Remote C–H Hydroxylation by an α-Ketoglutarate-Dependent

Dec 28, 2017 - Selective C–H functionalization at distal positions remains a highly challenging problem in organic synthesis. Though Nature has evol...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/JACS

Cite This: J. Am. Chem. Soc. 2018, 140, 1165−1169

Remote C−H Hydroxylation by an α‑Ketoglutarate-Dependent Dioxygenase Enables Efficient Chemoenzymatic Synthesis of Manzacidin C and Proline Analogs Christian R. Zwick, III and Hans Renata* Department of Chemistry, The Scripps Research Institute, 130 Scripps Way, Jupiter, Florida 33458, United States S Supporting Information *

ABSTRACT: Selective C−H functionalization at distal positions remains a highly challenging problem in organic synthesis. Though Nature has evolved a myriad of enzymes capable of such feat, their synthetic utility has largely been overlooked. Here, we functionally characterize an α-ketoglutarate-dependent dioxygenase (Fe/αKG) that selectively hydroxylates the δ position of various aliphatic amino acids. Kinetic analysis and substrate profiling of the enzyme show superior catalytic efficiency and substrate promiscuity relative to other Fe/αKGs that catalyze similar reactions. We demonstrate the practical utility of this transformation in the concise syntheses of a rare alkaloid, manzacidin C, and densely substituted amino acid derivatives with remarkable step efficiency. This work provides a blueprint for future applications of Fe/αKG hydroxylation in complex molecule synthesis and the development of powerful synthetic paradigms centered on enzymatic C−H functionalization logic.



INTRODUCTION Amino acids are among the most important classes of organic small molecules, serving as prized building blocks in the pharmaceutical industry and as versatile tools to probe biological systems. Given their importance, many synthetic methods have been developed to access novel amino acids beyond those found in naturally occurring proteins.1 Unfortunately, the presence of reactive free amino and carboxylate groups and the need to set multiple stereocenters render de novo synthesis of noncanonical amino acids (ncAAs) highly challenging and inefficient. C−H functionalization has recently emerged as a powerful paradigm in organic synthesis,2 and major strides have been made in the application of this logic in the preparation of ncAAs. Despite rapid progress in the field, activation at remote/distal positions of the amino acid substrate is still highly challenging; approaches using transition metals are typically limited to functionalization at the β position,3 due to the preferred formation of the five-membered metallacycle intermediate. There have been a few reports of functionalization at the δ position of an amino acid using variations of Hofmann−Löffler−Freytag (HLF) reaction;4a−c however, this strategy often suffers from limited amino acid substrate scope, and lack of control of the regio- and stereochemical outcomes of the reaction. © 2017 American Chemical Society

In comparison, Nature often employs chemo- and regioselective hydroxylation as a gateway transformation in the biosynthesis of ncAAs. One such example can be found in the biosynthesis of 4-methylproline, which arises from iterative oxidation of L-leucine at C5,5 followed by reductive amination of the 1-pyrroline product (Figure 1). At the outset of our studies, two leucine 5-hydroxylases, LdoA5a and EcdK,5b had been shown to participate in the biogenesis of (2S,4S)-4methylproline and (2S,4R)-4-methylproline, respectively. Another leucine 5-hydroxylase, GriE, was predicted for the biosynthesis of griselimycin in 2015.6 As C5-hydroxylated amino acids can be found in many bioactive peptide natural products, such as cyclomarin A, halipeptin, and amanitin, we became interested in exploring the synthetic potential of this transformation. Given the limited substrate scope of LdoA and the reported instability of EcdK during catalysis, we focused on characterizing GriE and examining its viability as practical biocatalyst for amino acid hydroxylation. Concurrent to our efforts, Müller and co-workers were able to confirm the hydroxylation activity of GriE through gene inactivation and in vitro studies,7 though its synthetic utility was not explored Received: December 6, 2017 Published: December 28, 2017 1165

DOI: 10.1021/jacs.7b12918 J. Am. Chem. Soc. 2018, 140, 1165−1169

Article

Journal of the American Chemical Society

Chart 1. Substrate Scope of Amino Acid Hydroxylation with GriEa,b

Figure 1. (A) Biosynthesis of (2S,4R)-4-methylproline motif in griselimycin, starting from a diastereoselective δ hydroxylation of Lleucine. (B) Contemporary approach for δ functionalization of amino acids, which relies on innate reactivity and lacks regio- and/or stereochemical control.

beyond the native substrate. Here, we demonstrate the synthetic utility of GriE as a practical biocatalyst for the hydroxylation of a range of amino acids (13 substrates in total) with useful turnover numbers. We further leverage the relaxed substrate specificity and the high activity of the enzyme in the synthesis of valuable small molecules, including the rare alkaloid manzacidin C and a suite of L-methylproline analogs. Overall, we show that harnessing Nature’s inventory of oxidation biocatalysts can greatly facilitate the preparation of complex molecules and valuable building blocks.

a

Conditions for total turnover number (TTN) measurement: free amino acid (20 mM, 1 equiv), αKG (60−80 mM, 3−4 equiv), Lascorbic acid (20 mM, 1 equiv), FeSO4 (1 mM, 0.05 equiv), GriE (0.01−0.2 mol % enzyme loading), kPi buffer (pH = 7.0, 50 mM, 1 mL total volume), 5 h at 20 °C. C5 stereochemistry of 8, 12, and 13 was assigned by analogy to 5. TTNs were determined by NMR analysis calibrating against an internal standard of known concentration. *Reaction produced a mixture of single and double oxidation products. bConditions for preparative scale in vitro reaction: free amino acid (20 mM, 0.1 mmol), αKG (24−80 mM, 0.12−0.4 mmol), L-ascorbic acid (10−20 mM, 0.05−0.1 mmol), FeSO4 (1 mM, 5.0 μmol), GriE (0.01−0.2 mol % enzyme loading), kPi buffer (pH = 7.0, 50 mM, 5 mL total volume), 5 h at 20 °C. Unless otherwise noted, yields refer to isolated yields after ion-exchange purification. #Isolated ̂ isolated yield after Fmoc yield after HPLC purification. Two-step derivatization.



RESULTS AND DISCUSSION Overexpression of GriE in E. coli as a C-His6-tagged protein was found to provide excellent protein yield (up to 100 mg/L). In the presence of αKG, ferrous salts, and ascorbic acid, the purified enzyme was found to catalyze selective C5 hydroxylation of L-leucine with complete diastereoselectivity and high (>7,500) total turnover number (TTN). The high TTN and subsequent kinetic analysis suggested that GriE is less susceptible to suicidal inactivation during turnover than EcdK (see Figure S2). To examine the broader utility of GriE, we next investigated the substrate scope of the reaction (Chart 1). In agreement with Müller’s work,7 no activity was observed with D-Leu, L-Val, and L-Ile. Interestingly, L-allo-Ile was found to undergo hydroxylation to yield 7, albeit with only moderate TTN. This observation suggests that the active site of the enzyme is sensitive to steric effects at the β position of the substrate. Productive reactions were observed with γ-methyl-Lleucine (Mleu) and 4-hydroxy-L-leucine to afford 4 and 10 respectively, demonstrating that substitution at the γ position of the substrate is well-tolerated. In fact, hydroxylation of Mleu proceeded with higher TTN than that of L-Leu, yielding a mixture of 4 and a cyclic imine arising from double oxidation at C5. The presence of an electron-withdrawing group strongly deactivates the adjacent carbon for C−H bond oxidation.8 Thus, the finding that 4-hydroxyleucine can undergo

hydroxylation at C5 suggests that geometric propensity of the C5 carbon to the iron center can partially override the negative inductive effect of neighboring tertiary alcohol. Increased bulk at the distal position of the substrate is accommodated as Lnorleucine (Nle) and 2-aminoheptanoic acid were found to be hydroxylated to produce 5 and 8 with 1900 and 460 TTNs, respectively. Furthermore, additional substitution at C5 and electron withdrawing groups at the ε position are tolerated (see products 11, 12, and 13). A small amount of hydroxylation product can also be observed with 2-aminooctanoic acid, but the poor solubility of the substrate prevented a more accurate TTN measurement. On the other hand, substantially larger amino acids such as cyclohexylalanine, or those containing Lewis basic group at the distal position (e.g., lysine derivatives) did not participate in the reaction. Remarkably, with the exception of L-methionine sulfoxidation (dr = ca. 3:1), biocatalytic oxidation with GriE proceeded with excellent regio- and/or diastereoselectivity. In all cases examined, the reaction is selective for the hydroxylation at the γ position and 1166

DOI: 10.1021/jacs.7b12918 J. Am. Chem. Soc. 2018, 140, 1165−1169

Article

Journal of the American Chemical Society products 2, 5, 8, 10, 12, and 13 were formed in >99:1 dr. This observation stands in stark contrast to contemporary chemical methods4a−c for amino acid functionalization, where regio- and diastereomeric mixtures are commonly observed. To unambiguously determine the stereochemical outcome of the oxidation, X-ray crystal structure analyses were performed on various derivatives of 2, 5, and 10 (see the Supporting Information) and structural information gained from these analyses was used to assign the stereochemistry at C5 for products 8, 12, and 13. The steady-state kinetic parameters for hydroxylation of representative amino acid substrates with GriE were also determined to more precisely characterize the observed reactivity trends (Figure S4). Hydroxylation of L-Leu was observed to proceed with higher catalytic efficiency than Mleu and Nle (kcat/KM (Leu) = 280 mM−1 min−1). This arises from the lower KM in the case of L-leu, as all three substrates exhibit comparable kcat. The kinetic parameters obtained for these substrates indicate that GriE is notably efficient relative to other Fe/αKGs capable of hydroxylating aliphatic amino acids.9 Hydroxylation of L-allo-Ile exhibited substantially reduced catalytic efficiency, which stands in agreement with the low TTN observed in the production of 7. The synthetic utility of this biotransformation was demonstrated via preparative (0.1 mmol) scale in vitro reactions with all the accepted substrates. On this scale, amino acid hydroxylation proceeded with isolated yields that closely mirrored the observed TTNs. Given the highly hydrophilic nature of the products, technical challenges associated with their isolation merit brief discussion. In cases with high substrate conversion (e.g., in the preparation of 2, 5, 8, 9, and 11), purification on cation exchange resin was sufficient to provide the desired products with high purity. When mixtures of amino acid products were obtained, the desired products could be isolated by reversed-phase HPLC or silica gel chromatography after Fmoc derivatization (e.g., products 6, 7, 10, and 13). If desired, the pure Fmoc derivatives could be readily converted to the free amino acids by treatment with NaOH. Finally, to gain structural basis for the substrate preference of GriE, we constructed a series of visualizations based on recently published crystal structures of the enzyme7 by placing various substrates into the binding pose adopted by L-leucine in the active site (Figure 2). On the basis of the structure of GriE-Leu complex, we posit that amino acids containing shorter side chains (no δ carbons) do not possess any C−H bonds within the requisite geometric proximity to the Fe center for hydroxylation to occur. Placement of L-Ile in the active site revealed potential steric clashes between its β-methyl group and residues V115 and W116 (3.2 and 2.8 Å away, respectively). Because L-Ile, out of the 20 canonical amino acids, bears the closest structural resemblance to Leu, we speculate that W116 and V115 act as gatekeeper residues to discriminate against the binding of Ile in the active site. The structural model generated for Nle indicated that its terminal methyl group could occupy a pocket of space created by H110, V115, and P172. This binding pose also oriented the pro-S hydrogen at C5 toward the Fe center, accounting for the observed stereochemical outcome of the hydroxylation. In contrast, a similar pose generated for Lys positioned its ε-amino group within distance to act as an L-type ligand to the Fe center. Alternatively, in its protonated state, the ε-amino group might induce unfavorable polar contacts within its local environment. In the case of D-Leu, binding pose that maintained the ionic interactions with the carboxylate and

Figure 2. (A) Crystal structure of GriE-Leu-αKG complex (PDB ID: 5NCI). (B) Model of L-Ile (teal) in the active site, showing potential clashes with V115 and W116. (C) Model of Nle (teal) in the active site, accounting for the observed stereochemical outcome of the reaction. (D) Models of Lys (teal and yellow) in the active site, showing the orientation of the ε-amino group relative to the Fe center.

amino motif could not be generated without causing severe steric clashes. Although an alternative binding mode could be constructed, this model placed the isobutyl group trans to His210, potentially impeding subsequent formation of the reactive Fe(IV)O species. Recent structural analyses on other Fe/αKGs have argued for large-scale structural rearrangements10 during the formation of the Fe(IV)O complex. Hence, it is possible that the active site plasticity of GriE actually allows for initial binding of some of these nonsubstrates, but the resulting complex precludes subsequent formation of the active Fe(IV)O species. In this regard, more in-depth analyses employing molecular dynamics simulations might be warranted. We next assessed the viability of large-scale preparative scale amino acid hydroxylation (Scheme 1) using crude E. coli lysate Scheme 1. Preparative-Scale Hydroxylation with Clarified Lysate

expressing GriE. Under these conditions, the biotransformation proceeded to virtually full conversion even at high substrate concentrations (up to 50 mM in L-leucine), and no decrease in conversion was observed when the reaction was performed on gram scale (10 mmol of L-leucine). A similar biotransformation was also performed with L-norleucine on 200 mg scale with excellent conversion to the desired product. Furthermore, a one-pot biocatalytic dihydroxylation cascade was developed by the sequential addition of crude lysate expressing Gox leucine 4-hydroxylase11 and crude lysate expressing GriE. On 100 mg 1167

DOI: 10.1021/jacs.7b12918 J. Am. Chem. Soc. 2018, 140, 1165−1169

Article

Journal of the American Chemical Society

for TBADT-catalyzed photofunctionalization and the first example of direct azidation of an unprotected amino acid. This reaction could be performed on 2 mmol scale to provide 49% yield of pure 20 after reverse phase HPLC purification. In agreement with our substrate profile model, azide 20 was found to undergo efficient δ-hydroxylation with GriE, affording >95% conversion to alcohol 21 on 130 mg scale with only minor amounts of overoxidized byproduct. No reduction to the corresponding amine at the C4 carbon was observed during the reaction, indicating that the azide moiety is fully compatible with the iron center of GriE. Next, a telescoped procedure consisting of (1) hydrogenation under basic conditions, (2) in situ Boc protection of the resulting amines, and (3) selective lactonization was devised to access lactone 22 in 41% yield. Conversion of 22 to manzacidin C is well-documented in the literature, proceeding in two facile steps.15 Overall, our approach constitutes a five-step formal synthesis of manzacidin C, and compares favorably to previous approaches to this alkaloid.16 More importantly, we showed that incorporation of enzymatic methods into the precepts of C−H functionalization logic17 can dramatically simplify a synthesis design. Although Müller did not note any formation of double oxidation product,7 we found that GriE was capable of performing iterative C5 oxidation of Mleu. In fact, reexamination of several L-Leu derived substrates revealed that at sufficiently high concentrations of GriE, iterative oxidation indeed took place and could even be driven to completion. This observation led us to investigate the possibility of developing a chemoenzymatic cascade to synthesize L-proline analogs containing stereodefined substituents. These motifs are prevalent in the acyldepsipeptide antibiotics,18 as well as synthetic inhibitors of the serine protease Factor Xa.19 To this end, we developed a one-pot protocol whereby the amino acid was first converted to the corresponding imine in the presence of GriE, followed by addition20 of NH3·BH3 to reduce the C5− N double bond (Scheme 2B). The efficiency of this transformation is striking, as various proline analogs (3, 23− 26) can be prepared in just one or two steps from commercial materials in very good yields. Furthermore, complete stereocontrol at C4 can be obtained by capitalizing on GriE’s ability to desymmetrize the two terminal methyl groups of the substrates while at the same time accommodating various substituents at the γ position, including the highly deactivating fluorine moiety, during the oxidation event. In contrast, contemporary chemical approaches to this motif often involve inefficient functional group interconversions and extraneous protection/deprotection sequences, and suffer from a lack of stereocontrol at C4. For example, chemical synthesis of 4methylproline requires conversion of protected 4-hydroxyproline to the corresponding ketone, followed by Wittig methylenation and hydrogenation with Crabtree’s catalyst, with reported dr ranging from 4:1 to 15:1.21 Taken together, this work further underscores the synthetic potential of Fe/ αKG oxidation in the preparation of valuable building blocks for modern drug discovery efforts.

scale, approximately 80% conversion to the dihydroxylated product was observed by LCMS analysis. We further showed that the desired hydroxylation products could be isolated from the crude cellular milieu by simple conversion to either the Boc or Fmoc derivatives, which also resulted in a facile intramolecular cyclization to yield lactones 14, 16, and 17. From the perspective of synthesis design, the propensity of the products to undergo self-lactonization is highly beneficial as it bypasses the need for additional protection steps of the free acid and alcohol. Furthermore, the δ-lactone product can be used directly in peptide bond formation, as illustrated by the conversion of 14 to 15 by simple heating with H-Gly-OEt·HCl in toluene in excellent yield. To date, little to no attention has been given to the use of Fe/αKGs in multistep synthesis. To further validate the practical utility of GriE, we next pursued a chemoenzymatic synthesis of manzacidin C (18, Scheme 2A), a natural product Scheme 2. (A) Application of GriE-Catalyzed Hydroxylation in the Formal Synthesis of Manzacidin C; (B) Chemoenzymatic Synthesis of Various Proline Analogs via GriE-Catalyzed Iterative C5 Oxidation

#

Contains ca. 10% of 2 as minor impurity

from Hymeniacidon sp. containing rare 3,4,5,6-tetrahydropyrimidine and bromopyrrolecarboxylic acid motifs.12 On the basis of the substrate profile model that we developed in Chart 1, we hypothesized that GriE could hydroxylate a derivative of Lleucine possessing a suitably masked amine precursor at C4. Inspired by a recent report of protecting-group-free fluorination of leucine by Britton,13 we developed a photocatalytic azidation of L-leucine at C4 using tetrabutylammonium decatungstate (TBADT) as catalyst and azide 19 as radical acceptor.14 This is the first report on the use of organic azides as trapping agents



CONCLUSION Chemoenzymatic synthesis of complex molecules featuring enzymatic C−H hydroxylation has mainly focused on the P450s,22 whereas Fe/αKGs have remained largely unexplored. Here, we have provided the first application of Fe/αKG enzyme in natural product synthesis and preparative-scale production of useful synthons as a proof-of-principle demonstration of the 1168

DOI: 10.1021/jacs.7b12918 J. Am. Chem. Soc. 2018, 140, 1165−1169

Article

Journal of the American Chemical Society

A.; Chiou, G.; Garg, N. K.; Tang, Y.; Walsh, C. T. J. Am. Chem. Soc. 2013, 135, 4457−4466. (6) Kling, A.; Lukat, P.; Almeida, D. V.; Bauer, A.; Fontaine, E.; Sordello, S.; Zaburannyi, N.; Herrmann, J.; Wenzel, S. C.; König, C.; Ammerman, N. C.; Barrio, M. B.; Borchers, K.; Bordon-Pallier, F.; Brönstrup, M.; Courtemanche, G.; Gerlitz, M.; Geslin, M.; Hammann, P.; Heinz, D. W.; Hoffmann, H.; Klieber, S.; Kohlmann, M.; Kurz, M.; Lair, C.; Matter, H.; Nuermberger, E.; Tyagi, S.; Fraisse, L.; Grosset, J. H.; Lagrange, S.; Müller, R. Science 2015, 348, 1106−1112. (7) Lukat, P.; Katsuyama, Y.; Wenzel, S.; Binz, T.; König, C.; Blankenfeldt, W.; Brönstrup, M.; Müller, R. Chem. Sci. 2017, 8, 7521− 7527. (8) Newhouse, T.; Baran, P. S. Angew. Chem., Int. Ed. 2011, 50, 3362−3374. (9) Hibi, M.; Ogawa, J. Appl. Microbiol. Biotechnol. 2014, 98, 3869− 3876. (10) (a) Zhang, Z.; Ren, J.-S.; Harlos, K.; McKinnon, C. H.; Clifton, I. J.; Schofield, C. J. FEBS Lett. 2002, 517, 7−12. (b) Mitchell, A. J.; Zhu, Q.; Maggiolo, A. O.; Ananth, N. R.; Hillwig, M. L.; Liu, X.; Boal, A. K. Nat. Chem. Biol. 2016, 12, 636−640. (c) Mitchell, A. J.; Dunham, N. P.; Martinie, R. J.; Bergman, J. A.; Pollock, C. J.; Hu, K.; Allen, B. D.; Chang, W.-C.; Silakov, A.; Bollinger, J. M., Jr.; Krebs, C.; Boal, A. K. J. Am. Chem. Soc. 2017, 139, 13830−13836. (11) Smirnov, S. V.; Sokolov, P. M.; Kodera, T.; Sugiyama, M.; Hibi, M.; Shimizu, S.; Yokozeki, K.; Ogawa, J. FEMS Microbiol. Lett. 2012, 331, 97−104. (12) Kobayashi, J.; Kanda, F.; Ishibashi, M.; Shigemori, H. J. Org. Chem. 1991, 56, 4574−4576. (13) Nodwell, M. B.; Yang, H.; Č olović, M.; Yuan, Z.; Merkens, H.; Martin, R. E.; Bénard, F.; Schaffer, P.; Britton, R. J. Am. Chem. Soc. 2017, 139, 3595−3598. (14) Huang, X.; Groves, J. T. ACS Catal. 2016, 6, 751−759. (15) Namba, K.; Shinada, T.; Teramoto, T.; Ohfune, Y. J. Am. Chem. Soc. 2000, 122, 10708−10709. (16) (a) Hashimoto, T.; Maruoka, K. Org. Biomol. Chem. 2008, 6, 829−835. (b) Tran, K.; Lombardi, P. J.; Leighton, J. L. Org. Lett. 2008, 10, 3165−3167. (c) Yoshimura, T.; Kinoshita, T.; Yoshioka, H.; Kawabata, T. Org. Lett. 2013, 15, 864−867. (d) Ichikawa, Y.; Okumura, K.; Matsuda, Y.; Hasegawa, T.; Nakamura, M.; Fujimoto, A.; Masuda, T.; Nakano, K.; Kotsuki, H. Org. Biomol. Chem. 2012, 10, 614−622. (e) Tong, T. M. T.; Soeta, T.; Suga, T.; Kawamoto, K.; Hayashi, Y.; Ukaji, Y. J. Org. Chem. 2017, 82, 1969−1976. (17) Gutekunst, W. R.; Baran, P. S. Chem. Soc. Rev. 2011, 40, 1976− 1991. (18) Osada, H.; Yano, T.; Koshino, H.; Isono, K. J. Antibiot. 1991, 44, 1463−1466. (19) Van Huis, C. A.; Casimiro-Garcia, A.; Bigge, C. F.; Cody, W. L.; Dudley, D. A.; Filipski, K. J.; Heemstra, R. J.; Kohrt, J. T.; Leadley, R. J., Jr.; Narasimhan, L. S.; McClanahan, T.; Mochalkin, I.; Pamment, M.; Peterson, J. T.; Sahasrabudhe, V.; Schaum, R. P.; Edmunds, J. J. Bioorg. Med. Chem. 2009, 17, 2501−2511. (20) For a related use of NH3·BH3 as water-compatible reductant: Ghislieri, D.; Green, A. P.; Pontini, M.; Willies, S. C.; Rowles, I.; Frank, A.; Grogan, G.; Turner, N. J. J. Am. Chem. Soc. 2013, 135, 10863− 10869. (21) Murphy, A. C.; Mitova, M. I.; Blunt, J. W.; Munro, M. H. G. J. Nat. Prod. 2008, 71, 806−809. (22) For recent examples: (a) Zhang, K.; Shafer, B. M.; Demars, M. D., II; Stern, H. A.; Fasan, R. J. Am. Chem. Soc. 2012, 134, 18695− 18704. (b) Kolev, J. N.; O’Dwyer, K. M.; Jordan, C. T.; Fasan, R. ACS Chem. Biol. 2014, 9, 164−173. (c) Loskot, S. A.; Romney, D. K.; Arnold, F. H.; Stoltz, B. M. J. Am. Chem. Soc. 2017, 139, 10196− 10199. (d) Lowell, A. N.; DeMars, M. D., II; Slocum, S. T.; Yu, F.; Anand, K.; Chemler, J. A.; Korakavi, N.; Priessnitz, J. K.; Park, S. R.; Koch, A. A.; Schultz, P. J.; Sherman, D. H. J. Am. Chem. Soc. 2017, 139, 7913−7920. (23) (a) Turner, N. J.; O’Reilly, E. Nat. Chem. Biol. 2013, 9, 285− 288. (b) de Souza, R. O. M. A.; Miranda, L. S. M.; Bornscheuer, U. T. Chem. - Eur. J. 2017, 23, 12040−12063.

potential utility of Fe/αKG-catalyzed hydroxylation in chemical synthesis. Central to our success is the serendipitous discovery of GriE’s broad substrate specificity, which was established via initial substrate-activity relationship and kinetic characterization studies on the enzyme. Capitalizing on this property, we next developed a synthesis strategy centered on the synergistic interplay of chemo- and biocatalytic C−H functionalization logic to arrive at a concise formal synthesis of manzacidin C and efficiently prepare a range of valuable pyrrolidine building blocks. The brevity of these syntheses is a testament to the transformative power of enzymatic oxidation in addressing challenging problems in chemical synthesis. Taken together, our work highlights the benefits of applying biocatalytic retrosynthesis23 and enzymatic C−H functionalization logic in synthesis planning. We anticipate more widespread applications of this approach in the preparation of therapeutically relevant molecules.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b12918. Experimental procedures and analytical data for all new compounds (PDF) Crystallographic information file for C13H15NO3 (CIF) Crystallographic information file for C20H19NO5 (CIF) Crystallographic information file for C11H19NO4 (CIF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Hans Renata: 0000-0003-2468-2328 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Financial support for this work was generously provided by The Scripps Research Institute. We thank K. M. Engle, R. A. Shenvi, and M. D. Disney for helpful discussions and assistance in manuscript preparation. We are grateful to A. L. Rheingold, C. E. Moore, and M. Gembicky (University of California, San Diego) for X-ray crystallographic analysis. We thank R. K. Zhang for technical assistance in macromolecular modeling with Coot. We thank the Shen lab, the Disney lab, the Kodadek lab, and the Roush lab for generous access to their instrumentations.



REFERENCES

(1) Nájera, C.; Sansano, J. M. Chem. Rev. 2007, 107, 4584−4671. (2) Yamaguchi, J.; Yamaguchi, A. D.; Itami, K. Angew. Chem., Int. Ed. 2012, 51, 8960−9009. (3) He, J.; Li, S.; Deng, Y.; Fu, H.; Laforteza, B. N.; Spangler, J. E.; Homs, A.; Yu, J.-Q. Science 2014, 343, 1216−1220. (4) (a) Quinn, R. K.; Könst, Z. A.; Michalak, S. E.; Schmidt, Y.; Szklarski, A. R.; Flores, A. R.; Nam, S.; Horne, D. A.; Vanderwal, C. D.; Alexanian, E. J. J. Am. Chem. Soc. 2016, 138, 696−702. (b) Liu, T.; Myers, M. C.; Yu, J.-Q. Angew. Chem., Int. Ed. 2017, 56, 306−309. (c) Reddy, L. R.; Reddy, B. V. S.; Corey, E. J. Org. Lett. 2006, 8, 2819− 2821. (5) (a) Hibi, M.; Kawashima, T.; Sokolov, P. M.; Smirnov, S. V.; Kodera, T.; Sugiyama, M.; Shimizu, S.; Yokozeki, K.; Ogawa, J. Appl. Microbiol. Biotechnol. 2013, 97, 2467−2472. (b) Jiang, W.; Cacho, R. 1169

DOI: 10.1021/jacs.7b12918 J. Am. Chem. Soc. 2018, 140, 1165−1169