cis-2,5-Diaminobicyclo[2.2.2]octane, a New Chiral Scaffold for

Article pubs.acs.org/accounts

cis-2,5-Diaminobicyclo[2.2.2]octane, a New Chiral Scaffold for Asymmetric Catalysis Subrata Shaw and James. D. White* Department of Chemistry, Oregon State University, Corvallis, Oregon 97331, United States CONSPECTUS: Catalysis of widely used chemical transformations in which the goal is to obtain the product as a pure enantiomer has become a major preoccupation of synthetic organic chemistry over the past three decades. A large number of chiral entities has been deployed to this end, many with considerable success, but one of the simplest and most effective catalytic systems to have emerged from this effort is that based on a chiral diamine, specifically trans-1,2-diaminocyclohexane. While there have been attempts to improve upon this scaffold in asymmetric synthesis, few have gained the recognition needed to take their place alongside this classic diamine. The challenge is to design a scaffold that retains the assets of trans-1,2-diaminocyclohexane while enhancing its intrinsic chirality and maximizing the scope of its applications. It occurred to us that cis-2,5-diaminobicyclo[2.2.2]octane could be such a scaffold. Synthesis of this diamine in enantiopure form was completed from benzoic acid, and the (1R,2R,4R,5R) enantiomer was used in all subsequent experiments in this laboratory. Condensation of the diamine with various salicyl aldehydes generated imine derivatives which proved to be excellent “salen” ligands for encapsulation of transition and other metals. In total, 12 salen−metal complexes were prepared from this ligand, many of which were crystalline and three of which, along with the ligand itself, yielded to X-ray crystallography. An advantage of this ligand is that it can be tuned sterically or electronically to confer specific catalytic properties on the salen−metal complex, and this feature was used in several applications of our salen−metal complexes in asymmetric synthesis. Thus, replacement of one of the tert-butyl groups in each benzenoid ring of the salen ligand by a methoxy substituent enhanced the catalytic efficiency of a cobalt(II)−salen complex used in asymmetric cyclopropanation of 1,1-disubstituted alkenes; the catalyst was employed in an improved synthesis of the cyclopropanecontaining drug candidate Synosutine. Reduction of the pair of imine functions of the ligand to secondary amines permitted formation of a copper(I)−salen complex that catalyzed asymmetric Henry (“nitroaldol”) condensation with excellent efficiency; this catalyst was applied in an economical synthesis of three drugs of the “beta-blocker” family including (S)-Propanolol. Chromium(II) and chromium(III) complexes were prepared from our bicyclooctane−salen ligand bearing a pair of tert-butyl groups in each benzenoid ring. These complexes were found to catalyze, respectively, enantioselective formation of homoallylic alcohols from Nozaki−Hiyama−Kishi allylation of aromatic aldehydes and dihydropyranones from hetero-Diels−Alder cycloaddition. Plausible reaction models emerging from knowledge of the absolute configuration of products from each of these reactions place the metal-coordinated substrate in a quadrant beneath the bicyclooctane scaffold so that one face of the substrate is blocked by an aryl ring of the salen ligand while the opposite face is left open to attack. The consistent and predictable stereochemical outcome from reactions catalyzed by salen−metal complexes derived from our diaminobicyclo[2.2.2]octane scaffold adds a valuable new dimension to asymmetric synthesis.

1. INTRODUCTION Enantioselective reactions in which the chirality of an asymmetric catalyst is conveyed efficiently to a reaction product are among the most valuable transformations in modern organic synthesis. Formation of a single enantiomeric product from an achiral substrate in an atom-economical process that requires only minimal intervention by a chiral catalyst has been a long-sought goal, and one of many paradigms to have emerged from efforts in the area is that diamines provide “privileged” chiral scaffolds for attaining that goal.1 The pioneering research which began in the early 1990s with the use of trans-1,2-diaminocyclohexane (1) as a chiral scaffold2 set in train a new era of asymmetric synthesis framed around Schiff bases of the diamine that are able to form complexes with transition metals. Ligands were developed, particularly bis(salicylidene) derivatives of 1 (“salen ligands”), whose metal © XXXX American Chemical Society

complexes showed good levels of stereoinduction in reactions such as epoxidation of alkenes3 and hetero-Diels−Alder cycloaddition.4 Refinements to the electronic and steric properties of these ligands greatly expanded the utility of salen catalysts derived from 1 but little variation in the diamine scaffold appeared until 2003 when Berkessel put forward cis2,5-diaminobicyclo[2.2.1]heptane (2, “DIANANE”) as a new template for asymmetric synthesis.5 Although the chromium(III) complex prepared from a salen derivative of 2 performed well as a catalyst in asymmetric hetero-Diels−Alder cycloaddition,6 its application to other processes such as the asymmetric Nozaki−Hiyama−Kishi reaction was less successful. This limitation and the tedious preparation of 2 in Received: June 9, 2016

A

DOI: 10.1021/acs.accounts.6b00286 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Scheme 1. Synthesis of Racemic cis-Bicyclo[2.2.2]octane2,5-dicarboxylic Acid, (±)-11, from Benzoic Acid

enantiomerically pure form led us to consider a new diamine, cis-2,5-diaminobicyclo[2.2.2]octane (3), as a chiral scaffold that would be easier to synthesize in quantity and which would have broad utility in asymmetric synthesis (Figure 1).

Figure 1. Chiral diamines trans-1,2-diaminocyclohexane (1), cis-2,5diaminobicyclo[2.2.1]heptane (DIANANE, 2), and cis-2,5diaminobicyclo[2.2.2]octane (3).

Racemic 3, a homologue of 2, was already known from work in our laboratory,7 and a computational comparison using B3LYP/6-31G(d) of structural features associated with 1, 2, and 3 suggested that a salen derivative of 3 should be a superior ligand for complexation with transition metals. A property of 3 that was especially appealing was the larger N−N separation (4.27 Å) compared to those in 2 (4.00 Å) and 1 (2.91 Å). This implied that salen derivatives of 3 would enclose a larger volume of chiral space in a well-defined asymmetric environment while preserving dimensions that would allow encapsulation of transition metal ions of variable ionic radius within a pocket below the bicyclic scaffold. A second structural feature of 3 that emerged from computational analysis was the large dihedral angle of 22° between CaHa and CbHb bonds (Figure 2). This appears to

Scheme 2. Resolution of Racemic cis-Bicyclo[2.2.2]octane2,5-dicarboxylic Acid (±)-11 via Diastereomeric Brucine Salts

Figure 2. Calculated dihedral angle between CaHa and CbHb bonds in 3.

reflect “twist” around the C2 axis of the bicyclic framework of 3 that is absent in 2, where the analogous dihedral angle (14°) is due solely to an eclipsing H−H interaction.

2. SYNTHESIS OF DIAMINE 3 AND ITS SALEN−METAL COMPLEXES The starting point for our synthesis of 3 was benzoic acid (4) which was reduced under Birch conditions to give initially 1,4cyclohexadiene-3-carboxylic acid (5, Scheme 1).8 Isomerization of 5 to conjugated diene 6 was followed by Diels−Alder addition of methyl acrylate under solvent free conditions. This reaction yielded a mixture of regioisomeric endo half-esters 7 and 8, which was hydrogenated over Adams’s catalyst to furnish 9 and 10 in a 3:1 ratio. Chromatographic separation of 9 from the mixture followed by ester saponification gave racemic cisbicyclo[2.2.2]octane-2,5-dicarboxylic acid, (±)-11. Resolution of racemic 11 was accomplished with (−)-brucine, and after three recrystallization cycles from water, diastereomeric salts of constant specific rotation were obtained (Scheme 2).9 The less soluble salt (−)-12 was beautifully crystalline, and its structure was determined by X-ray crystallography. The dicarboxylate portion of this salt was found to possess (1R,2R,4R,5R) absolute configuration. Concentration of the mother liquor produced diastereomeric

brucine salt (−)-13, and separate acidification of each salt gave dicarboxlic acids (−)-14 and (+)-15, respectively. Conversion of (−)-14 to diamine 3 was accomplished by first taking the dicarboxylic acid to diacyl azide (−)-16 via diacyl chloride (−)-17 (Scheme 3). Double Curtius rearrangement of 16 to bis-isocyanate (−)-18, followed by acidic hydrolysis and in situ decarboxylation of the intermediate, bis-carbamic acid, afforded (−)-3 in an overall yield of 67% from (−)-14.10 All subsequent experiments were performed with this enantiomer of the diamine. Following Jacobsen’s protocol with diamine 1,3 (−)-3 was condensed with 3,5-di-tert-butylsalicylaldehyde (19) to furnish crystalline bis salicylidene derivative (+)-20 (Scheme 4). The crystal structure of tetradentate ligand 20 confirmed that the pairs of nitrogen and oxygen atoms are well positioned for encapsulation of metal ions whose ionic radii match the size of the chiral pocket in the lower right quadrant under the bicyclic B


Article


graphic analysis. A representative crystal structure, that of nickel(II)−salen complex (+)-29, is shown in Figure 3.

Scheme 3. Synthesis of (1R,2R,4R,5R)-2,5Diaminobicyclo[2.2.2]octane, (−)-3, from Dicarboxylic Acid (−)-14

Scheme 4. Synthesis of Salen Ligand (+)-20 and Metal Complexes from Diamine (−)-3

Figure 3. X-ray crystal structure of nickel(II)−salen complex (+)-29.

Crystallographic details can be found elsewhere11 but it is clear from the data for 29−31 in Table 2 that the cavity created Table 2. Selected Crystallographic Dimensions of Metal− Salen Complexes (+)-29−31

scaffold. In practice, a large number of metal salts reacted with (+)-20 to form either crystalline or amorphous salen complexes (Table 1). Twelve of the complexes (21−32) were fully characterized, and three (29−31) yielded to X-ray crystallo-

reagents

solvent

temp

CH2Cl2

ambient

CH2Cl2

ambient

VO(acac)2

MeOH

reflux

CrCl2, Et3N, O2 Mn(OAc)2· 4H2O FeCl3, NaH Fe(acac)3 CoBr2, NaH Ni(OAc)2· 4H2O Cu(OAc)2· H2O Pd(OAc)2 AlCl3, NaH

THF

ambient

(+)-21, M = TiCl(OiPr) (+)-22, M = V(O) (Cl) (+)-23, M = V(O) (acac) (+)-24, M = CrCla

product

MeOH

reflux

(+)-25, M = MnCla,b brown

THF MeOH THF MeOH

reflux reflux reflux reflux

(+)-26, M = FeCl (+)-27, M = Fe(acac) (+)-28, M = Co (+)-29, M = Ni

MeOH

reflux

(+)-30, M = Cu

MeOH THF

reflux reflux

(+)-31, M = Pd (+)-32, M = AlCl

N1−M− N2 angle (deg)

M−N bond distance (Å)

M−O bond distance (Å)

N−N separation (Å)

(+)-29 (+)-30 (+)-31

0.692 0.737 0.865

96.78 98.38 97.70

1.893 1.963 2.026

1.862 1.903 2.002

2.830 2.972 3.051

by the four heteroatoms in ligand 20 is able to accommodate metals of varying ionic radius. In each structure, the metal center is slightly displaced from the plane formed by the four chelating heteroatoms resulting in a distorted square pyramidal array. These crystal structures reveal an open quadrant at the right front region (and the corresponding left rear space) under the bicyclic scaffold where a metal-catalyzed reaction can take place. This chiral architecture is used in postulating reaction pathways to explain the configuration of products resulting from our metal−salen catalysts. As expected, the infrared spectra of salen−metal complexes 21−32 all showed a shift of the CN absortion band to lower frequency compared to the uncomplexed ligand 20, indicating partial delocalization of C N π electrons into vacant d-orbitals of the metal. With these enantiopure salen−metal complexes available in quantity, the first task was evaluation of their catalytic activity in asymmetric synthesis. Since Jacobsen4 and Berkessel5 had prepared chromium(III)−salen systems analogous to 24 from their diamines 1 and 2, respectively, and had published data on those complexes as catalysts in hetero-Diels−Alder cycloaddition, calibration of our scaffold 3 could be made by comparing its catalytic efficiency in this reaction with those two diamines.

Table 1. Salen−Metal Complexes Formed by (+)-20 with Metal Salts Ti(OiPr)3, Et3N VOCl3, Et3N

metal complex

ionic radius of metal(II) ion (Å)

color yellow green green brown

brown violet orange redorange black red pale yellow

The flask was opened to air after the initial reaction. bThe reaction mixture was washed with brine during workup. a

C


Article


3. ASYMMETRIC HETERO-DIELS−ALDER CYCLOADDITION An attempt at the synthesis of chromium(III)−salen complex (+)-24 by direct reaction of (+)-20 with chromium(III) chloride or bromide was unsuccessful. Instead, an indirect method was used that first reduced chromium(III) to chromium(II) with superstoichiometric manganese. Ligand 20 formed an unstable complex with chromium(II) chloride, and its subsequent exposure to air resulted in rapid oxidation to stable complex (+)-24. Chromium(III) complex (+)-24 received its first test as a catalyst in the hetero-Diels−Alder (HDA) addition of benzaldehyde (33) to Danishefsky’s diene (34) (Scheme 5). With 5 mol % of (+)-24 and 4 Å molecular

lower. The absolute configuration of dihydropyranones in Table 3 was established as (S) in all cases, except for entries 6 and 9, by comparison of the measured specific rotation with the literature value. A few aldehydes, notably pyridine-3-carboxaldehyde, 4-methoxybenzaldehyde, and 4-(dimethylamino)benzaldehyde, failed to react with 34 in the presence of (+)-24. A reaction pathway that rationalizes the role of chromium catalyst (+)-24 in these asymmetric HDA cycloadditions is shown in Figure 4. The metal-coordinated aldehyde is

Scheme 5. HDA Reaction of Benzaldehyde with Danishefsky’s Diene, Catalyzed by (+)-24

sieves, the reaction cleanly gave cycloadduct 35, which, after treatment with trifluoroacetic acid, furnished (S)-dihydropyranone 36 in quantitative yield and 97% enantiomeric excess. This result compares favorably with the HDA reaction of 33 with 34, catalyzed by chromium(III)−salen complexes derived from diamines 1 and 2.4,6 Optimized conditions were developed for this prototype HDA reaction and were applied to cycloaddition of a range of aromatic aldehydes with diene 34. Dihydropyranones were obtained in excellent yield and in high enantiomeric excess after hydrolysis of the initial tetrahydropyran cycloadduct (Table 3). An exception was cyclohexanecarboxaldehyde (entry 10), where the enantiomeric excess of the product was significantly

Figure 4. Proposed model for the origin of stereoselectivity in the HDA reaction of aromatic aldehydes with Danishefsky’s diene, catalyzed by (+)-24.

positioned in the lower right-hand quadrant of the salen ligand below the bicyclic scaffold and is oriented to avoid a steric clash of the aldehyde hydrogen with a proximal tert-butyl substituent of the ligand. In this arrangement, the re face of the carbonyl is blocked by an aryl ring of the salen residue, leaving the si face exposed to approach by the diene. This approach trajectory leads to (S) configuration of cycloadduct 35 and therefore to (S)-dihydropyranone 36. A π-stacking interaction between the aryl aldehyde and a benzenoid ring of the salen ligand may contribute to stabilization of the complex since lower stereoselectivity is observed with the one aliphatic aldehyde substrate in Table 2 (entry 10).12

Table 3. Asymmetric HDA Cycloaddition of Danishefsky’s Diene with Aldehydes, Catalyzed by Chromium(III)−Salen Complex (+)-24a

4. ASYMMETRIC NOZAKI−HIYAMA−KISHI REACTION Development of an efficient catalytic, enantioselective version of the reaction of an allylic halide with an aldehyde (the Nozaki−Hiyama−Kishi (NHK) reaction) has presented a challenge, and success has been mixed.13 Our hope that the chromium(II) complex of ligand (+)-20 prepared en route to (+)-24 could provide a solution to this problem was now open to test. There were already examples in the literature for asymmetric NHK addition of allyl bromide to benzaldehyde using chromium(II)−salen complexes derived from 1 and 2,5,13 and those results again provided a comparative reaction for evaluating our scaffold 3. An initial experiment in which the chromium(II)−salen complex was prepared in situ by reduction of chromium(III) complex (+)-24 with manganese gave a 1:1 mixture of homoallylic alcohol (S)-37 and the pinacol product from reduction of benzaldehyde, but modification of the reaction conditions by including trimethylsilyl chloride in the medium led to a much improved ratio in favor of 37 (Scheme 6).8 After optimization of the reaction shown in Scheme 6, a series of aryl aldehydes was treated with allyl bromide to give homallylic silyl ethers. Acidic hydrolysis of the silyl ether

product entry

R

T (°C)

t (h)

yield (%)b

ee (%)c

1 2 3 4 5 6 7 8 9 10

2-CH3C6H4 3-CH3C6H4 3-CH3OC6H4 4-ClC6H4 4-O2NC6H4 3,5-(CH3O)2C6H3 1-naphthyl 2-furyl 3-furyl c-C6H11

−22 −22 −22 −22 −22 −30 −30 −30 −30 −22

24 24 36 24 24 48 40 48 42 36

98 99 99 98 99 97 96 93 96 91

92 92 96 94 94 94d 94 88 94d 68

a

Reactions were carried out on a 0.25 mmol scale in a 0.0625 M solution. bYield of isolated product. cDetermined by HPLC using a Chiralcel OD column. dAbsolute configuration of the product was not determined. D


Article


Figure 5. The allyl unit is first coordinated to the metal core through oxidative insertion of chromium(II) into the carbon−

Scheme 6. Asymmetric NHK Reaction of Allyl Bromide with Benzaldehyde, Catalyzed by the Chromium(II) Complex of (+)-20

afforded homoallylic alcohols of (S) configuration in excellent yield and in high enantiomeric excess (Table 4). Inclusion of 3 Å molecular sieves was found to be essential for best results from these reactions.

Figure 5. Proposed model for the asymmetric NHK reaction of an allylic halide with an aldehyde, catalyzed by (+)-24.

Table 4. Asymmetric NHK Reaction of Allyl Halides with Aromatic Aldehydes, Catalyzed by the Chromium(II)−Salen Complex from Reduction of (+)-24a

halogen bond, and the resultant chromium(III) species incorporates the aldehyde through coordination of the metal with the carbonyl oxygen. The chiral ligand directs intramolecular attack by the allyl residue toward the more accessible si face of the carbonyl to give an alcohol of (S) configuration. A catalytic cycle depicting the overall series of transformations, including roles played by manganese and trimethylsilyl chloride, is shown in Figure 6.

product entry d

1 2 3 4 5 6 7 8 9e

Ar

T (°C)

t (h)

yield (%)b

ee (%)c

C6H5 2-CH3C6H4 4-CH3C6H4 3-CH3OC6H4 4-ClC6H4 3,5-(CH3O)2C6H3 2,6-Cl2C6H3 1-naphthyl 3-furyl

10 10 10 10 20 10 20 20 10

15 6.5 6.5 6 5 8 7 4 3

92 92 97 96 96 95 92 97 93

89 92 95 89 96 93d 84 97 85

a

Reactions were carried out on 0.125 mmol scale at 0.625 M with 1.5 equiv of allyl bromide. bYield of isolated product. cDetermined by HPLC using a Daicel Chiralcel OD column. dAbsolute configuration of the product was not determined. eAllyl chloride was used instead of allyl bromide.

An experiment with vinyl iodide and 1-naphthaldehyde under conditions specified in Table 4, but with addition of a catalytic quantity of nickel(II) chloride, gave allylic alcohol (S)38 (Scheme 7). Although the yield and enantiomeric excess of

Figure 6. Catalytic cycle for asymmetric NHK allylation of aryl aldehydes with a chromium(II)−salen complex based on 3.

Scheme 7. Asymmetric NHK Reaction of 1-Naphthaldehyde with Vinyl Iodide, Catalyzed by the Chromium(II)−Salen Complex from (+)-24

5. ASYMMETRIC HENRY (“NITROALDOL”) CONDENSATION The condensation of a nitro alkane with an aldehyde or ketone, known as the Henry or “nitroaldol” reaction, is a C−C bond construction for which a catalyzed asymmetric version is especially valuable.14 Evans has shown that the Henry reaction can be catalyzed by copper(II) salts,15 and this precedent led us to examine copper(II)−salen complex (+)-30 as the catalyst for a nitroaldol condensation. An initial experiment with nitromethane and p-nitrobenzaldehyde in the presence of 5 mol% of (+)-30 gave an aldol product in poor yield and low enantiomeric excess, and attempts to adjust reaction conditions with this copper complex failed to improve the outcome. However, when bis-imine ligand (+)-20 was reduced to its tetrahydro derivative (+)-39 (Scheme 8), and when copper(II) was replaced by copper(I) at the core

38 are lower than those obtained with homoallylic alcohols, our result is comparable with previous attempts to use vinyl halides in asymmetric NHK reactions.5 Formation of homoallylic alcohols of (S) configuration in these NHK reactions is consistent with attack by the allyl nucleophile at the si face of the aldehyde carbonyl, as shown in E


Article

Accounts of Chemical Research of 39, a significant improvement was noted in both the yield and enantiomeric excess of the condensation product.16

in each synthesis is an asymmetric condensation of an α-aryloxy aldehyde with nitromethane (Scheme 9). Since the aryloxy

Scheme 8. Synthesis of Tetrahydrosalen Ligand (+)-39 by Reduction of (+)-20

Scheme 9. Syntheses of Beta-Blockers (S)-Toliprolol (40), (S)-Propanolol (41), and (S)-Moprolol (42) Using an Asymmetric Henry Reaction

Optimum conditions for the reaction of p-nitrobenzaldehyde with nitromethane were found using the copper(I) complex of tetrahydrosalen ligand (+)-39, prepared in situ from the toluene complex of copper(I) triflate, in warm methanol. The active catalyst under these conditions is unstable and could not be characterized, but the β-nitro alcohol product was formed in high enantiomeric excess at a catalyst loading as low as 1 mol %. This protocol was extended to a series of aromatic and aliphatic aldehydes, with the results shown in Table 5. The absolute Table 5. Asymmetric Synthesis of β-Nitro Alcohols with Ligand (+)-39 and Copper(I) Triflate.a aldehydes were unstable, they were prepared by oxidative cleavage of the corresponding allyl aryl ether and used in situ with nitromethane and the copper(I) complex of (+)-39. Hydrogenation of the product followed by condensation of the resultant amino alcohol with acetone and then a second hydrogenation, all in one flask, led to 40, 41, and 42 in excellent overall yield and high enantiomeric excess. The overall yield of (S)-Propanolol by this route was 75%; for comparison, the published synthesis of (S)-41 requires eight steps from commercial materials and occurs in 10.4% overall yield.18 Diastereoselectivity in condensations of aldehydes with more highly substituted nitro alkanes is generally poor and is reported to lead predominantly to the anti isomer.19 By contrast, the reaction of benzaldehyde and 1-naphthaldehyde with 1-nitropropane in the presence of the copper(I) complex of (+)-39 gave a syn/anti ratio of nitroaldol products, strongly favoring the syn isomer (Table 6). The relative configuration of these diastereomers, which were formed in high enantiomeric excess, was established by comparison of their 1H and 13C

product entry

R

t (h)

yield (%)b

ee (%)c

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

C6H5 3-CH3C6H4 2-CH3OC6H4 3-CH3OC6H4 4-CF3C6H4 2,6-Cl2C6H3 2-(OH)C6H4 3,5-(CH3O)2C6H3 2,4-(NO2)2C6H3 3-(4-CH3OC7H6O)-4-(NO2)C6H3 2-(OH)-3-(Br)-5-(Me3C)C6H2 1-naphthyl 2-furyl 3-furyl c-C6H11 Me3C CH3CCHC(CH3) (E) PhCHC(CH3) (E)

20 24 60 18 24 20 24 22 12 24 24 18 24 20 42 51 18 48

90 95 81 94 96 89 97 93 96 99 87 98 87 98 90 89 83 95

92 96 91 96 96 94d 92 96 95d 95d 98d 93 94 95 94 95 93 97d

Table 6. Asymmetric Henry Reactions of 1-Nitropropane with Benzaldehyde and 1-Naphthaldehyde, Catalyzed by the Copper(I) Complex of (+)-39a

a

Reactions were carried out on 0.2 mmol scale with 0.6 mL of nitromethane. bYield of isolated product. cDetermined by HPLC using Daicel Chiralcel AD, OD, OJ, and OD-H columns. dAbsolute configuration of the product was not determined.

configuration of each nitro alcohol was established as (R), except for entries 6, 9−11, and 18, by comparison of the measured specific rotation with the literature value. An application of this asymmetric Henry reaction was demonstrated with syntheses of three members of the family of beta-adrenergic receptor blocking agents, Toliprolol (40), Propanolol (41), and Moprolol (42).17 Each of these “betablockers” is characterized by a 1-aryloxy-3-isopropylamino-2propanol moiety, and all have (S) configuration. The key step

product entry

R

t [h]

dr syn/antib

yield (%)b

ee (%)c

1 2

C6H5 1-naphthyl

24 28

>20:1 >50:1

96 93

97 98

a

Reactions were carried out on 0.2 mmol scale with 0.6 mL of nitropropane. bDetermined by 1H NMR analysis. cCombined yields of syn and anti isomers. dDetermined by HPLC using Daicel Chiralcel OD-H and AS-H columns. F


Article


6. ASYMMETRIC CYCLOPROPANATION Asymmetric synthesis of cyclopropanes by reaction of alkenes with with α-diazo esters has been studied using a variety of chiral organo-metal systems with generally good stereoselectivity for monosubstituted alkenes.23 In contrast, 1,1disubstituted alkenes, where (E) and (Z) cyclopropanes can be formed, often give mixtures of geometrical isomers, and enantioselectivity can be poor as well.24 A solution to this problem was explored during investigation of cyclopropanation of α-methylstyrene with ethyl diazoacetate using a modified cobalt(II)−salen catalyst based on (+)-28.25 An initial experiment with 28 found that this complex was an active catalyst for cyclopropane formation with these reactants but that it gave a 1:1 mixture of (E) and (Z) isomers. We speculated that increased electron density at the cobalt core would make the catalyst more stereoselective, and a modification of salen ligand 20 was implemented that replaced a tert-butyl group in each benzenoid ring by a methoxy substituent. Synthesis of second generation cobalt complex (+)-43 began with with formylation of 2-(tert-butyl)-4-methoxyphenol to give an aldehyde which condensed with diamine (-)-3 in hot ethanol (Scheme 11). Cobalt(II) complex (+)-43 was obtained from ligand (+)-44 as an amorphous brown powder by reaction with anhydrous cobalt diacetate.

NMR spectra with published data, and their absolute configuration was confirmed as (R,R) by comparison of their specific rotations with literature values.20 The finding that the copper(I) complex of ligand (+)-39 gives nitroaldol products of (R) configuration requires attack by the nitro compound, as its nitronate tautomer, at the si face of the aldehyde carbonyl. A pathway that rationalizes this outcome is shown in Figure 7. We believe that a N−H hydrogen bond

Figure 7. Proposed model for the asymmetric Henry reaction of aldehydes with nitromethane and 1-nitropropane, catalyzed by the copper(I) complex of (+)-39.

between the tetrahydrosalen ligand and a nitronate oxygen plays a crucial role in this model, in which reactants are organized in a six-membered transition state. In this respect, our model parallels similar examples of the Henry reaction, where a N−H hydrogen bond between catalyst and nitronate assists organization of a transition state.21 However, a different “open mechanism” has been proposed by Feng for asymmetric Henry reactions catalyzed by a copper(I)−tetrahydrosalen complex based on 1,2-diphenylethylenediamine.22 Formation of syn products from the reaction of aldehydes with nitropropane implies that the copper-complexed nitronate has (Z) configuration. In this case, attack occurs at the si face of the aldehyde carbonyl through a cyclic transition state in which alkyl and aryl groups are oriented as shown for steric reasons. A catalytic cycle incorporating the key features of this condensation is shown in Scheme 10. After exchange of toluene

Scheme 11. Synthesis of Second-Generation Cobalt(II) Complex (+)-43

When the catalytic activity of (+)-43 was evaluated in cyclopropanation of α-methylstyrene with ethyl diazoacetate using 5 mol % of the catalyst, only a modest improvement in the (E:Z) ratio of cyclopropanes over the ratio obtained with 28 was observed. However, when potassium thioacetate was added to the mixture, a dramatic increase in the proportion of (E) cyclopropane was seen. Under optimized conditions, the (E:Z) ratio of trisubstituted cyclopropanes from this reaction was >30:1, the (E) isomer being formed in >90% enantiomeric excess. The absolute configuration of this cyclopropane was shown to be (1R,2R) by comparison of its optical rotation and NMR spectra with literature data. The conditions developed for cyclopropanation of αmethylstyrene using catalyst (+)-43 were extended to a series of 1,1-disubstituted ethylenes with the results shown in Table 7. All 1-aryl substituted alkenes gave a high proportion of the (E) cyclopropane which was formed in high enantiomeric excess in every case (entries 1−13). However, when a 1,1-dialkyl substituted alkene was subjected to cyclopropanation (entry 14) or when a heteroatom substituent was incorporated into the alkene (entry 15), both (E) selectivity and the enantioselectivity of the reaction were decreased. The favorable results in Table 7 led to an application of cobalt−salen catalyst (+)-43 in an improved synthesis of a cyclopropane-containing substance Synosutine (45, Figure 8) that is a dual re-uptake inhibitor of serotonin and norepinephrine transporter with activity in the 1−2 nM

Scheme 10. Catalytic Cycle for the Henry Reaction of Aldehydes with Nitroalkanes in the Presence of the Copper(I) Complex of (+)-39

at copper(I) triflate for tetrahydrosalen ligand (+)-39, the resultant complex reacts with the nitro alkane to give a (Z) copper nitronate. Intramolecular reaction of this species with the complexed aldehyde leads to a syn nitro alcohol of (R,R) configuration while regenerating the active catalyst. G


Article

Accounts of Chemical Research Table 7. Asymmetric Cyclopropanation of 1,1-Disubstituted Ethylenes, Catalyzed by (+)-43a,b

(E)-product entry

Ar

R

t (h)

ratio E/Z

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Ph Ph 2-OMeC6H4 2-Furyl Ph (2-CO2Me)-C6H4 3,4-di-OMeC6H3 2-thiophenyl 1-naphthyl 4-MeC6H4 3-(CO2Et)-5-C6H5-2-furyl 4-CF3C6H4 4-OMeC6H3 CH2-(1-naphthyl) SPh

Et n Bu Me Me CH2CO2Et Me Me Me Me (CH2)2CO2Me Me n Pr (CH2)3 Me n Pr

32 32 26 36 28 19 22 33 26 28 39 26 20 36 48

26:1 23:1 30:1 23:1 25:1 30:1 32:1 25:1 33:1 27:1 26:1 21:1 >50:1 18:1 16:1

d

yield (%)c

ee (%)e

94 91 90 89 91 92 96 93 97 96 92 95 90 95 87

92 90 96 92 91 95 94 90 96 94 97 92 98 83 88

a

The reactions were carried out on a 0.3 mmol scale in a 0.2 M solution with 1.5 equiv of alkene in the presence of (+)-43 and KSAc, each at 5 mol %. bCatalyst (+)-43 was stirred with KSAc for 1 h prior to the addition of alkene. cIsolated yields of (E) and (Z) isomers combined. dDetermined by 1 H NMR analysis. eDetermined by HPLC using Chiralcel OD, AD, OD-H, and AS-H columns.

which was obtained in 94% enantiomeric excess. Saponification of 49, conversion of the resultant carboxylic acid 50 to the corresponding N-methyl amide 51, and final reduction with hydride afforded enantiopure (R,R)-45 as its hydrochloride salt after acidification. Synthesis of 45 required six steps and proceeded in good overall yield from commercial materials. Scheme 13 shows a catalytic cycle for cyclopropanation of αmethylstyrene with ethyl diazoacetate using cobalt(II)−salen

Figure 8. Synosutine, a cyclopropane-containing dual reuptake inhibitor of serotonin and norepinephrine transporter.

Scheme 13. Catalytic Cycle for Cyclopropanation of αMethylstyrene with Ethyl Diazoacetate in the Presence of (+)-43 and Potassium Thioacetate (Y)

range.26 Our previous synthesis of (R,R)-45 installed the cyclopropane moiety via Charette asymmetric cyclopropanation of an allylic alcohol but data in Table 7, particularly entry 15, suggested a more direct approach to this target. The key step was asymmetric cyclopropanation of enol ether 46, prepared by acylation of 1-naphthol (47) with thiophen-2carbonyl chloride (48) followed by Tebbe methylenation (Scheme 12). Exposure of 46 and ethyl diazoacetate to (+)-43 and potassium thioacetate under the optimized conditions of Table 7 led to a 17:1 ratio in favor of (Z) cyclopropane 49, Scheme 12. Synthesis of Synosutine (+)-45 via Asymmetric Cyclopropanation with Catalyst (+)-43

complex (+)-43 with potassium thioacetate (Y) as an additive. Nucleophilic thioacetate is first incorporated into 43 to form an activated catalyst which reacts with ethyl diazoacetate to extrude nitrogen and form a cobalt-carbenoid. The alkene reacts at the cobalt center of the asymmetric carbenoid to generate enantioselectively a four-membered cobaltocycle with H


Article


provides potential entry to chiral organocatalysts, for example by reaction with isothiocyanates to generate bis-thiourea derivatives.32 Thus, the versatile platform represented by diamine 3 offers exciting prospects for future progress in asymmetric synthesis.

(E) configuration. A rationale for this configuration is presented in Figure 9, where the si face of α-methylstyrene approaches the

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

Figure 9. Proposed model for enantioselective cyclopropanation of αmethylstyrene with ethyl diazoacetate, catalyzed by cobalt(II)−salen complex (+)-43.

Biographies Subrata Shaw was born in West Bengal, India, and graduated with a B.Sc. from the University of Calcutta in 2005, with honors in chemistry. He obtained his M.Sc. degree at the Indian Institute of Technology Kanpur in 2007, and after briefly conducting research at McMaster University, Canada, he joined the graduate program at Oregon State University. His graduate studies were carried out under the supervision of Prof. James D. White, and he obtained his Ph.D. in 2014. He is currently doing postdoctoral research in the laboratory of Prof. Stephen Fesik at the Vanderbilt University Medical Center.

cobalt carbenoid from its open re face. Diastereoselection favoring the (E) metallocycle arises from a steric effect which places the phenyl group of the styrene anti to the ester substituent of the carbenoid while regioselection is decided by steric bulk around the congested cobalt which aligns the alkene to form a Co−C bond to the less substituted carbon. Reductive elimination of cobalt from the metallocyclobutane leads to a cyclopropane of (R,R) configuration as observed. Participation by thioacetate promoter Y in this cycle is believed to contribute to the catalytic function of 43 by the well-documented trans effect of sulfur ligands within the coordination sphere of transition metals.27

James D. White was born in Bristol, England, and obtained his undergraduate degree from Cambridge University in 1959. He completed a M.Sc. degree in chemistry at the University of British Columbia and obtained his Ph.D. at the Massachusetts Institute of Technology in 1965. In the same year, he joined the chemistry faculty of Harvard University and rose to Associate Professor in 1971, before moving to Oregon State University. He is currently Distinguished Professor Emeritus at Oregon State.

7. SUMMARY AND OUTLOOK Chirality embedded in cis-2,5-diaminobicyclo[2.2.2]-octane can be transmitted efficiently to enantioenriched products in certain reactions using salen−metal complexes and their reduced counterparts based on the diamine. Chromium(II), chromium(III), copper(I), and cobalt(II) complexes were found to catalyze asymmetric versions of hetero-Diels−Alder cycloaddition, the Nozaki−Hiyama−Kishi reaction of an allylic halide with aldehydes, the Henry nitroaldol condensation, and cyclopropanation of 1,1-disubstituted alkenes, all with high stereoselectivity and in good yield. Structural modifications to the salen ligand of these catalysts through steric and electronic tuning played an important role in optimizing these reactions. The structural feature of diamine 3 that we believe is responsible for the high level of stereoinduction in these reactions is a well-defined chiral pocket in a quadrant below the bicyclo[2.2.2]octane skeleton. In contrast to trans-1,2-diaminocyclohexane, the cis oriented amino substituents in 3 give the scaffold a directional character which augments its chiral properties. The pocket created by the salen ligand appended to this bicyclic diamine encloses a volume of chiral space that is large but flexible enough to encapsulate metals with a wide range of ionic radii. In addition to the catalytic entities employed in the reactions described above, other stable salen−metal complexes were prepared from (−)-3 that contain nickel(II), manganese(III), vanadium(III), aluminum(III), palladium(II), iron(III), and titanium(IV) as core metal ions. With the exception of iron(III)−salen complex (+)-26, for which good catalytic activity was demonstrated in enantioselective Michael addition of thiols to conjugated enones and in intramolecular Conia-ene synthesis of cyclopentanes,28−31 these complexes have not been explored. However, it is likely that they will have useful applications in asymmetric synthesis. The parent diamine 3 also

■

ACKNOWLEDGMENTS We are grateful to Prof. Peter Freeman, Oregon State University, for computational studies on 1, 2. and 3, and to Dr. Lev Zacharov, University of Oregon, for X-ray crystallographic analysis of many of the compounds prepared in the course of this work.

■

REFERENCES

(1) Yoon, T. P.; Jacobsen, E. N. Privileged Chiral Catalysts. Science 2003, 299, 1691−1693. (2) Bennani, Y. L.; Hanessian, S. trans-1,2-Diaminocyclohexane Derivatives as Chiral Reagents, Scaffolds, and Ligands for Catalysis: Applications in Asymmetric Synthesis and Molecular Recognition. Chem. Rev. 1997, 97, 3161−3196. (3) Jacobsen, E. N.; Zhang, W.; Muci, A. R.; Ecker, J. R.; Deng, L. Highly Enantioselective Epoxidation Catalysts Derived From 1,2Diaminocyclohexane. J. Am. Chem. Soc. 1991, 113, 7063−7064. (4) Schaus, S. E.; Branalt, J.; Jacobsen, E. N. Asymmetric HeteroDiels-Alder Reactions Catalyzed by Chiral (Salen)Chromium(III) Complexes. J. Org. Chem. 1998, 63, 403−405. (5) Berkessel, A.; Menche, D.; Sklorz, C. A.; Schroder, M.; Paterson, I. A Highly Enantioselective Catalyst for the Asymmetric NozakiHiyama-Kishi Reaction of Allylic and Vinylic Halides. Angew. Chem., Int. Ed. 2003, 42, 1032−1035. (6) Berkessel, A.; Vogl, N. DIANANE-CrIII-salen Complexes as Highly Enantioselective Catalysts for Hetero-Diels-Alder Reactions of Aldehydes with Dienes. Eur. J. Org. Chem. 2006, 2006, 5029−5035. (7) Racemic 3 was prepared from bicyclo[2.2.2]octane-2,5-dione by reduction of its bis oxime derivative: Werstiuk, N. H.; Yeroushalmi, S.; Guan-Lin, H. Synthesis of bicyclic diones and thiones. Facile methylation of the enolates of bicyclo[2.2.1]heptane-2,5-dione and I


Article

Accounts of Chemical Research bicyclo[2.2.2]octane-2,5-dione. An AM1 computational study of bicyclic enolates. Can. J. Chem. 1992, 70, 974−980. (8) White, J. D.; Shaw, S. Cis-2,5-Diaminobicyclo[2.2.2]-octane, a New Scaffold for Asymmetric Catalysis via Salen-Metal Complexes. Org. Lett. 2011, 13, 2488−2491. (9) Tichy, M.; Sicher, J. Stereochemical Studies. LXV. Synthetic Studies in the Tricyclo[4,4,0,03,8]decane and Tricyclo[4,3,1,03,7]decane Series: The Synthesis of Twistene. Collect. Czech. Chem. Commun. 1972, 37, 3106−3116. (10) A different synthesis of (−)-14 has recently been reported from the Corey laboratory: Reddy, K. M.; Bhimireddy, E.; Thirupathi, B.; Breitler, S.; Yu, S.; Corey, E. J. Cationic Chiral Fluorinated Oxazaborolidines. More Potent, Second-Generation Catalysts for Highly Enantioselective Cycloaddition Reactions. J. Am. Chem. Soc. 2016, 138, 2443−2453. (11) Shaw, S. Cis-2,5-diaminobicyclo[2.2.2]octane: a Novel Scaffold for Asymmetric Catalysis. Ph.D. Dissertation, Oregon State University, Corvallis, OR, January 2014. (12) It is recognized that this and other reaction pathways proposed as models in this Account serve primarily to rationalize a stereochemical result and may not constitute a kinetically accurate picture of the process. It is known, for example, that certain metal−salencatalyzed reactions exhibit non-first-order kinetics in the catalyst: Nielsen, L. P. C.; Stevenson, C. P.; Blackmond, D. G.; Jacobsen, E. N. Mechanistic Investigation Leads to a Synthetic Improvement in the Hydrolytic Kinetic Resolution of Terminal Epoxides. J. Am. Chem. Soc. 2004, 126, 1360−1362. A cooperative bimetallic mechanism can also be involved: Ford, D. D.; Nielsen, L. P. C.; Zuend, S. J.; Musgrave, C. B.; Jacobsen, E. N. Mechanistic Basis for High Stereoselectivity and Broad Substrate Scope in the (salen)Co(III)-Catalyzed Hydrolytic Resolution. J. Am. Chem. Soc. 2013, 135, 15595−15608. (13) Bandini, M.; Cozzi, P.; Melchiorre, P.; Umani-Ronchi, A. The First Catalytic Enantioselective Nozaki-Hiyama Reaction. Angew. Chem., Int. Ed. 1999, 38, 3357−3359. (14) Ananthi, N.; Velmathi, S. Asymmetric Henry Reaction Catalysed by Transition Metal Complexes: A Short Review. Indian J. Chem. 2013, 52B, 87−108. (15) Evans, D. A.; Seidel, D.; Rueping, M.; Lam, H. W.; Shaw, J. T.; Downey, C. W. A New Copper Acetate Bis(oxazoline)-Catalyzed, Enantioselective Henry Reaction. J. Am. Chem. Soc. 2003, 125, 12692− 12693. (16) White, J. D.; Shaw, S. A New Catalyst for the Asymmetric Henry Reaction: Synthesis of β-Nitroethanols in High Enantiomeric Excess. Org. Lett. 2012, 14, 6270−6273. (17) Dukes, M.; Smith, L. H. β-Adrenergic Blocking Agents. 9. Absolute Configuration of Propanolol and of a Number of Related Aryloxypropanolamines and Arylethanolamines. J. Med. Chem. 1971, 14, 326−328. (18) Sayyed, I. A.; Thakur, V. V.; Nikalje, M. D.; Dewkar, G. K.; Kotkar, S. P.; Sudalai, A. Asymmetric Synthesis of Aryloxypropanolamines via OsO4-Catalyzed Asymmetric Dihydroxylation. Tetrahedron 2005, 61, 2831−2838. (19) (a) Blay, G.; Domingo, L. R.; Olmos, V. H.; Pedro, J. R. New Highly Asymmetric Henry Reaction Catalyzed by CuII and a C1Symmetric Aminopyridine Ligand, and its Application to the Synthesis of Miconazole. Chem. - Eur. J. 2008, 14, 4725−4730. (b) Handa, S.; Nagawa, K.; Sohtome, Y.; Matsunaga, S.; Shibasaki, M. A Heterobimetallic Pd/La/Schiff Base Complex for anti-Selective Catalytic Asymmetric Nitroaldol Reactions and Applications to Short Syntheses of β-Adrenoceptor Agonists. Angew. Chem., Int. Ed. 2008, 47, 3230−3233. (20) Zhou, Y.; Dong, J.; Zhang, F.; Gong, Y. Synthesis of C1Symmetric Chiral Secondary Diamines and Their Applications in Asymmetric Copper(II)-Catalyzed Henry (Nitroaldol) Reactions. J. Org. Chem. 2011, 76, 588−600. (21) (a) Ube, H.; Terada, M. Enantioselective Henry (nitroaldol) reaction catalyzed by axially chiral guanidines. Bioorg. Med. Chem. Lett. 2009, 19, 3895−3898. (b) Taimoory, S. M.; Dudding, T. An Evolving Insight into Chiral H-Bond Catalyzed Aza-Henry Reactions: A

Cooperative Role for Noncovalent Attractive Interactions Unveiled by Density Functional Theory. J. Org. Chem. 2016, 81, 3286−3295. (22) Xiong, Y.; Wang, F.; Huang, X.; Wen, Y.; Feng, X. A New Copper(I)-Tetrahydrosalen-Catalyzed Asymmetric Henry Reaction and Its Extension to the Synthesis of (S)-Norphenylephrine. Chem. Eur. J. 2007, 13, 829−833. (23) Pellissier, H. Recent Developments in Asymmetric Cyclopropanation. Tetrahedron 2008, 64, 7041−7095. (24) (a) Bonaccorsi, C.; Bachmann, S.; Mezzetti, A. Electronic Tuning of the PNNP Ligand for the Asymmetric Cyclopropanation of Olefins Catalysed by [RuCl(PNNP)]+. Tetrahedron: Asymmetry 2003, 14, 845−854. (b) Chen, Y.; Zhang, X. P. Vitamin B12 Derivatives as Natural Asymmetric Catalysts: Enantioselective Cyclopropanation of Alkenes. J. Org. Chem. 2004, 69, 2431−2435. (c) Boldini, I.; Guillemot, G.; Caselli, A.; Proust, A.; Gallo, E. Polyoxometalates: Powerful Catalysts for Atom-Efficient Cyclopropanations. Adv. Synth. Catal. 2010, 352, 2365−2370. (25) White, J. D.; Shaw, S. A New Cobalt-Salen Catalyst for Asymmetric Cyclopropanation. Synthesis of the Serotonin-Norepinephrine Reuptake Inhibitor (+)-Synosutine. Org. Lett. 2014, 16, 3880−3883. (26) White, J. D.; Juniku, R.; Huang, K.; Yang, J.; Wong, D. T. Synthesis of 1,1-[1-Naphthyloxy-2-thiophenyl]-2-methylaminomethylcyclopropanes and Their Evaluation as Inhibitors of Serotonin, Norepinephrine, and Dopamine Transporters. J. Med. Chem. 2009, 52, 5872−5879. (27) For a review, see: Coe, B. J.; Glenwright, S. J. Trans-effects in Octahedral Transition Metal Complexes. Coord. Chem. Rev. 2000, 203, 5−80. (28) White, J. D.; Shaw, S. Iron Catalyzed Enantioselective SulfaMichael Addition: A Four Step Synthesis of the Anti-Asthma Agent Montelukast. Chem. Sci. 2014, 5, 2200−2204. (29) Shaw, S.; White, J. D. A New Iron(III)-Salen Catalyst for Enantioselective Conia-ene Carbocyclization. J. Am. Chem. Soc. 2014, 136, 13578−13581. (30) Shaw, S.; White, J. D. Regioselective and Enantioselective Addition of Sulfur Nucleophiles to Acyclic α,β,γ,δ-Unsaturated Dienones Catalyzed by an Iron(III)-Salen Complex. Org. Lett. 2015, 17, 4564−4567. (31) Shaw, S.; White, J. D. Asymmetric Catalysis with Iron-Salen Complexes. Synthesis 2016, DOI: 10.1055/s-0035-1562013. (32) De, C. K.; Klauber, E. G.; Seidel, D. Merging Nucleophilic and Hydrogen Bonding Catalysis: An Anion Bonding Approach to the Kinetic Resolution of Amines. J. Am. Chem. Soc. 2009, 131, 17060− 17061.

J


cis-2,5-Diaminobicyclo[2.2.2]octane, a New Chiral Scaffold for

Recommend Documents