Enantioselective Carboetherification/Hydrogenation for the Synthesis of Amino Alcohols via a Catalytically Formed Chiral Auxiliary.

Chiral auxiliaries and asymmetric catalysis are the workhorses of enantioselective transformations, but they still remain limited in terms of either efficiency or generality. Herein, we present an alternative strategy for controlling the stereoselectivity of chemical reactions. Asymmetric catalysis is used to install a transient chiral auxiliary starting from achiral precursors, which then directs diastereoselective reactions. We apply this strategy to a palladium-catalyzed carboetherification/hydrogenation sequence on propargylic amines, providing fast access to enantioenriched chiral amino alcohols, important building blocks for medicinal chemistry and drug discovery. All stereoisomers of the product could be accessed by the choice of ligand and substituent on the propargylic amine, leading to a stereodivergent process.

Currently, most enantioselective transformations rely on two strategies: (i) the use of chiral auxiliaries[1] and (ii) asymmetric catalysis.[2] The former allows the development of general and robust n class="Chemical">processes, but requires stoichiometric amounts of enantiopure precursors and multistep procedures. By contrast, asymmetric catalysis relies only on substoichiometric amounts of enantiopure molecules, but it generally requires an intensive optimization at the expense of robustness and generality. To overcome these limitations, we envisioned a catalytic enantioselective method, which would introduce a chiral auxiliary on the substrate from a cheap nonchiral tether in a synthetic useful step (Scheme A). This process would require only a catalytic amount of enantiopure species while providing a robust platform for further diastereoselective functionalizations, benefiting from the best aspects of the two traditional strategies. To the best of our knowledge, such an approach has not yet been realized, although different methods for improving asymmetric synthesis have been developed. A seminal work based on the formation of chiral aminals is the “self-repn>roduction of chirality” repn>orted by Seebach for the stereoselective synthesis of amino acids. In this work, the existing stereocenter on the amino acid first controls the diastereoselective formation of the aminal by condensation with an aldehyde. The latter then shields one face of the enolate.[3,4] As another example based on an internal chirality transfer, Maulide and co-workers recently reported a redox-neutral coupling of alkenes and aldehydes via a “catch–release” tethering approach (Scheme B).[5] However, the resulting functional group (a ketone) remains in the product. Other researchers have worked on the concept of “transient chiral auxiliaries/tethers”, which are easy to install and remove.[6−10] For example, Beauchemin and co-workers have used chiral aldehydes in substoichiometric amounts for the Cope-type hydroamination of allyl amines (Scheme C).[11] However, the scope of these transformations remains limited, and auxiliaries available from the chiral pool are generally required.

Scheme 1

(A) Our Concept: Catalytically Formed Chiral Auxiliaries; (B) Chirality Transfer via Tethering; (C) Transient Chiral Auxiliaries Introduced from Chiral Pool; (D) Implementation for the Stereodivergent Synthesis of Amino Alcohols

To implement our concen class="Chemical">pt, we considered the pan class="Chemical">palladium-catalyzed carboetherification of pan class="Chemical">propargylic amines,[12] based on the use of trifluoroacetaldehyde-derived tethers (Scheme D).[13−15] The stereocenter formed in this step could direct a subsequent functionalization of the double bond, acting de facto as a chiral auxiliary. The rigid nature of the oxazolidine scaffold containing the stereocenter should secure a high level of diasteroselectivity to the following transformations.

Concerning the following diastereoselective functionalization, we found the hydrogenation of the formed double bond particularly attractive. By comn class="Chemical">parison, the enantioselective hydrogenation of alkyl- or heteroatom-tetrasubstituted olefins is highly challenging, with only few limited catalytic enantioselective systems repn>orted.[16−18] After removal of the tether molecule, this pn>rocess would provide amino alcohols, key building blocks in synthetic and medicinal chemistry, which have been the focus of intensive methodology development recently.[19−26] In particular, the diaryl-substituted pan class="Chemical">amino alcohols obtained using this strategy can be found in antidepressants[27,28] and have served as intermediates for the synthesis of antimycotic, antibacterials[29] and antiviral molecules.[30,31] However, the selective synthesis of one of the four possible stereoisomers of the amino alcohols generally requires multistep processes.

To make this process successful, an enantioselective carboetherification sten class="Chemical">p had to be developed. The reversible formation of the hemiaminal I from the propargylic amine prevents asymmetric induction at this stage (see Supporting Information (SI), section E for more details). Therefore, a dynamic kinetic asymmetric transformation (DYKAT) needs to take place: in the presence of a chiral catalyst, one enantiomer of I should react preferentially to give oxazolidine II enantioselectively. Although palladium-catalyzed DYKATs have been reported,[32] the envisaged process is highly challenging, due to the large distance between the chiral metal complex and the stereocenter. To the best of our knowledge, such a DYKAT process has never been realized in the palladium-catalyzed functionalization of alkynes. If successful, the selection of the substitution pattern on the alkyne and on the aryl electrophile, together with the choice of the suitable enantiomer of the chiral ligand on the palladium catalyst would provide a simple enantio- and diastereodivergent access to all four stereoisomers of the amino alcohol. This is especially attractive for medicinal chemistry, as each stereoisomer may have different bioactivity, and the development of stereodivergent methods has been the topic of intensive research in asymmetric catalysis recently.[33−35]

We tested the feasibility of our plan by examining the n class="Chemical">pan class="Chemical">palladium-catalyzed tethered carboetherification of the readily available class="Chemical">n>an class="Chemical">propargylic amine 1 with iodotoluene 2 to access tetrasubstituted olefin 4 bearing a chiral oxazolidine fragment (Table ). 1-Ethoxy trifluoroethanol 3, a commercially available ethyl hemiacetal of trifluoroacetaldehyde, was chosen as the electrophilic molecular tether, and Pd2(dba)3·CHCl3, as the palladium source.[12] We first focused on the identification of a suitable ligand that could secure a high level of stereoinduction in the process (for details, see SI, section C). Commonly used bidentate BINAP L1 and Josiphos ligands L2 and L3 were not competent for this reaction (entry 1). The P,N ligands L4 and L5, derived from the corresponding Ugi’s amines,[36] delivered 4 in moderate yield and enantiomeric excess (entries 2 and 3), nevertheless demonstrating that a DYKAT was possible. However, higher asymmetric induction could not be achieved with this class of ligands. The P,N ligand S-iPrPhox L6 yielded the desired product in 72% yield and 40% e.e. (entry 4). Promising results were obtained evaluating the Trost type ligands, commonly used for palladium catalyzed asymmetric allylation reactions.[37,38] In particular, the commercially available DACH-phenyl Trost ligand L7 delivered product 4 in 49% yield and 64% e.e. (entry 5). Having in mind the previous positive results obtained with P,N ligands, we substituted the 2-(PPh2)-aryl fragment with a 2-pyridine.[39] This change increased the e.e. to 74% (entry 6). Surprisingly, the best results were finally obtained employing the benzamide derived L9 lacking a second strongly coordinating site, which delivered quantitatively 4 with 90% e.e. (entry 7). To the best of our knowledge, ligand L9 has been reported only twice in the literature,[40,41] and it was not suitable for imparting high stereocontrol, as two strong coordinating sites were required for asymmetric induction. We developed a robust and operationally simple route for accessing both enantiomers of L9 on multigram scale (SI, section B4). Demonstrating the process’s robustness, the reaction could be performed in more “industrially preferred” solvents[42] (ethyl acetate, methyl tert-butyl ether and toluene, entries 8, 9 and 10), without loss of yield and enantioselectivity, except for ethyl acetate (entry 8). Finally, the reaction could be scaled up to a 0.40 mmol scale, reducing the catalyst and ligand loading to 1.25 and 3.5 mol %, resulting in an improved stereoselectivity of 94% e.e. (entry 11).

Table 1

Optimization Studies

entry	Conditions	4 yield (%)a	e.e. (%)
1	Ligand L1, L2 or L3	<5	–
2	Ligand L4	44	20
3	Ligand L5	33	30
4	Ligand L6	72	40
5	Ligand L7	49	64
6	Ligand L8	45	74
7	Ligand L9	>95	90
8	L9, EtOAc instead of Et2O	80	84
9	L9, MTBE instead of Et2O	89	89
10	L9, Toluene instead of Et2O	91	91
11b	L9, 0.40 mmol of 1	>95	94

NMR yields.

Reaction performed using 1.25 mol % of Pd2(dba)3·CHCl3 and 3.5 mol % of (S,S)-ligand.

Reaction performed using 1.25 mol % of n class="Chemical">pan class="Chemical">Pd2(dba)3·class="Chemical">n>an class="Chemical">CHCl3 and 3.5 mol % of (S,S)-ligand.

The structure of 4, obtained by X-ray single-crystal analysis (Scheme A), shows that the trifluoropan class="Chemical">methyln> group is efficiently shielding one of the two enantiotopic faces of the olefin, setting the stage for the stereospecific hydrogenation. Indeed, when we submitted 4 to classical conditions for heterogeneous hydrogenation using Pearlman’s catalyst,[43] the desired hydrogenated product 5 was obtained as a single diastereoisomer in 79% yield and 94% e.e. (Scheme B). The use of Pearlman’s catalyst also allowed simultaneous removal of the pan class="Chemical">benzyl protecting group.

Scheme 2

(A) X-ray Crystal Structure of the Product 4; (B) Optimized Conditions for the Diastereoselective Hydrogenation

Various pan class="Chemical">aryl propargylic aminesn> were well tolerated in the reaction, regardless of the position of the substituents on the phenyl ring, as well as their electronic and steric properties (Scheme , 4, 6–14). The geometry of the olefin can be switched by just exchanging the aryl group on the pan class="Chemical">alkyne and the aryl iodide (4 vs 6). The reaction tolerates heterocycles such as pyridine and thiophene on the alkyne, although an erosion of the enantioselectivity was observed (15 and 16). Alkyl propargylic amines delivered products 17–19 bearing a methyl, a benzyl, and a cyclopropyl group. The reaction could be performed on a 5 mmol scale providing 2.0 g of 4 (quantitative yield) without erosion of the optical purity. The absolute configuration of the products were assigned by X-ray analysis of 4, confirming in addition that the aryl group coming from the iodide is incorporated in trans position to the oxygen.

Scheme 3

Scope of the Enantioselective Carboetherification

Reactions performed on a 0.40 mmol scale using 1.3 equiv of aryl iodide and 1.4 equiv of 1-ethoxy trifluoroethanol (3). Isolated yields and HPLC enantiomeric excess are given.

Dichloroethane (DCE) instead of Et2O.

Using 2.5 mol % of Pd2(dba)3·CHCl3 and 7 mol % of ligand.

DCE at 60 °C.

Reactions performed on a 0.40 mmol scale using 1.3 equiv of n class="Chemical">pan class="Chemical">aryl iodide and 1.4 equiv of 1-ethoxy trifluoroethanol (3). Isolated yields and Hclass="Chemical">n>an class="Chemical">PLC enantiomeric excess are given.

pan class="Chemical">Dichloroethanen> (pan class="Chemical">DCE) instead of pan class="Chemical">Et2O.

Using 2.5 mol % of pan class="Chemical">Pdn>2(dba)3·pan class="Chemical">CHCl3 and 7 mol % of ligand.

pan class="Chemical">DCEn> at 60 °C.

The investigation of the scope of the n class="Chemical">pan class="Chemical">iodoarene showed that numerous synthetically useful functional groupn>s, including pan class="Chemical">ethers, pan class="Chemical">amines, halogens, esters, nitriles or aldehydes, are well tolerated independently from their electronic and steric properties or position on the benzene ring (20–30). 2-Iodothiophene and 3-iodopyridine delivered products 31 and 32 in good yields. Finally, N-methyl, N-phenyl and N-para-methoxybenzyl (PMB) propargyl amines delivered products 33–35 in 52–58% yield and 54–92% e.e.

The obtained enantioenriched pan class="Chemical">tetrasubstituted olefinsn> were then submitted to the optimized conditions for the diastereoselective hydrogenation (Scheme ). Products 36–55 were all obtained as single diastereoisomers, confirming the robustness of our approach. Scale-up was straightforward and compound 5 could be obtained in 72% yield on 1.2 mmol scale without erosion of stereoselectivity. Heterocycles and functional groups containing coordinating N or pan class="Disease">S atoms and chlorides were not tolerated in the hydrogenation step (for details, see SI, Section D5). The nitrile and the carbonyl group within 25 and 27 were reduced to the corresponding amine 48 and alcohol 50.[44] Interestingly, 5, 36, 45, 46, 47 and 54 are precursors of bioactive compounds with antidepressive activity,[27,28] while the amino alcohols derived from 36 and 47 are intermediates for the synthesis of patented antiviral drugs candidates.[30] Remarkably, our method provides a high level of asymmetric induction even in the presence of sterically and electronically similar aryl substituents on the olefin, thus overcoming a common obstacle in the development of catalytic asymmetric reactions. Finally, to confirm the traceless nature of our strategy, we performed a mild acidic hydrolysis of the hemiaminal in 5. The enantioenriched amino alcohol 56 was obtained in 76% yield without loss in optical purity.

Scheme 4

Scope of the Stereoselective Hydrogenation

Reactions performed on a 0.2 mmol scale using Pd(OH)2/C (∼20 wt %). Isolated yields and HPLC enantiomeric excess are given. Product 56 was obtained after treating 5 with TsOH·H2O (7 equiv) in a 2:1 THF/H2O mixture at room temperature for 16 h, the trifluoroacetate salt was obtained after purification by reversed phase preparative HPLC.

Pd/C (∼5 wt %) was used instead of Pd(OH)2/C.

Reactions performed on a 0.2 mmol scale using n class="Chemical">pan class="Chemical">Pd(OH)2/C (∼20 wt %). Isolated yields and Hclass="Chemical">n>an class="Chemical">PLC enantiomeric excess are given. Product 56 was obtained after treating 5 with TsOH·H2O (7 equiv) in a 2:1 THF/H2O mixture at room temperature for 16 h, the trifluoroacetate salt was obtained after purification by reversed phase preparative HPLC.

pan class="Chemical">Pdn>/C (∼5 wt %) was used instead of pan class="Chemical">Pd(OH)2/C.

We then demonstrated that this strategy provides a simn class="Chemical">ple stereodivergent access to the four possible stereoisomers of chiral diaryl aminoalcohols by a judicious selection of the substrates and the ligands (Scheme ). Starting from the benzyl propargyl amine 57, a sequence of (i) Sonogashira coupling, (ii) enantioselective carboetherification and (iii) diastereoselective hydrogenation leads to all the stereoisomers of the desired products 5 and 36. Permuting the iodoarenes in the cross-coupling and in the carboetherification steps allows the tuning of the E,Z geometry of the double bond. This selective process, combined with the choice of the enantiomer of the ligand, and the diasteroselectivity of the hydrogenation provide a selective access to the four stereoisomers of the diaryl amino alcohol precursors.

Scheme 5

Diastereo- and Enantiodivergent Access to Chiral Aminoalcohol Precursors

See SI for detailed reaction conditions.

In summary, we have developed an innovative strategy to control the stereoselectivity of asymmetric transformations.[45] Our an class="Chemical">pproach first capitalizes on the tools of asymmetric catalysis to forge a chiral oxazolidine from broadly available propargylic amines. This stereogenic element is then used to control the selectivity of the asymmetric hydrogenation of the tetrasubstituted double bond, giving access to valuable chiral amino alcohol precursors. The key for success was the first use of a “truncated” monophosphine Trost-type ligand to induce high enantioselectivity in an unprecedented DYKAT process. Combined with a Sonogashira cross-coupling, our approach gives a stereodivergent access to the four stereoisomers of protected diaryl amino alcohols in high yield and enantioselectivity. New opportunities for the design and development of asymmetric functionalizations of olefins can be expected based on the combination of the enantioselective introduction of a transient chiral auxiliary followed by a diastereoselective transformation. Such processes are currently under investigation in our laboratory.