h.10
Scheme 10
Alternative reaction pathways for the evolution of gold(I) complex Id into the product of 1,5‐OR migration VId. Numbers in red correspond to the formation and evolution of IId“, the diastereomer of IId. DFT calculations (B3LYP, 6‐31G(d,p) (C, H, P, O) and SDD (Au), CH2Cl2). Values for free energies in kcal mol−1.
Alternative reaction pathways for the evolution of gold(I) complex Id into the product of 1,5‐OR migration VId. Numbers in red correspond to the formation and evolution of IId“, the diastereomer of IId. DFT calculations (B3LYP, 6‐31G(d,p) (C, H, P, O) and SDD (pan class="Chemical">Au), CH2Cl2). Values for free energies in kcal mol−1.
Conclusions
In general, 1,6‐enynes bearing OR groups at the propargyl position react through intermediates that can be formulated in a simplified manner as α,β‐unsaturated goldpan class="Chemical">(I) carbenes, which react in general with alkenes to form cyclopropanes or 1,3‐diketones to form products of α‐alkylation. We have found that among the various migrating OR groups, p‐nitrophenyl ether gives the best results. In addition, we have established that the gold(I)‐catalyzed enyne cyclization/1,5‐OR migration/cyclopropanation cascade takes place without racemization, which demonstrates that propargyl carbocations are not formed under the reaction conditions. This has been applied for the preparation of key intermediates for the synthesis of (+)‐schisanwilsonene A with higher enantiomeric purity. DFT calculations suggest that after the initial cyclization, the 1,5‐OR migration proceeds stepwise through a cyclic intermediate although the cleavage occurs through a very low barrier. However, additional mechanistic work is still required to understand why in cases in which propargyl group does not migrate an endo‐type single‐cleavage rearrangement is the most favorable reaction pathway.
As a service to our authors and readers, this journal provides supporting information supplied by the pan class="Chemical">authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.
SupplementaryClick here for additional data file.
Authors: Cristina Nieto-Oberhuber; Salomé López; M Paz Muñoz; Eloísa Jiménez-Núñez; Elena Buñuel; Diego J Cárdenas; Antonio M Echavarren Journal: Chemistry Date: 2006-02-08 Impact factor: 5.236
Authors: Javier Carreras; Madeleine Livendahl; Paul R McGonigal; Antonio M Echavarren Journal: Angew Chem Int Ed Engl Date: 2014-04-01 Impact factor: 15.336
Authors: Pilar Calleja; Óscar Pablo; Beatrice Ranieri; Morgane Gaydou; Anthony Pitaval; María Moreno; Mihai Raducan; Antonio M Echavarren Journal: Chemistry Date: 2016-08-16 Impact factor: 5.236
Authors: Pilar Calleja; Óscar Pablo; Beatrice Ranieri; Morgane Gaydou; Anthony Pitaval; María Moreno; Mihai Raducan; Antonio M Echavarren Journal: Chemistry Date: 2016-08-16 Impact factor: 5.236