Download PDF
Synlett 2010(11): 1694-1698  
DOI: 10.1055/s-0029-1219958
CLUSTER
© Georg Thieme Verlag Stuttgart ˙ New York

Creation of Quaternary Stereogenic Centers via Copper-Catalyzed Asymmetric Conjugate Addition of Alkenyl Alanes to α,β-Unsaturated Cyclic Ketones

Daniel Müller, Christine Hawner, Matthieu Tissot, Laëtitia Palais, Alexandre Alexakis*
Department of Organic Chemistry, University of Geneva, 30 Quai Ernest Ansermet, 1211 Geneve 4, Switzerland
Fax: +41(22)3793215; e-Mail: alexandre.alexakis@unige.ch;

Further Information

Publication History

Received 16 March 2010
Publication Date:
04 June 2010 (online)

 PDF Download Permissions and Reprints All articles of this category

Abstract

SimplePhos ligands proved to be very powerful ligands in the generation of quaternary stereogenic centers by Michael addition of alkenyl-aluminum reagents to cyclic enones. Using commercially available and easily accessible alkenylbromides as alane precursors the present procedure offers a facile access to β-alkenyl-substituted cyclohexanones with high enantioselectivities up to 96%.

Key words

aluminum - alkenes - copper - Michael addition - asymmetric catalysis

The landmark discovery about the rhodium-catalyzed conjugate addition of organoboron reagents to electron-deficient olefins by Miyaura and co-workers [¹] in 1997 gave rise to tremendous research efforts towards its asymmetric version. [²] Already in 1998 Hayashi showed that the use of a rhodium-BINAP complex achieved high enantio­selectivities for the ACA of boronic acids to acyclic and cyclic enones. [³] In his report he also employed alkenyl boronic acids which afforded the corresponding products with high enantioselectivity. In contrast to the advances in this fast-growing field of rhodium-catalyzed ACA, the copper-catalyzed ACA employing alkenyl nucleophiles has been rather neglected and only few examples have been reported so far. [4]

In our continuing interest in the copper-catalyzed ACA we are focusing our attention towards the creation of quaternary centers via Michael addition. [4d] [5] To this date no example has been published of the creation of quaternary stereogenic centers via the rhodium-catalyzed ACA with alkenyl nucleophiles and enones as substrates. [6] For ­copper-catalyzed ACA only some examples have been published so far; which include the drawback of using high catalyst loadings up to 30 mol%. [4c] [d] [f]

Mixed alanes as previously shown, [4d] [7] combine strong nucleophilicity as well as electrophilic activation via complexation of the carbonyl moiety and thus allow the addition of various nucleophiles to tertiary substituted α,β-unsaturated ketones. From our previous investigations we knew that iodoolefins as precursors underwent this transformation with good enantioselectivity. [4d] However, this protocol was quite sensitive to the presence of in situ generated LiI, and so we decided to experiment with other nucleophile precursors, namely bromoolefins, of which many are commercially available or easily accessible synthetically. [8] Using a variant of the procedure published by Seebach, [9] we employed two equivalents of t-BuLi to perform the bromo-lithium exchange and then quenched the corresponding alkenyllithium with Me2AlCl to generate the mixed alane species (Scheme  [¹] ). On the contrary the use of n-BuLi for the Br-Li exchange lead in the case of 2-bromopropene to acetylenic byproducts. [9]

Scheme 1 Generation of mixed alane species

It is well known, that in the case of mixed alane species the substituents with sp- and sp²-carbon centers attached to the aluminum are faster transferred than substituents bearing sp³-carbon centers. [¹0] [¹³] Therefore the two alkyl substituents act usually as ‘dummy substituents’ meaning in most cases they transfer in small quantities or not at all. [4d] To investigate the 1,4 addition to β-substituted cyclic enones we chose 3-methylcyclohex-2-enone as a Michael acceptor and dimethyl(prop-1-en-2-yl)aluminum as nucleophile. From our experiences with the conjugate addition employing trialkylalanes and mixed aryl­alanes, [4c] [d] [f] we knew that CuTC as copper source, diethyl ether as solvent, and temperatures about -30 ˚C were good parameters to start with and screened under these conditions various ligands (Table  [¹] , Figure  [¹] , Scheme  [²] ).

Table 1 Screening of the Ligandsa
Entry Ligand Conv.
(%)b 		Methyl/alkenyl transferb ee
(%)c
 1 L1  94  4:96 12 (R)
 2 L2  85 11:87  5 (S)
 3 L3  98  3/97 18 (R)
 4 L4 100  4:96 32 (R)
 5 L5  98  2:98 43 (R)
 6 L6 100  3:97 49 (R)
 7 L7  97 10:90 26 (R)
 8 L8  98  3:97 19 (R)
 9 L9  88  2:98 rac.
10 L10 100  2:98 59 (R)
11 L11 100  4:96 63 (R)
12 L12   2 n.d. n.d.
13 (R)-BINAP  57 n.d. 13 (R)
a Reaction conditions as described in Scheme  [²] .
b Determined by GC-MS.
c Determined by chiral GC.

The results of the ligand screening confirmed previous observations: An increase of the steric bulk on the BIPOL or aryl part and on the amine part promotes the enantioselectivity. [4f] This trend holds true for BIPOL (entries 3 vs. 4, Table  [¹] ), as well as for the SimplePhos ligands (entries 5 vs. 10, Table  [¹] ). To our delight SimplePhos ligands, a new class of ligands recently developed in our group, [4f] [¹¹] showed to be superior over all other classes of ligands we employed for the screening (Figure  [¹] ). SimplePhos ligand L11 which was already successfully used for the copper-catalyzed ACA using trialkyl- and mixed aryl-alanes proved to be the most selective of its class. [4d] [f] Interestingly, the electronic properties seem to play a major role in this transformation, the electron-poor ligand L9 gave almost racemic mixtures whereas the phenyl-substituted ligand L10 gave similar results as L11. For the majority of the optimization reactions we continued using the simplest SimplePhos ligand L5 due to its simple one-step synthesis. [4f] Gau and co-workers showed that the amount of alane used had a significant impact on the enantioselectivity in the addition of 2-thienylalane to 2′-acetophen­one, [¹²] therefore we considered investigating the dependency of the enantioselectivity on the amount of alane used for the reaction (Table  [²] ). To our surprise the amount of alane added had a huge influence on the enantioselectivity and caused significant changes up to 30% in difference.

Table 2 Study of the Effect of the Amount of Alane Added in the ACA as Described in Scheme  [²] a,b
Entry Alane
(equiv) 		Conv.
(%)c 		Methyl/alkenyl transferc ee
(%)d
1 3.0 100  3:97 49 (R)
2 2.8 100  3:97 52 (R)
3 2.3 100  6:94 62 (R)
4 2.0 100  5:95 66 (R)
5 1.7 100  9:91 68 (R)
6 1.4 100 20:80 73 (R)
7 1.2 100 25:75 78 (R)
a Reaction conditions as described in Scheme  [²] .
b Ligand L5 was used for the optimization.
c Determined by GC-MS.
d Determined by chiral GC.

Figure 1 Screened ligands

Scheme 2 Standard reaction used for optimization

Even though less equivalents of alane afforded higher enantioselectivity it also led to an increase of undesired methyl transfer. We considered therefore the use of 2.0 equivalents of alane as the best compromise between enantioselectivity and methyl transfer (entry 4, Table  [²] ). Screening of various copper sources led in a couple of cases to a significant increase in enantioselectivity but also to an increase in methyl transfer (entries 3 and 4, Table  [³] ). Therefore we decided to continue using CuTC as a copper source which afforded acceptable levels of enantioselectivities while reducing the amount of methyl transfer.

Table 3 Screening of Different Copper Sourcesa,b
Entry CuX Conv.
(%)c 		Methyl/alkenyl transferc ee
(%)d
1 CuTC 100  5:95 66 (R)
2 CuCN  63  4:96 32 (R)
3 Cu(OAC)2˙H2O 100 12:88 73 (R)
4 CuCl2  73 11:89 70 (R)
5 Cu(acac)2 100 21:79 66 (R)
6 Cu(OTf)2  93 19:81 63 (R)
7 CuBr  97  5:95 59 (R)
a Reaction conditions as described in Scheme  [²] , except the use of
2.0 equiv of alane.
b Ligand L5 was used for the optimization.
c Determined by GC-MS.
d Determined by chiral GC.

We also investigated the influence of the solvent on the reaction and found that THF led to almost racemic mixtures, CH2Cl2 to high methyl transfer and hexane to low conversion. Good results were only observed with methyl tert-butyl ether (MTBE) and diethyl ether (Table  [4] ).

Table 4 Investigation of Various Ratios of MTBE and Et2O on the Enantioselectivitya,b
Entry Solventc Solventd Conv.
(%)e 		Methyl/alkenyl transfere ee
(%)f
1 Et2O Et2O 100  3:97 89 (R)
2 Et2O Et2Og  96  6:94 93 (R)
3 Et2O Et2Oh 100 23:67 93 (R)
4 Et2O MTBE  91  3:97 95 (R)
5 MTBE Et2O 100  6:94 93 (R)
6 MTBE MTBE  58  5:95 96 (R)
a Reaction conditions as described in Scheme  [²] , except the use of
2.0 equiv of alane.
b Ligand L11 was employed.
c Solvent used for the generation of the alane.
d Solvent used for the complex formation (CuTC and ligand).
e Determined by GC-MS.
f The ee was determined by chiral GC.
g 15mol% of ligand used.
h Cu(OAc)2˙H2O was used as copper source.

Diethyl ether as a solvent for complex formation seems to be necessary for good conversions (entries 1 and 5, Table  [4] ), whereas MTBE favored high enantioselectivites (entries 5 and 6, Table  [4] ). To ensure complete conversion and at the same time to increase the enantioselectivity we prepared the complex in diethyl ether and the alane in MTBE which afforded the same levels of enantioselectivity as observed when employing 15 mol% of the ligand (entry 2). For all the above results the reactions were performed on a typical scale of 10.8 mmol, in order to ensure the consistency of the alane we prepared. To exclude a dependency of the reaction outcome on the amount of alane prepared we synthesized several propenyl-alane solutions in diethyl ether in a 1.8-4.5 mmol scale and observed no major difference in enantioselectivity by using these solutions. We also investigated the stability of the propenyl-alane stock solutions and observed that storing the solution at room temperature over one month afforded comparable levels of enantioselectivity and conversion. This finding shows that alkenylalanes possess the same levels of stability as previously observed for arylalanes. [7] Before testing the optimized conditions for different substrates we wanted to check if the addition of lithium salts and variations in the t-BuLi/Me2AlCl ratio would have an influence on the reaction (Table  [5] ).

Table 5 Influence of Lithium Salts and Variation of the t-BuLi/Me2AlCl Ratio on the Reaction Outcomea,b
Entry Ratio t-BuLi/Me2AlCl Conv. (%)c Methyl/alkenyl transferb ee (%)d
1 2.0:1.0 100  6:94 93 (R)
2e 2.0:1.0 100  4:96 83 (R)
3 1.8:1.0  91 30:70 83 (R)
4 2.0:0.9 100  4:96 77 (R)
a Reaction conditions as described in Scheme  [²] , except the use of
2.0 equiv of alane and complex formation in MTBE.
b Ligand L11 was employed.
c Determined by GC-MS.
d Determined by chiral GC.
e Addition of salts.

The addition of salts proved to be detrimental for the enantioselectivity (Table  [5] , entry 2). The investigation of the t-BuLi/Me2AlCl showed that ensuring a precise ratio of t-BuLi and Me2AlCl was critical to achieve good conversions and enantioselectivities. [¹³] Any excess of either t-BuLi or Me2AlCl led to a loss of enantioselectivity (entries 3 and 4) and in the case of an excess of Me2AlCl also to low conversion and high methyl transfer (entry 3). Therefore we used only fresh solutions of t-BuLi and Me2AlCl to ensure that the used concentration was as close to the indicated concentrations as possible. With the optimized conditions in hand we applied this methodology to various nucleophiles and substrates and carried out the reactions on a 1.5 mmol scale instead of a 0.3 mmol scale, as used for the optimization reactions, in order to prove its synthetic use (Table  [6] ). For methcyclohexenone 4 all the tested nucleophiles gave good to excellent levels of enantioselectivity. The difference in enantioselectivity for product 5 (entry 1, Table  [5] vs. entry 1, Table  [6] ) is probably due to a slightly different ratio in the t-BuLi/Me2AlCl ratio. Due to the sensitivity of the reaction towards the purity of the alkenyl alane slight variations on enantioselectivity are inevitable. In the case of product 6, n-hexenylbromide as nucleophile precursor contained 9% of an impurity which we could not remove by common purification methods. Whether this impurity is responsible for the drop in enantioselectivity is not clear yet. Product 8 has been isolated in poor yield due to the fact that commercially available α-bromostyrene contains approximately 4% of the trans-β-bromostyrene which was much faster transferred and thus led to a 8:92 ratio of 7/8. [¹4] Nevertheless we succeeded in separating compound 8 from 7 and the methyl adduct by column chromatography. This shows that, especially in the case of secondary bromoalkenes as nuclophile precursors, it is absolutely vital to avoid β-bromoalkene impurities which might lead to a large amount of undesired and difficult to separate side product.

Table 6 Screening of Alkenylalanes [¹5]

Entry Product Conv. (%)a Methyl/alkenyl transfera Yield (%)b [α]D ²0 ee (%)c
1 5 100  4:96 85 -62.5 91 (R)
2 6  94  2:98 50 -39.2 76 (R)
3 7  97  0:100 64 -70.8 91 (R)
4 8  94 11:89 27 -35.6 94 (R)
5 9  97  1:99 71 -92.1 96 (R)
a Determined by GC-MS.
b Isolated yield.
c The ee was determined by chiral GC and chiral SFC.

It should also be noted that encumbered nucleophiles led to higher levels of methyl transfer whereas in the case of unhindered nucleophiles, such as the nucleophiles leading to products 6, 7, and 9 almost no methyl transfer was observed. The absolute configuration of compound 6 has been determined by comparison of its optical rotation with previously reported data, [4d] whereas for compounds 5, 7-9 we suppose that the ligand directs the attack towards the same facial side, as demonstrated for SimplePhos ligands before. [4f]

Table 7 Screening of Substrates

Entry Substrate Conv.
(%)a 		Methyl/alkenyl transfera ee
(%)b
1

10 		49 n.d. 48
2

11 		19 0:100 43
3

12 		17 n.d. 83
a Determined by GC-MS.
b The ee was determined by chiral GC.

Having explored different kinds nucleophiles, we wanted to explore the reaction scope by choosing more challenging substrates such as 10-12. Unfortunately, in all cases a drop of conversion and enantioselectivity was observed. The difficulties encountered when using cyclopentenones as substrates for the copper-catalyzed ACA have been reported several times in the literature. [4d] [e] [5] We believe that the reason for the drop in reactivity is the diminished orbital overlap between the π-orbitals of the C=O bond and the C=C bond due to a smaller interior angle of the cyclopentenone scaffold. This would lead to a higher LUMO at the β-position and thus lower the reactivity. A conjugating group such as phenyl will probably also lead to a deactivation of the β-position of the conjugated system by decreasing the positive partial charge via π-donation. This would explain the drop in conversion for substrate 11. Substrate 12 shows that the reaction is much more sensitive to bulky substrates than to sterically demanding nucleophiles such as 8. The level of enantioselectivity, however, did only drop slightly (entry 3, Table  [7] ).

In summary, we have described an efficient copper-catalyzed 1,4-addition of vinyl alanes to cyclic enones. Reactions are promoted by SimplePhos ligands which afford the desired β-alkenylcycloalkanones in up to 96% enantioselectivity. The present methodology allows ACA reactions to create quaternary stereogenic centers employing commercially available or easily accessible alkenylbromides.

Challenging substrates, such as 10-12 evidently need more powerful ligands for better enantioselectivity and conversion. The development of such ligands and application of the methodology towards the synthesis of natural products are among the objectives being pursued in our laboratories.

Acknowledgment

We thank the Swiss National Research Foundation (grant N˚ 200020-126663) and COST action D40 (SER contract N˚ C07.0097) for financial support, as well as BASF for the generous gift of chiral amines.

15

General Procedure for the Cu-Catalyzed ACA Employing Alkenylalanes Exemplified for Product 5 To a solution of 2-propenylbromide (480 µL, 653 mg, 5.4 mmol, 1.0 equiv) in MTBE (6.0 mL) was added under inert atmosphere a fresh solution of t-BuLi (6.75 mL, 10.8 mmol, 1.6 M in pentane, 2.0 equiv) at -78 ˚C. The reaction was stirred for 30 min at this temperature. Then a fresh solution of Me2AlCl (6.0 mL, 5.4 mmol, 0.9 M in heptane, 1.0 equiv) was added, and the reaction mixture was stirred for another 2 h maintaining the temperature at -78 ˚C. Now the cooling bath was removed, and the reaction vessel was immediately submerged in a water bath. The alane was stirred over night at r.t., and 1 h before use of the solution for catalysis stirring was stopped to ensure precipitation of the salts. Then, 10.5 mL corresponding to 2.0 equiv of the supernatant solution of alane was taken out with a syringe and slowly added to the metal complex. In a separate flask, CuTc (28.5 mg, 0.15 mmol, 10 mol%), ligand L11 (93.5 mg, 0.17 mmol, 11 mol%), and Et2O (5.0 mL) were thoroughly stirred at r.t.
for 1 h. Then the flask was cooled to -30 ˚C, and the corresponding alane (10.5 mL, 3.0 mmol, 2.0 equiv) was added. After 15 min of stirring 3-methyl-2-cyclohexenone 4 (170 µL, 165 mg, 1.50 mmol, 1 equiv) was added, and the reaction mixture was stirred for 18 h at this temperature. Then the reaction mixture was quenched at -30 ˚C with MeOH (1.0 mL) and let warm to r.t. An aqueous solution of HCl (10%, 15 mL) was added, followed by Et2O (50 mL). Extraction of the aqueous phase with Et2O (2 × 50 mL) and addition of NaOCl solution (10%, 4 mL) to the combined organic solvents afforded a pale yellow suspension which after extensive shaking turned into a blue suspension.¹6 After removal of the aqueous phase the organic phase was dried over Na2SO4, and the solvent was removed in vacuo. The remaining crude oil was purified by flash chromatography (SiO2; pentane-Et2O = 7:1), and the pure compound 5 was afforded as a colorless oil with a pleasant eucalyptus like fragrance (194 mg, 1.27 mmol, 85%, R f  = 0.23 in pentane-Et2O = 9:1). The analytical data were in accord with the ones reported in the literature.¹7 Chiral separation: Chirasil
DEX-CB, 60-0-1-115-0-20-170, 50 cm/s, t R1 = 44.80 min, t R2 = 47.35 min.

16

NaOCl solution was added to oxidize remaining SimplePhos ligand L11 and thus facilitate the purification.

15

General Procedure for the Cu-Catalyzed ACA Employing Alkenylalanes Exemplified for Product 5 To a solution of 2-propenylbromide (480 µL, 653 mg, 5.4 mmol, 1.0 equiv) in MTBE (6.0 mL) was added under inert atmosphere a fresh solution of t-BuLi (6.75 mL, 10.8 mmol, 1.6 M in pentane, 2.0 equiv) at -78 ˚C. The reaction was stirred for 30 min at this temperature. Then a fresh solution of Me2AlCl (6.0 mL, 5.4 mmol, 0.9 M in heptane, 1.0 equiv) was added, and the reaction mixture was stirred for another 2 h maintaining the temperature at -78 ˚C. Now the cooling bath was removed, and the reaction vessel was immediately submerged in a water bath. The alane was stirred over night at r.t., and 1 h before use of the solution for catalysis stirring was stopped to ensure precipitation of the salts. Then, 10.5 mL corresponding to 2.0 equiv of the supernatant solution of alane was taken out with a syringe and slowly added to the metal complex. In a separate flask, CuTc (28.5 mg, 0.15 mmol, 10 mol%), ligand L11 (93.5 mg, 0.17 mmol, 11 mol%), and Et2O (5.0 mL) were thoroughly stirred at r.t.
for 1 h. Then the flask was cooled to -30 ˚C, and the corresponding alane (10.5 mL, 3.0 mmol, 2.0 equiv) was added. After 15 min of stirring 3-methyl-2-cyclohexenone 4 (170 µL, 165 mg, 1.50 mmol, 1 equiv) was added, and the reaction mixture was stirred for 18 h at this temperature. Then the reaction mixture was quenched at -30 ˚C with MeOH (1.0 mL) and let warm to r.t. An aqueous solution of HCl (10%, 15 mL) was added, followed by Et2O (50 mL). Extraction of the aqueous phase with Et2O (2 × 50 mL) and addition of NaOCl solution (10%, 4 mL) to the combined organic solvents afforded a pale yellow suspension which after extensive shaking turned into a blue suspension.¹6 After removal of the aqueous phase the organic phase was dried over Na2SO4, and the solvent was removed in vacuo. The remaining crude oil was purified by flash chromatography (SiO2; pentane-Et2O = 7:1), and the pure compound 5 was afforded as a colorless oil with a pleasant eucalyptus like fragrance (194 mg, 1.27 mmol, 85%, R f  = 0.23 in pentane-Et2O = 9:1). The analytical data were in accord with the ones reported in the literature.¹7 Chiral separation: Chirasil
DEX-CB, 60-0-1-115-0-20-170, 50 cm/s, t R1 = 44.80 min, t R2 = 47.35 min.

16

NaOCl solution was added to oxidize remaining SimplePhos ligand L11 and thus facilitate the purification.

   Permissions and Reprints All articles of this category

Scheme 1 Generation of mixed alane species

Figure 1 Screened ligands

Scheme 2 Standard reaction used for optimization