Coordinated conformational changes in P450 decarboxylases enable hydrocarbons production from renewable feedstocks

Generoso, Wesley Cardoso; Alvarenga, Alana Helen Santana; Simões, Isabelle Taira; Miyamoto, Renan Yuji; Melo, Ricardo Rodrigues de; Guilherme, Ederson Paulo Xavier; Mandelli, Fernanda; Santos, Clelton Aparecido; Prata, Rafaela; Santos, Camila Ramos dos; Colombari, Felippe Mariano; Morais, Mariana Abrahão Bueno; Pimentel Fernandes, Rodrigo; Persinoti, Gabriela Felix; Murakami, Mario Tyago; Zanphorlin, Leticia Maria

doi:10.1038/s41467-025-56256-4

Download PDF

Article
Open access
Published: 22 January 2025

Coordinated conformational changes in P450 decarboxylases enable hydrocarbons production from renewable feedstocks

Nature Communications volume 16, Article number: 945 (2025) Cite this article

9114 Accesses
4 Citations
40 Altmetric
Metrics details

Subjects

Abstract

Fatty acid peroxygenases have emerged as promising biocatalysts for hydrocarbon biosynthesis due to their ability to perform C-C scission, producing olefins - key building blocks for sustainable materials and fuels. These enzymes operate through non-canonical and complex mechanisms that yield a bifurcated chemoselectivity between hydroxylation and decarboxylation. In this study, we elucidate structural features in P450 decarboxylases that enable the catalysis of unsaturated substrates, expanding the mechanistic pathways for decarboxylation reaction. Combining X-ray crystallography, molecular dynamics simulations, and machine learning, we have identified intricate molecular rearrangements within the active site that enable the Cβ atom of the substrate to approach the heme iron, thereby promoting oleate decarboxylation. Furthermore, we demonstrate that the absence of the aromatic residue in the Phe-His-Arg triad preserves chemoselectivity for alkenes, providing a distinct perspective on the molecular determinants of decarboxylation activity. Ultimately, these findings enable the sustainable production of biohydrocarbons from industrial feedstocks.

Iron-catalyzed aliphatic C–H functionalization to construct carbon–carbon bonds

Article Open access 20 May 2025

Engineered cytochrome P450 for direct arylalkene-to-ketone oxidation via highly reactive carbocation intermediates

Article Open access 13 July 2023

Manganese catalyzed oximation of hydrocarbons to oximes

Article Open access 09 January 2026

Introduction

Olefins (alkenes) are key intermediates and commodity chemicals in the industrial production of polymers, plastics, surfactants, plasticizers, and fuels^1,2. Due to the C-C double bond, olefins play central roles in further synthesizing active pharmaceutical ingredients, biological probes, and advanced functional materials^2,3. Given the critical position of olefins derived from petroleum extraction, significant efforts have been directed toward developing strategies to synthesize these compounds from renewable sources^{1,3,4,5,6,7,8,9,10,11}.

Over the last decade, a distinctive class of enzymes termed P450 fatty acid decarboxylases have attracted great attention because of their ability to produce olefins^12,13,14,15, acting on the removal of the fatty acyl carboxyl group (Supplementary Fig. 1). These biocatalysts possess non-canonical and elaborate mechanisms, which involve hydrogen abstraction, controlled electrons and protons delivery, and bifurcated chemoselectivity, representing an eminent field for investigation^{13,15,16,17,18,19,20,21,22,23}.

Recently, a decarboxylase (OleTP_RN) active on unsaturated fatty acids, especially oleic acid, the most abundant fatty acid in nature²⁴ and inhibitor of previously characterized CYP152 peroxygenases^12,25 was discovered, representing an important step toward the biocatalytic synthesis of olefins¹². In addition, in previous work, it has been demonstrated that the productive substrate binding involves a motif named hydrophobic cradle that modulates enzyme specificity and facilitates product release¹². Despite the relevance of these findings for olefin biosynthesis by CYP152 peroxygenases, the understanding of the mechanism behind the non-canonical decarboxylation reaction by P450 enzymes remains elusive, primarily due to the absence of decarboxylase structures complexed with unsaturated substrates.

Here, we report the crystallographic structure of a CYP152 decarboxylase bound to different substrates, unveiling structural adaptations in the catalytic pocket involved in the modulation of (chemo)specificity. Using structure-guided mutagenesis, molecular dynamics simulations, and unsupervised machine learning, we elucidated the molecular trajectory of oleic acid as it reaches productive binding for the decarboxylation reaction, driven by conformational changes in phenylalanine within the hydrophobic cradle. These coordinated movements, involving the aromatic residue, modify the architecture at the end of the catalytic pocket to properly accommodate the substrate, then push the Cβ atom of oleic acid toward the catalytic heme-iron, thereby enabling alkene production. Our work also modifies the current paradigm regarding the strict requirement of the Phe-His-Arg triad, strategically positioned in the catalytic site near the heme group, for Cβ-regioselectivity and oxidative decarboxylation. We observed that a natural substitution of the phenylalanine by a valine retains the alkene formation capability for both saturated and unsaturated fatty acids. To demonstrate the potential application of these findings, we evaluated olefin production from distillers corn oil (DCO), a byproduct rich in oleic and linoleic acids, widely produced in corn mills in the USA and Brazil. Notably, the process yielded approximately 1.70 g/L of hydrocarbons with a turnover number (TTN) of 1470 for the decarboxylation reaction.

Taken together, these results provide unparalleled mechanistic insights into unsaturated fatty acid decarboxylation by P450 peroxygenases that can foster the development of sustainable and bio-based routes for the synthesis of olefins.

Results

Sequence similarity network uncovered fatty acid decarboxylases harboring distinct substrate specificities

Sequence similarity network (SSN) analysis was performed using the recently discovered CYP152 member from Rothia nasimurium (OleTP_RN) as a seed to systematically explore the sequence repertoire of fatty acid decarboxylases with activity against unsaturated fatty acids, resulting in four well-defined isofunctional clusters, as shown in Fig. 1a. Based on previously characterized CYP152 enzymes, members from cluster I were classified as primarily acting as hydroxylases, as demonstrated by the activity of CYP152s from Bacillus subtilis²⁶ and Sphingomonas paucimobilis²⁷. Members from clusters II and III primarily act as fatty acid decarboxylases, according to the peroxygenases OleTP_RN¹² and OleT_JE¹⁸, respectively (Supplementary Figs. 2, 3). By further investigating cluster II, which possesses putative oleate-consuming OleTs, we identified sequences from Corynebacterium and Kocuria species, which genera are known as alkene producers and/or diesel-contaminated soil degraders²⁸. Although the sequence similarity was relatively low with OleTP_RN, ranging from 36 to 58% (Supplementary Fig. 4 and Supplementary Table 1), these members showed conservation of some important residues considered essential for alkene production, such as F84, H90, and R246 (OleTP_RN numbering). Interestingly, CYP152 enzymes from Corynebacterium genus possess a natural variation, presenting a valine in substitution of F84 (Supplementary Fig. 5). Thus, we selected the member from C. lipophiloflavum to explore its potential for fatty acid decarboxylation. In addition, we also selected the member from K. marina, based on the high conservation of the hydrophobic cradle, a recently reported motif involved in oxidative decarboxylation. Despite the conservation of key molecular motifs for decarboxylation activity among P450 members, predicting their activity and specificity is challenging due to the low sequence similarity and the scarcity of functional data of orthologs. In addition, their genomic locations do not follow the same context as either OleTP_RN or OleT_JE (Supplementary Fig. 6).

**Fig. 1: Sequence similarity classification and substrate specificity.**

The peroxygenases, herein named as OleTP_CL (C. lipophiloflavum) and OleT_KM (K. marina), were recombinantly expressed and characterized (Supplementary Figs. 7–10 and Supplementary Table 2), disclosing that OleTP_CL binds fatty acids more tightly (Supplementary Table 2). Both enzymes primarily catalyzed fatty acid decarboxylation, producing alkenes as the main product for all tested substrates (Supplementary Table 3), followed by hydroxylation at C_β and C_α, respectively (Fig. 1b, c), further supporting the classification of these enzymes as fatty acid decarboxylases (cluster II). Surprisingly, while OleT_KM demonstrated a higher alkene ratio for saturated substrates, only OleTP_CL was able to decarboxylate oleic acid efficiently, producing 82% alkene among the recovered products (Supplementary Table 3). Furthermore, OleTP_CL is capable of efficiently converting linoleic (C18:2) and linolenic acids (C18:3), independently of the mixture within unsaturated substrates, yielding high (> 70%) alkene production (Supplementary Fig. 11). This finding is particularly relevant because oleic and linoleic acids are ubiquitous in nature, comprising over 70% of the fatty acids in most vegetable oils²⁹, and over 80% in the oleaginous yeast Yarrowia lipolytica ³⁰. In contrast, the presence of oleic acid impaired the activity of OleT_KM against myristic acid (Fig. 1d). Given that OleT_KM binds oleic acid with high affinity (Supplementary Table 2), the unsaturated substrate likely occupies the pocket without productive catalysis or product release, preventing the binding of other substrates and leading to competitive inhibition.

In summary, while OleTP_CL, OleT_KM, and OleTP_RN share similar substrate preference for medium- to long-chain fatty acids, distinct layers of molecular complexity appear to regulate substrate selectivity.

Aromatic residue at the Phe-His-Arg triad is not essential for OleTP_CL chemoselectivity

The crystal structures of the full-length OleTP_CL (1.8 Å) and OleT_KM (2.0 Å) were solved in complex with the saturated substrates, palmitic and myristic acids, respectively (Supplementary Fig. 12 and Supplementary Table 4). Both OleTP_CL and OleT_KM structures revealed a single-domain scaffold with a buried heme group as other fatty acid peroxygenases (Supplementary Fig. 13).

Although the crystallographic results did not indicate an evident oligomeric arrangement of these enzymes, hydrodynamic analysis revealed a prevalent monomeric arrangement for OleTP_CL, while OleT_KM is observed as a dimer in solution (Supplementary Fig. 14), such as OleTP_RN. OleTP_CL is found to be a monomer, most likely due to the natural substitutions of K299, Q404, and E407, essential in the OleTP_RN interface, by F293, A401, and I404 in OleTP_CL (Supplementary Fig. 14), respectively. The structural analysis also confirmed that OleT_KM harbors a well-defined hydrophobic cradle at the end of the catalytic pocket similar, but not identical, to that one of OleTP_RN, comprising the residues F16, F23, M43 and L326 (Supplementary Fig. 13). This lack of full conservation of the hydrophobic cradle may explain the narrower specificity across the different chain lengths in saturated fatty acids than OleTP_RN. In addition, the substrate positioning throughout the OleTP_CL and OleT_KM catalytic pockets are quite similar to those of other decarboxylases (Fig. 2a–c) and diverge from the configuration observed for hydroxylases, especially P450_BM and P450_SPα in which the canonical phenylalanine (OleTP_CL-F86 as reference) is not conserved (Fig. 2d–f).

**Fig. 2: Substrate binding and chemoselectivity.**

Notably, turnover data demonstrated that the native F86V substitution does not impair the fatty acid decarboxylation by OleTP_CL, and the reverse mutation to phenylalanine detains the alkene production mostly unaltered (Fig. 2g). This is intriguing since previous studies have shown that phenylalanine replacement impaired fatty acid decarboxylation of OleT_JE^17,18 and OleTP_RN¹². Indeed, the conservation of decarboxylase activity in the natural F86V variation of OleTP_CL challenges the previous understanding of the critical role played by the F86-H92-R248 (OleTP_CL numbering) triad, located close to the heme group, in alkene production by P450 enzymes. Furthermore, the presence of the V86 residue might be an evolutionary remnant of glutamine-containing CYP152, considering the higher conservation of phenylalanine at this position for the histidine-containing members (Supplementary Fig. 15). So far, investigations on the canonical histidine cannot fully attest to its role in fatty acid decarboxylation, considering that, while OleT_JE-H85Q mutant did not alter the enzyme chemoselectivity¹⁷, the corresponding mutation in OleTP_RN led to a significant predominancy of hydroxylation activity, mainly at the substrate α-carbon¹².

Interestingly, the H92Q replacement in OleTP_CL impaired chemoselectivity for saturated fatty acids, but oleic acid decarboxylation remained mostly unaffected (Fig. 2g). This suggests that residues other than glutamine may substitute histidine in proton donation, possibly through a well-organized water channel previously discussed³¹. In addition, since the decarboxylation reaction in the CYP152 family is believed to depend on both electronic factors and substrate positioning within the catalytic pocket, oleic acid may bind in a way that favors C_β regiochemistry over α-oxidation, while still maintaining the formation of the required water channel and electronic constraints for decarboxylation. Evidence for this is supported by the product profile of the wild-type enzyme, in which α-hydroxylates were not observed. Therefore, OleTP_CL demonstrates an unusual capacity to produce alkenes from unsaturated substrates, suggesting it may have evolved specifically for this purpose. These results also reinforce that the ability in decarboxylating fatty acids is not copiously dependent on the histidine and its interaction with the phenylalanine close to the heme group in CYP152 but might be influenced by additional structural elements.

Substrate complexes reveal conformational changes favoring decarboxylation

The repertoire of P450 fatty acid decarboxylases with elucidated structures is scarce, with only two enzymes complexed with either palmitic (OleTP_RN)¹² or arachidic acid (OleT_JE)¹⁸. However, arachidic acid is a rare biological substrate and no structural data are currently available for these enzymes with unsaturated substrates. In addition, OleTP_CL exhibits the capacity of producing alkenes from unsaturated substrates, despite having the natural substitution F86V of the Phe-His-Arg triad considered essential for decarboxylation reaction in P450 members. Therefore, structural studies aiming at obtaining such decarboxylase complexed with oleic acid would provide valuable insights into the molecular aspects that enable such biocatalytic reactions. Hence, the oleic acid-bound OleTP_CL structure was herein elucidated revealing that the substrate accommodation throughout the catalytic pocket as well as its configuration are substantially altered compared to the saturated substrate (Supplementary Fig. 16). Remarkably, such altered substrate conformation is associated with a large rotameric change of the aromatic residue F19 (~ 110° rotation) that leads to a subsequent rearrangement of the aliphatic residues L48 and L49 (Fig. 3a). Such conformational change of F19 is confirmed by the clear electron density maps in both crystal structure complexes obtained with saturated or unsaturated substrates (Supplementary Fig. 12).

**Fig. 3: Conformational changes upon unsaturated substrate binding.**

To further explore the relevance of these conformational changes, we then performed molecular dynamics (MD) simulations, which confirmed the intrinsic flexibility of the F19 residue in OleTP_CL. As shown in Fig. 4a–c, two distinct dihedral angles of OleTP_CL-F19 are visited, in contrast to OleT_KM which exhibited complete rigidity of the corresponding F16 residue, primarily due to its interaction with F23 (Figs. 3b and 4d–f). MD analysis revealed that these phenylalanine side chains remained close together in OleT_KM throughout all independent simulations (Supplementary Fig. 17). By contrast, OleTP_CL lacks this Phe-Phe interaction due to the native F26L substitution (Fig. 3). This aromatic interaction seems to have a key influence on protein stability since OleT_KM-F23 mutation into leucine, isoleucine, valine, or alanine could not be purified in soluble form. However, the L26F mutation in OleTP_CL significantly stabilizes protein folding in both in absence of substrate and in complex with oleic acid (Supplementary Figs. 18–20). Interestingly, in terms of catalytic performance, the OleTP_CL-L26F mutant presented around 30% reduction in total oleic acid conversion (Fig. 3c), indicating that a higher rigidity of F19 partially impairs the productive binding of unsaturated substrates in this enzyme. On the other hand, the consumption of saturated fatty acids, especially C12, was considerably enhanced, suggesting that bulkier residues contribute to increasing the hydrophobicity of the hydrophobic cradle, thereby shortening the catalytic pocket that favors the binding of shorter substrates. Nonetheless, this favored binding of shorter substrates does not appear to correlate with alkene formation, as the higher conversion led to an increased preference for the hydroxylation route by OleTP_CL-L26F in the presence of C12 (Fig. 3c).

**Fig. 4: Phenylalanine flexibility analysis by molecular dynamics (MD).**

This result is supported by functional data observed for OleT_KM, which displays high efficiency in metabolizing and producing alkenes from C12 and C10 substrates due to its shorter catalytic pocket promoted by the rigid conformation of the corresponding F19. Therefore, these results allude to the fact that the F19 mobility can modulate the substrate preference and, especially, it has a pivotal role in oleic acid decarboxylation by OleTP_CL. However, the rigidity imposed by Phe-Phe interactions is not a conserved feature associated with decreased activity on unsaturated substrates among CYP152s since OleTP_RN has such Phe-Phe interaction (Fig. 4g–i), and yet it can efficiently decarboxylate oleic acid into olefins.

Since OleTP_CL demonstrates the highest conversion of oleic acid into alkenes among the characterized P450 decarboxylases, we further explored whether this distinguished performance on unsaturated substrates is associated with the F19 flexibility in the hydrophobic cradle. As the mutations of this phenylalanine residue in both OleT_KM and OleTP_RN did not yield variants in the soluble and stable form, we then selected orthologs featuring the structural motif associated with phenylalanine flexibility including the P450s from Corynebacterium doosanense and Mycobacterium abscessus (Supplementary Fig. 21). Both OleTP_CD and OleTP_MA orthologs, which possess no Phe-Phe interaction through variation T28 and V24, respectively, were recombinantly expressed and purified, and the functional characterization revealed a high oleic acid conversion (> 90%) and alkene production similar to OleTP_CL (~ 70%). This result demonstrates that the increased flexibility of F19 is a conducive feature to alkene production from oleic acid among several P450 decarboxylases.

Hydrophobic cradle shortening impairs the accommodation of unsaturated fatty acids

The fact that only some members of the P450 superfamily are able to decarboxylate unsaturated substrates led us to in-depth investigate the structural determinants associated with substrate binding, mainly related to the pocket distal region. Concerning this, the distal F/G loop was demonstrated to play a key role in the correct binding and chemoselectivity of saturated fatty acids in OleT_JE¹³, whereas in OleTP_RN, this role is held by the hydrophobic cradle motif ¹².

The pocket architecture in OleT_KM is more similar to OleTP_CL complexed with palmitic acid than with oleic acid (Supplementary Fig. 22), implicating a shortened topology of the hydrophobic cradle (Fig. 5a). On the other hand, OleTP_CL complexed with oleic acid exhibited a prolonged pocket, which is a consequence of F19 mobility. By comparison, OleTP_RN, which is also effective in converting unsaturated substrates, encompasses a pocket dimensionally similar to the oleic acid-bound OleTP_CL (Supplementary Fig. 22).

**Fig. 5: Binding pocket (hydrophobic cradle) length and hydrophobicity.**

Considering that OleT_KM and OleTP_RN both have rigid phenylalanine (F16 and F17, respectively) in the pocket, the shorter pocket topology of OleT_KM likely compromises the performance in oleic acid decarboxylation. To verify this hypothesis, we designed the OleT_KM-L45V mutation, which naturally occurs in OleTP_RN, resulting in an increase in the total oleic acid conversion, albeit still presenting a similar alkene amount as the wildtype (Fig. 5b). However, this mutation considerably reduced the activity for saturated substrates, highlighting the importance of the hydrophobic cradle in modulating catalysis (Supplementary Fig. 23).

Further mutations aiming at enlarging the hydrophobic cradle of OleT_KM were designed including M303A and F16A. However, they did not improve alkene production (Fig. 5b), which can be attributed to the expressive reduction in hydrophobicity of the hydrophobic cradle in OleT_KM, a decarboxylase that is not rich in aromatic residues as OleTP_RN. Supporting this, the respective mutation in OleTP_RN (F17A) did not alter the oleic acid conversion into alkenes (Supplementary Fig. 24). To further explore the importance of the distal pocket topology of OleT_KM, the L326F variant was designed, which leads to increased hydrophobicity and a shortening of the hydrophobic cradle. As a result, both oleic acid conversion and alkene production were compromised (Fig. 5b). This mutational analysis indicates a clear trade-off between the pocket elongation and hydrophobicity of the hydrophobic cradle in enabling these enzymes to decarboxylate principally unsaturated substrates.

Taken together, these results reinforce the critical role of the hydrophobic cradle motif in modulating substrate binding, fatty acid conversion, and alkene formation. Notably, the mechanism for (un)saturated fatty acids decarboxylation differs even among iso-functionally clustered decarboxylases, highlighting the intricate molecular complexity underlying substrate recognition and decarboxylation in these enzymes.

Coordinated enzyme-substrate movements drive the decarboxylation of unsaturated substrates

To explore the molecular basis of substrate specificity in CYP152 decarboxylases, we analyzed MD trajectories of the OleTP_CL complexed with oleic acid (Supplementary Fig. 25). By projecting these MD trajectories into a low-dimensional space using principal component analysis (PCA), we captured highly correlated molecular motions (Supplementary Video 1). Our analysis revealed a pseudo-trajectory characterized by coordinated conformational changes in the F19 residue, oleic acid, and the heme co-factor. Specifically, the shift in the F19 dihedral angle was accompanied by a movement of the Cβ atom of oleic acid toward the catalytic heme group. Clustering the first two principal components (PC1 and PC2) identified distinct structural conformation clusters (Fig. 6a). Notably, two of these clusters featured concerted changes in substrate conformation, Fe–Cβ distance, and the rotameric configuration of F19 (Fig. 6b, c). This finding suggests that oleic acid decarboxylation is driven by substrate re-orientation within the catalytic pocket, particularly toward Cβ-regiochemistry, which is essential for hydrogen abstraction and subsequent alkene formation. The productive substrate binding involves the entrance of oleic acid into the catalytic pocket, facilitated by the opened F19 rotameric configuration (58°), which allows the accommodation of longer fatty acids. As F19 rotates to 153° (closed configuration), it imposes a substrate re-orientation into the active site, positioning the Cβ atom optimally toward the heme iron, thereby facilitating catalysis.

**Fig. 6: Principal component and clustering analyses of OleTPCL in complex with oleic acid.**

OleTP_CL-mediated bioconversion of distillers corn oil into renewable alkenes

Given the promising role of enzymatic decarboxylation in olefin production principally with the discovery of P450 decarboxylases active on unsaturated fatty acids¹², we investigated the use of the enzyme OleTP_CL to convert distillers corn oil (DCO) into alkenes. DCO is a lipid-rich byproduct produced in corn ethanol biorefineries (Supplementary Fig. 26), consisting mostly of unsaturated fatty acids (> 80%) (Fig. 7a and Supplementary Table 5). To achieve this, a two-step enzyme cascade, including hydrolysis and decarboxylation reactions, was evaluated. In addition, we employed a Plackett-Burman (PB) design to investigate key parameters (Supplementary Table 6), resulting in a total alkene titer of 1.57 g/L, predominantly composed of long-chain polyenes (C17:3), dienes (C17:2), and terminal olefins (C15:1) (Fig. 7b and Supplementary Table 7). To further assess the potential of OleTP_CL in biotransforming an industrial feedstock, we conducted this enzyme cascade in a 50 mL reaction system, a 100-fold volume increase, yielding 1.70 g/L of total alkenes and a total turnover number (TTN) of 1470 (Fig. 7c and Supplementary Table 7). This result is remarkably higher than those reported for cell-free systems employing the model enzyme OleT_JE, which exhibits much lower efficiency, especially in converting unsaturated fatty acids¹². To date, the highest performance with vegetable oils has been achieved with coconut oil (primarily composed of saturated C12 fatty acids), yielding a total hydrocarbon titer of 0.49 g/L in a 10 mL reaction volume (Supplementary Table 8). Therefore, these findings demonstrate that OleTP_CL can efficiently convert industrially relevant feedstocks rich in unsaturated fatty acids into alkenes, highlighting the biotechnological potential of the fatty acid decarboxylases to produce sustainable petroleum-like compounds.

**Fig. 7: Biotransformation of corn ethanol biorefinery byproduct into alkenes.**

Discussion

Since the discovery of the biocatalytic pathways for hydrocarbon production³², efforts have been directed towards unraveling the molecular mechanisms governing alkene formation by P450 decarboxylases^{12,13,15,17,18,19,22,23,31,33,34,35}. The divergence between hydroxylation and decarboxylation remains one of the central unresolved questions in the field. Structural variations in the active site of decarboxylases with differing specificities may impact substrate conformation and accessibility to the heme iron center, thereby influencing enzymatic bifurcation between decarboxylation and hydroxylation.

Another key question to be addressed concerns the decarboxylation of unsaturated fatty acids, the most abundant and industrially relevant feedstocks. These unsaturated substrates are inhibitors of the enzyme OleT_JE, that has been served as the primary model for studying fatty acid decarboxylation for over a decade. Recently, a P450 decarboxylase that effectively converts unsaturated fatty acids into alkenes was discovered; however, the lack of structural data in complex with these unsaturated substrates has hampered the elucidation of the molecular mechanism underlying this biotechnologically relevant activity of the P450 superfamily.

With this regard, we solved the complex of a CYP152 enzyme with oleic acid, which along with computational simulations, site-directed mutagenesis, bioinformatics, enzyme assays, and unsupervised machine learning, provided a mechanistic model for unsaturated fatty acid decarboxylation. These results led to the identification of an intricate conformational itinerary that depends on coordinated movements at the end of the binding pocket and the substrate, enabling decarboxylation via β-carbon oxidation. This mechanism is closely linked to a rotameric shift of the residue F19 in OleTP_CL, which has been identified as a conserved feature in certain P450 orthologs, including the decarboxylases from C. doosanense and M. abscessus, as shown in this study. Interestingly, F19 is located at a similar position to L176 at the OleT_JE longer F/G loop (Supplementary Fig. 27) that was previously postulated to be crucial for saturated fatty acid decarboxylation¹³. In this framework, similarly, F19 is closely related to OleTP_CL ability in efficiently decarboxylating fatty acids. Unlike saturated fatty acid, oleic acid contains a double bond within its chain, making its conformation distinct and, therefore, requiring a certain level of plasticity of the catalytic pocket for proper substrate accommodation.

In this context, although belonging to the same cluster, the P450 peroxygenase from K. marina (OleT_KM) is inefficient in decarboxylating oleic acid. This inefficiency is attributed to the shortened hydrophobic cradle with low hydrophobicity and the rigidity of the phenylalanine residue corresponding to F19 in OleTP_CL. Both stereochemical limitations do not allow the unsaturated substrate to achieve productive positioning, leading to an obstructed binding pocket that results in substrate inhibition (Supplementary Fig. 28). In OleTP_RN, the corresponding F19 residue is also unable to change its rotameric configuration; however, this enzyme contains a natively prolonged binding pocket with high hydrophobicity, acting as a facilitator in product release¹². These findings highlight the importance of the distinct structural determinants associated with the binding pocket, particularly the hydrophobic cradle, for driving fatty acid decarboxylation and substrate specificity in P450 decarboxylases.

Furthermore, we showed that the Phe-His-Arg triad located near the heme group, previously identified as strictly necessary for decarboxylation, is not entirely conserved among P450 decarboxylases, modifying the current mechanistic paradigm. Here, we have demonstrated that OleTP_CL possesses a native substitution of F86 into a valine that preserves its chemoselectivity. In addition, the H92Q substitution maintained the predominance of alkene formation from unsaturated substrates but not from saturated substrates. These results indicate that the oxidative decarboxylation of unsaturated fatty acids by OleTP_CL, is regardless of the presence of histidine and its interaction with the adjacent phenylalanine.

Given the importance of olefins in petrochemical and bioenergy industries, we next assessed the utilization of P450 decarboxylases in the production of olefins from an industrial byproduct, the DCO from an ethanol corn biorefinery. DCO is the second most utilized biodiesel feedstock in the USA; however, its high acidity poses challenges for biodiesel production, which primarily relies on chemical catalysts, resulting in reduced product recovery and yield. In contrast, DCO is advantageous for biological processes, as free fatty acids are the natural substrates for P450 decarboxylases. Notably, OleTP_CL resulted in promising alkene titers from a residual low-cost feedstock, pointing to the potential of these enzymes for promoting the production of advanced biofuels and other petroleum-like chemicals by integrating byproducts of established economic sectors with renewable hydrocarbon biomanufacturing.

Taken together, these results shed light on the decarboxylation mechanism of unsaturated substrates and demonstrate their potential for converting abundant and industrially relevant feedstocks into hydrocarbons. These findings can contribute to addressing the urgent need to reduce fossil resource consumption, given the significance of olefins in the petrochemical and bioenergy industries.

Methods

Sequence similarity network analysis

The web-based Enzyme Function Initiative-Enzyme Similarity tool was used to build a Sequence Similarity Network (SSN) for members of the P450 superfamily, specifically the CYP152 family. The sequence of OleTP_RN was used as a seed to retrieve potential decarboxylase sequences from the Uniprot database (https://www.uniprot.org/). The initial SSN was constructed using an e-value cutoff of 1e-¹⁰ and connected sequences shared an alignment score of at least 90. The SSN was further refined, and an additional threshold of sequence identity of 50% was used, resulting in the clear definition of four isofunctional groups. The final SSN was analyzed and visualized using Cytoscape software (https://cytoscape.org/).

Reagents and plasmids

The cytochrome P450 decarboxylase encoding ORFs from Corynebacterium lipophiloflavum, Kocuria marina, Corynebacterium doosanense and Mycobacterium abscessus cloned into the pET28a(+) vector was synthesized by Genscript (Piscataway, USA). Fatty acids (C10:0, C12:0, C14:0, C16:0, C17:0, C18:0, C20:0, C18:1, C18:2 and C18:3), alkenes (1-nonene, 1-undecene, 1-tridecene, 1-pentadecene and 1-hexadecene) and methyl pentadecanoate were purchased from Sigma Aldrich Co. (St Louis, USA). α and β-hydroxyl acids derived from C10:0, C12:0, C14:0, C16:0, C17:0, C18:0, and C20:0 were purchased from Larodan AB (Solna, Sweden). Hydrogen peroxide was obtained from Sigma Aldrich. Ni-NTA and Q-Sepharose resins were obtained from GE Healthcare Biosciences (Chicago, USA).

Heterologous expression and purification of OleTs and mutants

Recombinant protein expression and purification were performed as in Rade, et al.¹²., employing E. coli BL21(DE3) transformed with pET-28a based vectors and the pG-TF2 plasmid (Takara Bio, Kusatsu, JPN). Succinctly, recombinant E. coli strains were grown in 500 mL Terrific Broth medium (TB) supplemented with 125 mg/L thiamine hydrochloride, 500 µL trace metals solution (containing 50 mM FeCl₃, 20 mM CaCl₂, 10 mM MnCl₂, 10 mM ZnCl₂, 2 mM CoCl₂, 2 mM CuCl₂, 2 mM NiSO₄) and selective antibiotics, 25 µg/mL chloramphenicol and 50 µg/mL kanamycin. Cells were grown up to the optical density (OD_600nm) 0.6 at 37 °C, then temperature was reduced to 20 °C, and protein expression was trigged with 200 µM isopropyl-β-D-thiogalactopyranoside (IPTG) (Invitrogen), 5 µM FeCl₃, 100 µM δ-aminolevulinic acid (ALA), and 10 µg/mL tetracycline addition. After 16 h incubation, cells were harvested by centrifugation at 7.000 × g (10 min at 4 °C) and submitted to disruption by sonication in a Vibracell VCX 500 device (Sonics and Materials, Newtown, USA) in buffer A [100 mM potassium phosphate (pH 7.5), 300 mM NaCl, 10 mM imidazole, and 5 % (v/v) glycerol]. Five rounds of 2 min sonication (40% amplitude and no pause), followed by 5 min stirring at 4 °C were performed, and then the cell lysate was centrifugated at 12.000 ×g for 25 min, and the clear supernatant was loaded onto a 5 mL His-Trap chelating HP column (GE Healthcare Biosciences), pre-equilibrated with buffer A. Recombinant protein elution was reached with an imidazole gradient (5 to 300 mM) and the eluted samples containing protein were combined and dialyzed overnight against 100 mM sodium phosphate (pH 7.5), 50 mM NaCl and 5% (v/v) glycerol buffer. Recombinant proteins were further purified onto a fast-flow Q-sepharose column (Sigma-Aldrich Co., St Louis, USA) by a NaCl gradient (50 mM to 1 M). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, 12%) and spectrophotometric analysis were performed with protein fractions to evaluate purification.

Site-directed mutagenesis

OleTP_CL and OleT_KM mutants were generated by inverse PCR (Supplementary Table 9). Primer pairs were constructed to hold a complementary sequence longer than 15 nt and/or 50 °C, in which the codon changes were introduced. PCR amplicons were circularized by Gibson assembly. Mutations were validated by Sanger sequencing. OleTP_RN mutants were retrieved from Rade, et al.¹².

Steady-state Turnover- Fatty acid substrate conversion

Activity assays were carried out with different fatty acid substrates (C10-C20 and C18:1). For this, a 500 μL reaction was prepared by mixing 500 μM fatty acid substrate, 2 μM enzyme, and 600 μM H₂O₂ in the best buffer for each enzyme. Hydrogen peroxide was added in six steps of 1 μL hydrogen peroxide (10 mM stock) every 10 min on positive control over the course of 1 hr. For the negative control, buffer was added instead of H₂O₂. Reactions were carried out at the best temperature and 700 rpm. The reaction was immediately stopped with 5 μL of 50% HCl. Then, the internal standards 1-tetradecene and heptadecanoic acid (final concentration of 0.98 mg/mL) were added to the reaction mixtures. The products were extracted with 500 μL chloroform by homogenization in a tube revolver rotator (20 min) followed by centrifugation at 3000; × g for 5 min. 400 μL of the organic phase was transferred to a 2 mL glass vial for the derivatization step, in which 25 μL HCl (2 M in methanol) and 75 μL of a second internal standard solution (1-hexadecene and methyl pentadecanoate at 50 µg/mL final concentration in methanol) were added. The sample was incubated at 50 °C for 30 min and then quenched with 40 μL NaOH solution (1 M in methanol). 500 μL of pure water was added for phase separation, and, after centrifugation at 3000 × g for 5 min, the organic phase was analyzed by gas chromatography.

Enzyme cascade to convert DCO oil into alkenes

The enzymatic cascade for alkene biosynthesis from distillers corn oil (DCO) was established using a Plackett-Burman (PB) design^36,37 to assess the combined impact of five key reaction parameters: OleTP_CL, lipase, hydrogen peroxide (H₂O₂), Triton X100, and agitation speed (Supplementary Tables 6, 7). DCO oil, used as a source of free fatty acids, was kindly provided by the company Inpasa Brasil. The enzymatic cascade consisted of two stages. In the first stage, DCO oil hydrolysis (3.57% w/v) was carried out using commercial lipase from Aspergillus oryzae (Supplementary Table 7) in a 100 mM sodium phosphate buffer (pH 7.5) for 17 h at 45 °C with 200 rpm agitation. In the second stage, decarboxylation was performed in a final volume of 500 μL, containing 1% w/v (~ 17.27 mM FFAs) of the hydrolyzed DCO oil, OleTP_CL, Triton X100, and hydrogen peroxide (H₂O₂), which was added every 5 min to maintain OleTP_CL activation (Supplementary Table 7). The reactions, buffered with 100 mM sodium phosphate buffer (pH 7.5), were carried out at 45 °C for 240 min with constant stirring (Supplementary Table 7) and were terminated by adding 1 mL of chloroform. Internal standards (520 µg/mL of 1-tetradecene and 720 µg/mL of heptadecanoic acid) were added for analysis and mixed with 100 µL saturated sodium chloride solution. The mixture was shaken vigorously for 20 min and then centrifuged at 3000 × g for 5 min. The upper aqueous phase was discarded, and 900 µL of the lower organic phase was transferred to a 2 mL glass vial. In the vial, an internal standard solution (600 µg/mL 1-hexadecene and 680 µg/mL methyl pentadecanoate in methanol) and 50 µL 1 M NaOH in methanol were added. The mixture was shaken for 5 min, incubated at 60 °C for 10 min, and then cooled to room temperature. Subsequently, 300 µL ammonium chloride and sulfuric acid solution in methanol were added, and the mixture was shaken and incubated again at 60 °C for 10 min. After cooling, 200 µL water was added, and the mixture was shaken for 3 min before centrifugation at 3000 × g for 5 min to separate the organic and aqueous phases. The organic phase was collected and analyzed by gas chromatography.

The alkene biosynthesis reaction from DCO oil was scaled up using the best conditions identified by the PB design. Briefly, the decarboxylase reaction was conducted in a final volume of 50 mL, consisting of 1% DCO oil hydrolyzed for 17 h at 45 °C using commercial lipase from Aspergillus oryzae (1.12 mg/mL), 5 µM of OleTP_CL, and 36 mM of hydrogen peroxide (H₂O₂), which was added every 5 min (Supplementary Table 7). The decarboxylation reactions were buffered with 100 mM sodium phosphate buffer (pH 7.5) and carried out at 45 °C for 240 min. After this, 500 μL of the reaction mixture was sampled and stopped by adding 1 mL of chloroform. The samples were then derivatized as described in the Steady-state Turnover-Fatty acid substrate conversion section, and the organic phase was collected and analyzed using gas chromatography.

Analytical analyses for fatty acid and product quantification

The reaction products were analyzed using gas chromatography analysis (7890 A, Agilent Technologies, Santa Clara, USA). The following conditions were used: RTx-5MS column (30 m x 0.25 mm × 0.25 μm), with a detector temperature of 290 °C, injector temperature of 230 °C, split ratio 1:10, and program: starting at 30 °C for 2 min, then increasing to 220 °C at a rate of 10 °C/min, then to 240 °C at a rate of 2 °C/min, and finally to 350 °C at 50 °C/min. Peaks were identified by comparing retention times with authentic standards of fatty acids, alkenes, and hydroxylate products. Quantification of compounds was done using internal standards (1-tetradecene for alkene quantification and heptadecanoic acid for fatty acids and hydroxylates quantification), and the exact quantification of the internal standards added was determined using calibration curves for each internal standard (1-hexadecene and methyl pentadecanoate).

For DCO oil chromatographic analyses, an Agilent 7890 A gas chromatograph equipped with a flame ionization detector (FID) was employed coupled to a DB-23 column (60 m x 0.250 mm, 0.25 µm). Helium served as the carrier gas with a column flow rate of 1.0 mL/min. The injection was carried out in split mode (1:10) with an injection volume of 1 µL. The injector and FID temperatures were maintained at 250 °C. The GC oven was programmed as follows: the initial temperature was set to 50 °C and held for 5 min, then ramped to 250 °C at a rate of 5 °C/min and held for 5 min. The total analysis time was 50 min.

Protein crystallization, X‑ray data collection, and structure determination

Protein purified samples were concentrated to 15, 30, and 60 mg/mL and submitted to crystallization trials without substrate and with 1 mM C12:0, C14:0, C16:0, C18:0 or C18:1, using the vapor diffusion method in sitting drops at 4 and 18 °C. OleTP_CL crystals were obtained with C16:0 and C18:1 from a solution containing 30% (w/v) PEG4000, 0.2 M MgCl_2, and 0.1 M Hepes (pH 7.5), and OleT_KM crystals with C14:0 from 10% (w/v) PEG400, 4 M NaCl and 0.1 M Hepes (pH 7.5). No apo-enzyme crystal could be achieved, and substrate-complexed crystals were only obtained at 60 mg/mL concentration and 18 °C. Diffraction data were collected at the Manacá beamline from the Brazilian Synchrotron Light Laboratory (LNLS, Campinas, Brazil), using a Dectris Pilatus 2 M detector at 100 K. Data were processed using XDS 62³⁸, and the structure was solved by the molecular replacement method using the program MOLREP³⁹. The atomic coordinates were fitted based on the electron density using COOT⁴⁰ and refined using PHENIX⁴¹ refine and REFMAC⁴². The final atomic model was verified with MolProbity^43,44 and PDBRedo server⁴⁵. Structural comparisons were performed with the Dali Server. Detailed statistics of data collection and refinement are provided in Supplementary Table 4.

Molecular dynamics simulations

The structures of OleTP_CL, OleT_KM, and OleTP_RN enzymes covalently bonded to HEME groups were considered and simulated in the absence of substrate and with both OLA and PLM substrates by adding or deleting the required carbon atoms of the ligands from complexed structures. Protonation states for the enzymes were obtained at their optimum pH (7.5 for OleTP_CL and OleT_KM, and 7.0 for OleTP_RN), using the propKa 3.0 software⁴⁶. Ligands were modeled in their deprotonated forms, and crystallographic water molecules were kept at their initial positions. All structures were then centered in a cubic box (10 nm size), and a total of 30000 TIP3P water molecules were added. Total charges were neutralized by the addition of Na⁺ ions in replacement of random solvent molecules.

All systems were subjected to energy minimization steps to remove repulsive contacts. An initial Steepest Descent minimization procedure was performed, keeping the protein complex heavy atoms harmonically restrained at their initial positions with a 1000 kJ/mol/nm² force constant, until RMS forces reached values below 500 kJ/mol/nm². An additional Conjugated Gradient minimization procedure without restraints was performed until RMS forces reached values below 100 kJ/mol/nm². Systems were then equilibrated in multiple stages as follows: i) 100 ps heating up to 100 K using the same protein complex harmonic restraints as before; ii) additional 100 ps at 100 K without restraints; iii) 100 ps heating up to 200 K without restraints and iv) 100 ps heating to target temperature (318.15 K for OleTP_CL; 313.15 for OleT_KM; 308.15 K for OleTP_RN) without restraints. After the heating procedure at the NVT ensemble, an equilibration of 500 ps was performed at the NpT ensemble at 1 bar followed by an additional 500 ps equilibration at the NVT ensemble. Data analyses presented were carried from the production of 250 ns at the NVT ensemble since all systems reached convergence (Supplementary Fig. 29). Calculations were performed using the July 2022 revision of Charmm36⁴⁷ force field and the Gromacs 2022.3 package^48,49. The Charmm36 parameter set includes a specific residue for the HEME cofactor (including the central Fe bonded to the four coplanar porphyrin nitrogen atoms) and for a cysteine residue in thiolate form (named CYM). Further, it also contains information regarding bonding between the residues if the required atoms (Fe and S) are within 0.25 nm. Thus, by renaming the required CYS to CYM in the original PDB file prior to using the Gromacs’ pdb2gmx tool, the Fe-S axial bond was added to the final topologies. Simulations were performed considering the resting state since control simulations of OleTP_CL performed with the oxygen atom bonded to HEME iron (CPD I) displayed similar results both for time convergence and F19 flexibility (Supplementary Fig. 30).

Coulomb interactions were calculated in real space up to 1.2 nm, and the PME algorithm^50,51 was used to treat long-range interactions. Lennard-Jones interactions were calculated up to 1.0 nm with a switching function that brings the potential to zero at 1.2 nm. All molecular dynamics calculations were performed using 2 fs timestep to integrate Newton equations of motion altogether with LINCS algorithms⁵² to keep all bonds containing hydrogen atoms rigid. The V-rescale scheme was used to control the temperature in all equilibration steps whereas the Nose-Hoover thermostat was used in all production runs. The C-rescale scheme was used to control the pressure in the NpT step. Four independent simulations were launched for each system, each one starting from the initial heating step, in which atomic velocities were assigned randomly according to the Maxwell-Boltzmann distribution.

Unsupervised machine-learning analysis: PCA and Clustering

Principal Component Analysis (PCA) was performed on 1 µs trajectories resulting from the concatenation of the four 250 ns replicas of each system. Translational and rotational degrees of freedom were removed by centering and aligning resulting in 50000 frames to the backbone of the protein structure from the first frame. The covariance matrix calculation and diagonalization were then performed considering the system containing the protein backbone atoms, all heavy atoms from both HEME and ligand molecules, and from other specific residues of interest (such as the F19). This step was performed using Gromacs 2022.3 covar tool, and resulting eigenvalues and eigenvectors were analyzed using Gromacs 2022.3 anaeig tool.

The resulting 50000 data points of the two principal components with higher variance (PC1 and PC2, Supplementary Fig. 25) were then used to build a smaller dataset (25000 data points) by considering only the even-indexed points of the original data frame. Finally, clustering analyses of PC1 x PC2 projections were performed using the HDBSCAN algorithm, as implemented at the Scikit-learn 1.5.2 package^53,54. This method defines clusters as dense regions in the data space, separated by regions of lower density, using a hierarchical approach to handle density variations and find the most stable clusters. The kd_tree algorithm was employed along with the leaf cluster selection method to provide the most fine-grained and homogeneous clusters. The Manhattan distance metric was set, and the following hyperparameters were tuned to perform a conservative clustering, yielding well-defined dense regions: i) the minimum number of points for a region to be considered dense (min_samples, set to 125), ii) the minimum number of samples to consider the group a cluster (min_cluster_size, set to 250). Further, representative structures were extracted from the original trajectory by calculating the medoids – samples from the original data which minimize the distance to all other points within each cluster.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

Structural data is deposited in the PDB database under codes 8W1J, 8W1K, and 8VWK. Additional P450 structural data employed in this work are available under the accession codes 8D8P, 4L40, 1IZO, 6FYJ, and 3AWM. P450 decarboxylase encoding ORFs employed in this work are available under the accession codes Corynebacterium doosanense (A0A097IEQ1 [https://www.uniprot.org/uniprotkb/A0A097IEQ1]), Corynebacterium lipophiloflavum (C0XPZ5 [https://www.uniprot.org/uniprotkb/C0XPZ5]), Kocuria marina (A0A0B0D9P4 [https://www.uniprot.org/uniprotkb/A0A0B0D9P4]), Mycobacterium abscessus (UPI000929683D [https://www.uniprot.org/uniparc/UPI000929683D]) and Rothia nasimusium (A0A1Y1RQ53 [https://www.uniprot.org/uniprotkb/A0A1Y1RQ53]. All initial/final coordinates, topology files, MD parameter files, and forcefield parameter files were uploaded to Zenodo [https://doi.org/10.5281/zenodo.14265678]. The sequence logo representation in Supplementary Fig. 15 can be reproduced by inputting the amino acid multiple sequence alignment data into a web server [https://weblogo.berkeley.edu/logo.cgi], however, the process is automatic, meaning the numerical source data is inaccessible. Unless otherwise stated, all data supporting the results of this study can be found in the article, supplementary, and source data files. Source data are provided in this paper.

References

Wang, C. et al. Direct synthesis of extra-heavy olefins from carbon monoxide and water. Nat. Commun. 14, 1857 (2023).
Article CAS PubMed PubMed Central MATH Google Scholar
Wu, D. et al. Alkene 1,1-difunctionalizations via organometallic-radical relay. Nat. Catal. 6, 1030–1041 (2023).
Article CAS Google Scholar
Nguyen, V. T. et al. Alkene synthesis by photocatalytic chemoenzymatically compatible dehydrodecarboxylation of carboxylic acids and biomass. ACS Catal. 9, 9485 (2019).
Article CAS PubMed PubMed Central MATH Google Scholar
Cai, P. et al. Microbial synthesis of long-chain α-alkenes from methanol by engineering Pichia pastoris. Bioresour. Bioprocess. 9, 1–8 (2022).
Article Google Scholar
Zhang, Y., Nielsen, J. & Liu, Z. Metabolic engineering of Saccharomyces cerevisiae for production of fatty acid–derived hydrocarbons. Biotechnol. Bioeng. 115, 2139–2147 (2018).
Article CAS PubMed MATH Google Scholar
Wu, S., Zhou, Y., Gerngross, D., Jeschek, M. & Ward, T. R. Chemo-enzymatic cascades to produce cycloalkenes from bio-based resources. Nat. Commun. 10, 5060 (2019).
Article ADS PubMed PubMed Central Google Scholar
Rui, Z. et al. Microbial biosynthesis of medium-chain 1-alkenes by a nonheme iron oxidase. Proc. Natl. Acad. Sci. USA 111, 18237–18242 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Sorigué, D. et al. An algal photoenzyme converts fatty acids to hydrocarbons. Science 357, 903–907 (2017).
Article ADS PubMed MATH Google Scholar
Zhou, Y. J. et al. Production of fatty acid-derived oleochemicals and biofuels by synthetic yeast cell factories. Nat. Commun. 7, 11709 (2016).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Xu, Y.-Y. et al. Cellulosic hydrocarbons production by engineering dual synthesis pathways in Corynebacterium glutamicum. Biotechnol. Biofuels Bioprod. 15, 29 (2022).
Article CAS PubMed PubMed Central Google Scholar
Howard, T. P. et al. Synthesis of customized petroleum-replica fuel molecules by targeted modification of free fatty acid pools in Escherichia coli. Proc. Natl. Acad. Sci. USA 110, 7636–7641 (2013).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Rade, L. L. et al. Dimer-assisted mechanism of (un)saturated fatty acid decarboxylation for alkene production. Proc. Natl. Acad. Sci. USA 120, e2221483120 (2023).
Amaya, J. A., Rutland, C. D., Leschinsky, N. & Makris, T. M. A distal loop controls product release and chemo- and regioselectivity in cytochrome P450 decarboxylases. Biochemistry 57, 344–353 (2018).
Article CAS PubMed Google Scholar
Jiang, Y. et al. Biochemical characterization of three new α-olefin-producing P450 fatty acid decarboxylases with a halophilic property. Biotechnol. Biofuels 12, 1–14 (2019).
Article Google Scholar
Rude, M. Terminal olefin (1-alkene) biosynthesis by a novel P450 fatty acid decarboxylase from Jeotgalicoccus species. Appl. Environ. Microb. 77, 1718–1727 (2011).
Article ADS CAS MATH Google Scholar
Pickl, M. et al. Mechanistic studies of fatty acid activation by CYP152 peroxygenases reveal unexpected desaturase activity. ACS Catal. 9, 565–577 (2019).
Article CAS PubMed Google Scholar
Matthews, S. et al. Catalytic determinants of alkene production by the cytochrome P450 peroxygenase OleTJE. J. Biol. Chem. 292, 5128–5143 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Belcher, J. et al. Structure and biochemical properties of the alkene producing cytochrome p450 OleTJE (CYP152l1) from the jeotgalicoccus sp. 8456 bacterium. J. Biolo Chem. 289, 6535–6550 (2014).
Article CAS MATH Google Scholar
Munro, A. W., McLean, K. J., Grant, J. L. & Makris, T. M. Structure and function of the cytochrome P450 peroxygenase enzymes. Biochem. Soc. Trans. 46, 183–196 (2018).
Article CAS PubMed PubMed Central MATH Google Scholar
Jiang, Y. et al. Biochemical characterization of three new α-olefin-producing P450 fatty acid decarboxylases with a halophilic property. Biotechnol. Biofuels 12, 79 (2019).
Article PubMed PubMed Central Google Scholar
Xu, H. et al. In vitro oxidative decarboxylation of free fatty acids to terminal alkenes by two new P450 peroxygenases. Biotechnol. Biofuels 10, 208 (2017).
Article PubMed PubMed Central Google Scholar
Hsieh, C. H. et al. The enigmatic P450 decarboxylase OleT is capable of, but evolved to frustrate, oxygen rebound chemistry. Biochemistry 56, 3347–3357 (2017).
Article CAS PubMed MATH Google Scholar
Bharadwaj, V. S., Kim, S., Guarnieri, M. T. & Crowley, M. F. Different behaviors of a substrate in P450 decarboxylase and hydroxylase reveal reactivity-enabling actors. Sci. Rep. 8, 12826 (2018).
Article ADS PubMed PubMed Central Google Scholar
Castillo-Quan, J. I. et al. An antisteatosis response regulated by oleic acid through lipid droplet-mediated ERAD enhancement. Sci. Adv. 9, eadc8917 (2023).
Article CAS PubMed PubMed Central Google Scholar
Zachos, I. et al. Photobiocatalytic decarboxylation for olefin synthesis. Chem. Commun. 51, 1918–1921 (2015).
Article CAS MATH Google Scholar
Lee, D. S. et al. Substrate recognition and molecular mechanism of fatty acid hydroxylation by cytochrome P450 from Bacillus subtilis. Crystallographic, spectroscopic, and mutational studies. J. Biol. Chem. 278, 9761–9767 (2003).
Article CAS PubMed MATH Google Scholar
Matsunaga, I., Sumimoto, T., Ueda, A., Kusunose, E. & Ichihara, K. Fatty acid-specific, regiospecific, and stereospecific hydroxylation by cytochrome P450 (CYP152B1) from Sphingomonas paucimobilis: substrate structure required for alpha-hydroxylation. Lipids 35, 365–371 (2000).
Article CAS PubMed Google Scholar
Papik, J. et al. Legacy effects of phytoremediation on plant-associated prokaryotic communities in remediated subarctic soil historically contaminated with petroleum hydrocarbons. Microbiol. Spectr. 11, e0444822 (2023).
Article PubMed MATH Google Scholar
Medeiros Vicentini-Polette, C., Rodolfo Ramos, P., Bernardo Gonçalves, C. & Lopes De Oliveira, A. Determination of free fatty acids in crude vegetable oil samples obtained by high-pressure processes. Food Chem. X 12, 100166 (2021).
Article CAS PubMed PubMed Central Google Scholar
Michalik, B., Biel, W., Lubowicki, R. & Jacyno, E. Chemical composition and biological value of proteins of the yeast Yarrowia lipolytica growing on industrial glycerol. Can. J. Anim. Sci. 94, 99–104 (2014).
Article CAS Google Scholar
Yadav, S., Shaik, S. & Dubey, K. D. Decarboxylation and protonation enigma in the H85Q mutant of cytochrome P450OleT. J. Phys. Chem. B 127, 2927–2933 (2023).
Article CAS PubMed Google Scholar
Schirmer, A., Rude, M. A., Li, X., Popova, E. & Del Cardayre, S. B. Microbial biosynthesis of alkanes. Science 329, 559–562 (2010).
Article ADS CAS PubMed Google Scholar
Grant, J. L., Mitchell, M. E. & Makris, T. M. Catalytic strategy for carbon-carbon bond scission by the cytochrome p450 olet. Proc. Natl. Acad. Sci. USA 113, 10049–10054 (2016).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Grant, J. L., Hsieh, C. H. & Makris, T. M. Decarboxylation of fatty acids to terminal alkenes by cytochrome P450 compound I. J. Am. Chem. Soc. 137, 4940–4943 (2015).
Article CAS PubMed MATH Google Scholar
Matthews, S. et al. Production of alkenes and novel secondary products by P450 OleT JE using novel H 2 O 2 -generating fusion protein systems. FEBS Lett. 591, 737–750 (2017).
Article CAS PubMed MATH Google Scholar
PLACKETT, R. L. & BURMAN, J. P. The design of optimum multifactorial experiments. Biometrika 33, 305–325 (1946).
Rodrigues, M. I. & Iemma, A. F. Experimental Design and Process Optimization (Crc Press, 2014).
Kabsch, W. XDS. Acta Crystallogr. D Biol. Crystallogr. 66, 125 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Vagin, A. & Teplyakov, A. MOLREP: an Automated program for molecular replacement. J. Appl. Cryst. 30, 1022–1025 (1997).
Article ADS CAS MATH Google Scholar
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501 (2010).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Liebschner, D. et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr. D Struct. Biol. 75, 861–877 (2019).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Murshudov, G. N. et al. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr. D Biol. Crystallogr. 67, 355–367 (2011).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Chen, V. B. et al. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D Biol. Crystallogr. 66, 12–21 (2010).
Article ADS CAS PubMed MATH Google Scholar
Williams, C. J. et al. MolProbity: More and better reference data for improved all‐atom structure validation. Protein Sci. 27, 293 (2018).
Article CAS PubMed MATH Google Scholar
Joosten, R. P., Long, F., Murshudov, G. N. & Perrakis, A. The PDB_REDO server for macromolecular structure model optimization. IUCrJ 1, 213–220 (2014).
Article CAS PubMed PubMed Central Google Scholar
Olsson, M. H. M., SØndergaard, C. R., Rostkowski, M. & Jensen, J. H. PROPKA3: Consistent treatment of internal and surface residues in empirical p K a predictions. J. Chem. Theory Comput. 7, 525–537 (2011).
Article CAS PubMed MATH Google Scholar
Huang, J. & Mackerell, A. D. CHARMM36 all-atom additive protein force field: Validation based on comparison to NMR data. J. Comput. Chem. 34, 2135–2145 (2013).
Article CAS PubMed PubMed Central MATH Google Scholar
Van Der Spoel, D. et al. GROMACS: Fast, flexible, and free. J. Comput. Chem. 26, 1701–1718 (2005).
Article PubMed MATH Google Scholar
Abraham, M. J. et al. GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 1–2, 19–25 (2015).
Article ADS MATH Google Scholar
Darden, T., York, D. & Pedersen, L. Particle mesh Ewald: An N⋅log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993).
Article ADS CAS MATH Google Scholar
Essmann, U. et al. A smooth particle mesh Ewald method. J. Chem. Phys. 103, 8577–8593 (1995).
Article ADS CAS MATH Google Scholar
Hess, B. et al. LINCS: A linear constraint solver for molecular simulations. J. Comput. Chem. 18, 1463–1472 (1997).
Campello, R. J. G. B., Moulavi, D. & Sander, J. Density-Based Clustering Based on Hierarchical Density Estimates. In: Advances in Knowledge Discovery and Data Mining. PAKDD 2013. (eds Pei, J., Tseng, V. S., Cao, L., Motoda, H. & Xu, G.) Lecture Notes in Computer Science, vol 7819. (Springer, Berlin, Heidelberg, 2013).
Pedregosa, F. et al. Scikit-learn: Machine Learning in Python. J. Mach. Learn. Res. 12, 2825–2830 (2011).

Download references

Acknowledgements

We gratefully thank the Brazilian Synchrotron Light Laboratory (LNLS, CNPEM, Campinas, Brazil) for the use of the Manacá beamline; the Biosciences National Laboratory (LNBio, CNPEM, Campinas, Brazil) for the automated crystallization (Robolab) and the Spectroscopy and Calorimetry facilities; the Biophysics of Macromolecules and Metabolomics of LNBR (CNPEM, Campinas, Brazil) for the use of facilities; the National Laboratory for Scientific Computing (LNCC/MCTI, Brazil) for providing HPC resources of the SDumont supercomputer; and Antonio Kaupert Neto for the laboratory support. This work was supported by grants from the São Paulo Research Foundation (FAPESP) #2018/02865-2 (R.Y.M.); #2019/08855-1 (L.M.Z.); #2021/14410-2 (I.T.S.). This work was supported by Sinochem Brasil and LNBR – CNPEM (Campinas, Brazil).

Author information

Authors and Affiliations

Brazilian Biorenewables National Laboratory, Brazilian Center for Research in Energy and Materials, Campinas, SP, Brazil
Wesley Cardoso Generoso, Alana Helen Santana Alvarenga, Isabelle Taira Simões, Renan Yuji Miyamoto, Ricardo Rodrigues de Melo, Ederson Paulo Xavier Guilherme, Fernanda Mandelli, Clelton Aparecido Santos, Rafaela Prata, Camila Ramos dos Santos, Felippe Mariano Colombari, Mariana Abrahão Bueno Morais, Gabriela Felix Persinoti, Mario Tyago Murakami & Leticia Maria Zanphorlin
Sinochem Petróleo Brasil Ltda, Rio de Janeiro, RJ, Brazil
Rodrigo Pimentel Fernandes

Authors

Wesley Cardoso Generoso
View author publications
Search author on:PubMed Google Scholar
Alana Helen Santana Alvarenga
View author publications
Search author on:PubMed Google Scholar
Isabelle Taira Simões
View author publications
Search author on:PubMed Google Scholar
Renan Yuji Miyamoto
View author publications
Search author on:PubMed Google Scholar
Ricardo Rodrigues de Melo
View author publications
Search author on:PubMed Google Scholar
Ederson Paulo Xavier Guilherme
View author publications
Search author on:PubMed Google Scholar
Fernanda Mandelli
View author publications
Search author on:PubMed Google Scholar
Clelton Aparecido Santos
View author publications
Search author on:PubMed Google Scholar
Rafaela Prata
View author publications
Search author on:PubMed Google Scholar
Camila Ramos dos Santos
View author publications
Search author on:PubMed Google Scholar
Felippe Mariano Colombari
View author publications
Search author on:PubMed Google Scholar
Mariana Abrahão Bueno Morais
View author publications
Search author on:PubMed Google Scholar
Rodrigo Pimentel Fernandes
View author publications
Search author on:PubMed Google Scholar
Gabriela Felix Persinoti
View author publications
Search author on:PubMed Google Scholar
Mario Tyago Murakami
View author publications
Search author on:PubMed Google Scholar
Leticia Maria Zanphorlin
View author publications
Search author on:PubMed Google Scholar

Contributions

L.M.Z. conceived and supervised this project. W.C.G. and L.M.Z designed research analyzed all generated data and wrote the manuscript; W.C.G., A.H.S.A., and I.T.S. performed functional characterization of wildtypes and mutants; R.R.M., E.P.X.G., and R.P. performed DCO experiments; F.M. and C.A.S. performed biophysical characterizations; R.Y.M., C.R.S., and M.T.M. performed the structural analysis; G.F.P. and R.F. performed sequence analysis; F.M.C. and M.A.B.M. performed molecular dynamics analysis. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Leticia Maria Zanphorlin.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Kshatresh Dubey and the other 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

Supplementary Information (download PDF )

Description of Additional Supplementary Files (download PDF )

Supplementary Movie 1 (download MP4 )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source Data (download XLSX )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, 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 you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. 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-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Generoso, W.C., Alvarenga, A.H.S., Simões, I.T. et al. Coordinated conformational changes in P450 decarboxylases enable hydrocarbons production from renewable feedstocks. Nat Commun 16, 945 (2025). https://doi.org/10.1038/s41467-025-56256-4

Download citation

Received: 10 May 2024
Accepted: 14 January 2025
Published: 22 January 2025
Version of record: 22 January 2025
DOI: https://doi.org/10.1038/s41467-025-56256-4