Asymmetric Nucleophilic Catalysis with an Octahedral Chiral-at-Metal

Research Article pubs.acs.org/acscatalysis

Asymmetric Nucleophilic Catalysis with an Octahedral Chiral-atMetal Iridium(III) Complex Thomas Cruchter,† Michael G. Medvedev,‡,§ Xiaodong Shen,† Thomas Mietke,† Klaus Harms,† Michael Marsch,† and Eric Meggers*,† †

Fachbereich Chemie, Philipps-Universität Marburg, Hans-Meerwein-Strasse 4, 35043 Marburg, Germany X-ray Structural Laboratory, A.N. Nesmeyanov Institute of Organoelement Compounds RAS, Vavilova St. 28, 119991 Moscow, Russian Federation § N.D. Zelinsky Institute of Organic Chemistry RAS, Leninsky Prospect 47, 119991 Moscow, Russian Federation ‡

S Supporting Information *

ABSTRACT: Herein, we report about the design, synthesis, and application of a nucleophilic octahedral chiral-only-atmetal iridium(III) complex. We demonstrate that the enantiopure form of this complex serves as an efficient catalyst for the asymmetric Steglich rearrangement of O-acylated azlactones (up to 96% ee and 99% yield) and the related asymmetric Black rearrangement of O-acylated benzofuranones (up to 94% ee and 99% yield). We provide insight into the mechanisms of these two acyl migration reactions and the catalyst’s manner of chiral recognition with crystal structures of the active catalyst and a catalysis intermediate analog, as well as with quantum chemical calculations based on them. Furthermore, we demonstrate that the presented catalyst also efficiently catalyzes the asymmetric reaction between aryl alkyl ketenes and 2-cyanopyrrole to give the corresponding α-chiral N-acyl pyrroles (up to 95% ee and 99% yield). KEYWORDS: nucleophilic catalyst, Lewis base catalysis, asymmetric catalysis, iridium, chiral-at-metal, acyl transfer

■

INTRODUCTION Over the past 20 years, since the introduction of the first “chiral DMAP” reagent by Vedejs and Chen in 1996 (A, Figure 1),1 a significant amount of work has been performed by many research groups in developing highly active and at the same time highly selective nucleophilic catalysts capable of various asymmetric transformations, such as (dynamic) kinetic resolutions of racemic compounds and desymmetrizations of meso compounds,2−9 Steglich-type acyl migration reactions,9−15 reactions of ketenes with various substrates,16,17 and others.18 In large part, the developed chiral catalysts rely on 4dimethylaminopyridine (DMAP) or closely related 4-pyrrolidinopyridine (PPY) as their catalytically active entities due to their high intrinsic activity.2,3,11,12,16,19−21 However, powerful non-DMAP-type nucleophilic catalysts have been developed as well, which rely on amidines,4,9 isothioureas,5,9,13 imidazoles,6,14 O-nucleophiles,7,15 phosphines,8 N-heterocyclic carbenes (NHCs),17 and others.18 A straightforward approach to achieve enantiocontrol in asymmetric transformations is to introduce a steering stereocenter in proximity to the catalyst’s active sitein case of DMAP, in the 2-position. However, as efficient steering stereocenters usually comprise sterically demanding moieties, this typically goes along with a significant loss of catalytic © XXXX American Chemical Society

activity, as is the case for acylation reagent A from Vedejs and Chen, which has to be employed in its N-acylated form in stoichiometric amounts.1 Hence, from the standpoint of catalytic activity, it is desirable to have a remote stereocenter, but this, on the other hand, often goes along with a loss of enantiocontrol.3a,d,12a,b,g,13b,14a This phenomenon has aptly been called the selectivity−reactivity dilemma by Fuji et al.3a and is also encountered in nonplanar systems, such as catalyst D (Figure 1), although, due to the lack of planarity, it is less pronounced there.4,13,14a With respect to nucleophilic catalysts, several interesting strategies have been reported to circumvent low catalytic activity when aiming for high selectivity including strategies that go beyond “classic” carbon-centered tetrahedral chirality. In 1996, shortly after Vedejs and Chen’s seminal publication, Fu and co-workers introduced planar-chiral ferrocene-fused DMAP derivatives as chiral nucleophilic catalysts (B, Figure 1),2a which have, up to date, been demonstrated to efficiently catalyze a broad array of asymmetric transformations and also been extended to analogous ruthenocyl-fused derivatives.21 Due to their unrivaled versatility,2,11,16,19 Fu’s ferrocene-fused DMAP derivatives B have Received: April 22, 2017 Revised: June 19, 2017 Published: June 23, 2017 5151

DOI: 10.1021/acscatal.7b01296 ACS Catal. 2017, 7, 5151−5162

Research Article

ACS Catalysis

Figure 1. Vedejs’ asymmetric acylation reagent (A) and various chiral nucleophilic catalysts (B−E) relying on different frameworks and types of chirality.

Figure 2. Evaluation of chiral iridium(III) complexes Λ-Ir1−Λ-Ir5 in the Steglich rearrangement of substrate 1a.

Figure 3. From left to right: (1) Structural formula of precatalyst Λ-Ir5. (2) Structural formula of active catalyst Λ-Ir5′ and its three-dimensional structure (obtained from a crystal structure of rac-Ir5′; positional disorder and cocrystallized solvent molecules omitted for clarity; 50% probability of thermal ellipsoids).30 (3) Structural formula of enantiomeric catalyst Δ-Ir5′.

emerged as a reference standard in asymmetric nucleophilic catalysis.4d,22a Meanwhile, the concept of creating planar-chiral nucleophilic catalysts by ferrocene fusion has been extended by others to non-DMAP-type catalysts, such as NHC- and amidine-based systems.4d,23 In contrast, Spivey and co-workers made use of axial chirality by employing axial-chiral DMAP derivatives in the kinetic resolution and desymmetrization of alcohols.3b,c,e,f Ooi and co-workers developed axial-chiral BINOL-derived O-nucleophilic ammonium betaines, which they utilized in Steglich rearrangements of O-acylated azlactones.15 And quite recently, Suga and co-workers reported about an axial-chiral BINOL-derived DMAP-type catalyst (C, Figure 1) which they employed in the Steglich-type rearrangement of O-acylated oxindoles, in the kinetic resolution of alcohols, and in the desymmetrization of diols.12g Carbery and co-workers developed a helicenoidal DMAP derivative which they used in the kinetic resolution of alcohols.3h,i And in this context it is worth mentioning that a photoswitchable helicalchiral DMAP-type catalyst has recently been reported by Chen and co-workers allowing enantiodivergent Steglich rearrangement of O-acylated azlactones with a single catalyst.12h

Inert octahedral metal complexes were introduced and established as structural templates for asymmetric catalysis first by the research groups of Belokon, Fontecave, Gladysz, and Ohkuma.24 Our group contributed to this area with the design of bis-cyclometalated iridium complexes as hydrogenbonding catalysts,25 Brønsted base catalysts,26 and enamine catalysts.27,28 In this work, we present the development of a robust, kinetically inert, octahedral complex with metalcentered chirality featuring a nucleophilic ligand that can be used as a versatile nucleophilic catalyst in various asymmetric transformations.

■

RESULTS AND DISCUSSION Inspired by our results with chiral-at-metal hydrogen-bonding catalysts;25 structurally related Brønsted base catalysts;26 and chiral amidine-,4,9 isothiourea-,5,9,13 and imidazole-based6,14 nucleophilic catalysts reported by Birman and others (e.g., D, Figure 1), we reasoned that it might be possible to utilize a biscyclometalated octahedral iridium(III) complex comprising a metal-located stereocenter and a bidentate 3H-imidazo[4,5h]quinoline ligand as a nucleophilic catalyst (E, Figure 1).28 5152


Research Article

ACS Catalysis

Scheme 1. Synthesis of Precatalyst Λ/Δ-Ir5 and Catalyst Λ/Δ-Ir5′ Starting from Benzoxazole 3 and IrCl3·3H2O

a

Yield of deprotonation step D: 99%. bAlternatively, wash CH2Cl2 solution of Λ/Δ-Ir5 with aqueous 0.3 M NaOH and water.

Figure 4. Left: CD spectra of enantiomeric precatalysts Λ- and Δ-Ir5. Right: Superposition of absorption-normalized excerpts of the HPLC traces of catalysts rac-, Λ-, and Δ-Ir5′.

Indeed, when we added 3 mol % of complexes Λ-Ir1−Λ-Ir5 together with 6 mol % of cesium carbonate to a solution of Oacylated azlactone 1a, we observed in all cases smooth Steglich rearrangement of substrate 1a to product 2a (Figure 2).10 However, choice of the catalyst’s residues R1−R3 proved to be crucial in order to obtain 2a with high enantioselectivity. Under nonoptimized conditions, we found the highest selectivity for complex Λ-Ir5 with R1 = 2,4,6-(iPr)3-Ph, R2 = 2,6-(Me)2-Ph, and R3 = Cl providing product 2a with 85% ee (Figure 2). Complex Λ-Ir5 is the HPF6 salt of actual in situ generated catalytically active species Λ-Ir5′, which is generated from precatalyst Λ-Ir5 by reaction with a weak Brønsted base, for example cesium carbonate.29 The entire structural formulas of precatalyst Λ-Ir5 and active catalyst Λ-Ir5′ are depicted in Figure 3, together with the three-dimensional structure of catalyst Λ-Ir5′ and the structural formula of enantiomeric catalyst Δ-Ir5′, which was concomitantly prepared with Λ-Ir5′ (see Scheme 1) and also employed as a catalyst in the course of this study. Synthesis of enantiopure precatalyst Λ/Δ-Ir5 and catalyst Λ/ Δ-Ir5′ is depicted in Scheme 1 and based on methodology previously developed in our group.31 First, reaction of benzoxazole ligand 3 with IrCl3·3H2O in 2-ethoxyethanol results in the formation of the corresponding cyclometalated

iridium(III) dimer [IrIII(C^N)2(μ-Cl)]2 (Scheme 1, A; dimer structure not drawn; C^N represents cyclometalated 3; crystal structure of this dimer is provided in the Supporting Information), which is cleaved with auxiliary ligand (S)-432 in the presence of silver triflate to give diastereomeric iridium(III) complexes Λ-(S)-Ir5 and Δ-(S)-Ir5, which are purified and resolved by standard silica flash chromatography (Scheme 1, B). Auxiliary ligand (S)-4 is then cleaved off with triflic acid in acetonitrile and replaced by imidazoquinoline ligand 5 (Scheme 1, C, I−II).33 After anion exchange, chromatographic purification yields precatalysts Λ-Ir5 and Δ-Ir5 (Scheme 1C, III).34 As shown in Figure 2, Λ/Δ-Ir5 can be directly used in case an excess of a Brønsted base relative to the catalyst is present in the reaction mixture, leading to in situ formation of active Λ/Δ-Ir5′.29 We later found that active catalyst Λ/Δ-Ir5′ can be conveniently prepared by passing a CH2Cl2 solution of Λ/Δ-Ir5 through a plug of piperidinomethylated polystyrene beads35 or, alternatively, by washing a CH2Cl2 solution of Λ/ΔIr5 with diluted aqueous sodium hydroxide followed by thorough washing with water (Scheme 1, D). Chiral HPLC analysis revealed >99% ee for both Λ- and Δ-Ir5′, and the absolute configurations were assigned by circular dichroism spectroscopy (Figure 4). Detailed preparation procedures and 5153


Research Article

ACS Catalysis Table 1. Catalytic Enantioselective Steglich Rearrangement of O-Acylated Azlactonesa

entry c, d

1 2c,d 3c,d 4c 5 6 7 8 9 10e,f 11e 12e,f 13e,g 14e,g,h

t (°C)

time (h)

R1

R2

R3

yield (%)b

ee (%)b

product

0 0 0 −30 −30 −30 −30 −30 −30 −30 −30 −30 −30 −30

12 12 12 72 72 72 72 72 72 72 72 72 72 72

Bn i-Bu Me Bn Bn Bn Bn Bn Bn Bn Bn Bn Bn Bn

Bn Bn Bn Bn Me C(CH3)2CCl3 Bn Bn Bn Bn Bn Bn Bn Bn

p-MeO-Ph p-MeO-Ph p-MeO-Ph p-MeO-Ph p-MeO-Ph p-MeO-Ph p-(t-Bu)-Ph 1-Naphtyl 2-Naphtyl 2-Naphtyl p-MeO-Ph p-MeO-Ph t-Bu t-Bu

97 91 80 95 68 99 89 88 90 94 97 93 91 90

88 70 45 93 88 90 91 91 95 94 93 93 96 96

2a (R) 2b (R) 2c (R) 2a (R) 2d (R) 2e (R) 2f (R) 2g (R) 2h (R) 2h (R) 2a (R) 2a (R) ent-2i (S)g ent-2i (S)g

Standard reaction conditions: substrate 1 (1.0 equiv; approximately 90 μmol) with precatalyst Λ-Ir5 (1 mol %) and Cs2CO3 (3 mol %) in EtOAc (c = 0.65 M) at the indicated temperature. bIsolated yields; enantioselectivities determined from isolated products by chiral HPLC analysis. c0.5 mol % of catalyst was used. dc = 0.5 M. eActive catalyst Ir5′ was used instead of precatalyst Ir5 (no addition of Cs2CO3). fScaled-up reactions, entry 10: 400 mg (919 μmol) of substrate 1h; entry 12: 381 mg (917 μmol) of substrate 1a. gΔ-configured catalyst Δ-Ir5′ used (major product: S-configuration). h CD2Cl2 solution of Δ-Ir5′ left standing for 7 days at 25 °C before usage. a

had meanwhile discovered that both methods, using HPF6 salt Λ/Δ-Ir5 in combination with Cs2CO3 or using active catalyst Λ/Δ-Ir5′ alone, led to identical results, which is visible from entries 4 and 11. All reactions discussed further below were carried out using catalyst Λ/Δ-Ir5′ instead of precatalyst Λ/ΔIr5. In order to demonstrate the catalyst’s stability, we left a dichloromethane solution of catalyst Δ-Ir5′ standing for 7 days at 25 °C before usage, which then delivered product ent-2i with the same selectivity and yield as freshly prepared Δ-Ir5′ (entry 14, compare with entry 13; details in SI). In this context it is worth mentioning that we observed neither any decomposition nor any racemization of precatalysts Λ/Δ-Ir5 and catalysts Λ/ Δ-Ir5′ when we stored them as solids for about 6 months at 4 °C under standard atmospheric conditions. Exemplarily, we scaled up two reactions with substrates 1a and 1h by a factor of 10 (∼1 mmol scale) and obtained similar yields and basically identical selectivities as in the case of the corresponding standard scale examples (entries 10 and 12, compare with entries 9 and 11). It is noteworthy that all Steglich rearrangements from Table 1 were set up under a normal atmosphere without any special precautions against moisture and oxygen. Stereochemistry assignment details for C-acylated azlactones 2 are provided in the Supporting Information. Having demonstrated its suitability in the Steglich rearrangement of O-acylated azlactones, we became curious whether Λ/ Δ-Ir5′ could also be employed in the related Black rearrangement of O-acylated benzofuranones.36 A short reaction conditions screening revealed that tert-amyl alcohol (TAA) was the solvent of choice for this reaction in terms of selectivity and activity. However, for reaction temperatures considerably lower than 0 °C, we found a 2:1 mixture of TAA (mp ∼ −10 °C) and cyclopentyl methyl ether (CPME) more suitable (details in SI). Again, we found low catalyst loadings of 1−2 mol % as sufficient. Next, we examined the scope of this reaction (Table 2). First, we tested substrate 6a with R1, R2 =

characterizations of all presented iridium(III) complexes are provided in the Supporting Information (SI). Having identified Λ/Δ-Ir5 as a promising precatalyst, we continued focusing on the Steglich rearrangement of O-acylated azlactones (1 → 2), and so we first performed a reaction conditions screening (details in SI). On the basis of this screening, we identified ethyl acetate as the solvent of choice, as it already provided a high enantioselectivity at 0 °C and further allowed us to lower the temperature to −30 °C, resulting in a significantly enhanced selectivity at a still reasonable reaction rate. We found loadings of 0.5−1 mol % of precatalyst Λ-Ir5 and 3 mol % of cesium carbonate as sufficient for this type of reaction. Next, we investigated its scope (Table 1). With just 0.5 mol % of precatalyst Λ-Ir5 in the presence of 3 mol % Cs2CO3 at 0 °C, substrate 1a (R1 = Bn, R2 = Bn, R3 = pMeO-Ph) gave product 2a with 88% ee and 97% yield within 12 h (Table 1, entry 1). We then changed over to substrate 1b with R1 = i-Bu and obtained product 2b with moderate 70% ee (entry 2). Subsequently, we switched to 1c with R1 = Me, and the selectivity dropped to 45% ee (2c, entry 3). For 1a being the best substrate so far, we now lowered the temperature to −30 °C and prolonged the reaction time to 72 h, which led to an increase from 88% ee to 93% ee for product 2a (entries 1 and 4). Keeping these conditions, we now exchanged the migrating acyl group R2 = Bn for R2 = Me (1d, entry 5) and R2 = C(CH3)2CCl3 (1e, entry 6), which gave the products 2d and 2e in 88% ee and 90% ee, respectively, indicating that the catalyst is less sensitive to variation of R2 compared to R1. Finally, we screened R3 (entries 7−9, 11, 13), and the best results were obtained here with substrate 1h with R3 = 2naphtyl (2h, 95% ee, 90% yield, entry 9) and substrate 1i with R3 = t-Bu (ent-2i, 96% ee, 91% yield, entry 13; Δ-configured catalyst used). Importantly, the reactions from entries 11 and 13 were, in contrast to all others discussed so far, carried out using active catalyst Λ/Δ-Ir5′ in the absence of Cs2CO3, as we 5154


Research Article

ACS Catalysis Table 2. Catalytic Enantioselective Black Rearrangement of O-Acylated Benzofuranonesa

entry

t (°C)

1c,d 2d 3d 4d,e 5d 6 7 8 9 10f 11 12 13g 14g 15i

0 °C 0 °C 0 °C 0 °C 0 °C −15 °C −15 °C −15 °C −30 °C −30 °C −30 °C −30 °C −30 °C −30 °C −30 °C

time (h) 18 10 18 48 20 48 48 48 48 48 48 48 48 48 48

h h h h h h h h h h h h h h h

R1

R2

yield (%)b

ee (%)b

product

Ph Me Me Me Bn Me i-Bu Bn Me Me Me Me Me Me Me

Ph Ph Bn i-Pr C(CH3)2CCl3 Ph Ph Ph Ph Ph Me C(CH3)2CCl3 C(CH3)2CCl3 Ph Ph

N.D.c 99 99 98 98 98 99 99 99 99 93 99 99 99 (73h) 99

30 91 80 82 89 92 90 91 94 94 70 93 93 92 (98h) 93

7a (R) 7b (R) 7c (R) 7d (R) 7e (S)j 7b (R) 7f (R) 7g (R) 7b (R) ent-7b (S)f 7h (R) 7i (S)j 7i (S)j 7b (R) 7b (R)

Standard reaction conditions: substrate 6 (1.0 equiv; 70−90 μmol) with catalyst Λ-Ir5′ (1 mol %) in TAA/CPME 2:1 (c = 0.2 M) at the indicated temperature. bIsolated yields; enantioselectivities determined from isolated products by chiral HPLC analysis (except entry 1: ee of 7a from crude mixture). cNot determined, incomplete after 18 h; ∼ 50% conv. acc. to TLC. dSolvent: TAA, c = 0.3 M. e2 mol % catalyst used. fΔ-configured catalyst used. gScaled-up reactions, entry 13: 252 mg (717 μmol) of substrate 6i; entry 14: 830 mg (3.09 mmol) of substrate 6b. hYield and ee after recrystallization from methanol. iRecovered catalyst used (>99% ee; recovered from entry 14). jActually no inversion of the stereocenter; product is S-configured due to higher priority of the exocyclic acyl fragment in this particular case. a

Ph at 0 °C in TAA in the presence of 1 mol % of Λ-Ir5′ (Table 2, entry 1). Unfortunately, we only found approximately 50% conversion after 18 h and 30% ee for 7a. However, when we replaced R1 = Ph with R1 = Me, we could isolate product 7b after just 10 h under identical conditions in 91% ee and 99% yield (entry 2). Next, we replaced the migrating acyl group R2 = Ph with R2 = Bn (entry 3) and R2 = i-Pr (entry 4), which resulted in a decreased selectivity for both products (80% ee for 7c and 82% ee for 7d) and in the case of the latter in a slower reaction requiring 2 mol % of the catalyst to run to completion within 48 h. However, when we tested substrate 6e at 0 °C (R1 = Bn and R2 = C(CH3)2CCl3), we obtained product 7e in 89% ee (entry 5). For 6b, being the best substrate so far, we now lowered the temperature to −15 °C, switched over to TAA/ CPME 2:1, and increased the reaction time to 48 h, which pushed the ee for product 7b to 92% ee (entry 6). Under these conditions, we also tested substrates 6f (R1 = i-Bu; entry 7) and 6g (R1 = Bn; entry 8), which gave products 7f and 7g with good ee’s (90% and 91%, respectively) and in excellent yields (both 99% yield). Decreasing the temperature further from −15 °C to −30 °C, under otherwise identical conditions, pushed the ee of product 7b to 94% with no drop in isolated yield (99%; entries 6 and 9). With enantiomeric catalyst Δ-Ir5′, we obtained, as expected, the same result with an inverted R/S product ratio (product ent-7b (S), 99% yield, 94% ee, entry 10). Under identical conditions, we tested another two substrates with different acyl groups R2 (entries 11, 12). While substrate 6h (R2 = Me) with its small methyl group gave product 7h with only moderate 70% ee (entry 11), substrate 6i (R2 = C(CH3)2CCl3) with its more bulky group gave product 7i with virtually the same selectivity and yield (99% yield, 93% ee, entry 12) as substrate 6b. Again, we exemplarily employed ΛIr5′ in two scaled-up reactions with substrates 6i (0.72 mmol; entry 13) and substrate 6b (3.1 mmol; entry 14) giving us product 7i with identical and product 7b with slightly

diminished selectivity and both products with identical yields compared to the standard scale experiments (entries 9 and 12− 14). It is noteworthy that product 7b from the scaled-up reaction (entry 14) could be pushed to 98% ee by one recrystallization while maintaining an acceptable yield of 73% (details in SI). We recovered catalyst Λ-Ir5′ from this scaled-up experiment (entry 14) in 94% yield and >99% ee (procedure and details in SI) and used recovered catalyst Λ-Ir5′ again in a standard scale reaction with substrate 6b where we virtually obtained the same result as with freshly prepared Λ-Ir5′ (99% yield, 93% ee, entry 15). Like the Steglich rearrangements from Table 1, all Black rearrangements from Table 2 were set up under a normal atmosphere without any special precautions against moisture and oxygen. Stereochemistry assignment details for C-acylated benzofuranones 7 and an X-ray structure of (S)-7i are provided in the Supporting Information. Having demonstrated that catalyst Λ/Δ-Ir5′ is capable of catalyzing stereogenic O → C acyl migrations with high levels of activity and selectivity,37 we wondered if Λ/Δ-Ir5′ would also be a suitable catalyst for a completely different type of reaction. So, we focused on the asymmetric reaction between aryl alkyl ketenes (8) and 2-cyanopyrrole (9) yielding α-chiral N-acyl pyrroles (10), a reaction first reported by Fu and coworkers with catalyst B (see Figure 1).16c In the course of a short reaction conditions screening (details in SI), we identified toluene and methyl tert-butyl ether (MTBE) as the solvents of choice38 and a loading of 2 mol % of Λ-Ir5′, an excess of 2.0 equiv of 2-cyanopyrrole (9),39 a reaction temperature of 0 °C, and a general reaction time of 48 h as suitable reaction parameters. Next, we examined the scope of this reaction (Table 3). Starting with simple phenyl ethyl ketene (8a), product 10a was obtained in 89% ee (Table 3, entry 1). The addition of a methyl group in the ortho-position of the ketene’s phenyl ring caused the ee to drop to 84% (entry 2; product 10b). However, 5155


Research Article

ACS Catalysis

oxygen (details in SI). Stereochemistry assignment details for N-acyl pyrroles 10 are provided in the Supporting Information. Fu and co-workers have demonstrated that α-chiral N-acyl pyrroles (10) are useful compounds from a synthetic point of view, as they can be easily converted into α-chiral acids, esters, amides, aldehydes, and alcohols.16c Taking this and the large number of nucleophilic catalysts reported to date into account, we were surprised to find no other report apart from Fu’s initial publication dealing with this transformation when this publication was written.40 Now, we turned our attention to the mechanisms of the three investigated reactions. The anticipated mechanism for the Black rearrangements (6 → 7) is depicted in Figure 5 using the example of substrate 6i.11b,36 At first, the imidazoquinoline ligand’s nucleophilic nitrogen atom reacts with the carbonate group of substrate 6i, resulting in the formation of N-acylated intermediate Λ-Ir5-Int, which is a salt that consists of a +1 charged iridium(III) complex cation and its associated benzofuranone enolate counteranion (Figure 5, step I). In order to support the formation of anticipated ΛIr5-Int experimentally, we prepared its analog rac-Ir5-TfO, in which we replaced the reactive enolate counteranion with an inert triflate counteranion, and, to our delight, obtained a crystal structure for rac-Ir5-TfO, which enabled us to elucidate the three-dimensional structure of Λ-Ir5-TfO (Figure 5; details in SI). After the formation of Λ-Ir5-Int, product 7i is finally formed by nucleophilic attack of the reactive enolate onto the N-acylated complex cation, which goes along with regeneration of catalyst Λ-Ir5′ (Figure 5, step II). The stereochemical outcome of the reaction depends on with which of its prochiral faces the enolate approaches and reacts with the catalyst-bound acyl group. In this context, it is important to note that the enolate is only able to approach the acyl group from one of both imidazoquinoline faces as the opposite face is completely shielded by a bulky 2,4,6-triisopropylphenyl group (see Figures 5 and 6). In the case of Λ-configured catalyst Λ-Ir5′, the approach and reaction of the enolate with its Si-face oriented toward the catalyst-bound acyl group is obviously favored, leading to predominant formation of the S-enantiomer in case of product 7i. We anticipate that the Steglich rearrangement of O-acylated azlactones (1 → 2) follows an analogous mechanism.11a,12h It is noteworthy that we verified with a crossover experiment that catalysis intermediates of type Λ-Ir5-Int interchange their

Table 3. Catalytic Enantioselective Addition of 2Cyanopyrrole to Aryl Alkyl Ketenesa

entry

R1

R2

yield (%)b

ee (%)b

product

c

Et Et i-Pr i-Pr i-Pr i-Pr i-Pr Cyclopentyl Cyclopentyl Cyclohexyl

Ph o-Tol Ph p-Cl-Ph p-OMe-Ph m-Tol m-Tol Ph Ph Ph

96 99 99 99 99 99 98 99 99 99

89 84 93 90 91 95 95 93 93 93

10a (S) 10b (S) 10c (S) 10d (S) 10e (S) 10f (S) 10f (S) 10g (S) 10g (S) 10h (S)

1 2c 3 4 5 6 7d 8 9d 10

Standard reaction conditions: ketene (8; 1.0 equiv; 100−130 μmol) and 2-cyanopyrrole (9; 2.0 equiv) with catalyst Λ-Ir5′ (2 mol %) in toluene (c = 12 mM) at 0 °C for 48 h. bIsolated yields; enantioselectivities determined from isolated products by chiral HPLC analysis. cSolvent: MTBE used instead of toluene. dScaled-up reactions, entry 7: 194 mg (1.11 mmol) of ketene 8f; entry 9: 209 mg (1.12 mmol) of ketene 8g. a

replacement of phenyl ethyl ketene (8a) by phenyl isopropyl ketene (8c) improved the ee to 93% (entry 3). Introduction of a methoxy group and a chloro substituent in the para-position of phenyl isopropyl ketene’s phenyl ring yielded corresponding products 10d (entry 4) and 10e (entry 5) with slightly diminished but still high selectivity (90% ee and 91% ee, respectively). The highest selectivity was found for m-tolyl isopropyl ketene (8f) giving product 10f in 95% ee (entry 6). Finally, phenyl isopropyl ketene’s isopropyl group was replaced by a cyclopentyl (substrate 8g; entry 8) and a cyclohexyl group (substrate 8h; entry 10), which yielded the corresponding products (10g,h) both in 93% ee. Again, we exemplarily scaled up two reactions with ketenes 8f and 8g as substrates by a factor of about 10 (∼1 mmol scale) and obtained also in the case of the scaled-up reactions excellent yields (98% and 99%) and identical selectivities as in the case of the corresponding standard scale examples (entries 7 and 9, compare with entries 6 and 8). Due to the sensitive nature of ketenes, all reactions from Table 3 were set up under strict exclusion of moisture and

Figure 5. Left: Anticipated mechanism for the Black rearrangements of O-acylated benzofuranones (6) catalyzed by Λ-Ir5′ using the example of substrate 6i. Right: Three-dimensional structure of catalysis intermediate analog Λ-Ir5-TfO obtained from a crystal structure of rac-Ir5-TfO (altered perspective to provide a clearer view; positional disorder and cocrystallized solvent molecules omitted for clarity; 60% probability of thermal ellipsoids).30 Preparation and crystallographic data of rac-Ir5-TfO are provided in the Supporting Information. 5156


Research Article

ACS Catalysis

Figure 6. From left to right: (1) Molecular surface of the complex cation of Λ-Ir5-TfO (Connolly surface, probe radius: 1.4 Å)41,42 with a view on the sterically accessible face of the N-acylated imidazoquinoline ligand. (2) Same structure as left, tilted by 90°, resulting in a view on the edge of the N-acylated imidazoquinoline ligand. The blue arrows in both representations point into the chiral molecular cavity (details in text). (3) Representations of both possible π-conjugated conformers and their calculated relative energies (details in text).

Figure 7. Quantum chemical modeling of the Black rearrangement’s stereogenic step with substrate 6h and a model50 of catalyst Λ-Ir5′. Top: Superposition of all four transition states42 and their Newman projections with their relative energies and their percentaged Curtin−Hammett-based contributions (C). Bottom: Main contributing transition state TS-Si-A (Connolly surface, probe radius 1.4 Å;41 angles of view identical with Figure 6)42 and structural formula of the simplified model50 of catalyst Λ-Ir5′ used for the quantum chemical calculations (removed fragments shown in orange).

cation in the absence of crystal effects at the PBE043-D344/ IMCP-SR145,46//PBE0-D3/SBKJC47 SMD48(butanol) level of theory (full computational details in SI). These calculations revealed that conformer A is indeed 5.3 kcal·mol−1 lower in energy than B, which affirms that conformer A is clearly preferred in solution as well.49 Having ruled out excessive conformational freedom of the N-bound acyl group, we were now able to shed light on the underlying manner of chiral recognition of intermediates of type Λ-Ir5-Int by the addition

electrostatically associated enolate counterions, which leads to partial formation of scrambled products (details in SI). Importantly, the N-bound acyl group of Λ-Ir5-TfO and ΛIr5-Int gives, in theory, rise to two coplanar conformers A and B (Figure 6, right), both allowing favorable conjugation with the neighboring π-system. However, only conformer A is found in the crystal structure obtained for rac-Ir5-TfO. To exclude a misleading crystal packing effect, we carried out a quantum chemical modeling of both conformations of the rac-Ir5-TfO 5157


Research Article

ACS Catalysis

Figure 8. Reasonable pathways for the reaction of aryl alkyl ketenes (8) with 2-cyanopyrrole (9) catalyzed by Λ-Ir5′ using the example of ketene substrate 8c.

states are similar in nature, so systematic errors in them largely cancel out.53 Representations of all transition states, their relative energies, and their Curtin−Hammett-based contributions (C) to the reaction are depicted in Figure 7. The calculations confirm our underlying idea of a rigid molecular cavity which must be entered by an enolate in order to react with the acyl group’s carbonyl moiety to finally form the product and correctly predict the stereochemical outcome. Furthermore, they indicate that the unique rigid framework of the catalyst with its constrained, well-defined active site should facilitate (computationally supported) rational design of new related catalysts and tailoring of catalysts (in silico) to certain substrates or the other way around.54 It is noteworthy that a comparison of Figures 6 and 7 clearly illustrates that a small migrating acyl group results in a wider molecular cavity, which understandably can result in diminished selectivity (e.g., 6h). A closer look at the four transition states also renders the poor result for substrate 6a consequential as the occurring all-planar, rigid, phenyl-substituted enolate is hardly able to enter the resulting small molecular cavity, which manifests in a sluggish reaction and a poor selectivity. For the asymmetric addition of 2-cyanopyrrole (9) to aryl alkyl ketenes (8), two pathways are reasonable, which are depicted in Figure 8 using the example of ketene substrate 8c.16c,22 The first reasonable pathway starts with proton transfer from 2-cyanopyrrole (9) to Λ-Ir5′ followed by addition of the generated pyrrolate to ketene 8c, which eventually results in the formation of Λ-Ir5-Int2, a salt consisting of an N-protonated +1 charged iridium(III) complex cation and its associated ketene enolate counteranion (Figure 8, step A). On the assumption that this pathway is operational, the observed stereochemical outcome originates from the circumstance that chiral Brønsted acid Λ-Ir5-Int2 preferentially protonates the prochiral ketene enolate from the Re-face, which leads to predominant formation of the S-enantiomer of product 10c (Figure 8, step B). In Figure 8, the E-enolate is arbitrarily drawn; however, in case this pathway is operational, the ketene enolate might be existent as an E/Z mixture.22 As discussed for the acyl migration reactions, the ketene enolate can only approach the protonated imidazoquinoline ligand from one face as the other one is sterically inaccessible. The second reasonable pathway starts with the addition of catalyst Λ-Ir5′ to ketene 8c, which results in the formation of zwitterionic iridium(III) complex Λ-Ir5-Int3, which might again exist as an E/Z mixture (Figure 8, step C, E-form

of a molecular surface to the complex cation of intermediate analog Λ-Ir5-TfO (Figure 6).41 As depicted in Figure 6, the target carbonyl moiety of the acyl group is nestled down in a chiral molecular pocket or molecular cavity (blue arrows in Figure 6), which a prochiral enolate has to enter in order to react with the acyl group. The “bottom” of this cavity consists of a part of the N-acylated imidazoquinoline ligand including the target electrophilic carbonyl moiety. And the “walls” of this cavity are shaped of, from left to right, a 2,4,6-triisopropylphenyl moiety, the cyclometalated benzoxazole ligand the latter is bound to, a 2,6-dimethylphenyl moiety being also part of this benzoxazole ligand, and finally the tail of the N-bound acyl group, which in the case of Λ-Ir5-TfO is the CMe2CCl3 fragment. Figure 6 clearly shows that the second face of the imidazoquinoline ligand is, in contrast, effectively shielded by another 2,4,6-triisopropylphenyl moiety. Keeping these considerations in the back of one’s mind, the catalyst screening from Figure 2 becomes more intelligible as Figure 6 clearly visualizes that an appropriate choice of R1 and R2 in the catalyst is essential for both high selectivity and activity. The slight advantage of Λ-Ir5 (Figure 2, R3 = Cl) over Λ-Ir4 (R3 = H) remains somewhat elusive; chlorine introduction might result in a subtle tilting of the overlying 2,4,6-triisopropylphenyl moiety (see Figure 6) or might just cause a subtle change of the electronic properties of the system. To confirm our mechanistic considerations and to establish the mechanism in greater detail, we performed a quantum chemical modeling of the Black rearrangement’s stereogenic step with substrate 6h and a simplified model50 of catalyst ΛIr5′ (Figure 7). In the course of these calculations, the four transition states depicted in Figure 7 were located using the approach recently developed by some of us51 at the PBE043D344/IMCP-SR145,46//PBE0-D3/SBKJC47 SMD48(butanol) level of theory (see SI for details). The PBE0 functional was chosen because it is known to be accurate for organic chemistry calculations43b and has recently been shown to be wellgrounded in theory,43c and the IMCP-SR1 basis set was chosen as it incorporates scalar-relativistic effects, which are important for heavy elements such as iridium in our case,46 and retains the correct nodal structures of atoms.45a The Curtin−Hammett principle52 was invoked to predict the stereochemical outcome of the rearrangement to product 7h, which was calculated to be 48% ee (R), which, in consideration of the applied approximations, is in a very good agreement with the observed 70% ee (R). Favorably, our calculations benefit from an error cancellation which rises from the fact that the studied transition 5158


Research Article

ACS Catalysis arbitrarily drawn).22 In case this pathway is operational, preferred proton transfer to the Re-face of the catalyst-bound ketene enolate from 2-cyanopyrrole (9) eventually leads to predominant formation of the S-enantiomer of product 10c (Figure 8, step D). Again, 2-cyanopyrrole (9) can only approach the imidazoquinoline ligand from one face as the other one is sterically inaccessible. In an attempt to clarify which of both cycles (Figure 8) is in operation, we made the following observations: When we mixed rac-Ir5′ in dry CD2Cl2 with excess p-chlorophenyl isopropyl ketene (8d) at room temperature, we only found 1H NMR signals belonging to 8d and rac-Ir5′ but no indication for the formation of a zwitterionic species like Λ-Ir5-Int3. This result is in agreement with observations made by Fu and coworkers for their DMAP-derived catalyst B (see Figure 1) and speaks, at first, against the right cycle.16c However, we also did not observe ion pair formation when we mixed rac-Ir5′ with excess 2-cyanopyrrole (9) under identical conditions, we only observed a weak hydrogen-bonding interaction between the Nbound proton of 9 and the nitrogen atom of rac-Ir5′ (details of both experiments in the Supporting Information). This is in contrast to observations made by Fu and co-workers for catalyst B, who report about ion pair formation for this experiment.16c On the basis of their NMR observations together with additional experimental data, primarily kinetic data, Fu and co-workers carefully assumed that the Brønsted base/acid mechanism should be in operation in the case of catalyst B.16c In contrast, a recent thorough computational study from Cheong and co-workers arrived at the conclusion that the nucleophilic pathway should be in operation in the case of B.22 For the reaction between phenyl methyl ketene and 2cyanopyrrole (9) catalyzed by B, Cheong and co-workers revealed that (1) the decisive rate-determining step (RDS) of the nucleophilic pathway is favored by 8.3 kcal·mol−1 over the RDS of the Brønsted base/acid pathway; (2) the NMRobserved ion pair formed between 9 and B is merely the unproductive resting state of the catalyst; (3) the NMR-based nonobservation of a zwitterionic intermediate formed between a ketene and B is consequential due to its endergonic nature (see ref 16c and 22 for a detailed discussion).16c,22 Being aware of the different structures and properties of Fu’s catalyst B and our catalyst Λ/Δ-Ir5′, we at the moment carefully assume, in consideration of the recent calculations from Cheong and coworkers for B, that for catalyst Λ/Δ-Ir5′ the nucleophilic catalysis pathway is in operation as well (Figure 8, right cycle). A study to clarify the situation for catalyst Λ/Δ-Ir5′ is under way in our laboratory. At least we can conclude that the lone pair of the imidazoquinoline’s nitrogen atom is indispensable for this reaction either way, as mixing of N-methylated catalyst (Λ-Ir5-Me) with ketene 8c and 2-cyanopyrrole (9) results in no formation of product 10c (preparation of Λ-Ir5-Me and experimental details in the Supporting Information).29

related Black rearrangement of O-acylated benzofuranones to C-acylated benzofuranones going along with the formation of an all-carbon quaternary stereocenter (up to 99% yield and 94% ee), and finally the addition of 2-cyanopyrrole to aryl alkyl ketenes giving rise to α-chiral N-acyl pyrroles (up to 99% yield and 95% ee), a reaction for which catalyst Λ/Δ-Ir5′ is, to the best of our knowledge, after Fu’s planar-chiral catalyst B the second reported chiral catalyst so far. Notably, we were able to carry out all these transformations with only 0.5−2 mol % of precatalyst Λ/Δ-Ir5 or catalyst Λ/Δ-Ir5′, partially at temperatures as low as −30 °C, which underlines the catalyst’s high activity. Both the Λ- and Δ-enantiomer of the catalyst are accessible in enantiopure form (>99% ee) by simple flash chromatography of intermediary diastereomers. We have demonstrated that the catalyst can be recovered in high yield after reaction completion without deterioration of its enantiopurity and reused again. With crystal structures of active catalyst rac-Ir5′ and catalysis intermediate analog racIr5-TfO (Figures 3 and 5) as well as with analysis of the active site accessibility of the latter (Figure 6), we have shed light on the catalyst’s manner of chiral recognition, and in the case of the Black rearrangements, we have confirmed our mechanistic suggestions with explicit quantum chemical calculations (Figure 7). Whether the reaction between 2-cyanopyrrole and aryl alkyl ketenes follows a Brønsted base/acid pathway or a nucleophilic pathway (Figure 8) remains somewhat elusive for now, and a separate study to clarify the situation is under way in our laboratory. For the time being, we assume that a nucleophilic catalysis pathway is active. Future experimental studies with Λ/ Δ-Ir5′ or derivatives thereof will aim at identifying other asymmetric transformations being suitable for this type of chiral nucleophilic catalyst and will also seek to shed light on the iridium center’s electronic influence on the imidazoquinoline ligand and hence its influence on the catalytic activity of this ligand and the catalyst as a whole.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscatal.7b01296. Preparation and characterization of the iridium(III) complexes and catalysis substrates, conditions screenings, experimental procedures for the catalyzed reactions (Figure 1, Tables 1−3), characterization of the catalysis products, absolute stereochemistry assignments, mechanistic experiments, quantum chemical calculation details, CD spectra of iridium(III) complexes, 1H and 13C NMR spectra, and HPLC traces of iridium(III) complexes and catalysis products (PDF) Crystallographic data for rac-Ir5′ (CCDC 1536878) (CIF) Crystallographic data for rac-Ir5-TfO (CCDC 1536879) (CIF) Crystallographic data for [IrIII(C^N)2(μ-Cl)]2 (C^N is cyclometalated 3; CCDC 1536877) (CIF) Crystallographic data for (S)-7i (CCDC 1545358) (CIF)

■

CONCLUSION In conclusion, we have developed an air- and moisture-stable nucleophilic catalyst (Λ- and Δ-Ir5′) featuring a remote, iridium(III)-located octahedral stereocenter as its exclusive element of chirality and an imidazoquinoline ligand serving as the catalyst’s nucleophilic site. The versatility of this catalyst was demonstrated for three different enantioselective transformations: the Steglich rearrangement of O-acylated azlactones to C-acylated azlactones going along with the formation of a quaternary stereocenter (up to 99% yield and 96% ee), the

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Website: http:// www.uni-marburg.de/fb15/ag-meggers. 5159


Research Article

ACS Catalysis ORCID

Chem. Soc. 2012, 134, 17605−17612. (g) Bumbu, V. D.; Yang, X.; Birman, V. B. Org. Lett. 2013, 15, 2790−2793. (6) Examples for kinetic resolutions with imidazole-derived catalysts: (a) Miller, S. J.; Copeland, G. T.; Papaioannou, N.; Horstmann, T. E.; Ruel, E. M. J. Am. Chem. Soc. 1998, 120, 1629−1630. (b) Jarvo, E. R.; Copeland, G. T.; Papaioannou, N.; Bonitatebus, P. J., Jr.; Miller, S. J. J. Am. Chem. Soc. 1999, 121, 11638−11643. (c) Ishihara, K.; Kosugi, Y.; Akakura, M. J. Am. Chem. Soc. 2004, 126, 12212−12213. (d) Müller, C. E.; Wanka, L.; Jewell, K.; Schreiner, P. R. Angew. Chem., Int. Ed. 2008, 47, 6180−6183. (7) Notte, G. T.; Sammakia, T. J. Am. Chem. Soc. 2006, 128, 4230− 4231. (8) (a) Vedejs, E.; Daugulis, O.; Diver, S. T. J. Org. Chem. 1996, 61, 430−431. (b) Vedejs, E.; Daugulis, O. J. Am. Chem. Soc. 1999, 121, 5813−5814. (c) Vedejs, E.; Daugulis, O.; MacKay, J. A.; Rozners, E. Synlett 2001, 2001, 1499−1505. (d) Vedejs, E.; Daugulis, O. J. Am. Chem. Soc. 2003, 125, 4166−4173. (9) Recent review articles about chiral amidine- and isothioureaderived catalysts: (a) Merad, J.; Pons, J. M.; Chuzel, O.; Bressy, C. Eur. J. Org. Chem. 2016, 2016, 5589−5610. (b) Birman, V. B. Aldrichimica Acta 2016, 49, 23−33. (10) (a) Steglich, W.; Höfle, G. Angew. Chem., Int. Ed. Engl. 1968, 7, 61. (b) Steglich, W.; Höfle, G. Chem. Ber. 1969, 102, 883−898. (c) Steglich, W.; Hö fle, G. Chem. Ber. 1969, 102, 899−903. (d) Steglich, W.; Höfle, G. Tetrahedron Lett. 1970, 11, 4727−4730. (11) Steglich-type rearrangements with Fu’s planar-chiral DMAPderived catalysts: (a) Ruble, J. C.; Fu, G. C. J. Am. Chem. Soc. 1998, 120, 11532−11533. (b) Hills, I. D.; Fu, G. C. Angew. Chem., Int. Ed. 2003, 42, 3921−3924. (12) Steglich-type rearrangements with other DMAP-derived catalysts: (a) Shaw, S. A.; Aleman, P.; Vedejs, E. J. Am. Chem. Soc. 2003, 125, 13368−13369. (b) Shaw, S. A.; Aleman, P.; Christy, J.; Kampf, J. W.; Va, P.; Vedejs, E. J. Am. Chem. Soc. 2006, 128, 925−934. (c) Nguyen, H. V.; Butler, D. C. D.; Richards, C. J. Org. Lett. 2006, 8, 769−772. (d) De, C. K.; Mittal, N.; Seidel, D. J. Am. Chem. Soc. 2011, 133, 16802−16805. (e) Poisson, T.; Oudeyer, S.; Levacher, V. Tetrahedron Lett. 2012, 53, 3284−3287. (f) Mandai, H.; Fujiwara, T.; Noda, K.; Fujii, K.; Mitsudo, K.; Korenaga, T.; Suga, S. Org. Lett. 2015, 17, 4436−4439. (g) Mandai, H.; Fujii, K.; Yasuhara, H.; Abe, K.; Mitsudo, K.; Korenaga, T.; Suga, S. Nat. Commun. 2016, 7, 11297. (h) Chen, C.-T.; Tsai, C.-C.; Tsou, P.-K.; Huang, G.-T.; Yu, C.-H. Chem. Sci. 2017, 8, 524−529. (13) Steglich-type rearrangements with isothiourea-derived catalysts: (a) Joannesse, C.; Johnston, C. P.; Concellón, C.; Simal, C.; Philp, D.; Smith, A. D. Angew. Chem., Int. Ed. 2009, 48, 8914−8918. (b) Viswambharan, B.; Okimura, T.; Suzuki, S.; Okamoto, S. J. Org. Chem. 2011, 76, 6678−6685. (14) Steglich-type rearrangements with imidazole-derived catalysts: (a) Zhang, Z.; Xie, F.; Jia, J.; Zhang, W. J. Am. Chem. Soc. 2010, 132, 15939−15941. (b) Wang, M.; Zhang, Z.; Liu, S.; Xie, F.; Zhang, W. Chem. Commun. 2014, 50, 1227−1230. (15) Uraguchi, D.; Koshimoto, K.; Miyake, S.; Ooi, T. Angew. Chem., Int. Ed. 2010, 49, 5567−5569. (16) Ketene reactions catalyzed by Fu’s planar-chiral DMAP-type catalysts: (a) Hodous, B. L.; Ruble, J. C.; Fu, G. C. J. Am. Chem. Soc. 1999, 121, 2637−2638. (b) Hodous, B. L.; Fu, G. C. J. Am. Chem. Soc. 2002, 124, 1578−1579. (c) Hodous, B. L.; Fu, G. C. J. Am. Chem. Soc. 2002, 124, 10006−10007. (d) Wilson, J. E.; Fu, G. C. Angew. Chem., Int. Ed. 2004, 43, 6358−6360. (e) Wiskur, S. L.; Fu, G. C. J. Am. Chem. Soc. 2005, 127, 6176−6177. (f) Lee, E. C.; Hodous, B. L.; Bergin, E.; Shih, C.; Fu, G. C. J. Am. Chem. Soc. 2005, 127, 11586−11587. (g) Schaefer, C.; Fu, G. C. Angew. Chem., Int. Ed. 2005, 44, 4606− 4608. (h) Lee, E. C.; McCauley, K. M.; Fu, G. C. Angew. Chem., Int. Ed. 2007, 46, 977−979. (i) Dai, X.; Nakai, T.; Romero, J. A. C.; Fu, G. C. Angew. Chem., Int. Ed. 2007, 46, 4367. (j) Berlin, J. M.; Fu, G. C. Angew. Chem., Int. Ed. 2008, 47, 7048−7050. (17) Ketene reactions catalyzed by chiral NHC-derived catalysts: (a) Zhang, Y.-R.; He, L.; Wu, X.; Shao, P.-L.; Ye, S. Org. Lett. 2008, 10, 277−280. (b) Wang, X.-N.; Lv, H.; Huang, X.-L.; Ye, S. Org. Biomol.

Michael G. Medvedev: 0000-0001-7070-4052 Eric Meggers: 0000-0002-8851-7623 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Deutsche Forschungsgemeinschaft (ME 1805/11-1). We thank Vladimir A. Larionov and Alexander F. Smolyakov for establishing the contact to M.G.M., who performed the quantum chemical calculations. M.G.M. is grateful to the Russian Science Foundation grant #17-13-01526 for financial support and acknowledges the computational resources on which the quantum chemical calculations were performed: the IBM Blue Gene/P supercomputer at Moscow State University’s (MSU) Faculty of Computational Mathematics and Cybernetics, the Lomonosov supercomputer55 at the Supercomputing Center of MSU, and the HPC2 supercomputer at the federal center for collective usage at NRC “Kurchatov Institute” (http://computing.kiae.ru/).

■

REFERENCES

(1) Vedejs, E.; Chen, X. J. Am. Chem. Soc. 1996, 118, 1809−1810. (2) Kinetic resolutions with Fu’s planar-chiral DMAP-derived catalysts: (a) Ruble, J. C.; Fu, G. C. J. Org. Chem. 1996, 61, 7230− 7231. (b) Ruble, J. C.; Latham, H. A.; Fu, G. C. J. Am. Chem. Soc. 1997, 119, 1492−1493. (c) Liang, J.; Ruble, J. C.; Fu, G. C. J. Org. Chem. 1998, 63, 3154−3155. (d) Tao, B.; Ruble, J. C.; Hoic, D. A.; Fu, G. C. J. Am. Chem. Soc. 1999, 121, 5091−5092. (e) Lee, S. Y.; Murphy, J. M.; Ukai, A.; Fu, G. C. J. Am. Chem. Soc. 2012, 134, 15149−15153. (f) Díaz-Á lvarez, A. E.; Mesas-Sánchez, L.; Dinér, P. Angew. Chem., Int. Ed. 2013, 52, 502−504. (3) Kinetic resolutions with other DMAP-derived catalysts: (a) Kawabata, T.; Nagato, M.; Takasu, K.; Fuji, K. J. Am. Chem. Soc. 1997, 119, 3169−3170. (b) Spivey, A. C.; Fekner, T.; Spey, S. E. J. Org. Chem. 2000, 65, 3154−3159. (c) Spivey, A. C.; Zhu, F.; Mitchell, M. B.; Davey, S. G.; Jarvest, R. L. J. Org. Chem. 2003, 68, 7379−7385. (d) Priem, G.; Pelotier, B.; Macdonald, S. J. F.; Anson, M. S.; Campbell, I. B. J. Org. Chem. 2003, 68, 3844−3848. (e) Spivey, A. C.; Leese, D. P.; Zhu, F.; Davey, S. G.; Jarvest, R. L. Tetrahedron 2004, 60, 4513−4525. (f) Spivey, A. C.; Arseniyadis, S.; Fekner, T.; Maddaford, A.; Leese, D. P. Tetrahedron 2006, 62, 295−301. (g) Nguyen, H. V.; Motevalli, M.; Richards, C. J. Synlett 2007, 2007, 725−728. (h) Crittall, M. R.; Rzepa, H. S.; Carbery, D. R. Org. Lett. 2011, 13, 1250−1253. (i) Crittall, M. R.; Fairhurst, N. W. G.; Carbery, D. R. Chem. Commun. 2012, 48, 11181−11183. (4) Kinetic resolutions with amidine-derived catalysts: (a) Birman, V. B.; Uffman, E. W.; Jiang, H.; Li, X.; Kilbane, C. J. J. Am. Chem. Soc. 2004, 126, 12226−12227. (b) Birman, V. B.; Jiang, H. Org. Lett. 2005, 7, 3445−3447. (c) Birman, V. B.; Li, X.; Jiang, H.; Uffman, E. W. Tetrahedron 2006, 62, 285−294. (d) Hu, B.; Meng, M.; Wang, Z.; Du, W.; Fossey, J. S.; Hu, X.; Deng, W.-P. J. Am. Chem. Soc. 2010, 132, 17041−17044. (e) Li, X.; Jiang, H.; Uffman, E. W.; Guo, L.; Zhang, Y.; Yang, X.; Birman, V. B. J. Org. Chem. 2012, 77, 1722−1737. (f) Jiang, S.-S.; Xu, Q.-C.; Zhu, M.-Y.; Yu, X.; Deng, W.-P. J. Org. Chem. 2015, 80, 3159−3169. (g) Jiang, S.-S.; Gu, B.-Q.; Zhu, M.-Y.; Yu, X.; Deng, W.-P. Tetrahedron 2015, 71, 1187−1191. (5) Kinetic resolutions with isothiourea-derived catalysts: (a) Birman, V. B.; Li, X. Org. Lett. 2006, 8, 1351−1354. (b) Birman, V. B.; Jiang, H.; Li, X.; Guo, L.; Uffman, E. W. J. Am. Chem. Soc. 2006, 128, 6536− 6537. (c) Birman, V. B.; Guo, L. Org. Lett. 2006, 8, 4859−4861. (d) Yang, X.; Birman, V. B. Adv. Synth. Catal. 2009, 351, 2301−2304. (e) Bumbu, V. D.; Birman, V. B. J. Am. Chem. Soc. 2011, 133, 13902− 13905. (f) Yang, X.; Bumbu, V. D.; Liu, P.; Li, X.; Jiang, H.; Uffman, E. W.; Guo, L.; Zhang, W.; Jiang, X.; Houk, K. N.; Birman, V. D. J. Am. 5160


Research Article

ACS Catalysis Chem. 2009, 7, 346−350. (c) Douglas, J.; Ling, K. B.; Concellón, C.; Churchill, G.; Slawin, A. M. Z.; Smith, A. D. Eur. J. Org. Chem. 2010, 2010, 5863−5869. (d) Jian, T.-Y.; Shao, P.-L.; Ye, S. Chem. Commun. 2011, 47, 2381−2383. (18) Books and review articles dealing with nucleophilic catalysis/ Lewis base catalysis (also in a broader sense): (a) France, S.; Guerin, D. J.; Miller, S. J.; Lectka, T. Chem. Rev. 2003, 103, 2985−3012. (b) Spivey, A. C.; Arseniyadis, S. Angew. Chem., Int. Ed. 2004, 43, 5436−5441. (c) Methot, J. L.; Roush, W. R. Adv. Synth. Catal. 2004, 346, 1035−1050. (d) Wurz, R. P. Chem. Rev. 2007, 107, 5570−5595. (e) Gaunt, M. J.; Johansson, C. C. C. Chem. Rev. 2007, 107, 5596− 5605. (f) Denmark, S. E.; Beutner, G. L. Angew. Chem., Int. Ed. 2008, 47, 1560−1638. (g) Spivey, A. C.; Arseniyadis, S. Top. Curr. Chem. 2010, 291, 233−280. (h) Müller, C. E.; Schreiner, P. R. Angew. Chem., Int. Ed. 2011, 50, 6012−6042. (i) Lewis Base Catalysis in Organic Synthesis, 1st ed.; Vedejs, E., Denmark, S. E., Eds.; Wiley-VCH: Weinheim, 2016. (19) (a) Fu, G. C. Acc. Chem. Res. 2000, 33, 412−420. (b) Fu, G. C. Acc. Chem. Res. 2004, 37, 542−547. (20) Articles which focus on the high catalytic activity of DMAP and its derivatives (in addition to ref 18b): (a) Litvinenko, L. M.; Kirichenko, A. I. Dokl. Chem. 1967, 763−766. Litvinenko, L. M.; Kirichenko, A. I. Dokl. Akad. Nauk SSSR, Ser. Khim. 1967, 176, 97− 100. (b) Steglich, W.; Höfle, G. Angew. Chem., Int. Ed. Engl. 1969, 8, 981. (c) Grondal, C. Synlett 2003, 10, 1568−1569. (21) Garrett, C. E.; Fu, G. C. J. Am. Chem. Soc. 1998, 120, 7479− 7483. (22) (a) Pattawong, O.; Mustard, T. J. L.; Johnston, R. C.; Cheong, P. H.-Y. Angew. Chem., Int. Ed. 2013, 52, 1420−1423. (b) Walden, D. M.; Ogba, O. M.; Johnston, R. C.; Cheong, P. H.-Y. Acc. Chem. Res. 2016, 49, 1279−1291. (23) Check, C. T.; Jang, K. P.; Schwamb, C. B.; Wong, A. S.; Wang, M. H.; Scheidt, K. A. Angew. Chem., Int. Ed. 2015, 54, 4264−4268. (24) (a) Belokon, Y. N.; Bulychev, A. G.; Maleev, V. I.; North, M.; Malfanov, I. L.; Ikonnikov, N. S. Mendeleev Commun. 2004, 14, 249− 250. (b) Hamelin, O.; Rimboud, M.; Pécaut, J.; Fontecave, M. Inorg. Chem. 2007, 46, 5354−5360. (c) Ganzmann, C.; Gladysz, J. A. Chem. Eur. J. 2008, 14, 5397−5400. (d) Kurono, N.; Arai, K.; Uemura, M.; Ohkuma, T. Angew. Chem., Int. Ed. 2008, 47, 6643−6646. (e) Kurono, N.; Nii, N.; Sakaguchi, Y.; Uemura, M.; Ohkuma, T. Angew. Chem., Int. Ed. 2011, 50, 5541−5544. (f) Belokon, Y. N.; Maleev, V. I.; North, M.; Larionov, V. A.; Savel’yeva, T. F.; Nijland, A.; Nelyubina, Y. V. ACS Catal. 2013, 3, 1951−1955. (g) Sakaguchi, Y.; Kurono, N.; Yamauchi, K.; Ohkuma, T. Org. Lett. 2014, 16, 808−811. (h) Maleev, V. I.; North, M.; Larionov, V. A.; Fedyanin, I. V.; Savel’yeva, T. F.; Moscalenko, M. A.; Smolyakov, A. F.; Belokon, Y. N. Adv. Synth. Catal. 2014, 356, 1803−1810. (i) Lewis, K. G.; Ghosh, S. K.; Bhuvanesh, N.; Gladysz, J. A. ACS Cent. Sci. 2015, 1, 50−56. (j) Kumar, A.; Ghosh, S. K.; Gladysz, J. A. Org. Lett. 2016, 18, 760−763. (k) Ghosh, S. K.; Ganzmann, C.; Bhuvanesh, N.; Gladysz, J. A. Angew. Chem., Int. Ed. 2016, 55, 4356−4360. (l) Ehnbom, A.; Ghosh, S. K.; Lewis, K. G.; Gladysz, J. A. Chem. Soc. Rev. 2016, 45, 6799−6811. (25) (a) Chen, L.-A.; Xu, W.; Huang, B.; Ma, J.; Wang, L.; Xi, J.; Harms, K.; Gong, L.; Meggers, E. J. Am. Chem. Soc. 2013, 135, 10598− 10601. (b) Chen, L.-A.; Tang, X.; Xi, J.; Xu, W.; Gong, L.; Meggers, E. Angew. Chem., Int. Ed. 2013, 52, 14021−14025. (c) Xu, W.; Arieno, M.; Löw, H.; Huang, K.; Xie, X.; Cruchter, T.; Ma, Q.; Xi, J.; Huang, B.; Wiest, O.; Gong, L.; Meggers, E. J. Am. Chem. Soc. 2016, 138, 8774−8780. (d) Xu, W.; Shen, X.; Ma, Q.; Gong, L.; Meggers, E. ACS Catal. 2016, 6, 7641−7646. (26) (a) Ma, J.; Ding, X.; Hu, Y.; Huang, Y.; Gong, L.; Meggers, E. Nat. Commun. 2014, 5, 4531. (b) Ding, X.; Lin, H.; Gong, L.; Meggers, E. Asian J. Org. Chem. 2015, 4, 434−437. (c) Hu, Y.; Zhou, Z.; Gong, L.; Meggers, E. Org. Chem. Front. 2015, 2, 968−972. (d) Ding, X.; Tian, C.; Hu, Y.; Gong, L.; Meggers, E. Eur. J. Org. Chem. 2016, 2016, 887−890. (27) Huo, H.; Fu, C.; Wang, C.; Harms, K.; Meggers, E. Chem. Commun. 2014, 50, 10409−10411.

(28) We would like to note here that we have previously reported a chiral-at-metal enamine catalyst featuring a secondary amine (ref 27). In a broader sense, that catalyst can also be referred to as “nucleophilic.” However, catalysts involving enamine or iminium ion formation should be clearly distinguished from “classic” nucleophilic catalysts, such as DMAP, as they catalyze different reactions. See refs 18a and 18f for a detailed discussion. (29) In the case where cesium carbonate is absent, precatalyst Ir5 is not catalytically active at all, and in the case where N-methylated Ir5 (= Ir5-Me; details in SI) is employed instead of Ir5 in the presence of cesium carbonate, no reaction is observed as well. Detailed evidence for the acid/base behavior of Ir5 and Ir5′ and the necessity for a deprotonated imidaqozuinoline ligand for catalytic activity are provided in detail in the Supporting Information. (30) (a) Mercury, version 3.7 RC1; The Cambridge Crystallographic Data Centre: Cambridge, UK, 2015. (b) POV-Ray for Windows, version 3.7.0.mscv10.win64; Persistance of Vision Pty. Ltd.: Williamstown, Victoria, Australia, 2013. (31) Helms, M.; Lin, Z.; Gong, L.; Harms, K.; Meggers, E. Eur. J. Inorg. Chem. 2013, 2013, 4164−4172. (32) (a) Bolm, C.; Weickhardt, K.; Zehnder, M.; Ranff, T. Chem. Ber. 1991, 124, 1173−1180. (b) Takemoto, Y.; Kuraoka, S.; Hamaue, N.; Aoe, K.; Hiramatsu, H.; Iwata, C. Tetrahedron 1996, 52, 14177−14188. (33) In principle, (S)-4 can be replaced with 5 using a one-pot protocol with NH4PF6 as a weak acid in MeCN (see ref 25a for details). However, we faced lower yields and problems with the purification of Λ/Δ-Ir5 using this protocol. Therefore, we first formed an intermediary bisacetonitrile complex before adding ligand 5 (details in SI). (34) Chromatographic purification of the HPF6 salt (Λ/Δ-Ir5) was found to be considerably easier than purification of the corresponding TfOH salt (the latter strongly tails on silica gel). (35) Piperidine, polymer-bound, 200−400 mesh, extent of labeling: 3.0−4.0 mmol/g loading, 1% cross-linked with divinylbenzene; SigmaAldrich product number: 494615. (36) (a) Black, T. H.; Arrivo, S. M.; Schumm, J. S.; Knobeloch, J. M. J. Chem. Soc., Chem. Commun. 1986, 1524−1525. (b) Black, T. H.; Arrivo, S. M.; Schumm, J. S.; Knobeloch, J. M. J. Org. Chem. 1987, 52, 5425−5430. (37) It is noteworthy that we had also shortly investigated the related Steglich-type acyl migration reaction N,O-diacylated oxindole S31 → C,N-diacylated oxindole S32 with Λ-Ir5′ as a catalyst. In a first, unoptimized attempt, we had obtained S32 in 96% yield and 59% ee (details in SI). Simultaneously, we had obtained promising results with ketenes as substrates (see Table 3), so we discontinued working on this reaction. (38) MTBE was found to give slightly better ee’s for the screened ethyl substituted ketenes (details in SI). (39) It is noteworthy that we have also attempted this type of reaction with pyrrole, 2-acetylpyrrole, indole, and carbazole. In contrast to 2-cyanopyrrole (9), we did not detect any conversion with pyrrole, indole, and carbazole, and in the case of 2-acetylpyrrole the reaction was very sluggish, so we finally stuck to 2-cyanopyrrole (9). (40) In a thorough literature search with CAS SciFinder and Elsevier Reaxys, we could only find the mentioned publication from Fu and coworkers (ref 16c; we searched for asymmetric reactions between NHheterocycles and ketenes yielding the corresponding addition products). (41) Connolly, M. L. Science 1983, 221, 709−713. (42) PyMOL, version 0.99rc6; DeLano Scientific LLC: Palo Alto, CA, 2006. (43) (a) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158− 6170. (b) Peverati, R.; Truhlar, D. G. Philos. Trans. R. Soc., A 2014, 372, 20120476. (c) Medvedev, M. G.; Bushmarinov, I. S.; Sun, J.; Perdew, J. P.; Lyssenko, K. A. Science 2017, 355, 49−52. (44) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys. 2010, 132, 154104. 5161


Research Article

ACS Catalysis (45) (a) Lovallo, C. C.; Klobukowski, M. J. Comput. Chem. 2003, 24, 1009−1015. (b) Lovallo, C. C.; Klobukowski, M. J. Comput. Chem. 2004, 25, 1206−1213. (46) Smith, A. R. G.; Riley, M. J.; Lo, S.-C.; Burn, P. L.; Gentle, I. R.; Powell, B. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 83, 041105. (47) (a) Stevens, W. J.; Basch, H.; Krauss, M. J. Chem. Phys. 1984, 81, 6026−6033. (b) Stevens, W. J.; Krauss, M.; Basch, H.; Jasien, P. G. Can. J. Chem. 1992, 70, 612−630. (48) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2009, 113, 6378−6396. (49) A similar conformational preference has been observed in related systems, see refs 2d, 11b, 13a, 21, 22, and: Velázquez-Ponce, M.; Salgado-Zamora, M.; Jiménez-Vázquez, H. A.; Campos-Aldrete, M. E.; Jiménez, R.; Cervantes, H.; Hadda, T. B. Chem. Cent. J. 2013, 7, 20. (50) Initially, we intended to perform the quantum chemical calculations with entire catalyst Λ-Ir5′ and substrate 6i. However, the computational complexity of this system forced us to switch over to the simplified catalyst model depicted in Figure 7 and substrate 6h. Importantly, in this catalyst model, the shape of the “active center” remains entirely unchanged (details in SI). (51) Medvedev, M. G.; Panova, M. V.; Chilov, G. G.; Bushmarinov, I. S.; Novikov, F. N.; Stroganov, O. V.; Zeifman, A. A.; Svitanko, I. V. Mendeleev Commun. 2017, 27, 224−227. (52) (a) Seeman, J. I. J. Chem. Educ. 1986, 63, 42−48. For a deeper discussion, see: (b) Seeman, J. I. Chem. Rev. 1983, 83, 83−134. (53) (a) Medford, A. J.; Wellendorff, J.; Vojvodic, A.; Studt, F.; AbildPedersen, F.; Jacobsen, K. W.; Bligaard, T.; Nørskov, J. K. Science 2014, 345, 197−200. (b) Medvedev, M. G.; Zeifman, A. A.; Novikov, F. N.; Bushmarinov, I. S.; Stroganov, O. V.; Titov, I. Y.; Chilov, G. G.; Svitanko, I. V. J. Am. Chem. Soc. 2017, 139, 3942−3945. (54) Quite recently, we demonstrated for hydrogen-bonding catalysts that rigid octahedral metal-templated catalysts are well-suited for rational improvements due to their constrained geometry (see refs 25c and 25d for details). (55) Sadovnichy, V.; Tikhonravov, A.; Voevodin, V.; Opanasenko, V. “Lomonosov”: Supercomputing at Moscow State University. In Contemporary High Performance Computing: From Petascale toward Exascale; Vetter, J. S., Eds.; Chapman & Hall/CRC Computational Science; CRC Press: Boca Raton, FL, 2013; pp 283−307.

5162


Asymmetric Nucleophilic Catalysis with an Octahedral Chiral-at-Metal

Recommend Documents