Dynamic Kinetic Resolution of Biaryl Atropisomers via Peptide-Catalyzed Asymmetric Bromination

See allHide authors and affiliations

Science  04 Jun 2010:
Vol. 328, Issue 5983, pp. 1251-1255
DOI: 10.1126/science.1188403


Despite the widespread use of axially chiral, or atropisomeric, biaryl ligands in modern synthesis and the occurrence of numerous natural products exhibiting axial chirality, few catalytic methods have emerged for the direct asymmetric preparation of this compound class. Here, we present a tripeptide-derived small-molecule catalyst for the dynamic kinetic resolution of racemic biaryl substrates. The reaction proceeds via an atropisomer-selective electrophilic aromatic substitution reaction using simple bromination reagents. The result is an enantioselective synthesis that delivers chiral nonracemic biaryl compounds with excellent optical purity and good isolated chemical yields (in most cases a >95:5 enantiomer ratio and isolated yields of 65 to 87%). A mechanistic model is advanced that accounts for the basis of selectivity observed.

The stereochemical implications of hindered rotation in nonplanar molecules, termed atropisomerism, have intrigued chemists for at least 89 years (1, 2). Atropisomeric compounds exhibit an axis of chirality (Fig. 1A), rather than a stereogenic atom, such as an sp3-hybridized carbon with four distinct substituents (Fig. 1B). The capacity of the single bond between two aromatic rings to freely rotate is the basis of racemization for many atropisomeric compounds (3). Yet in naturally occurring compounds, atropisomeric molecules are often found in single isomeric form because of substituents on aryl rings that raise the barriers to racemization. Also, the localization of aryl rings within multicyclic ring systems can constrain single bond rotations, preventing isomerization and the observation of mixtures (4, 5). These properties no doubt contribute to the remarkable structures and functions of numerous biologically active compounds that contain single atropisomers as part of their structure (6). Perhaps the glycopeptide antibiotic vancomycin is the signature bioactive natural product of this type (Fig. 1C). The chiral ligand BINAP [2,2′-bis(diphenylphosphino)-1,1′-binaphthyl] (Fig. 1D), a venerable ligand for enantioselective catalysis, may be the best-known designed example (7).

Fig. 1

(A) Biaryl atropisomers are isolable if the barrier to rotation about the single bond linking the rings is high. The enantiomers interconvert via racemization if the barrier is low. R groups are listed in Fig. 2. (B) sp3-hybridized carbon atoms with four different substituents form generally stable enantiomers. (C) Vancomycin is a natural product that exists as a single atropisomer. Me, methyl group. (D) BINAP [2,2′-bis(diphenylphosphino)-1,1′-binaphthyl] is a widely used chiral ligand. Ph, phenyl group.

Fig. 2

Initial screen of catalysis for asymmetric bromination of 1. Isolated yields correspond to the anisole methyl ester after treatment with 4 equivalents of trimethylsilyl (TMS)–diazomethane (2 M in Et2O) for 15 min in 0.2 M toluene:MeOH (3:1). This work-up assists in purification and e.r. determination by chiral high performance liquid chromatography (HPLC). See (30) for details. Enantiometer ratios were determined by chiral HPLC.

Despite the prevalence and importance of atropisomerism in organic structures, the field of asymmetric catalysis has not yet recorded extensive success in the development of catalysts that control this stereochemical feature. In several excellent cases, asymmetric metal-catalyzed formation of the biaryl bond has resulted in high enantioselectivity for certain substrates (812). Pioneering work by Bringmann (13) and Clayden (14) focused instead on stoichiometrically inducing the selective reaction of a single atropisomer in a dynamic mixture of freely rotating, rapidly racemizing biaryls. Catalytic reactions of this nature are presently rare, and only modest atropisomer selectivity has been observed (15). Here, we report a simple chiral catalyst that mediates highly enantioselective electrophilic aromatic substitution reactions, thereby promoting atropisomer-selective functionalization of rapidly racemizing biaryl compounds.

We began our project by identifying a substrate class that would exhibit axial chirality with atropisomers that might rapidly interconvert. Compound 1 fulfills this criterion, with a barrier to atropisomer interconversion that may be estimated to be ~7 kcal/mol (16). Upon reaction with N-bromosuccinimide (NBS) in the presence of catalytic N-methylpyrrolidine [10 mole % (mol %) 2], compound 1 was triply brominated to yield 3 (Scheme 1), which exhibits much more restricted rotation about the bond connecting the aromatic rings. In fact, the barrier to racemization for compound 3 may be estimated to be ~30 kcal/mol (17). This barrier is sufficiently high so that in principle, if an enantioselective catalyst could be found for the conversion of 1 to 3, isolation of optically enriched, nonracemizing products could be obtained at room temperature. Notably, singly or doubly brominated intermediates were not observed under the reaction conditions that we used.

Scheme 1

We then turned our attention to the issue of enantioselectivity. Our choice of catalyst for these studies followed an empirical path, given the paucity of precedents for this type of asymmetric catalytic reaction. Nonetheless, we were guided by the principle that Lewis base catalysis of electrophilic bromination reactions is possible (1823). Furthermore, we were driven by the recognition that simple peptide-based catalysts have substantial capacity to mediate a wide variety of mechanistically diverse enantioselective transformations (24, 25). Moreover, we had recently shown that peptide-based catalysts exhibit remarkable enantioselectivity in the derivatization of unusual aromatic compounds (26).

We chose peptide 4 as a starting place for catalyst screening (Fig. 2), with the rationale that the chiral environment of the peptide β turn (27) could lead to atropisomer selection during the electrophilic aromatic substitution reaction. β−N,N-Dimethylamino alanine (Dmaa) was introduced as the N-terminal residue with hope that it might function as an additional basic site, favoring catalyst-substrate contacts. This first-generation catalyst provided encouraging results, with methyl ester (1a) giving the product in nonracemic form (entry 1). Swapping 1a for benzyl amide (1b) improved enantioselectivity to nearly 2:1 (entry 2). Nitro group substitution led to lower selectivity (entry 3). However, carboxylic acid (1d) proved a very promising substrate for the reaction, affording 3d with a 75:25 enantiomeric ratio (e.r.) (entry 4). We were particularly encouraged by the mathematical implications of this result for dynamic kinetic resolution (28). A 75:25 ratio of enantiomers at high levels of conversion requires that substrate racemization occur (29), at least to some extent, during the course of the overall reaction. Moreover, we observed no erosion of e.r.s upon extended storage or heating (100°C, 15 hours), consistent with the expected high barriers to biaryl bond rotation/racemization required for our fundamental premise to operate.

We then turned our attention to the optimization of the catalyst structure. Tripeptide catalysts were initially examined with the Dmaa residue at the N terminus and chiral α-methylbenzylamides (αMba) at the C terminus (Fig. 3). Simple replacement of L-Pro in 4 for L-pipecolinic acid (Pip) led to a notable improvement in selectivity (catalyst 5, entry 2). Substitution of a range of amino acids in the i + 2 position reduced enantioselectivity slightly (entries 3 to 5), whereas altering the stereogenic sense at this position (entry 6) or the adjacent C-terminal position (entry 7) led to an even greater loss in selectivity. Even so, catalyst 11a, with a simplified N,N-dimethyl amide at the C terminus, delivered 3d in the highest observed enantioselectivity under the initial reaction conditions (entry 8). Ester substitution in place of the N,N-dimethyl amide lowered selectivity substantially (entry 9). We therefore declared peptide 11a our lead catalyst for further study.

Fig. 3

Optimization of the catalyst structure. Data determined as in Fig. 2. Boc, tert-butoxycarbonyl.

We then evaluated additional reaction parameters, including solvents, concentration, temperature, and brominating agent. Among these factors, the bromine source proved particularly influential. N-Bromophthalimide (NBP) consistently gave the best results. In combination with optimized solvent and concentration conditions (30), catalyst 11a (10 mol %) was found to promote conversion of racemic (±)-1d to an 80% yield of 3d, with an e.r. of 97:3 (Scheme 2). These results were obtained at room temperature on a 0.5 mmol scale, and we observed no variation over a range of reaction scales up to 9 mmol (~2 g).

Scheme 2

With an effective catalyst and mild reaction conditions in hand, we then examined substrate scope. The substrates shown in Fig. 4 were readily prepared by using established aryl-aryl bond-forming cross-coupling reactions (31). A range of substitutents on the aryl rings of the substrate proved compatible with the excellent results obtained for substrate 1d. For example, electron-withdrawing NO2-group substitution para to the phenolic arene, as in compound 12, led to the isolation of 13 with a 97:3 e.r. (entry 2). Compound 14, with meta-NO2 substitution, gave a similar result (entry 3), as did electron-rich anisoles 16 (entry 4) and 18 (entry 5). Given the importance of fluorinated aromatic compounds, we assessed meta-substituted fluorobenzene (20) and para-substituted fluorinated substrate 22 (32) and found that they also were excellent substrates for the reaction (entries 6 and 7). Alkyl substitution ortho to the carboxylic acid leads to somewhat lower enantioselectivity (entry 8).

Fig. 4

Substrate scope for enantioselective bromination. Data determined as in Fig. 2. Results represent the average of two to three runs per substrate.

It appears that this catalytic methodology could be quite useful for asymmetric synthesis of a range of heteroarene compounds containing functionalities of relevance to bioactive natural product substructures (33). For example, pyrrole analog 26 was converted to 27 with a moderate but encouraging 85:15 e.r. (entry 9). Catechol 28, a potential substructure in the stegane natural products (34), was converted to 29 with an e.r. of 95:5 (entry 10), and can be enriched to exhibit e.r.s of >98:2 by a single recrystallization. The results shown in Fig. 4, taken together, reveal that catalyst 11a and, perhaps more importantly, its analogs could be quite useful conceptually for synthesis of a broad range of optically enriched biaryl-type compounds.

We have also gained some insight into the mechanism of these intriguing reactions. Several experiments with stripped down, potentially catalytic moieties confirm the possibility of amide catalysis (20), perhaps via a type of [O-Br]-cationic species. For example, when the reaction is conducted in the absence of a catalyst, bromination is sluggish, and 1 is converted to 3 in only 15% yield after 18 hours (Fig. 5). Tertiary amines such as N,N-diisopropylethylamine also provide only slight rate acceleration, and the yield of 3 is 30% under analogous conditions. However, the use of the N,N-dimethylamide 30 leads to 91% isolated yield of 3. These results point to a functional role for one of the several amides resident in catalyst 11a, including the terminal N,N-dimethylamide. Neither mono- nor dibrominated products were observed in substantial quantities under these conditions.

Fig. 5

Assessment of the catalytic efficiency of simple functional groups. i-Pr, iso-propyl group.

These observations, in combination with our knowledge of stereochemical aspects of the reaction, have allowed us to posit a possible explanation for the stereochemical outcome. Single-crystal, heavy-atom x-ray analysis allowed assignment of the absolute configuration of the major product of enantioenriched 3d as the (R)-atropisomer (Fig. 6). An accessible conformation of catalyst 11a is likely as shown in Fig. 6, with axial disposition of the pipecolinic acid substituent, because of the well-known preference of N-acyl piperidines to adopt conformations with axial 2-substituents to avoid allylic strain (35). Docking of the substrate 1d through salt bridge formation between the Dmaa tertiary amine and the substrate carboxylic acid disposes a putative O-bromonium ion toward formation of the observed stereoisomer of 3d. Notably, free rotation of partially brominated (e.g., mono- and dibromonated) species may be possible until the ortho-ortho′ substituents are each installed, leading to a barrier to rotation high enough to preclude product racemization. On the other hand, hydrogen bonds between the phenolic proton and the catalyst amide groups may prevent such bond rotation at intermediate stages of the reaction. Although other mechanisms may also be operative, models such as that offered in Fig. 6 provide one starting place for evaluation.

Fig. 6

X-ray structure of the major enantiometer of 3d (right) and a possible docking model explaining selectivity (left). Structure shown is an Oak Ridge thermal ellipsoid plot.

Synthesis of optically enriched biaryl compounds using enantioselective catalysts and dynamic kinetic resolution should enable improved access to stereodefined atropisomeric materials. More broadly, the approach described herein may also stimulate related research involving selective reactions of other interconverting, axially chiral compounds, promoted by simple peptide-based catalysts.

Supporting Online Material

Materials and Methods

Figs. S1 to S4

Tables S1 to S10

References and Notes

References and Notes

  1. Experimental procedures and compound characterization are available as supporting material on Science Online.
  2. We are grateful to the National Institute of General Medical Sciences of the NIH for support (GM068649). Metrical parameters for the solid state structure of compound 3d are available free of charge from the Cambridge Crystallographic Data Centre under CCDC-773872.
View Abstract

Stay Connected to Science

Navigate This Article