Report

Biocatalytic synthesis of planar chiral macrocycles

See allHide authors and affiliations

Science  21 Feb 2020:
Vol. 367, Issue 6480, pp. 917-921
DOI: 10.1126/science.aaz7381

Enzymes lock in planar chirality

Molecules with very large rings—macrocycles—are often conformationally constrained, and some exhibit planar chirality when substituents of the ring cannot rotate freely. Restricted rotation is generally valued in macrocycles because it can hold the molecule in functional conformations. Using a well-established lipase enzyme, Gagnon et al. developed a synthesis of planar chiral macrocycles with handles that can be easily functionalized. Computational docking suggests how using an enzyme as the catalyst for sequential acylation reactions can impart the observed stereochemistry.

Science, this issue p. 917

Abstract

Macrocycles can restrict the rotation of substituents through steric repulsions, locking in conformations that provide or enhance the activities of pharmaceuticals, agrochemicals, aroma chemicals, and materials. In many cases, the arrangement of substituents in the macrocycle imparts an element of planar chirality. The difficulty in predicting when planar chirality will arise, as well as the limited number of synthetic methods to impart selectivity, have led to planar chirality being regarded as an irritant. We report a strategy for enantio- and atroposelective biocatalytic synthesis of planar chiral macrocycles. The macrocycles can be formed with high enantioselectivity from simple building blocks and are decorated with functionality that allows one to further modify the macrocycles with diverse structural features.

Macrocycles are capable of simultaneously displaying extended molecular frameworks while retaining some conformational bias (1, 2). The cyclic skeleton can even impart severe restrictions on bond rotations that can lock functional groups or other molecular fragments in conformations that would be otherwise unfavorable in an acyclic analog (3, 4). Such is the case in planar chiral cyclophanes, a subset of macrocycles for which conformation (or size) limits the rotation of an aromatic unit within the skeleton.

The presence of planar chirality in natural product terpenes as well as macrocyclic peptides is well documented, and atropisomerism has become increasingly apparent in drug discovery (Fig. 1A) (57). As such, methods for forming peptidic cyclophanes have attracted increased attention. However, in most cases, synthetic methods face the steep challenge of having to form the rigidified and often strained macrocycle itself while simultaneously imparting high levels of enantioselection. Consequently, atroposelective macrocyclizations are rare and can be classified into two synthetic strategies (Fig. 1B). Most protocols employ auxiliaries that enforce conformations of the acyclic precursor through noncovalent interactions (8, 9). Much more rare are instances in which catalysis is exploited to induce asymmetry during the formation of the cyclophane (10, 11).

Fig. 1 Planar chirality in macrocycles.

(A) Examples of planar chiral macrocycles in natural products and pharmaceuticals. (B) Methods for installing planar chirality in macrocycles. (C and D) Notable concepts for a proposed chemoenzymatic synthesis of planar chiral macrocycles. Me, methyl; p-tol, p-toluene; Cy, cyclohexyl; Ph, phenyl; Ts, tosyl; R, H, aryl; R1, alkyl, aryl, halogen; R2, alkyl, aryl; rt, room temperature; krac, rate of racemic reaction.

Although macrocyclization has been examined via biocatalysis (1216), the preparation of prevalent planar chiral macrocycles has largely been ignored. This is surprising given that biocatalysis has had a profound impact on the synthesis of crucial chiral building blocks such as secondary alcohols and amines (1721). In particular, the dynamic kinetic resolution (DKR) developed by Bäckvall and co-workers is a strategy involving transition metal catalysis and biocatalysis working in concert. The process involves a transition metal complex that catalyzes racemization of a substrate (in general, a secondary alcohol or amine) and an enzyme (typically a lipase) that selectively acylates one enantiomer forming an ester or amide (22). A subsequent step is then required to access the desired alcohol and amine through deacylation (Fig. 1C). In examining whether an analogous process could be applied to the preparation of planar chiral macrocycles, several differences from the standard DKR are apparent (Fig. 1D). First, in DKR protocols, the acylation is only temporary, as the free alcohol or amine is typically desired. In addition, the acylating agent can be added in excess to improve reaction rates and yields. In a macrocyclization process, the acylation is inherent in the final product and the stoichiometry between alcohol and acylating agent is naturally fixed. Second, in the absence of secondary alcohols, a different racemization process must be used. Despite the challenges, a biocatalytic DKR process to access planar chiral cyclophanes has considerable potential: The thermal stabilities and high enantioselectivities observed with commercially available lipases make them ideal for macrocyclization processes, and the simple building blocks required allow one to rapidly build complexity in a chiral architecture.

A key goal in the development of a chemoenzymatic synthesis of planar chiral macrocycles was to permit synthesis from common and simple building blocks. As such, we envisioned exploiting common diacids or diesters as aliphatic linkers (A, Fig. 1D). The chemoenzymatic macrocyclization would take place via sequential acylations using a lipase on an aromatic diol B (Fig. 1D) that possessed core functionality amendable to diversification and applicability to drug discovery efforts.

In contrast to the classic DKR process, racemization of intermediate C occurs through free rotation of the aromatic ring. Several challenges exist for the chemoenzymatic macrocyclization. The ring-closing event would result in a rigidified macrocycle, and the enzyme must be able to promote such a ring closure. We expected that elevated temperatures could be used to promote macrocyclization, but we were conscious of possibly degrading the enzyme or affecting the conformational stability of the cyclophane (i.e., at what temperature the ansa-bridge would be able to freely rotate and racemize the desired macrocycle). To help favor macrocyclization, longer diesters A could be employed, but the aromatic substituents R1 would have to be bulkier to restrict the rotation of the ansa-bridge. The size of the R1 substituents is also critical, as they must affect the selectivity of the enzyme but not negatively influence the reactivity.

Given the previous success of the serine hydrolase Candida antarctica lipase B (CALB) (23, 24) in DKR of secondary alcohols (25, 26), we used the enzyme in our synthesis of [13]paracyclophanes by macrolactonization of benzylic diols (Fig. 2A). We tested a molecule with an unsubstituted aromatic core (1a) and isolated the desired achiral macrocycle 3a in reasonable yield. Having demonstrated that CALB could promote the macrocyclization, we examined a subsequent cyclization having an aromatic core substituted with OMe (methoxy) groups (1b), but the yield of the corresponding [13]paracyclophane 3b was only 10%. In addition, variable temperature nuclear magnetic resonance (VT NMR) analysis of the benzylic proton signals (highlighted in green, Fig. 2A) showed coalescence of the signals at 50°C. Macrocyclization employing larger bromo substituents (diol 1c) was even less successful, most likely due to an unfavorable steric clash between the ortho-substituted benzylic diol and the enzyme active site. We redesigned the starting diol with an inserted methylene group next to the aromatic core (5, Fig. 2B). With the extended diol, we obtained the desired [14]paracyclophane 6 in good yield and high enantioselectivity. VT NMR analysis of the benzylic proton signals of 6 showed no coalescence of the signals even at 100°C. Although lowering the temperature decreased the yield of 6, raising the temperature had no beneficial effect on yield and did not promote rotation of the ansa-bridge, which would lower the overall enantiopurity.

Fig. 2 Biocatalytic synthesis of planar chiral macrocycles employing a lipase CALB.

(A and B) Development of the biocatalytic macrocyclization. Green highlighted areas indicate methylene units monitored by VT NMR. aIsolated yields (0.1-mmol scale). bDetermined by chiral SFC high-performance liquid chromatography analysis. See supplementary materials for details. cExtending reaction time to 48 hours: 81% 6, >99% enantiomeric excess (ee). dUsing MeCN as solvent: 16% 6, >99% ee. e[2.5 mM]. f[10 mM]. (C) Computational docking of products and intermediates to CALB.

Although CALB tends to favor acylation of R-centered carbon chiral centers (27), it was unknown how the CALB active site, which has naturally evolved to differentiate the geometry of tetrahedral carbons centers, would accommodate the prochiral aromatic plane of the forming macrocycle. To better understand how the active site environment might engage different conformations of the cyclophane substrate, we performed docking with the program Fitted (28, 29) from the Forecaster computational platform (Fig. 2C) (30). Lee and co-workers previously reported an x-ray crystal structure of CALB bound to a phosphonate inhibitor (PDB ID 5GV5) (31). We replaced the phosphonate inhibitor with each atropisomer of the macrocycle and examined the binding mode suggested by docking (32). The major product (−)-6 is oriented with its carboxyl groups toward the nearby catalytic serine residue (Ser105), with one of the bromine substituents pointing toward the exterior of the active site. The opposite enantiomer is translated by >2.5 Å in the docking model, which suggests that the serine-catalyzed reaction would be geometrically challenging. The projected translation of (+)-6 results from a clash between a bromine atom and Leu140, precluding binding of the bromine atom into the hydrophobic site delineated by Leu140, Ala141, and Leu144. In a docking of the starting dibromo diol 5, it fits into the cavity with its alcohol extending toward the catalytically active serine. Indeed, even a monoesterified intermediate orients itself within the active site with its carboxylate toward the serine residue and the bromine substituents in a conformation mirroring that of the desired cyclophane. The biocatalytic synthesis of [14]paracyclophane 6 could be easily reproduced on the gram scale (see supplementary materials), and we proceeded to explore the substrate scope with regard to ring size. Although the dibromo cyclophane 6 could be obtained by using a diacid with a 6-methylene spacer [-(CH2)6-], reducing the spacer to four or five methylene units did not substantially increase ring strain, and the resulting [12]- and [13]paracyclophanes were obtained in good yields and high enantioselectivities (Fig. 3). This series of macrocycles could be prepared with chlorine or iodine atoms replacing the bromine substituents. The resulting [12]-, [13]-, and [14]paracyclophanes (1012) were all isolated with comparable yield and enantiopurity, despite variation in the size of the halogen substituent.

Fig. 3 Scope of the biocatalytic macrocyclization to afford planar chiral cyclophanes.

All macrocyclizations were performed under standard conditions. Variations in reaction time are indicated when necessary. Detailed reaction conditions for the derivatizations of 6 are presented in the supplementary materials.

Extending the ansa-bridge by an additional methylene spacer provided a good yield of [15]paracyclophane 9; however, the product was isolated as a racemic mixture, suggesting that the larger aliphatic ring no longer constrained rotation of the planar cyclophane. The larger size of the iodine allowed for the synthesis of the enantioenriched [15]paracyclophane 17. To investigate whether the active site of the enzyme could tolerate more functionalized ansa-bridges, we prepared two different macrocycles that have rigidified 1,3-diynes in their backbones with phenyl-substituted (18) and alkynyl-substituted (19) cores, as well as a [14]paracyclophane 20 that has a disulfide bridge, a common motif found in bioactive macrocyclic peptides (33). A series of functionalized aromatic diols were also well tolerated in the macrocyclization process. The terphenyl-based macrocycle 21 could be formed via macrocyclization as could similarly substituted p-anisoyl and m-anisoyl cyclophanes (22 and 23, respectively) with high enantioselectivity. We were also able to synthesize [14]paracyclophanes 24 and 25, which have cores with either phenylalkynyl or hexynyl substituents. Heteroatom-containing functional groups such as the p-methylaniline substituents present in macrocycle 26 could also be installed within the chiral macrocyclic frameworks. Finally, C1-symmetric derivatives were formed in high enantioselectivity. Macrocycle 27 was isolated with one iodide substituent and one alkynyl unit, whereas macrocycle 28 was isolated with one bromide substituent and a Csp3-hybridized motif (benzyl). Notably, the halogen-containing planar chiral macrocycles can act as a platform for the synthesis of other derivatives through modern cross-coupling techniques (34). For example, the macrocycle 29 was prepared having Bpin (pinacolatoboron) functionality via Miyaura borylation. The bromo-substituted cyclophane 6 could be subjected to Heck coupling to form macrocycle 30. The macrocycle 30 could be formed via the biocatalytic route in high selectivity but lower yield (19%). As such, the ability to exploit the “platform” macrocycles is particularly powerful for accessing other substitution patterns that may not be compatible with the CALB enzyme (3537). Other cross-coupling techniques were also viable for diversification of the macrocycles: Suzuki-coupling formed the terphenyl-based macrocycle 21 from the Br-substituted cyclophane 6, whereas Sonogashira coupling on the iodo-substituted cyclophane 16 afforded macrocycle 25 in high yield with little loss of enantiopurity, exemplifying the resistance of the macrocycles to cleavage even when heated in basic aqueous conditions. The preparation of a series of functionalized planar chiral macrocycles that have halogen or borylated substituents opens avenues for diversification outside the boundaries of what may be tolerated by the enzyme active site. The challenge in the preparation of planar chiral macrocycles in drug discovery is now well recognized. With the pervasive awareness of environmental issues, it would seem apt that biocatalysis provides an innovative solution.

Supplementary Materials

science.sciencemag.org/content/367/6480/917/suppl/DC1

Materials and Methods

NMR Spectra

References (3852)

References and Notes

  1. For full experimental conditions, see the supplementary materials.
Acknowledgments: We thank N. Moitessier for help with the Forecaster program. Funding: This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC, Discovery grant 1043344), American Chemical Society Petroleum Research Fund (ACS PRF 60211-ND1), Université de Montréal, and the Fonds de Recherche Nature et Technologie via the Centre in Green Chemistry and Catalysis (grant FRQNT-2020-RS4-265155-CCVC). Author contributions: C.G., E.G., C.M., J.S., and C.P. synthesized precursors and macrocycles. C.G. performed docking experiments. S.K.C. designed and directed the investigations. S.K.C., C.G., J.S., and E.G. wrote the manuscript. Competing interests: None declared. Data and materials availability: All data needed to evaluate the conclusions in the paper are available in the main text or the supplementary materials.

Stay Connected to Science

Navigate This Article