Advances in Organic Synthesis: Volume 9

By Bentham Science Publishers

Advances in Organic Synthesis is a book series devoted to the latest advances in synthetic approaches towards challenging structures. The series presents comprehensive reviews written by eminent authorities on different synthetic approaches to selected ta

• Language: English
• Publisher: Bentham Science Publishers
• Release date: Aug 1, 2018
• ISBN: 9781681086958

Book preview

Asymmetric Hydrogenation of Tetrasubstituted Olefins

Kristen Ryan¹, Robert B. Kargbo², *

¹ Drug Discovery Services, Medicinal Chemistry, AMRI, 26 Corporate Circle, Albany, NY 12201-5098, USA

² Usona Institute, 277 Granada Drive, San Luis Obispo, CA 93401, USA

Abstract

The asymmetric hydrogenation of tetrasubstituted olefins has been one of the most formidable challenges in the reduction of olefins. While the reduction of di- and tri-substituted olefins has grown at a rapid pace, the reduction of tetrasubstituted olefins has grown at a much slower rate. However, there have been a number of pivotal breakthroughs, from the reports of Zhou and Buchwald to the advent of the discovery of the chiral phosphoramidite ligands. In addition, high throughput experimentation has been very beneficial in the discovery of rapid and useful reaction conditions to affect remarkable selectivities. This led to the discovery of rhodium-Josiphos based catalysts, which have been frequently used in the asymmetric reduction of tetrasubstituted olefins.

Keywords: Asymmetry, Breakthroughs, Enantiotopic, Environmental, Functionalized, High throughput screening, Hydrogenation, Hydrogenolysis, Indenes, Iridium, Josiphos ligand, Ligands, Olefins, Palladium, Phosphoramidite ligands, Rhodium, Ruthenium, Selectivity, Tetrasubstituted, Unfunctionalized, Zirconium.

---

* Corresponding authors Robert B. Kargbo: Usona Institute, 277 Granada Drive, San Luis Obispo, CA 93401, USA; Tel: [phone]; E-mail: [e-mail]

INTRODUCTION

The growing regulatory challenges in drug discovery programs have necessitated the need to utilize the once inaccessible 1,2-contiguous stereocenter via direct asymmetric hydrogenation [1]. From the landmark report of the asymmetric hydrogenation by Noyori and Knowles, which culminated in the award of the Nobel Prize in Chemistry in 2001, there have been numerous reports of asymmetric hydrogenation of di- and tri-substituted olefins [2]. The reduction of tetrasubstituted olefins, on the other hand, has grown at a much slower pace due

to the difficulties associated in hydrogenating such substrates. Nonetheless, significant progress has been made in the development of asymmetric hydrogenation of tetrasubstituted olefins (AHTOs).

One interesting aspect of AHTO is the possibility of influencing the reactivity and selectivity by the choice of ligands used in the reactions. We have chronicled the discoveries of these ligands and given examples of AHTO reactions in which they were used successfully. In addition, we have categorized the review by metal types and the various ligands used for functionalized and unfunctionalized substrates. We have also highlighted the difficulties in identifying a metal-ligand type that is universal to a wide range of substrates. Nonetheless, high throughput experimentations (HTEs) have been extremely useful for the quick identification of reaction conditions that are suitable for different categories of substrates.

ASYMMETRIC HYDROGENATION OF UNFUNCTIONALIZED OLEFINS

Chiral Zirconocene Catalysts

Unfunctionalized tetrasubstituted olefins are particularly challenging for asymmetric hydrogenation due to steric hindrance and the lack of neighboring coordinating groups, which otherwise aid in the olefin’s ability to bind to transition metal complexes [3]. In 1999, Buchwald and co-workers [4] were initially successful in selectively reducing unfunctionalized trisubstituted olefins using the catalyst (S,S)-(EBTHI)TiH (EBTHI = ethylenebistetrahydroindenyl) (Scheme 1). However, the reduction of trisubstituted substrates bearing bulkier substituents were generally quite slow (required multiple days), even at high pressures, and it was assumed that tetrasubstituted olefins would fare even worse. Therefore, cationic titanocene and zirconocene catalysts were investigated as alternatives based on the idea that their high electrophilicity would more likely allow binding of tetrasubstituted olefins. When subjected to an atmosphere of hydrogen, the catalysts are converted into cationic metallocene hydrides. The group published a series of asymmetric reductions of tetrasubstituted indenes to the corresponding indanes using a chiral zirconocene catalyst generated from (EBTHI)ZrMe2 and the non-coordinating acid [PhMe2NH]+[B(C6F5)4]- (Scheme 1). Interestingly, this same catalyst system was previously applied to the reduction of 1,1-disubstituted olefins but the ee’s were <50%. Presumably, the additional steric hindrance of the tetrasubstituted substrates allowed for more effective discrimination of the catalyst between the enantiotopic faces of those olefins.

In this report, the hydrogenations were carried out at room temperature in aromatic hydrocarbon solvents using either (R,R)-3 or (S,S-3) (Scheme 1) under pressures that varied from 80 psi to 1000-2000 psi. Enantiomeric excesses for

Scheme 1)

Chiral zirconocene catalyzed AHTO.

most of the examples were generally high but, in some cases, the pressure had a profound impact on the rates and enantioselectivities of the reactions. It was initially assumed that the reaction proceeded via a pathway similar to that proposed for the corresponding trisubstituted olefins where the double bond first inserts into the zirconium hydride. The resultant intermediate then undergoes hydrogenolysis, providing the product and regenerating zirconium hydride, resulting in formation of the cis products when applied to cyclic substrates. As shown in Scheme 1, increasing the steric bulk of the benzylic methyl substituent to an ethyl (Scheme 1, 4 vs 8) resulted in an expected decrease in enantio- selectivity due to the more unfavorable steric interaction in the preferred transition state. However, at the higher-pressure range, the bulkier benzylic butyl group (7) recovered the high level of enantioselectivity. This suggested the mechanism was more complex than originally thought and highly substrate dependent.

The poor predictability of the reaction outcomes, high catalyst loadings, and high sensitivity of the catalyst could have prevented the widespread application of the protocol over the years. Consequently, reported development in this area stalled following Buchwald’s publication.

Chiral Iridium Catalysts

In 2007, Pfaltz and co-workers used iridium complexes with chiral N,P ligands to address the issues that surrounded the use of the highly reactive zirconocene catalysts (Scheme 2) [5].

Scheme 2)

Iridium-oxazoline asymmetric hydrogenation.

Iridium catalysts based on chiral phosphanyl oxazoline ligands (Scheme 2, 12a–c) showed high activity in the asymmetric hydrogenation of unfunctionalized tetrasubstituted alkenes at a catalyst loading of only 2 mol%. A library of phosphanyl oxazoline and phosphinite oxazoline ligands were prepared and, among the phosphanylmethyloxazolines, ligand 12c (Scheme 2) produced ee’s exceeding 90% at lower pressures (13–18). While an increase in the size of the substituent at the benzylic position (14 and 15) did not affect the high conversion and enantioselectivity, changing R² from a methyl to a phenyl resulted in a substantial decrease in conversion (16–18). In addition to ligand 12c, a number of the other ligands examined such as 12a and 12b provided high enantioselectivities for a variety of different substrates including phenyl-substituted dihydronaphthalene 19, para-fluorophenyl-substituted alkene 20, and tricyclic systems such as 21.

Several years later, in 2013, Busacca and co-workers reported the use of cationic iridium (COD)BArF complexes using BIPI ligands for the asymmetric hydrogenation of tetrasubstituted olefins (Scheme 3) [6]. The BIPI ligands were invented in the late 1900s and their modular nature allowed for systematic tuning at each region of the ligand. A series of structurally diverse BIPI ligands were converted to the corresponding iridium complexes and all were stable to silica gel chromatography for purification. One common feature amongst the entire set was an N-acyl functionality which was determined to be required for successful application to AHTO. Otherwise, four substituent subclasses were chosen with either aryl or alkyl groups on the phosphorus and imidazoline substituents. All of the metal-ligand complexes were subsequently used for the asymmetric reduction of dimethylindene 22 under 1 bar of hydrogen in dichloromethane (Scheme 3).

Scheme 3)

Structurally diverse BIPI ligands.

Complexes that incorporated ligands with dialkylphosphines and diarylimidazolines, such as BIPI 1, showed low conversion, low selectivity, or both. Ligands with aryl groups on both the phosphorus and carbons, such as BIPI 2, all gave full conversion but selectivities were moderate (BIPI 2 gave the highest ee of the set with 78%). The combination of alkyl groups on the carbons and aryl groups on the phosphorus, such as BIPI 3, resulted in full conversion but lower enantioselectivity (68%). Finally, the fourth class of ligands with alkyl substituents on the phosphorus and carbons, such as BIPI 4, resulted in full conversion and higher ee’s as long as small alkyl groups were avoided. When BIPI 4 was modified to a naphthyl core (BIPI 5), the ee increased slightly from 88% to 90%. Changing the substituent at the naphthyl peri position (C-8) from a hydrogen to a fluorine (BIPI 6), resulted in the highest enantioselectivity for this particular reduction (96%). The presence of the naphthyl peri proton (or fluorine) may have caused conformational restriction of the phosphine substituents leading to the increase in selectivity.

Though iridium complexes have generally been used for the reduction of unfunctionalized tetrasubstituted olefins, Zhou reported the use of chiral spiro iridium catalysts for the asymmetric hydrogenation of functionalized tetrasubstituted α,β-unsaturated carboxylic acids in 2013 [7]. In designing the catalyst, a benzyl group on the oxazoline was found to be optimal and the substituents on the P-phenyl rings greatly affected the enantioselectivity of the reaction. Bulky 3,5-di-tert-butylphenyl groups on the phosphorus led to high conversion and ee values for a number of substrates (Scheme 4). Basic additives to the reaction were also examined and, although cesium carbonate was found to promote the reaction, triethylamine gave higher enantioselectivities. In terms of the substrates, a series of α-aryl-β,β-dimethyl acrylic acids were examined in the hydrogenation using the optimized catalyst. Substituents on the phenyl ring of the substrates had only weak influence on the outcome of the reactions and all of the intended products were formed in high yields and enantioselectivities (27, Scheme 4).

Scheme 4)

Iridium-SIPHOX mediated asymmetric reduction.

In extending the substrate scope, α-methyl-β,β-diethyl acrylic acid and 2-cyclopentylidene-2-phenylacetic acid were also successfully reduced using the optimized conditions to give products 28 and 29 (Scheme 4) in high ee’s. However, the substrate leading to carboxylic acid 29 was less reactive and required 4 mol % catalyst to drive the reaction to completion rather than the standard 1 mol%. The conditions failed to reduce cyclic substrate 30, however, bearing a tetrasubstituted endocyclic olefin.

Zhou recently reported the application of another spiro iridium catalyst complex, Ir-SpiroPAP, for the asymmetric reduction of tetrasubstituted cyclic enones to form chiral cycloalkanols with three contiguous chiral centers (Scheme 5) [8].

Scheme 5)

Application of Ir-SpiroPAP catalysts.

The process involved a single-step sequential reduction of the C=C bond and the C=O bond and led to the successful formation of a series of cyclopentanols and cyclohexanols with high enantioselectivity and diastereoselectivity. A series of different β-alkyl groups were examined and found to have little influence on the outcome of the reactions. Interestingly, the cyclopentanols were selectively produced with cis,trans stereochemistry while the cyclohexanol products were predominantly cis,cis with >99% diastereoselectivity in all cases. This method was then applied to the first catalytic asymmetric synthesis of all the stereoisomers of the antiulcer drug rosaprostol.

Chiral Ruthenium Catalysts

Although the asymmetric hydrogenation of functionalized tetrasubstituted olefins is more developed, it poses its own set of challenges such as designing appropriate directing and protecting groups. The directing group could potentially hinder the reaction sterically and/or electronically or a strong binding affinity of the directing group could possibly deactivate the catalyst. In 2003, Zhang reported the first asymmetric hydrogenation of cyclic tetrasubstituted β-(acylamino)acrylates through the use of a ruthenium catalyst (Scheme 6) [9]. These substrates would be especially useful for the synthesis of chiral β-peptides.

Scheme 6)

Application of Ru-Josiphos catalysts.

Initially, the group applied conditions previously used to successfully reduce trisubstituted β-(acylamino)acrylates, employing Rh-(S,S,R,R)-TangPhos. However, when applied to the reduction of the corresponding tetrasubstituted substrates, no reactivity was observed. The team shifted their focus on preparing in situ Ru catalysts from a monomeric Ru precursor, a chiral phosphorus ligand, and HBF4. A mixture of Ru(COD)-(methylallyl)2 and a chiral biphosphorus ligand was protonated with two equivalents of HBF4•Me2O in dichloromethane. The solvent was then evaporated and the residue was used directly in the hydrogenation. Several different chiral ligands were explored and a range of enantioselectivities were observed. DIOP, Me-DuPhos, and TangPhos gave only moderate ee’s (34%-69%), while MeO-BIPHEP and BINAP both gave 99% ee. The effect of the dihedral angle of the chiral biaryl ligand on the enantioselectivity of the reaction was then investigated by examining a series of TunaPhos ligands with varying dihedral angles. The C1-TunaPhos and C6-TunaPhos provided slightly lower ee’s (98% and 97%) but all others provided at least 99% ee. It was also determined that alcoholic solvents such as methanol and ethanol were optimal for the reaction.

A series of tetrasubstituted β-(acylamino)acrylates were then reduced using C3-TunaPhos as the ligand and the results are shown in Scheme 6. All of the five-membered ring substrates resulted in high ee’s including a substrate containing a Boc-NH (49). A heterocyclic example providing product 50 was highly enantioselective as well. As the ring size was expanded, there was a corresponding decrease in the enantioselectivity of the reaction (51–53). An acyclic example (reaction leading to 54) suffered a significant decrease in ee as well.

In yet another example of successful AHTO employing a ruthenium catalyst, Christensen and co-workers described the asymmetric reduction of α-methyl-β-cyclopropyldihydrocinnamates (Scheme 7) [10]. Among the challenges associated with this work was the presence of the hydrogenation-labile cyclopropyl group. The team used high throughput experimentation (HTE) to quickly evaluate the reaction space and identify promising reaction conditions. Two transition metal precursors previously used for AHTO were chosen for the investigation, (NBD)2RhBF4 and (Me-allyl)2RuCOD/HBF4. A series of chiral phosphine ligands were screened along with two different reaction solvents, 2-Me-THF and methanol. Combinations of metal complexes, ligands, and solvents resulted in a total of 96 experiment permutations, which were all run at 500 psi hydrogen and at 60 °C overnight. The rhodium catalysts generally gave low conversion due to substrate decomposition, likely a result of cyclopropane ring opening. Most of the ruthenium catalysts also gave low conversion to the desired product, except for the combination of (Me-allyl)2Ru(COD)/HBF4 and Josiphos, which resulted in 83% conversion and 94% ee. Additional commercial chiral phosphine ligands were then screened in an effort to optimize the reaction and five (mainly in the Josiphos family of ligands) improved the conversion and resulted in approximately the same or better enantioselectivity (Scheme 7). While rhodium catalysts led to low conversion due to possible decomposition via cyclopropyl ring opening, ruthenium on the other hand, afforded the product in high conversion and selectivity. The team synthesized α-methyl-β-cyclopropylcinnamate analogs on 200 g to kilogram scales with excellent enantioselectivities.

Scheme 7)

Application of Ru-Josiphos catalysts.

Chiral Palladium Catalysts

In the search for new catalytic systems for AHTO, Zhou’s team applied palladium complexes that had previously been used to reduce trisubstituted enesulfonamides to the reduction of tetrasubstituted cyclic β-(arylsulfonamido)acrylates (Scheme 8) [11]. A previous mechanistic study suggested that the hydrogenation of enesulfonamides occurred via Brønsted acid catalyzed tautomerization to N-sulfonylimine intermediates, followed by a dynamic kinetic resolution during the hydrogenation. Given that the cyclic substrates of interest could readily isomerize to the corresponding sulfonylimine intermediates in the presence of a Brønsted acid, the same catalytic system was explored. Based on previous success, Pd(OCOCF3)2/(R,Sp)-Josiphos was chosen as a starting point. However, low conversion and ee values were observed. To boost the conversion and enantioselectivity, a number of Brønsted acid additives were tested and it was determined that the chirality of the additive had no effect on the reaction outcome. Addition of trifluoroacetic acid as the additive resulted in full conversion and a significant improvement in the ee values. Many commercially available biphosphine ligands were examined. Ferrocenyl and axial chiral ligands gave full conversion but only moderate ee’s. (1R,1'R,2S,2'S)-DuanPhos (56, Scheme 8), developed by Zhang, proved to result in the highest overall yields and enantioselectivities.

A series of cyclic β-(arylsulfonamido)acrylates were prepared and reduced using the optimized conditions (Scheme 8). The size of the ester functionality did not impact the outcome of the reaction and both phenyl and 4-methyl-phenyl substituted sulfonamides were well tolerated. One of the reduced intermediates was ultimately used in the preparation of a potential Alzheimer’s drug developed by GSK.

Scheme 8)

Application of Pd-DuanPhos catalysts.

Chiral Rhodium Catalysts

The AHTO for functionalized olefins is more developed compared to the unfunctionalized olefins. This is due in part to the efficient and convenient access to chiral compounds such as α-amino acids, β,β-disubstituted α-amino acids, chiral carboxylic acids, 1,2-chiral cyclic α-amino acids and so forth [12]. These compounds are structural motifs of significant interest to the pharmaceutical, perfumery, biochemical and agrochemical industries. The asymmetric synthesis of these compounds represent a considerable synthetic challenge due to the difficulties of controlling vicinal stereogenic centers, appropriate directing and protecting groups, and developing chiral catalysts that synergistically balance the steric and electronic factors of the substrates synchronously [13]. In addition, a severe steric environment of the substrates renders them inaccessible to the ligand-metal complex, no matter how reactive the chiral catalyst might be. Furthermore, strong binding affinity of a chelating/directing group would deactivate the catalyst and render the process noncatalytic [14].

In spite of the synthetic challenges posed by the asymmetric hydrogenation of tetrasubstituted olefin substrates, Burk and co-workers in 1995 achieved arguably the first truly successful asymmetric hydrogenation of β,β-disubstituted α-enamides [15]. As shown in Scheme 9, the catalyst tolerated either E or Z-substituents. Optimal conditions were cationic Me-BPE-Rh catalyst, MeOH and 60 psi H2 pressure. The flexibility of the Me-BPE ligand allowed greater range of β-substituents on α-enamides demonstrating unique advantages of ligand design and providing higher enantioselectivities compared to the Me-DuPHOS-Rh catalyst.

Scheme 9)

Application of rhodium complexes of DuPHOS and BPE

It is worth noting the sensitivity of the catalyst in