Abstract
The C−H bond functionalization has been widely used in chemical synthesis over the past decade. However, regio- and stereoselectivity still remain a significant challenge, especially for inert aliphatic C−H bonds. Here we report the mechanism of three Fe(II)/α-ketoglutarate-dependent dioxygenases in bicyclomycin synthesis, which depicts the natural tactic to sequentially hydroxylate specific C−H bonds of similar substrates (cyclodipeptides). Molecular basis by crystallographic studies, computational simulations, and site-directed mutagenesis reveals the exquisite arrangement of three enzymes using mutually orthogonal strategies to realize three different regio-selectivities. Moreover, this programmable selective hydroxylation can be extended to other cyclodipeptides. This evidence not only provides a naturally occurring showcase corresponding to the widely used methods in chemical catalysis but also expands the toolbox of biocatalysts to address the regioselective functionalization of C−H bonds.
Similar content being viewed by others
Introduction
Carbon-hydrogen (C−H) functionalization has been established as a powerful strategy to construct diverse natural products and organic molecules in synthetic chemistry1,2,3,4,5. However, achieving site- and stereo-selective functionalization remains a formidable challenge, especially among aliphatic (sp3) C−H bonds (the dissociation energy of C−H bond shown in Supplementary Fig. 1)6,7,8,9,10,11. Two chemical procedures are developed to address these challenges. One relies on the substrate which is based on inherent reactivity or contains a directing group, or the reaction is conducted intramolecularly. The other relies on chemical catalysts, which use ligand architecture or spatial orientation to control selectivity, often involving transition metals (Fig. 1a)6,9,12,13,14,15. In most cases, the site-selectivity is predictable according to the reaction mechanism, but it is still hard to achieve high enantio-selectivity. By contrast, enzymatic catalysis often shows excellent selectivity by fitting a specific C−H bond of substrate in the active site10,13,14,16,17,18,19,20,21,22,23,24. However, it is difficult to predict this selectivity without precise geometrical information from enzyme-substrate complex structure and detailed energetic calculations. Thus, studying the mechanism of enzymatic C−H activation to provide a predictable selectivity model is highly desirable.
a, The general chemical synthetic approach of C−H activation. b, Examples of enzymes to control C−H activation in natural product biosynthesis. 1) Two highly homologous Rieske oxygenases hydroxylate the substrate on different sites by the shape of the active pocket. 2) Two αKGDs (non-heme Fe(II)/α-ketoglutarate-dependent dioxygenase) chlorinate lysine at C-4 and C-5, respectively. 3) Three homologous αKGDs catalyze three different types of C−H bond activation, BCM is the abbreviation of biclomycin.
Numerous enzymes have been identified in various natural product biosynthetic pathways, which perform different C−H functionalization reactions such as hydroxylation, desaturation, halogenation, and alkylation. Enzymatic C−H hydroxylation that exhibits extraordinary selectivity is often catalyzed by several metalloenzymes, including non-heme Fe(II)/α-ketoglutarate-dependent dioxygenase (αKGD), cytochrome P450 monooxygenase, Rieske oxygenase and copper monooxygenase7,16,17,19,25,26,27,28,29,30,31,32,33,34,35,36. Some of them were well characterized, for example, two Rieske oxygenases SxtT and GxtA, which share 88% sequence identity, catalyze site-selective C−H hydroxylation reactions in paralytic shellfish toxin biosynthesis (Fig. 1b1)37,38. Two αKGDs, BesD and HalB, could chlorinate lysine at C-4 and C-5, respectively (Fig. 1b2)39. Several αKG-dependent amino acid hydroxylases can hydroxylate the same amino acid at different sites. For example, VioC, OrfP, CmnC, and EFEs catalyze the hydroxylation of l-Arg or its analogs, while P3H and P4H selectively oxidize the C−H bonds of proline40,41,42,43,44,45,46,47,48,49. The substrates involved in the above reactions are highly functionalized or serve as prerequisites for subsequent oxidation reactions. Chemically, precise spatial and chiral recognition of a directing template now enables the iterative C−H editing of special substrate (quinoline pharmacophores) at any desired site and order50. In contrast, the example of homologous enzymes that catalyze C−H hydroxylation with different site-selectivities in the same biosynthetic pathway is extremely rare. Three αKGDs (BcmE/BcmC/BcmG) catalyze site- and stereo-selective hydroxylation of inert C−H bonds in BCM’s (bicyclomycin) cyclic dipeptide scaffold (Fig. 1b3), one of the simplest peptide derivatives without additional functional groups51,52,53,54,55. Despite extensive prior characterization of the BCM pathway51,52,53,54,55, the selectivity mechanisms of these enzymes remain unexplored.
In this work, we constructed truncated thoezyme models based on reported spectroscopic, crystallographic and computational data in αKGDs22,56,57,58,59,60,61,62,63,64,65 to investigate the inherent site selectivity of cognate substrates (1–3). Contrary to expectations, our analysis uncover divergent selectivity profiles between substrates 1 and 3 relative to thoezymatic system. Through integrating crystallographic studies, computational simulations and biochemical experiments, we identify key residues modulating selectivity of three αKGDs. Our hypothesis posits distinct regulatory mechanisms: BcmE employs steric hindrance, BcmC relies on inherent reactivity, while BcmG utilizes a directing group to control C–H hydroxylation. Substrate scope validation confirms this framework, each enzyme exhibits characteristic substrate preferences consistent with their proposed strategies. This study provides a systematic mechanistic basis for understanding how αKGDs (BcmE/BcmC/BcmG) achieve orthogonality in site selectivity during BCM biosynthesis.
Results
BcmE and BcmG employ different strategies to control site selectivity compared to BcmC
The mechanism of aliphatic C−H hydroxylation by αKGDs has been investigated thoroughly, with processing via hydrogen abstraction and radical rebound, and the rate-limiting step is the former22,42,66,67,68,69,70,71,72. To reveal the inherent site selectivity of hydroxylation, we conducted density functional theory (DFT) calculations on the cognate substrates (1–3) catalyzed by a truncated theozyme model, which contained FeIV-oxo, two methylimidazoles for two histidine residues, two acetate anions in aspartic acid and succinate, as well as one water (Fig. 2a). Given the activation free energy of hydrogen abstraction on the C-2’ (5.1 kcal mol−1) is the lowest for BcmC, we predicted the site selectivity of BcmC-catalyzed hydroxylation depends primarily on the innate reactivity of substrate. Indeed, BcmC selectively hydroxylates the C-2’ position of 2. However, the lowest energy barrier on the C-2’ (6.4 kcal mol−1) is lower than that on the C-7 for BcmE, while the lowest energy barrier on the C-5 (5.3 kcal mol−1) is lower than that on the C-3’ for BcmG (Fig. 2b and Supplementary Figs. 2–4). As expected, substrate 1-3, catalyzed by a theozyme without the protein scaffold, has the highest activity of tertiary carbon, C-2’, C-2’, and C-5, respectively. In fact, BcmC targets the most reactive position (C-2’) but BcmE and BcmG react with sites that are intrinsically less reactive (C-7 not C-2’, C-3’ not C-5), indicating that the specific enzyme scaffold is another essential factor influencing the regioselectivity of the substrate39,45,47,61,73,74,75,76
a, Computations at the CPCM (chlorobenzene)-B3LYP-D3/6-311 + G(2 d,p)[SDD for Fe]//CPCM-B3LYP-D3/6-31 G(d,p)[LanL2DZ for Fe] level of theory and transition state structures (carbon: gray, hydrogen: white, oxygen: red, nitrogen: blue, iron: cyan). b, The energy barriers of every C−H bond are listed in the table, and the reaction sites are highlighted.
Structural characterizations of BcmE, BcmC and BcmG
To investigate the effect of the microenvironment of these three enzymes on the site selectivity, we attempted to determine the crystal structures of SsBcmE, SsBcmC, and SsBcmG from Streptomyces sapporonensis. Despite the extensive screening, we only obtained the binary or ternary complexes crystals of SsBcmE and SsBcmC, with Fe(II) or/and αKG, but failed for that of SsBcmG. We subsequently selected the homologous enzymes and finally obtained the crystals of the ternary complexes of PaBcmG (from Pseudomonas aeruginosa), which also catalyzed the conversion of substrate 3 to 4 (Supplementary Fig. 5). As a result, we determined a collection of X-ray crystal structures including ternary complexes of SsBcmE, SsBcmC, PaBcmG (with Fe(II) or/and αKG) (Supplementary Table 1). These three structures shared the conserved “jelly-roll” core42,66 and possessed the HXD-Xn-H motif42,77 coordinated with iron (Supplementary Figs. 6a–c). Although these three enzymes share only 35–42% amino acid identities (Supplementary Fig. 7), the overall structures are similar, and r.m.s.d. values are 0.70 Å (SsBcmE vs SsBcmC), 1.07 Å (SsBcmE vs PaBcmG) and 0.87 Å (SsBcmC vs PaBcmG), respectively (Supplementary Fig. 6d). The primary distinction among the three structures lies in the C-terminal helix and the shape of active pocket (Supplementary Fig. 6d). The C-terminal helix is too flexible to be modeled, especially that of SsBcmE. And the B-factor of that is also much higher than that of other parts of structures (Supplementary Fig. 6d). To further determine the key residues controlling the site selection, we tried to solve the quaternary complex structures of SsBcmE, SsBcmC and PaBcmG with their corresponding substrates. Except for solving the quaternary complex of PaBcmG•FeII•αKG•3, we did not get the other two complex structures using SsBcmE and SsBcmC.
We tried many ways to determine the quaternary complex structure of SsBcmE•FeII•αKG•1 by various approaches: expressing the protein in distinct expression host, utilizing diverse fusion tags, homologous enzymes and setting crystallization experiments at a range of temperatures with multiple concentration gradients and mutant variants. Eventually, we solved the structure of SsBcmET307A•FeII•αKG•1 (PDB ID: 8XHY) (Fig. 3a, Supplementary Fig. 8a and Supplementary Table 1), which could catalyze the hydroxylation at C-5a (vide infra, Supplementary Fig. 9). The carbonyl group of the amide part of the DKP ring is stabilized by a π-π stacking interactions with the aromatic ring F273. Additionally, Y308 forms hydrogen bonds with the NH group of the DKP ring and the hydroxyl group of Y120. Two water molecules mediate the hydrogen interactions between carbonyl groups of DKP and residues R111 and D275. S198 interacts with D275 by a hydrogen bond, too. Another water molecule mediates the hydrogen bonding network between D275 and H175. The 5-methyl group of the DKP ring is oriented towards the iron center, suggesting that hydrogen abstraction is likely to occur preferentially at the C-5a position. This observation is consistent with the activity of SsBcmET307A (Supplementary Fig. 9). The presence of a rotatable single bond between C-4 and C-5 brings the possibility of C-7 closer to the iron center, thereby preserving the reactivity observed in the wild-type SsBcmE (Supplementary Fig. 9). Given that the mutant complex structure did not fully exhibit the characteristics of the wild-type protein complex, we modeled the structure of wild-type SsBcmE based on the mutant quaternary complex structure and docked substrate 1 to the active sites. We found the steric hindrance of T307 could bring C-7 closer to the iron center (Supplementary Figs. 9, 10a). Residues M177, V304, and A300 form the hydrophobic portion of the cavity to accommodate the hydrophobic portion of 1 (Fig. 3a). In our molecular dynamics (MD) snapshot, we observed that specific residues are arranged around substrate 1, forming steric hindrance (Supplementary Fig. 10a). This arrangement suggests that these residues may play a crucial role in substrate binding and significantly influence the selectivity of hydroxylation. Therefore, we hypothesize that the selectivity of BcmE-catalyzed C − H oxidation stems from the steric effect formed by the catalytic cavity.
a–c The active pocket of the complex structures of SsBcmET307A•FeII•αKG•1 (cyan), SoBcmC•FeII•αKG•2 (pink) and PaBcmG•FeII•αKG•3 (blue), respectively. The above is the overall cartoon structure and the below is the detailed interactions of corresponding substrates and residues (iron: orange, nitrogen: blue, oxygen: red) shown as sticks. The omit map (blue mesh) for substrates in active site is contoured to 3.0σ. The carbon atoms of αKG, 1, 2 and 3 are colored in yellow, magenta, cyan and orange, respectively.
Similarly to above, we determined the quaternary complex structure of SoBcmC•FeII•αKG•2 (PDB ID: 8XHQ) (Figs. 3b, Supplementary Fig 8b and Supplementary Table 1) by crystallizing the homologous enzyme SoBcmC from Streptomyces ossamyceticus. The DKP ring of substrate 2 is locked in the hydrophobic cavity formed by F245, F276 and F288. Two carbonyl groups of the DKP ring are involved in the formation of direct hydrogen bonds with the imino group of W281 and the water-mediated hydrogen bonds with the amide group of N277, respectively. Steric interaction between W149 and Ile moiety of 2 causes the 7-OH of 2 to move upward, forming a hydrogen bond with T170 at a distance of 2.7 Å. Notably, the distance between C-3’ and the iron center (5.0 Å) seems closer than that of C-2’ (5.7 Å), which contradicts the hydroxylation selectivity of BcmC (Fig. 2b). We speculated that the crystal structure only captured one state of 2, and the single C-1’ − C-2’ bond of 2 is rotatable which could rotate to the state that C-2’ directs to the iron center, facilitating the hydroxylation reaction. This is in agreement with the results of DFT (Supplementary Fig. 3) and MD calculations (Supplementary Fig. 10b).
The quaternary complex structure of PaBcmG•FeII•αKG•3 (PDB ID: 8XHX) (Fig. 3c, Supplementary Fig. 8c and Supplementary Table 1) reveals that the DKP ring in substrate 3 is securely positioned above F96, L280 and L284 through van der Waals interactions. The DKP ring interacts with F248 via π-π stacking, while its two carbonyl groups form hydrogen bonds with T156 and S175, respectively. The 7-OH of 3 forms water-mediated hydrogen bonds with H154 and R81, whereas the 2’-OH group forms a hydrogen bond with the phenolic hydroxyl group of Y288 which couples with the amine group of Q94. Furthermore, the C-3’ is closer to the iron center, which is consistent with the result that BcmG catalyzed the C-3’ hydroxylation (Supplementary Fig. 10c). Likewise, the comparison between PaBcmG•FeII•αKG and PaBcmG•FeII•αKG•3 structures revealed that the C-terminal helix may function as a gate or lid for the active pocket39,78, and the binding of substrate triggers the closure of this gate (Supplementary Fig. 11e, f). Our results suggest that the interactions between PaBcmG residues and substrate 3 control the substrate to be in a specific orientation for hydroxylation, which effectively reverses its inherent site selectivity.
BcmE, BcmC and BcmG use three different strategies to control the site selectivity of C − H hydroxylation
Based on the structural analyzes above, we employed MD simulations and site-directed mutagenesis experiments to investigate the mechanisms behind the site selectivity and stereoselectivity of C−H hydroxylation by SsBcmE, SoBcmC and PaBcmG. We supposed that these three enzymes controlled the site selectivity by modulating the active conformation to favor the desired site of hydroxylation. According to the DFT-optimized structures of transition states, we defined the active conformation that would achieve site-specific hydroxylation as O − H distance is ≤ 2.8 Å (the sum of van der Waals atomic radii of carbon and oxygen atoms) and angle O − H − C is in the range of 170 ± 15°. We constructed the reactive species configurations of SsBcmE•FeIV-oxo•succinate•1, SoBcmC•FeIV-oxo•succinate•2 and PaBcmG•FeIV-oxo•succinate•3, and obtained the corresponding representative MD snapshots (Supplementary Fig. 10d–f).
Among 5000 conformations of 1 in the SsBcmE•FeIV-oxo•succinate•1 complex, there are only ten active conformations for C-7 hydroxylation, while there are no active conformations for other sites (Supplementary Fig. 10a–d). We hypothesized that a constrained cavity composed of bulky residues Y120, F273, V304, T307, and Y308 forced C-7 in 1 into a pre-reactive conformation through van der Waals interactions. Consistent with our assumption, site-directed mutagenesis showed that Y105A almost lost the enzyme activity, as well as Y120A, M177A, V304A, and Y308A drastically decreased it, indicating the critical role of these residues (Supplementary Fig. 9). Conversely, mutations Y120F, V304T, T307A, and Y308F enhanced total activities (72%–95% total conversion yields) and produced a new C-5a-hydroxylated product 1a (2%–27% conversion yields, Supplementary Fig. 9). Surprisingly, the mutants T307L (62% conversion yield, 100% selectivity) and Y308M (21% conversion yield, 100% selectivity) completely reverse the site selectivity of C − H hydroxylation, yielding a C-2’-hydroxylated product 1b, akin to the reaction catalyzed by BcmC. MD simulations (Supplementary Fig. 12c, d) indicated that the larger side chain of T307L increased steric hindrance with substrate 1, altering its binding and favoring the reaction at the lower energy barrier of C-2’. Given the structural proximity of Y308M to T307L, we hypothesize that Y308M may similarly affect substrate binding. Additionally, the mutation F273A completely terminates the C-7 hydroxylation activity but exhibits the C-6 hydroxylation to afford 1c (58% conversion yield, 100% selectivity, Supplementary Fig. 9). These findings suggest that the residues F273, T307 and Y308 in SsBcmE play a crucial role in regulating the site selectivity. We also modeled mutants F273A, T307L, and T307A complexes by using a pre-equilibrated SsBcmE•FeIV-oxo•succinate•1 complex structure as a template (Supplementary Fig. 12a–c, e). As expected, the dominant active conformations for hydroxylation in the F273A, T307L and T307A systems are C-6, C-2’ and C-5a (shown in the dotted rectangle: 49/5000, 90/5000 and 14/5000, Supplementary Fig. 12b–d, f), respectively, which is consistent with our mutagenesis results and structures, particularly the crystal structure of SsBcmET307A•FeII•αKG•1. The active conformations for C-7 in these three mutants account for only 3/5000, 3/5000, and 4/5000, respectively (Supplementary Fig. 12b–d, f), further supporting the idea that steric hindrance mediated by these residues controls site selectivity in SsBcmE-catalyzed hydroxylation.
Together, we propose that BcmE regulates site selectivity by constructing a specific hydrophobic cavity that accommodates the orientation of substrate 1, employing a steric hindrance control strategy similar to those used in chemical synthesis (Fig. 4a). For instance, mutating the bulky F273 to alanine reduces the steric hindrance on the leucine side of substrate 1, positioning the preferred reactive site, C-6, closer to the center and facilitating hydroxylation. This observation is consistent with the reactivity predicted in Fig. 2b.
a–d The corresponding mechanisms (left, cutview of active pockets) and their concept maps (right). In the right panels, we present simplified models of the active pockets to clarify the selectivity mechanisms. It’s important to note that these models do not represent the full enzymes, and the iron centers are embedded within the active pockets, not exposed on the enzyme surface.
In SoBcmC•FeIV-oxo•succinate•2 complex, we found that C-2’ hydroxylation is the dominant active conformation (shown in the dotted rectangle: 122/5000, in Supplementary Fig. 10e), which is accordant with the DFT calculation result that the activation free energy of hydrogen abstraction on the C-2’ is the lowest for BcmC-catalyzed hydroxylation reaction. Structural analyzes demonstrate that residues F245, F276 and F288 orient the DKP ring in substrate 2 into the active cavity via hydrophobic interactions. However, the single-site mutations of F245A, F276A and F288A have almost no effect on the activity and selectivity (Supplementary Fig. 13), probably because the cavity of the SoBcmC is too large to affect the substrate binding and orientation (volume of SoBcmC is 252 Å3, but that of SsBcmE and PaBcmG are 139 Å3 and 210 Å3, respectively). The W149A mutation dramatically reduces the activity, owing to W149 may lock the substrate in the active pocket (Fig. 3b). In addition, mutagenesis of T170, N277 and W281, which constituted the hydrogen bond network with substrate 2, resulted in the reduction of C-2’ hydroxylation activity (19%–75% conversion yields) but performed the C-5 hydroxylation activity (4%–23% conversion yields) to generate product 2a (Supplementary Fig. 13), which is also the inherently more reactive site with the lowest activation free energy (Fig. 2b). Consequently, these findings suggest that the site selectivity of BcmC-catalyzed hydroxylation mainly depends on the innate selectivity of substrate 2, in which one C − H bond is inherently more reactive than the other bonds. Such a substrate control strategy (Fig. 4c) is commonly used in chemical synthesis (Fig. 1a2).
In PaBcmG•FeIV-oxo•succinate•3 complex, the active C-3’ hydroxylation conformation is dominant (shown in the dotted rectangle: 327/5000, Supplementary Fig. 10f). However, the above DFT calculations manifested that the activation free energy of hydrogen abstraction on the C-3’ is the highest in the BcmG theozyme model (Fig. 2b). Combined with the crystal structure analyzes, we speculated that the hydrogen bonding interactions medicated by R281, Q94, H154, T156, S175 and Y288 and the hydrophobic interactions mediated by F96, F248, L280 and L284 stabilize and orient the substrate 3 in a suitable conformation to favor the C-3’ hydroxylation. To investigate our assumptions of the putative key residues, we performed systematic site-directed mutagenesis of PaBcmG (Supplementary Fig. 14). The mutations F96A, T156A, S175A, L280A and F248A substantially decrease the activities, and the mutant L284A completely lost its activity (Supplementary Fig. 14), suggesting these residues play a crucial role in enzyme activity. Similarly, the H154A and R81A variants reduce the activity, indicating that the water-mediated hydrogen bonds with 7-OH contribute to the substrate binding for hydroxylation. Moreover, the hydrogen bonding interactions between Y288 (also interact with Q94) and 2’-OH drives C-3’ in substrate 3 to adopt a catalytically active conformation for hydroxylation (Supplementary Fig. 14). Expectedly, the activity of Y288A is completely abolished and the activity of Q94A is markedly diminished. More importantly, we found that the substitution of Y288 to F and W reversed the C-2’ diastereoselectivity to produce a diastereomeric product 3a (100% selectivity) and its five-membered derivates 3b/3c (Supplementary Fig. 14). All these bioactive results further confirmed the role of these key residues. Collectively, we speculate that BcmG performs exquisite control over the site selectivity and stereoselectivity mainly through the directed guidance mediated by the hydrogen bond between Y288 and the 2’-OH of 3, similar to a directing group control strategy in chemical catalysis (Fig. 4d).
Mechanism-guided selective C–H hydroxylation of 2,5-diketopiperazines
Based on the above results, we have obtained three enzymes that can hydroxylate the inert C–H bond from CH, CH2 and CH3 (Fig. 5a and Supplementary Fig. 15), respectively. The different site selectivity strategies of BcmE/BcmC/BcmG also inspired an investigation of whether these αKGDs would activate C–H bonds of other cyclodipeptides. To this end, several cyclodipeptides condensed from hydrophobic amino acids were used as substrates to test the catalytic selectivity of these three enzymes (Fig. 5b, c). As expected, BcmE only recognized isoleucyl-containing cyclodipeptides including 5, 6, and 7, and hydroxylated the C-7 position of the isoleucine moiety of these compounds to yield 5a (59% conversion yield), 6a (12% conversion yield) and 7a (16% conversion yield), indicating that BcmE exhibited high regioselectivity (up to 100%) and substrate specificity toward these unactivated isoleucyl-containing DKP scaffolds (Figs. 5b and Supplementary Fig. 16). BcmC could specifically hydroxylate the tertiary carbons (C-2’) of the leucine moiety in leucyl-containing cyclodipeptides (1 and 8–10) with high catalytic activity (up to 100% conversion yield) and excellent selectivity (up to 100%) (Figs. 5c and Supplementary Fig. 17). As calculated above (Fig. 2), these results also confirm that the site-selectivity of BcmC-catalyzed hydroxylation depends mainly on the innate reactivity of the substrate. For BcmG, it preferentially catalyzes the hydroxylation of DKP compounds (2 and 8a–10a) with a hydroxyl group in vicinal position with high regioselectivity (up to 100%) and activity (Figs. 5c and Supplementary Fig. 18). For example, BcmG was able to hydroxylate the natural intermediate 2 at C-6 to form 2b, further supporting the importance of the directing group (i.e., the ortho hydroxyl group) for the hydroxylation site-selectivity of BcmG (Figs. 5c and Supplementary Fig. 15 and Supplementary Fig. 18). Depending on the reaction characteristics, we speculated that the C−H hydroxylation of 2,5-diketopiperazines could be programmed by these mutually orthogonal enzymes to obtain various hydroxylated products (Fig. 5c). Collectively, these data further support the proposed regulatory mechanism that BcmE, BcmC, and BcmG precisely tune site-selective C−H hydroxylations by using different strategies (Fig. 4a, c and d). This also implies that the αKGDs BcmE, BcmC, and BcmG diversified the enzyme toolkit, providing excellent biocatalysts for site-selective and stereoselective C–H hydroxylation of 2,5-diketopiperazines.
a SsBcmE-, SsBcmEF273A- and SoBcmC-catalyzed site-selective C–H activations. b SsBcmE-catalyzed regioselective C–H hydroxylation of 2,5-diketopiperazines. c SsBcmE-, SoBcmC-, and PaBcmG-catalyzed the programmable C–H hydroxylation for the synthesis of diverse 2,5-diketopiperazines. aThe percentage refers to regioselectivity. bThe percentage refers to diastereoselectivity.
Discussion
Activating inert C–H bonds while controlling stereo- and regioselectivity is a significant challenge in chemical synthesis25,50. Many catalysts have been developed to tackle this issue, but they typically rely on substrate control, such as the intrinsic reactivity of the substrate (including electronic, steric, and stereoelectronic effects) or pre-installed directing groups2,10, or the architecture of chemical catalyst. All this procedure involves the installation and removal of the catalyst’s directing group or coordinated ligands, which adds steps to experiments2. In contrast, natural enzymes employ different strategies to discriminate between similar C–H bonds and activate them during natural product biosynthesis. Arguably, nature exemplifies the power of these approaches, where a cumulative set of interactions between the functionalities present in the substrate and the active site imbues nature’s enzymes with their exquisite selectivity79.
In this study, we investigated the mechanisms employed by three homologous enzymes, BcmE, BcmC, and BcmG, which catalyze the hydroxylation of 1–3 sequentially. We specially investigated the residues in the second coordinate sphere of these three enzymes and found out it was consistent with previous research, that the second coordinate sphere determines the substrate recognition and selectivity80,81,82,83. Our analysis revealed that these enzymes utilize mutually orthogonal strategies to achieve regio- and stereo-selectivity: steric hindrance control, substrate innate control, and directing group control. Further experiments with expanded set of substrates demonstrated that BcmE can preferentially activate CH3 group, BcmC targets the C–H bond with the lowest bond dissociation energy, and BcmG favors the site near the directing hydroxy group. The active assays of mutants showed that engineered BcmE or BcmG could gain new selectivity against WT.
These findings provide valuable insights into how the αKGDs distinguish closely related substrates and catalyze regioselective hydroxylation in BCM biosynthesis. Moreover, the mutually orthogonal mechanisms of these three enzymes ensure the predictable and programable aliphatic C–H activation to hydroxylate different sites on substrates. This offers a promising approach for future applications in chemical synthesis and the development of novel biocatalysts. Continued advancement of this approach may lead to the ultimate realization of molecular editing: the freedom to modify organic molecules at any site, in any order.
Methods
General
Chemicals, biological reagents, media, enzymes and Kits were purchased from standard commercial sources unless otherwise stated. Compounds 1–4, 1b, 1 d, 4a, and 4b were previously reported by our laboratory51,84,85. Compounds 5–10 were purchased from Leon (Nanjing) Biotechnology Co., Ltd. Oligonucleotide primers were synthesized, and DNA was sequenced by TsingKe Biotech Co., Ltd. (Shanghai, China). PCR amplification was performed using a C1000 Touch™ Thermal cycler (Bio-Rad, USA). All the protein quantification was measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific Inc., USA). High performance liquid chromatography (HPLC) analysis was performed on an Agilent 1260 instrument (Agilent Technologies, Germany). Liquid chromatography coupled with mass spectrometry (LC-MS) analysis was performed on a Thermo Fisher LTQ Fleet mass spectrometer (Thermo Fisher Scientific) equipped with an electrospray ionization (ESI) source. Enzymatic products were isolated by semi-preparative HPLC on a Shimadzu LC-20-AT system (Shimadzu Corporation, Japan). High resolution (HR) ESI-MS data were obtained on a Bruker maXis 4 G instrument. NMR spectra were recorded on Agilent 500/54 Premium Shielded 500 MHz NMR spectrometer and Bruker AscendTM 600 MHz spectrometer with CD3OD as the solvent. Chemical shifts (δ) are given in parts per million (ppm) and coupling constants (J) are given in hertz (Hz).
Protein expression and purification
The glycerol stock (E. coli BL21(DE3) strain) harboring plasmid pET-37b-SsbcmE was previously reported by our laboratory51,84,85. The gene encoding SobcmC from Streptomyces ossamyceticus and PabcmG from Pseudomonas aeruginosa were codon-optimized for heterologous expression in E. coli BL21 (DE3) and synthesized by TsingKe Biotech Co., Ltd. (Shanghai, China). The synthesized genes were subsequently cloned into the pET-37b(+) vector (pre-linearized with Nde I and Xho I). The verified recombinant plasmids were transformed into E. coli BL21 (DE3) competent cells (TransGen Biotech, China) for protein expression. A single clony was inoculated into 10 mL Luria-Bertani (LB) medium supplemented with 50 μg mL−1 kanamycin at 37 °C overnight. Then, 8 mL of an overnight seed culture was used to inoculate 800 mL of LB media supplemented with 50 μg mL−1 kanamycin. Cells were grown at 37 °C with shaking (220 rpm) until an optical density at 600 nm (OD600) reached 0.6–0.8. Protein expression was induced by addition 0.4 mM isopropyl β-D-thiogalactoside (IPTG), followed by incubation at 16 °C for another 20 h with shaking (200 rpm). The cells were harvested by centrifugation at 7,600 g for 5 min at 4 °C, resuspended in 30 mL of pre-cooled lysis buffer (50 mM Tris-HCl, 300 mM sodium chloride, and 10 mM imidazole; pH 7.5), and disrupted by sonication in an ice bath. After centrifugation at 63,000 g for 15 min at 4 °C, the clarified supernatant was incubated with 3 mL of Ni-NTA resin for 1 h at 4 °C with gently shaking. The mixture was then loaded onto a column, and the resin was washed with 30 mL of lysis buffer. The bound recombinant protein was eluted using a stepwise gradient of imidazole (20 mM, 50 mM, and 300 mM) in elution buffers (50 mM Tris-HCl, 300 mM sodium chloride). The purified protein was concentrated and buffer-exchanged into storage buffer (50 mM Tris-HCl, 150 mM sodium chloride; pH 7.5) using an Amicon Ultra-30K (Millipore, USA). Protein purity was confirmed by SDS-PAGE, and protein concentration was determined by measuring absorbance at 280 nm using a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific). The concentration values were corrected using the theoretical extinction coefficient calculated by the ProtParam tool on the Expasy server (https://web.expasy.org/protparam/). Finally, the purified proteins were flash-frozen in liquid nitrogen and stored at –80 °C for further experiments.
Enzyme activity assays and HPLC and LC-MS analyzes
Each enzymatic reaction (50-μL scale) consisted of 50 mM Tris-HCl buffer (pH 7.5), 0.6 mM substrate, 2 mM α-Ketoglutaric acid (αKG), 2 mM l-ascorbic acid, 0.1 mM FeSO4·7H2O, and 50 μg of purified enzyme. The reactions were performed at 37 °C for SoBcmC and PaBcmG, and at 16 °C for SsBcmE. The reaction times were 12 h for SsBcmE, 2 h for SoBcmC, and 1 h for PaBcmG. A heat-inactivated enzyme (boiled at 100 °C for 15 min) served as the negative control. After the reactions, 100 μL of pre-cooled methanol was added to quench the reactions, followed by centrifugation at 13,800 g for 30 min. The supernatants were then analyzed by HPLC and LC-MS. HPLC and LC-MS analyzes were performed according to our previous methods84,85.
Site-directed mutagenesis and enzyme activity assay of mutants
Site-directed mutants were constructed using a Fast Mutagenesis System kit (TransGen Biotech) with the corresponding primers listed in Supplementary Table 2–4. Protein expression, purification, and enzyme activity assays of all these mutants were performed according to the same protocol described for the wild-type (WT) enzymes.
Protein expression and purification for crystallization
Protein SsBcmE was expressed in E. coli BL21(DE3) in 2×YT medium supplemented with 50 μg mL−1 kanamycin at 37 °C. When OD600 reached 0.8–1.0, protein expression was induced by 0.4 mM IPTG at 16 °C for 18 h. Cells were harvested by centrifugation at 7600 g for 3 min at 4 °C. The cells were resuspended in pre-cooled buffer A containing 25 mM Tris (pH 8.0), 500 mM NaCl, 5 mM β-mercaptoethanol, and 1 mM phenylmethylsulfonyl fluoride (PMSF), and lysed by French press with a high-pressure homogenizer (500–600 bar). Cell debris was removed by centrifugation at 63,000 g for 15 min at 4 °C. The supernatant was applied to a pre-equilibrated Ni-NTA affinity column (GE Healthcare, USA), sequentially washed with 100 mL buffer A, followed by an additional wash with a buffer consisting of 25 mM Tris (pH 8.0), 500 mM NaCl, 30 mM imidazole, and 5 mM β-mercaptoethanol. The His-tagged protein was eluted using buffer B, which consisted of 25 mM Tris (pH 8.0), 500 mM NaCl, 300 mM imidazole, and 5 mM β-mercaptoethanol. The purified protein was concentrated using an Amicon Ultra-30KDa (Millipore), and subsequently loaded to a pre-equilibrated Hi Load Superdex 75 column (GE Healthcare) with a running buffer of 25 mM Tris (pH 8.0), 150 mM NaCl, and 2 mM dithiothreitol (DTT). The pure fractions were then collected and concentrated for crystallization. Proteins SsBcmC, SoBcmC, SsBcmG, PaBcmG and mutant SsBcmET307A were overexpressed and purified following the same procedure described above.
Protein crystallization
The initial crystal screening ultilized commercial kits (QIAGEN, Hampton Research, Rigaku, and Molecular Dimensions), while all optimization reagents were obtained from Sigma-Aldrich. Crystals were obtained using the sitting drop vapor diffusion method at 20 °C. In general, 1 μL protein sample (10 mg mL–1) was mixed with 1 μL reservoir solution and equilibrated against 50 μL reservoir solution. The stored protein was diluted by buffer containing 25 mM Tris (pH 8.0), 150 mM NaCl and 2 mM DTT. The substrate free crystals was gained by incubated with 5 mM αKG and the complex crystals was incubated with 5 mM αKG and 1 mM corresponding substrate on ice for 30 min. The binary complex SsBcmE (with FeII) crystals were grown in 2.5 M Sodium chloride and 0.1 M imidazole hydrochloric acid pH 8.0, and the crystals appeared after almost two months at 20 °C. After trying a lot but failed to determine the complex of SsBcmE, we crystalized several of its mutants and homologous BcmE, and gained mutant SsBcmET307A complex crystals in anaerobic box after almost two weeks, the condition was 1.26 M Sodium phosphate monobasic monohydrate, 0.14 M Potassium phosphate dibasic, pH 5.6. Crystallization of the SsBcmET307A complex involved the following steps: First, the aerobically purified protein was placed in an anaerobic chamber. Then, its buffer was replaced with an oxygen-free solution of the same composition. Additionally, the cosubstrate αKG and substrate 1 were prepared by dissolving them in water and DMSO, respectively, ensuring that the oxygen was removed before use. The screening process was the same as above mentioned. The ternary SsBcmC (with FeII and αKG) was grown in 0.1 M MES/imidazole pH6.5, 0.03 M of each divalent cation (0.3 M magnesium chloride, 0.3 M calcium chloride) and 10% PEG 20000, 20% PEG MME 550. We could not get any high quality SsBcmC complex crystals, so we tried the homolog proteins from S. ossamyceticus and P. aeruginosa. The crystals of SoBcmC•FeII•αKG•2 were grown in 0.2 M Ammonium chloride, 0.1 M Na acetate pH 5, 20% PEG 6000. The crystals of BcmG were also gained by homologous protein from P. aeruginosa. The crystals of ternary PaBcmG (with Fe(II) and αKG)were grown in 10% w/v PEG 4000, 20% v/v glycerol 0.03 M of each NPS (0.3 M sodium nitrate, 0.3 M disodium hydrogen phosphate, 0.3 M ammonium sulfate), 0.1 M MES/imidazole pH 6.5. The crystals of PaBcmG•FeII•αKG•3 were grown in 0.1 M MES pH 6.5, 1.6 M magnesium sulfate. All crystals were flash-frozen in liquid nitrogen using the reservoir solution containing 20% (v/v) glycerol as cryo-protectant.
Data collection and structure determination
All X-ray diffraction data were collected at the Shanghai Synchrotron Radiation Facility (SSRF). For SsBcmE•FeII data were collected at the wavelength of 0.9792 Å in beamline BL17U1. Data of SsBcmET307A•FeII•αKG•1, SoBcmC•FeII•αKG•2, SsBcmC•FeII•αKG and PaBcmG•FeII•αKG were collected at the wavelength of 0.9785 Å in beamline BL19U1. Data of PaBcmG•FeII•αKG•3 were collected at the wavelength of 0.9792 Å in beamline BL18U1. Data reduction and integration were achieved with HKL300086 and the XDS87 software package. The statistics for data collection are listed in Supplementary Table 1. The phase of SsBcmE•FeII was solved by molecular replacement in the Phenix-Phaser program88 using 7V3O as initial searching model. The complex structures SsBcmET307A•FeII•αKG•1, SsBcmC•FeII•αKG and PaBcmG•FeII•αKG were determined by molecular replacement using Phaser with SsBcmE•FeII structure as the initial search model. The quaternary complex structure of SoBcmC and PaBcmG were determined by molecular replacement using Phaser with corresponding substrate free crystal structure as the initial search model. Iterative cycles of model building and refinement were performed in Coot89 and Phenix90, respectively. Refinement statistics for each final model are recorded in Supplementary Table 1. Structure Figures were drawn using PyMol 2.4.1 (Schrodinger, LLC)91. For the crystal structure of apo SsBcmE was not intact with some loops and fragments were missing, we built its structure by AlphaFold2 (AF2) (https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb) and found it was very similar to both apo and complex SsBcmE, with r.m.s.d. value 0.57 Å and 0.82 Å, respectively (Supplementary Figs. 6e–g). So we using the AF2 model as the apo structure for the next docking or comparison work. The complex model of SsBcmEF273A•FeII•αKG•1 and SsBcmE•FeII•αKG•1 was built by Autodock vina92. The volume of active pockets was calculated by online sever (http://altair.sci.hokudai.ac.jp/g6/service/pocasa/).
Quantum chemical computations
Quantum chemical calculations were performed using Gaussian 16 package93. Geometry optimizations were conducted at the function of B3LYP-D394, and Fe was described via the effective core potential (ECP) of the LanL2DZ95 and all the other atoms with the 6-31 G(d,p) basis set. Vibrational frequency calculations were carried out at the same level to ensure that all of the stationary points were transition states (one imaginary frequency) or minima (no imaginary frequency) and to evaluate zero-point vibrational energies (ZPVE) and thermal corrections at 298 K. Intrinsic reaction coordinates (IRC) calculations were conducted for all transition states to confirm their connectivity between corresponding minima. Single-point energy calculations were carried out at the B3LYP-D3 level with the SDD96 for Fe, and the 6-311 + G(2 d,p) basis set for all the other atoms. The B3LYP functional within density functional theory (DFT) has been extensively validated for studies of αKGDs57,60,62,64,97,98,99,100. Notably, high-level ab initio calculations were also performed and compared with the DFT results by Neese101 and Borowski22,102 groups to study the dioxygen activation process of αKGDs. Neese et al.60,101 found that B3LYP results show excellent agreement with the ab initio calculations. And Borowski et al.102 revealed that CCSD(T) benchmark calculations support the dispersion corrected B3LYP (B3LYP-D3) spin-state energies. The CPCM103,104,105 solvation model was used with chlorobenzene as the solvent that approximates protein environment99,106,107.
Molecular dynamics simulations
All molecular dynamics (MD) simulations were conducted using the Amber 16 package108. The force field parameters for the iron-containing active site were generated using the Metal Center Parameter Builder (MCPB.py)109 as implemented in Amber 16 package. Bond and angle force constants were calculated using the Seminario method110, while atomic point charges were derived from electrostatic potentials using the ChgModB method. The MCPB tools have been successfully used for description of mononuclear non-heme iron center57,64,111,112. Substrates 1–3 were optimized at the B3LYP-D3/6-31 G(d,p) level using Gaussian 16. The partial charges for these ligands were fitted with HF/6-31 G(d) calculations and the restrained electrostatic potential (RESP)113,114 protocol implemented by the Antechamber module in Amber 16 package. The force field parameters for these three ligands were adapted from the standard general amber force field 2.0 (gaff2)115 parameters. Meanwhile, the protein was described using the standard Amber14SB force field. Each system was neutralized by the adding of Na+ ions and subsequently solvated in a rectangular TIP3P116 water box, with a buffer distance of 10 Å maintained on each side. Each system underwent equilibration through a series of minimizations interspersed with short MD simulations. During this process, restraints on the protein backbone heavy atoms were gradually released, with force constant decreasing from 10, 2, 0.1 and finally 0 kcal mol–1 Å–2. The system was then gradually heated from 0 to 300 K over 50 ps, during which a restraint of 10 kcal mol–1 Å–2 was maintained on the protein backbone heavy atoms. Subsequently, a 10 ns MD simulation was conducted at constant temperature and pressure, with the protein backbone heavy atoms restrained by a force constant of 5 kcal mol–1 Å–2. Finally, a standard 10 ns MD simulation was performed without restraints. The pressure was maintained at 1 atm using isotropic position scaling, while the temperature was kept at 300 K through Berendsen thermostat method. Long-range electrostatic interactions were handled using the Particle Mesh Ewald (PME)117 method, with a cutoff distance of 12 Å applied to both PME and van der Waals (vdW) interactions. A time step of 2 fs was employed, in conjunction with SHAKE algorithm applied to hydrogen atoms. Periodic boundary condition was also utilized. The stability of each system, including structure, energy, and temperature fluctuations, was thoroughly examined. Additionally, the convergence of the simulations was assessed by monitoring the root-mean-square deviation (r.m.s.d.) of the structures.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Atomic coordinates of SsBcmET307A•FeII•αKG•1, SsBcmC•FeII•αKG, SoBcmC•FeII•αKG•2, PaBcmG•FeII•αKG, PaBcmG•FeII•αKG•3, have been deposited in the Protein Data Bank (PDB) under accession codes 8XHY, 8XHP, 8XHQ, 8XHT, 8XHX (https://www.rcsb.org/). The PDB 7V3O could be download from PDB Database (https://doi.org/10.2210/pdb7V3O/pdb). The DNA sequences of genes SsbcmE, SsbcmC, SsbcmG, SobcmC and PabcmG are downloaded from GenBank with accession numbers MG018995, WP_055518629 and HGP0145857 and the codon-optimized DNA sequences are listed in Supplementary Information. Other data supporting the conclusions of this study are presented in the main text, Supplementary Information and from corresponding author(s) upon request. Source data are provided with this paper.
References
Abrams, D. J., Provencher, P. A. & Sorensen, E. J. Recent applications of C−H functionalization in complex natural product synthesis. Chem. Soc. Rev. 47, 8925–8967 (2018).
Sinha, S. K. et al. Toolbox for distal C−H bond functionalizations in organic molecules. Chem. Rev. 122, 5682–5841 (2022).
Guillemard, L., Kaplaneris, N., Ackermann, L. & Johansson, M. J. Late-stage C−H functionalization offers new opportunities in drug discovery. Nat. Rev. Chem. 5, 522–545 (2021).
Fazekas, TimothyJ. et al. Diversification of aliphatic C–H bonds in small molecules and polyolefins through radical chain transfer. Science 375, 545–550 (2022).
Labinger, J. A. & Bercaw, J. E. Understanding and exploiting C−H bond activation. Nature 417, 507–514 (2002).
Goswami, N., Bhattacharya, T. & Maiti, D. Transient directing ligands for selective metal-catalysed C–H activation. Nat. Rev. Chem. 5, 646–659 (2021).
Newmister, S. A. et al. Molecular basis of iterative C−H oxidation by tami, a multifunctional P450 monooxygenase from the tirandamycin biosynthetic pathway. ACS Catal. 10, 13445–13454 (2020).
Sarkar, S., Cheung, K. P. S. & Gevorgyan, V. C−H functionalization reactions enabled by hydrogen atom transfer to carbon-centered radicals. Chem. Sci. 11, 12974–12993 (2020).
Liao, K. et al. Site-selective and stereoselective functionalization of non-activated tertiary C−H bonds. Nature 551, 609–613 (2017).
Zhang, K. et al. Biocatalytic enantioselective beta-hydroxylation of unactivated C−H bonds in aliphatic carboxylic acids. Angew. Chem. Int. Ed. 61, e202204290 (2022).
Hong, B., Luo, T. & Lei, X. Late-stage diversification of natural products. ACS Cent. Sci. 6, 622–635 (2020).
Li, J. et al. Site-specific allylic C−H bond functionalization with a copper-bound N-centred radical. Nature 574, 516–521 (2019).
Roy, L. Theoretical identification of the factors governing the reactivity of C−H bond activation by non-heme iron(IV)-oxo complexes. Chempluschem 84, 893–906 (2019).
He, J., Wasa, M., Chan, K. S. L., Shao, Q. & Yu, J. Q. Palladium-catalyzed transformations of alkyl C−H bonds. Chem. Rev. 117, 8754–8786 (2017).
Lam, N. Y. S. et al. Empirical guidelines for the development of remote directing templates through quantitative and experimental analyses. J. Am. Chem. Soc. 144, 2793–2803 (2022).
Chakrabarty, S., Wang, Y., Perkins, J. C. & Narayan, A. R. H. Scalable biocatalytic C−H oxyfunctionalization reactions. Chem. Soc. Rev. 49, 8137–8155 (2020).
Borovik, A. S. Role of metal-oxo complexes in the cleavage of C−H bonds. Chem. Soc. Rev. 40, 1870–1874 (2011).
Lewis, J. C., Coelho, P. S. & Arnold, F. H. Enzymatic functionalization of carbon-hydrogen bonds. Chem. Soc. Rev. 40, 2003–2021 (2011).
Huang, X. & Groves, J. T. Beyond ferryl-mediated hydroxylation: 40 years of the rebound mechanism and C−H activation. J. Biol. Inorg. Chem. 22, 185–207 (2017).
Li, J., Tang, X., Awakawa, T. & Moore, B. S. Enzymatic C−H oxidation-amidation cascade in the production of natural and unnatural thiotetronate antibiotics with potentiated bioactivity. Angew. Chem. Int. Ed. 56, 12234–12239 (2017).
Narayan, A. R. H. et al. Enzymatic hydroxylation of an unactivated methylene C−H bond guided by molecular dynamics simulations. Nat. Chem. 7, 653–660 (2015).
Wojdyla, Z. & Borowski, T. Properties of the reactants and their interactions within and with the enzyme binding cavity determine reaction selectivities. the case of Fe(II)/2-oxoglutarate dependent enzymes. Chem. Eur. J. 28, e202104106 (2022).
Tao, H. et al. Molecular insights into the unusually promiscuous and catalytically versatile Fe(II)/alpha-ketoglutarate-dependent oxygenase SptF. Nat. Commun. 13, 95 (2022).
Siitonen, V. et al. Divergent non-heme iron enzymes in the nogalamycin biosynthetic pathway. Proc. Natl Acad. Sci. USA 113, 5251–5256 (2016).
Münch, J., Püllmann, P., Zhang, W. & Weissenborn, M. J. Enzymatic hydroxylations of sp3-carbons. ACS Catal. 11, 9168–9203 (2021).
Fasan, R. Tuning P450 enzymes as oxidation catalysts. ACS Catal. 2, 647–666 (2012).
Sydor, P. K. et al. Regio- and stereodivergent antibiotic oxidative carbocyclizations catalysed by Rieske oxygenase-like enzymes. Nat. Chem. 3, 388–392 (2011).
Peng, J. et al. A mechanistic understanding of the distinct regio- and chemoselectivity of multifunctional P450s by structural comparison of IkaD and CftA complexed with common substrates. Angew. Chem. Int. Ed., 62, e202310728 (2023).
Wang, Z. et al. Cytochrome P450 mediated cyclization in eunicellane derived diterpenoid biosynthesis. Angew. Chem. Int. Ed., 62, e202312490 (2023).
Zhang, R. K., Huang, X. & Arnold, F. H. Selective C−H bond functionalization with engineered heme proteins: new tools to generate complexity. Curr. Opin. Chem. Biol. 49, 67–75 (2019).
Matthews, M. L. et al. Substrate-triggered formation and remarkable stability of the C−H bond-cleaving chloroferryl intermediate in the aliphatic halogenase, SyrB2. Biochemistry 48, 4331–4343 (2009).
Ushimaru, R. et al. Substrate conformation correlates with the outcome of hyoscyamine 6 beta-hydroxylase catalyzed oxidation reactions. J. Am. Chem. Soc. 140, 7433–7436 (2018).
Kim, H. J. et al. Biosynthesis of spinosyn in saccharopolyspora spinosa: synthesis of permethylated rhamnose and characterization of the functions of SpnH, Spnl, and SpnK. J. Am. Chem. Soc. 132, 4996–4997 (2010).
Dunham, N. P. & Arnold, F. H. Nature’s machinery, repurposed: expanding the repertoire of iron-dependent oxygenases. ACS Catal. 10, 12239–12255 (2020).
Voss, M., Malca, S. H. & Buller, R. Exploring the biocatalytic potential of Fe/alpha-ketoglutarate-dependent halogenases. Chemistry 26, 7336–7345 (2020).
Espinoza, R. V. et al. Engineering P450 TamI as an iterative biocatalyst for selective late-stage C−H functionalization and epoxidation of tirandamycin antibiotics. ACS Catal. 11, 8304–8316 (2021).
Lukowski, A. L., Liu, J., Bridwell-Rabb, J. & Narayan, A. R. H. Structural basis for divergent C−H hydroxylation selectivity in two Rieske oxygenases. Nat. Commun. 11, 2991 (2020).
Liu, J. X. et al. Design principles for site-selective hydroxylation by a Rieske oxygenase. Nat. Commun. 13, 255 (2022).
Kissmana, ElijahN. et al. Biocatalytic control of site-selectivity and chain lengthselectivity in radical amino acid halogenases. Proc. Natl Acad. Sci. USA 120, e2214512120 (2023).
Zwick, C. R. & Renata, H. Overview of amino acid modifications by iron- and α-ketoglutarate-dependent enzymes. ACS Catal. 13, 4853–4865 (2023).
Wu, L. et al. Molecular Insights into the regioselectivity of the Fe(II)/2-ketoglutarate-dependent dioxygenase-catalyzed C–H hydroxylation of amino acids. ACS Catal. 12, 11586–11596 (2022).
Islam, M. S., Leissing, T. M., Chowdhury, R., Hopkinson, R. J. & Schofield, C. J. 2-oxoglutarate-dependent oxygenases. Annu. Rev. Biochem. 87, 585–620 (2018).
de Visser, S. P. et al. Negative catalysis/non-bell-evans-polanyi reactivity by metalloenzymes: examples from mononuclear heme and non-heme iron oxygenases. Coord. Chem. Rev. 439, 213914 (2021).
Klein, C. & Huttel, W. Tertiary alcohol preferred: Hydroxylation of trans-3-methyl-L-proline with proline hydroxylases. Beilstein J. Org. Chem. 7, 1643–1647 (2011).
Ali, H. S. & de Visser, S. P. Catalytic divergencies in the mechanism of L-arginine hydroxylating nonheme iron enzymes. Front. Chem. 12, 1365494 (2024).
Hsiao, Y. H. et al. Crystal structure of the α-ketoglutarate-dependent non-heme iron oxygenase CmnC in capreomycin biosynthesis and its engineering to catalyze hydroxylation of the substrate enantiomer. Front. Chem. 10, 1001311 (2022).
Ali, H. S., Henchman, R. H., Warwicker, J. & de Visser, S. P. How do electrostatic perturbations of the protein affect the bifurcation pathways of substrate hydroxylation versus desaturation in the nonheme iron-dependent viomycin biosynthesis enzyme? J. Phys. Chem. A 125, 1720–1737 (2021).
Mitchell, A. J. et al. Visualizing the reaction cycle in an iron(II)- and 2-(Oxo)-glutarate-dependent hydroxylase. J. Am. Chem. Soc. 139, 13830–13836 (2017).
Peters, C. & Buller, R. M. Industrial application of 2-oxoglutarate-dependent oxygenases. 9, 221 (2019).
Fan, Z. et al. Molecular editing of aza-arene C−H bonds by distance, geometry and chirality. Nature 610, 87–93 (2022).
Meng, S. et al. A Six-oxidase cascade for tandem C–H bond activation revealed by reconstitution of bicyclomycin biosynthesis. Angew. Chem. Int. Ed. 57, 719–723 (2018).
Patteson, J. B., Cai, W., Johnson, R. A., Maria, K. C. S. & Li, B. Identification of the biosynthetic pathway for the antibiotic bicyclomycin. Biochemistry 57, 61–65 (2018).
Chekan, J. R. & Moore, B. S. Biosynthesis of the antibiotic bicyclomycin in soil and pathogenic bacteria. Biochemistry 57, 897–898 (2018).
Vior, N. M. et al. Discovery and biosynthesis of the antibiotic bicyclomycin in distantly related bacterial classes. Appl. Environ. 84, 2828–2844 (2018).
Witwinowski, J. et al. Study of bicyclomycin biosynthesis in Streptomyces cinnamoneus by genetic and biochemical approaches. Sci. Rep. 9, 20226 (2019).
Rifayee, S. et al. Catalysis by KDM6 histone demethylases - a synergy between the non-heme iron(II) center, second coordination sphere, and long-range interactions. Chemistry 29, e202301305 (2023).
Chaturvedi, S. S. et al. Catalysis by the non-heme iron(II) histone demethylase phf8 involves iron center rearrangement and conformational modulation of substrate orientation. ACS Catal. 10, 1195–1209 (2020).
Yeh, C. C. G. et al. Cluster model study into the catalytic mechanism of α-ketoglutarate biodegradation by the ethylene-forming enzyme reveals structural differences with nonheme iron hydroxylases. ACS Catal. 12, 3923–3937 (2022).
Chaturvedi, S. S. et al. Dioxygen binding is controlled by the protein environment in non-heme Fe(II) and 2-oxoglutarate oxygenases: a study on histone demethylase PHF8 and an ethylene-forming enzyme. Chemistry 29, e202300138 (2023).
Aboelnga, M. M. & Gauld, J. W. Establishing a substrate-assisted mechanism for the pre-transfer editing in SerRS and IleRS: a QM/QM investigation. Struct. Chem. 35, 519–530 (2024).
Ali, H. S., Warwicker, J. & de Visser, S. P. How does the nonheme iron enzyme napi react through l-arginine desaturation rather than hydroxylation? a quantum mechanics/molecular mechanics study. ACS Catal. 13, 10705–10721 (2023).
de Visser, S. P., Mukherjee, G., Ali, H. S. & Sastri, C. V. Local charge distributions, electric dipole moments, and local electric fields influence reactivity patterns and guide regioselectivities in α-ketoglutarate-dependent non-heme iron dioxygenases. Acc. Chem. Res. 55, 65–74 (2022).
Hardy, F. J. et al. Probing ferryl reactivity in a nonheme iron oxygenase using an expanded genetic code. ACS Catal. 14, 11584–11590 (2024).
Waheed, S. O. et al. Role of structural dynamics in selectivity and mechanism of non-heme Fe(II) and 2-oxoglutarate-dependent oxygenases involved in DNA Repair. ACS Cent. Sci. 6, 795–814 (2020).
Thomas, M. G. et al. The unique role of the second coordination sphere to unlock and control catalysis in nonheme Fe(II)/2-oxoglutarate histone demethylase KDM2A. Inorg. Chem. 63, 10737–10755 (2024).
Martinez, S. & Hausinger, R. P. Catalytic mechanisms of Fe(II)-and 2-oxoglutarate-dependent oxygenases. J. Biol. Chem. 290, 20702–20711 (2015).
Hausinger, R. P. FeII/alpha-ketoglutarate-dependent hydroxylases and related enzymes. Crit. Rev. Biochem. Mol. Biol. 39, 21–68 (2004).
Bugg, T. D. H. Oxygenases: mechanisms and structural motifs for O2 activation. Curr. Opin. Chem. Biol. 5, 550–555 (2001).
Dunham, N. P. et al. Two distinct mechanisms for C–C desaturation by Iron(II)- and 2-(Oxo)glutarate-dependent oxygenases: importance of alpha-heteroatom assistance. J. Am. Chem. Soc. 140, 7116–7126 (2018).
Herr, C. Q. & Hausinger, R. P. Amazing diversity in biochemical roles of Fe(II)/2-oxoglutarate oxygenases. Trends Biochem. Sci. 43, 517–532 (2018).
Gao, S. S., Naowarojna, N., Cheng, R., Liu, X. & Liu, P. Recent examples of α-ketoglutarate-dependent mononuclear non-haem iron enzymes in natural product biosyntheses. Nat. Prod. Rep. 35, 792–837 (2018).
Wu, L. F., Meng, S. & Tang, G. L. Ferrous iron and alpha-ketoglutarate-dependent dioxygenases in the biosynthesis of microbial natural products. Biochim. Biophys. Acta -Proteins Proteom. 1864, 453–470 (2016).
Wenger, E. S. et al. Optimized substrate positioning enables switches in the C–H cleavage site and reaction outcome in the hydroxylation-epoxidation sequence catalyzed by hyoscyamine 6β-hydroxylase. J. Am. Chem. Soc. 146, 24271–24287 (2024).
Mitchell, A. J. Structural basis for alternative reaction outcome in iron and 2-oxoglutarate-dependent oxygenases. PhD thesis, Pennsylvania State University (2017).
Karamzadeh, B., Kumar, D., Sastry, G. N. & de Visser, S. P. Steric factors override thermodynamic driving force in regioselectivity of proline hydroxylation by prolyl-4-hydroxylase Enzymes. J. Phys. Chem. A 114, 13234–13243 (2010).
Matthews, M. L. et al. Substrate positioning controls the partition between halogenation and hydroxylation in the aliphatic halogenase, SyrB2. Proc. Natl Acad. Sci. USA 106, 17723–17728 (2009).
Schofield, Zhihong Zhang, C. J. Structural and mechanistic studies on 2-oxoglutarate-dependent oxygenases and related enzymes. Curr. Opin. Struct. Biol. 9, 722–731 (1999).
Helmetag, V., Samel, S. A., Thomas, M. G., Marahiel, M. A. & Essen, L. O. Structural basis for the erythro-stereospecificity of the L-arginine oxygenase VioC in viomycin biosynthesis. FEBS J. 276, 3669–3682 (2009).
Meng, G. et al. Achieving site-selectivity for C–H activation processes based on distance and geometry: a carpenter’s approach. J. Am. Chem. Soc. 142, 10571–10591 (2020).
Fernandez, R. L., Juntunen, N. D. & Brunold, T. C. Differences in the second coordination sphere tailor the substrate specificity and reactivity of thiol dioxygenases. Acc. Chem. Res. 55, 2480–2490 (2022).
Roy, B. & Govindaraju, T. Enzyme-mimetic catalyst architectures: the role of second coordination sphere in catalytic activity. Bulletin of the Chemical Society of Japan 97, bcsj.20230224 (2023).
Hangasky, J. A., Taabazuing, C. Y., Valliere, M. A. & Knapp, M. J. Imposing function down a (cupin)-barrel: secondary structure and metal stereochemistry in the αKG-dependent oxygenases. Metallomics 5, 287–301 (2013).
Schenk, G., Neidig, M. L., Zhou, J., Holman, T. R. & Solomon, E. I. Spectroscopic characterization of soybean lipoxygenase-1 mutants: the role of second coordination sphere residues in the regulation of enzyme activity. Biochemistry 42, 7294–7302 (2003).
He, J.-B. et al. Enzymatic catalysis favours eight-membered over five-membered ring closure in bicyclomycin biosynthesis. Nat. Catal. 6, 637–648 (2023).
Liu, Y.-P. et al. Establishing modular cell-free expression system for the biosynthesis of bicyclomycin from a chemically synthesized cyclodipeptide. Chin. J. Chem. 42, 7 (2024).
Minor, W., Cymborowski, M., Otwinowski, Z. & Chruszcz, M. HKL-3000: the integration of data reduction and structure solution-from diffraction images to an initial model in minutes. Acta Crystallogr D. Biol. Crystallogr 62, 859–866 (2006).
Kabsch, W. XDS. Acta Crystallogr D. Biol. Crystallogr 66, 125–132 (2010).
Sheldrick, G. M. Experimental phasing with SHELXC/D/E: combining chain tracing with density modification. Acta Crystallogr D. Biol. Crystallogr 66, 479–485 (2010).
Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr D. Biol. Crystallogr 60, 2126–2132 (2004).
Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D. Biol. Crystallogr 66, 213–221 (2010).
PyMoL Molecular Graphics System, v.2 release 2020-5 (Schrödinger, 2020).
Eberhardt, J., Santos-Martins, D., Tillack, A. F. & Forli, S. AutoDock Vina 1.2.0: new docking methods, expanded force field, and python bindings. J. Chem. Inf. Model 61, 3891–3898 (2021).
Frisch. M. J. et al. Gaussian 16 Rev. A.03 (Wallingford, CT, 2016).
Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 132, 154104 (2010).
Hay, P. J. & Wadt, W. R. Ab initio effective core potentials for molecular calculations. potentials for K to Au including the outermost core orbitals. J. Chem. Phys. 82, 299–310 (1985).
Dolg, M., Wedig, U., Stoll, H. & Preuss, H. Energy‐adjusted ab initio pseudopotentials for the first row transition elements. J. Chem. Phys. 86, 866–872 (1987).
Song, X., Lu, J. & Lai, W. Mechanistic insights into dioxygen activation, oxygen atom exchange and substrate epoxidation by AsqJ dioxygenase from quantum mechanical/molecular mechanical calculations. Phys. Chem. Chem. Phys. 19, 20188–20197 (2017).
Su, H., Sheng, X., Zhu, W., Ma, G. & Liu, Y. Mechanistic insights into the decoupled desaturation and epoxidation catalyzed by dioxygenase AsqJ involved in the biosynthesis of quinolone alkaloids. ACS Catal. 7, 5534–5543 (2017).
Ghafoor, S., Mansha, A. & de Visser, S. P. Selective hydrogen atom abstraction from dihydroflavonol by a nonheme iron center is the key step in the enzymatic flavonol synthesis and avoids byproducts. J. Am. Chem. Soc. 141, 20278–20292 (2019).
Álvarez-Barcia, S. & Kästner, J. Atom tunneling in the hydroxylation process of taurine/α-ketoglutarate dioxygenase identified by quantum mechanics/molecular mechanics simulations. J. Phys. Chem. B 121, 5347–5354 (2017).
Ye, S. et al. Electronic structure analysis of the oxygen-activation mechanism by Fe(II)- and α-ketoglutarate (αKG)-dependent dioxygenases. Chem. Eur. J. 18, 6555–6567 (2012).
Wójcik, A., Radoń, M. & Borowski, T. Mechanism of O2 activation by α-ketoglutarate dependent oxygenases revisited. a quantum chemical study. J. Phys. Chem. A 120, 1261–1274 (2016).
Barone, V. & Cossi, M. Quantum calculation of molecular energies and energy gradients in solution by a conductor solvent model. J. Phys. Chem. A 102, 1995–2001 (1998).
Takano, Y. & Houk, K. N. Benchmarking the conductor-like polarizable continuum model (CPCM) for aqueous solvation free energies of neutral and ionic organic molecules. J. Chem. Theory Comput. 1, 70–77 (2005).
Cossi, M., Rega, N., Scalmani, G. & Barone, V. Energies, structures, and electronic properties of molecules in solution with the C-PCM solvation model. J. Comput. Chem. 24, 669–681 (2003).
Schyman, P., Lai, W., Chen, H., Wang, Y. & Shaik, S. The directive of the protein: how does cytochrome P450 select the mechanism of dopamine formation? J. Am. Chem. Soc. 133, 7977–7984 (2011).
Wu, X., Chen, Y., Wang, X., Wei, W. & Liang, Y. Origin of site selectivity in toluene hydroxylation by cytochrome P450 Enzymes. J. Org. Chem. 86, 13768–13773 (2021).
Case, D. A. et al. Amber 2016, University of California, San Francisco. (2016).
Li, P. & Merz, K. M. Jr. MCPB.py: a python based metal center parameter builder. J. Chem. Inf. Model 56, 599–604 (2016).
Seminario, J. M. Calculation of intramolecular force fields from second-derivative tensors. Int. J. Quantum Chem. 60, 1271–1277 (1996).
Singh, W., Quinn, D., Moody, T. S. & Huang, M. Reaction mechanism of histone demethylation in αKG-dependent non-heme iron enzymes. J. Phys. Chem. B 123, 7801–7811 (2019).
Liu, G. et al. Structure-guided insights into heterocyclic ring-cleavage catalysis of the non-heme Fe (II) dioxygenase NicX. Nat. Commun. 12, 1301 (2021).
Jakalian, A., Jack, D. B. & Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. parameterization and validation. J. Comput. Chem. 23, 1623–1641 (2002).
Jakalian, A., Bush, B. L., Jack, D. B. & Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: I. Method. J. Comput. Chem. 21, 132–146 (2000).
Wang, J., Wolf, R. M., Caldwell, J. W., Kollman, P. A. & Case, D. A. Development and testing of a general amber force field. J. Comput. Chem. 25, 1157–1174 (2004).
Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W. & Klein, M. L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 926–935 (1983).
Darden, T. A., York, D. M. & Pedersen, L. G. Particle mesh Ewald: An N⋅log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993).
Acknowledgements
This work was supported by grants from Jiangsu Basic Research Center for Synthetic Biology Grant (BK20233003 to J.Z.), the National Key Research and Development Program of China (2022YFC2303100 to G.-L.T. and 2022YFA1503200 to Y.L.), the National Natural Science Foundation of China (22207117 to J.-B.H., 22077062 to Y.L. and 21825804 to S.Y.), the Natural Science Foundation of Jiangsu Province (BK20230018 to Y.L.) and the Shenzhen Science and Technology Program (ZDSYS20210623091810032 to J.Z.). The authors also thank the staff members of beam lines BL17U1 of Shanghai Synchrotron Radiation Facility, BL18U1 (https://cstr.cn/31129.02.NFPS.BL18U1) and BL19U1 (https://cstr.cn/31129.02.NFPS.BL19U1) of National Facility for Protein Science in Shanghai for access and help with the X-ray data collection. We thank the High Performance Computing Center (HPCC) of Nanjing University for doing the numerical calculations in this paper on its blade cluster system.
Author information
Authors and Affiliations
Contributions
S.Y., Y.L., G.-L.T. and J.Z. conceived the project. L.W. performed all crystallographic studies. J.-B.H. performed all in vitro biochemical experiments. W.W. conducted all computational studies. H.-X.P. and X.W. performed protein purification. All authors analyzed and discussed the results. L.W., J.-B.H., W.W., S.Y., Y.L., G.-L.T., and J.Z. prepared the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Source data
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Wu, L., He, JB., Wei, W. et al. Three distinct strategies lead to programmable aliphatic C−H oxidation in bicyclomycin biosynthesis. Nat Commun 16, 4651 (2025). https://doi.org/10.1038/s41467-025-58997-8
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41467-025-58997-8
