Electron transfer engineering of artificially designed cell factory for complete biosynthesis of natural products

doi:10.21203/rs.3.rs-4718218/v1

Download PDF

Article

Electron transfer engineering of artificially designed cell factory for complete biosynthesis of natural products

https://doi.org/10.21203/rs.3.rs-4718218/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Biosynthesis of natural products (NPs) by artificially designed cell factories often involves numerous nicotinamide adenine dinucleotide phosphate (NADPH)-dependent enzymes that mediate electron transfer reactions. However, the mechanisms of electron transfer from regeneration to the final delivery to the active centers of various types of NADPH-dependent enzymes remain elusive, limiting our ability to systematically engineer electron transfer (ETE) to improve NPs production. Here, we elucidated the electron transfer mechanisms of NADPH-dependent enzymes, which were further contributed to systematically ETE of Saccharomyces cerevisiae, including step-by-step engineering the electron transfer residues of 7-Dehydrocholesterol reductase (DHCR7) and P450 sterol side chain cleaving enzyme (P450_scc), electron transfer components for directing carbon flux, and NADPH regeneration pathways, for high-level production of the cholesterol (1.78 g/L) and pregnenolone (0.83 g/L). Additionally, computational chemical analysis indicated that the ETE process make the electron transfer chains shorter and more stable which significantly accelerated proton coupled electron transfer process. This study underscores the significance of ETE strategies in NPs biosynthesis and expands synthetic biology approaches.

Biological sciences/Chemical biology/Synthetic biology

Biological sciences/Biotechnology/Metabolic engineering

Biological sciences/Biological techniques/Genetic engineering

Biological sciences/Chemical biology/Enzyme mechanisms

Electron

natural products

steroids

synthetic biology

cell factories

Electron-requiring reactions, such as the catalytic processes of cytochrome P450¹ and reductases², are essential for numerous biosynthesis steps of natural products (NPs). Generally, it starts from nicotinamide adenine dinucleotide phosphate (NADPH) and other electron donor regeneration, which are then transferred to NADPH-dependent enzymes by electron transfer components, and finally reach the catalytic center by residues with electron transferring function³. To increase the availability of electron donors for NPs productions, existing synthetic biology strategies in metabolic⁴, cofactor^{3, 5}, and enzyme engineering¹ primarily focus on modifying cofactor regeneration pathways, or reduce the physical distance between NADPH-dependent enzymes and electron transfer components by using linker peptides. However, these strategies often overlook the fact that electron transfer to the active center requires a series of residues to facilitate the process through proton-coupled electron transfer (PCET) and the electron transfer in biological systems is typically a multi-step and coordinated process^6–8. Therefore, systematically engineering the above whole electron transfer process (ETE strategy) is considered as a novel synthetic biology strategy for overcoming multiple electron-requiring rate-limiting steps in the biosynthesis of NPs.

Steroids and their pharmaceutical derivatives^{9, 10}, a class of NPs with high pharmaceutical value¹¹, involve numerous electron-requiring rate-limiting steps in their biosynthesis from simple carbon sources in Saccharomyces cerevisiae^{9, 12}. For instance, the biosynthesis of cholesterol (Cho) and pregnenolone (Prn) required reducing the C7 = C8 double bond (catalyzed by 7-Dehydrocholesterol reductase, DHCR7), cleaving the side chain (catalyzed by P450 sterol side chain cleaving enzyme, P450_scc), and other electron-requiring rate-limiting steps^{2, 9}. P450_scc is composed of CYP11A1 and cytochrome P450 reductase (CPR)⁹. CPR captures electrons from NADPH and subsequently transfers them to CYP11A1 via the electron transfer chain¹³, while DHCR7 can transfer electrons independently without CPR². However, the internal electron transfer mechanism and the residues of the electronic transfer chain of DHCR7 remain unclear. Moreover, the acidic residues of P450_scc's CYP11A1 could seize a proton from the hydroxyl group of 20α,22(R)-dihydroxycholesterol, thereby promoting the side-chain cleavage through a two-electron transfer to the heme center¹⁴. Given that they represent two distinct typical modes of the electron transfer chain², we thus selected these two processes corresponding to two NPs (Cho and Prn) to demonstrate the effectiveness of ETE in improving NPs production.

In this study, to establish standardized steps for ETE process, S. cerevisiae with rate-limiting DHCR7 and P450_scc was systematically engineered by engineering the electronic transport residues of NAPDH-dependent enzymes, electronic transport elements, and electronic donor regeneration pathway for efficient complete biosynthesis of Cho and Prn. For ETE of DHCR7 from Bos taurus, introducing polar residues into the reductase domain and replacing tyrosine residues into the electron transfer chain shortened the distance for capturing electrons and brought more stable electron transfer and efficient substrate capture, respectively. This results in a 68% reduction in the length of the electron transfer chain and a 9.6-fold increase in the activity of the mutation. For ETE of CYP11A1 (P450_scc) from Rehmannia glutinosa, using linker to link reductase and introducing acidic residues into catalytic pocket facilitated proton capture and accelerated electron transfer. Finally, building upon strengthening the NADPH regeneration pathway, electron transfer partners from Homo sapiens were further integrated to redirect more carbon flux toward Cho and Prn in S. cerevisiae. These ETE methods collectively enabled the high-level production of 1.78 g/L of Cho and 0.83 g/L of Prn in a 5-L bioreactor. Thus, our comprehensive and systematic ETE processes provide a paradigm for improving NPs biosynthesis.

DHCR7 is rate-limiting step for complete biosynthesis of Cho in S. cerevisiae

First, we constructed the biosynthetic pathway to produce 7-dehydrocholesterol (Dhc), the key precursor of Cho. Considering the endogenous biosynthetic pathway for ergosterol in S. cerevisiae competes for carbon flux with the pathway for Cho de novo synthesis (Fig. 1A)¹⁵, DHCR24 from Gallus with the promoter P_TEF was inserted into ERG5 and ERG6 sites to achieve two copies, thereby facilitating the accumulation of Dhc and reducing carbon flux toward ergosterol (Fig. 1A). The resulting strain D7-1 produced 203 mg/L of Dhc (Fig. 1B). Subsequently, to convert Dhc into Cho, we screened DHCR7 from four species (Bos taurus, Pangasianodon hypophthalmus, Sus scrofa, and Danio rerio). The resulting strain D7-1 with pY26-BtDhcr7 could produce a maximum Cho production of 31 mg/L (Fig. 1B) but 146 mg/L Dhc was still accumulated (Fig. 1C). Obviously, DHCR7 was the rate-limiting heterologous enzyme and needed for engineering.

ETE of BtDHCR7 for efficient complete biosynthesis of Cho

Elucidating the catalytic mechanism of DHCR7 and its electron transfer mechanism was crucial for the ETE process. First, BtDHCR7 was modeled by AlphaFold2 and docked with both Dhc and NADPH to determine the substrate-binding domain and NADPH-binding domain. (Fig. 2A). Dhc could both bind to the substrate-binding domain with Y317 and the surface at the entrance of the catalytic pocket with F430 and F434 (Fig. 2A). The nicotinamide group of NADPH could bind to Y54, which was formed the NADPH-binding domain of DHCR7. However, how electrons were transferred from the NADPH-binding domain to the active center still remained unknown. Aromatic residues, such as phenylalanine and tyrosine, are excellent carriers of electron transfer and could facilitate electron transfer through the resonance stabilization of radical intermediates^{6, 16}. We happened to discover that there is a series of aromatic residues (Y54, F56, F430, F434, and Y317) between the NADPH binding domain and the catalytic center, which might form the electron transfer chain of DHCR7 (Fig. 2A). To verify our above hypothesis, Y54, F56, F430 and F434 were mutated to alanine while Y317 was mutated into phenylalanine for disturbing electron transfer (Fig. 2B). It showed that a sharp decline in Cho production was observed when one of residues involved in electron transfer was disrupted, and DHCR7 even completely lost its activity with no Cho production when two residues involved in electron transfer were disrupted (Fig. 2B). Moreover, Y317F-DHCR7 mutation completely lost its activity (Fig. 2A and 2B), suggesting its essential role in finally transferring the electron to the substrate. In conclusion, free Dhc in the cytoplasm was captured by DHCR7 through π-π stacking interactions with F434 and F430, subsequently entering the active center composed of Y317. Simultaneously, electrons from NADPH are captured by Y55 and transferred through the electron transfer chain (measured as 29.4 Å) with F56, F430, and F434 to the active center Y317, facilitating the electrophilic addition at the C7 = C8 of Dhc to synthesize Cho (Fig. 2A).

Based on the catalytic and electron transfer mechanism, substrate-recognizing domain (SRD), substrate-binding domain (SBD), and electron transfer chain (ETC) of BtDHCR7 were gradually engineered. First, for engineering SRD, we noticed that compared to the conformation of Dhc in the catalytic pocket, the side chain of Dhc in the substrate recognition domain was in a twisted state (Fig. 2C). This resulted in higher binding energy (-4.17 kcal/mol) than that of catalytic pocket (-9.01 kcal/mol), which prevented DHCR7 effectively capture the substrate. Thus, L67 and L426 of BtDHCR7 closed to the side chain of Dhc (Fig. 2C) were mutated into hydrophobic residue of small conformations (G/A/V/I). It showed that the BtDHCR7 -L67V-L436A (M1) variant increased the production of Cho to 76.6 mg/L (Fig. 2D). Subsequently, based on M1, SBD was further engineered. Considering Dhc only formed a π-π stacking interaction with Y317 in SBD of M1, and its ²C-OH did not form stronger hydrogen bonds with other residues. Thus, W280 and Y290, located within 4 Å of the ²C-OH, were subjected to saturation mutagenesis. It showed that M1- W281Y-Y291W (M2) variant increased the production of Cho to 92.3 mg/L (Fig. 2D). Finally, after iteratively engineering the DHCR7's SRD and SBD, resulting in M2, its electron transfer chain needed to be engineered. The ETE of BtDHCR7 involved two steps: shortening the distance for NADPH electron capture and engineering the electron transfer residues. Starting from M2, A49 and I53 of BtDHCR7 were gradually engineered into polar residues for accelerating obtaining the electrons of NADPH. It showed that only M2- I53T (M3) variant increased the production of Cho to 124.3 mg/L (Fig. 2D). The mechanistic analysis of BtDHCR7 has confirmed that F430 and F434 in the electron transfer chain not only transferred electrons but also captured the Dhc from the medium. Additionally, we also found that the binding position of F430 with Dhc was similar to that of Y317 with Dhc in the SBD. However, F430 can only facilitate electron transfer and not act as an electron donor, so it was further mutated to tyrosine. It showed that only M3- F430Y variant increased the production of Cho to 179.0 mg/L (Fig. 2D). Tyrosine could also increase the π-electron cloud density of the conjugated system, thereby further enhancing electron transfer efficiency⁶. Thus, F434 of electron transfer chain was further mutated into tyrosine, which resulted in M3-F430Y-F434Y (M4, MuDHCR7) variant could increase the production of Cho to 191.7 mg/L, representing a 5.93-fold improvement over wild-type BtDHCR7.

Computational simulation analysis for ETE of BtDHCR7 process

ETE of BtDHCR7 has achieved remarkable results and still needed to employ computational simulations to deeply analyze the specific electron transfer mechanisms. The electron transfer chains of wild-type BtDHCR7 and MuDHCR7 are composed of Y55-F56-F434-F430-Y317 and T53-Y430-Y434-Y317, respectively. Firstly, we used 100-ns molecular dynamics simulations to analyze the dynamic changes of distances between the electron transfer residues in BtDHCR7 and M4 (Fig. 3A), respectively. Starting from the capture of electrons from NADPH by the electron transfer chain, the distances from the nicotinamide group of NADPH to T53 (wild-type BtDHCR7) and Y55 (MuBtDHCR7) were 5.8 ± 0.3 Å and 3.9 ± 0.4 Å, respectively (Fig. 3B). Subsequently, the distances between residues in the electron transfer chain of wild-type BtDHCR7 remain within 5–10 Å, whereas that of M4 can be stably maintained within 2–5 Å (Fig. 3A). As a result, the electron transfer chain length of WT-DHCR7 and MuDHCR7 was 29.4 and 21.3 Å, respectively (Fig. 3A).

Enhancing the efficiency of electron transfer from the terminal residues of the electron transfer chain to the substrate was also the crucial step in ETE process. We have confirmed that Y317 was the key electron-donating residue in the active pocket. MD simulation results showed that the distance between Y317 of MuBtDHCR7 and the C7 = C8 double bond of the substrate is 3.9 ± 0.14 Å, with the length and fluctuation range of this distance being lower than that of wild-type BtDHCR7 (4.1 ± 0.52 Å) (Fig. 3B). This indirectly indicated that the Y317 of M4 could provide electrons to the substrate more effectively. Notably, the ETE of BtDHCR7 also involved engineering the SRD. F430 of wild type BtDHCR7 SRD only facilitated electron transfer without providing electrons to the substrate and was further mutated into tyrosine for possessing both electron transfer and electron donation capabilities to the substrate, which significantly enhanced the activity of DHCR7. To verify our conclusion as Y430 was the novel catalytic center, Y317 and Y430 were further mutated to alanine, respectively (Fig. S20). Consistent with our conclusion, only M4- Y317A-Y430A variant lost activity. We noticed that the two active pockets respectively formed by Y317 and Y430 both have the similar binding conformation with Dhc. Thus, we further extracted the catalytic pockets involving Y317-Y280 and Y430-Y434, along with the substrate, to create a cluster model for QM calculations (Fig. 3C). In step 1, the C8-Dhc toke a proton from Y317 or Y430 to generate intermediate (Int1) via the transition state (TS1), which requires an activation-free energy of 21.2 and 26.5 kcal/mol (Fig. 3C). In step 2, Y317 or Y430 respectively attracted protons from Y280 or Y434 and transfer them to C7-Dhc to generate intermediate (Int2) via the transition state (TS2), which requires an activation-free energy of 20.5 and 23.0 kcal/mol (Fig. 3C). QM calculations confirmed that the catalytic center composed of dual tyrosine residues effectively facilitates the double bond reduction process. Due to the introduction of Y430 to form a new active center, the new electron transfer chain was further shortened from 21.3 Å to 9.5 Å. As a result, ETE process of BtDHCR7 brought novel catalytic site, which further shortened the electron transfer chain by 68%, which significantly increased the activity.

Optimizing the elements of P450 _scc systems for complete biosynthesis of Prn.

Constructing an efficient de novo biosynthesis pathway for Prn requires obtaining effective P450_scc catalytic components (Fig. 4A)¹⁷. Firstly, we screened CYP11A1 within the P450_scc system. CYP11A1 sequences from 28 different species were retrieved from the National Center for Biotechnology Information (NCBI) database and used to conduct a phylogenetic tree analysis (Fig. S15). However, analyzing 28 CYP11A1 enzymes represented a substantial workload. Thus, the deep learning technique¹⁸ was employed to calculate their turnover numbers (k_cat values). Based on their DLk_cat values, the top ten CYP11A1 enzymes were selected to introduce into the D7-1 with pY26-MuDHCR7 strain for further catalytic validation (Table S4). Moreover, ADR and ADX from H. sapiens and ATR from Arabidopsis thaliana were selected to serve as redox partners. It showed that the actual catalytic abilities of different forms of CYP11A1 closely matched with the calculated k_cat values derived from DLk_cat (Table S4), and CYP11A1 from R. glutinosa (RgCYP11A1) give the best Prn production, reaching 93.5 mg/L. (Fig. 4B). Thus, RgCYP11A1 was selected for further engineered.

ETE of RgCYP11A1 for efficient complete biosynthesis of Prn and mechanism analysis

CYP11A1 can utilize its heme center to hydroxylate substrates and in combination with residues near the heme achieved side chain cleavage¹⁴. However, the side chain cleavage residues of RgCYP11A1 remain unknown. First, residues of RgCYP11A1 within 4 Å region of Cho were selected for alanine scanning to determining the side chain cleavage residues (Fig. 5A). RgCYP11A1- Y105A variant completely lost the ability of side chain cleavage and was identified as a side chain cleavage residue. Moreover, RgCYP11A1-V118A variant could increase the Prn production from 161 to 197.0 mg/L by adding 300 mg/L Cho as the substrate (Fig. 5B) and was selected for next generation of engineering process. Based on the docking results, we found that Y105 is close to the C22-OH group, allowing it to abstract the proton and facilitate the subsequent side chain cleavage. Subsequently, QM calculations determined that the C22-OH toke a proton from Y105 to generate intermediate (Int1) via the transition state (TS1), which requires an activation-free energy of 28.2 kcal/mol (Fig. 3C). The high energy barrier limits the efficiency of side chain cleavage by CYP11A1. We noticed that the naturally alkaline microenvironment in R. glutinosa benefited CYP11A1 for facilitating the final step of Prn synthesis by capturing the proton from the hydroxyl group of 20α,22(R)-dihydroxycholesterol to form an ionic bond¹⁹, thereby promoting electron transfer and enabling side-chain cleavage. Thus, starting from RgCYP11A1-V118A variant, A281 and T293 within 4 Å of C22-OH-Cho were further engineered into acidic residues for accelerating PECT process, which was the second-generation modification of CYP11A1 and belong to the ETE process. To this end, we found that the A281D-T293E variant of V118A-RgCYP11A1 (MuRgCYP11A1) could produce 243 mg/L of Prn (Fig. 5C).

The biosynthesis of Prn by cutting the side chain of Cho with P450_scc was a complex process involving two hydroxylation as the first step, followed by the loss of two protons from the two hydroxyl groups and ultimately the formation of a ketone group (Fig. 5D) as the second step¹⁴. For the first step, ensuring that the two carbon positions requiring hydroxylation were stably positioned above the heme center and in similar distances were the key to efficient dihydroxylation²⁰. 100-ns MD simulation of RgCYP11A1 and Cho complexes indicated that C-20 and C22 were more stably exposed above the center of the heme (Fig. 5E and 5F), which further accelerated dihydroxylation process. Additionally, free energy calculations revealed that the D281 and E285 of MuRgCYP11A1 could contribute intermolecular forces of -48 kJ/mol and − 51 kJ/mol, respectively (Fig. S21). The stronger binding forces allow Cho to bind more stably in the substrate pocket of CYP11A1 for higher catalytic efficiency. For the second step, QM calculation of wild type RgCYP11A1 suggests that the lack of acidic residues within 4-Å distance around 20α,22(R)-dihydroxycholesterol to capture its proton limited PCET process. Thus, we further carried out QM calculation for examining the first step of proton abstraction from the 22-hydroxyl group in MuRgCYP11A1, which was engineered by ETE (Fig. 5E). Introducing T293E into the catalytic pocket of CYP11A1, compared to the original Y105, allows for closer proximity to the 22-OH proton (reduced from 2.0 Å to 1.6 Å), which further lowered the reaction energy barrier (from 28.2 kcal/mol to 14.6 kcal/mol) (Fig. 5G). In fact, ETE process effectively promoted the PCET process and accelerating electron transfer efficiency, which significantly improved the catalytic efficiency of CYP11A1.

ETE of S. cerevisiae for efficient complete biosynthesis of Cho and Prn

Last, the above obtained BtDHCR7 and RgCYP11A1 variants with the highest catalytic efficiency (MuDHCR7 and MuRgCYP11A1) were used to construct de novo biosynthesis pathway of Cho and Prn in S. cerevisiae. Before this, the precursor metabolic pathway of Cho synthesis was strengthened (Fig. 6A). First, the mevalonate (MVA) pathway was strengthened by integrating tHMG1 and ID11 into the TY1 multiple copy sites of D7-1, resulting in D7-2. D7-2 with pY26-MuDHCR7 produced 228 mg/L Cho (Fig. 6B). ERG2 and ERG3 were further integrated into TY3 multiple copy sites of D7-2 for increasing the conversion rate of squalene to Cho, resulting in D7-3. D7-3 with pY26-MuDHCR7 produced 267 mg/L Cho (Fig. 6B). Subsequently, MuDHCR7 was integrated into the D7-3 genome in single, double, and multiple copy numbers, respectively (Fig. S15). As a result, D7-3 with two copies of MuDHCR7 (D7-4) could produce 303 mg/L Cho (Fig. 6B).

After optimizing the carbon flux through metabolic engineering, D7-4 was continuously engineered by ETE strategy (Fig. 6B). Given that the animal-derived electron transfer system supplied electrons to the endogenous ERG11 and ERG25 in S. cerevisiae^21–24, thereby directing more carbon flux toward the subcellular structures involved in animal sterol synthesis, two electron transfer systems (adrenodoxin reductase (ADR)-adrenodoxin (ADX) from H. sapiens and human cytochrome b5, respectively) were integrated into D7-4, resulting in D7-5 and D7-6, respectively. D7-5 with b5 could produce 282 mg/L Cho, while D7-6 with both ADR and ADX produced 382 mg/L Cho (Fig. 6B). Then, a glucose 6-phosphate dehydrogenase (GDH) from Bacillus Subtilis was integrated into D7-6 for recycling NADPH and enhancing the electron transfer efficiency, resulting in D7-7. This strain could produce 425 mg/L Cho (Fig. 6B). Besides enhancing the cofactor regeneration efficiency, FPK and PTA were introduced into the PPP for NADPH regeneration (D7-8). Consequently, D7-8 could synthesize 460 mg/L of Cho (Fig. 6B). In order to test whether the ETE process improves the efficiency of yeast NADPH regeneration, the ratio of NADPH/NADP⁺ of D7-6, D7-7, and D7-8 were measured, respectively. It showed that the ratio of NADPH/NADP⁺ from D7-6 to D7-8 increased from 0.3 to 0.8 using the electron transfer strategy (Fig. 6C). The strains D7-8 were used to produce Cho through fermentation in a 5-L bioreactor (Fig. 6D). After optimization, the titers of Cho reached 1.78 g/L (Fig. 6E).

For Prn production, pY26-MuRgCYP11A1-ATR were introduced into D7-8, resulting in D7-9 could produce 161.84 mg/L Prn. Subsequently, wild type RgCYP11A1-ATR and MuRgCYP11A1-ATR were integrated into the genome of strain D7-8 to achieve a more stable fermentation synthesis, resulting in D7-10-1 and D7-10-2. D7-10-2 with MuRgCYP11A1 could produce 193.77 mg/L of Prn, an increase of 138% compared to the wild type. Reducing the physical distance between redox partners and CYP450 enzymes is a strategy to enhance the electron transfer efficiency (ETE) of P450 enzymes¹. Thus, to enhance electron transfer efficiency, MuRgCYP11A1 and ATR were linked using the flexible peptide GGGGS (Fig. 6E), resulting in D7-10-3. As a result, D7-10-3 could produce 261 mg/L Prn (Fig. 6E), an increase of 37% compared to the D7-10-2. Finally, the strains D7-10-3 were used to produce Prn through fermentation in a 5-L bioreactor (Fig. 6F). After optimization, the titers of Prn reached 0.83 g/L (Fig. 6F).

The ETE strategies established in this study included engineering electronic transfer residues of two types of NADPH-dependent enzymes, engineering electronic transport element for directing carbon flux, and engineering electronic donor regeneration pathway. Advancing beyond traditional synthetic biology strategies, the ETE approach in this study refines electron transfer engineering down to the residue level^{1, 3}. Starting from the capture of electron from NADPH, rationally mutating residues binding the nicotinamide group of NADPH in the reductase domain into the polar residues have shortened the initial electron capture distance in the electron transfer chain. Subsequently, our study confirmed that aromatic residues can utilize π-electron cloud overlap resonance to further transfer electrons to the active center^{6, 16}, forming an internal electron transfer chain. MD simulation and QM calculation revealed that substituting phenylalanine with tyrosine in the electron transfer residues increased the π-electron cloud density of the conjugated system⁶, which resulted in the evolution of the electron transfer residues to donate electrons to the substrate and a shorter-more stable electron transfer chain.

Our study also confirmed the ETE process was also effective for oxidation-reduction partner-dependent enzymes, such as P450_scc^{9, 17}. A combination of deep learning, phylogenetic analysis, and empirical catalytic experiments revealed that plant-derived P450_scc possessed a longer evolutionary history and higher activity compared with animal-derived P450_scc. QM calculations in this study revealed that the rate-limiting step in the side chain cleavage process by RgCYP11A1 was the proton abstraction from C22-OH-Cho¹⁴. Interestingly, the naturally alkaline microenvironment in R. glutinosa benefited CYP11A1 for facilitating the final step of Prn synthesis by capturing the proton from the hydroxyl group of 20α,22(R)-dihydroxycholesterol to form an ionic bond¹⁹, thereby promoting electron transfer and enabling side-chain cleavage. Unlike the ionic bonds that captured protons, introducing the acidic residues into the RgCYP11A1 catalytic center by ETE could form covalent bonds with the C22-OH-Cho to accelerate electron transfer. Additionally, introducing the flexible peptide GGGGS to decrease the physical distance between CYP11A1 and its redox partner enhanced electron transfer efficiency, further boosting catalytic activity^{1, 25}.

The complete biosynthesis of animal sterols from simple carbon sources in a cellular factory is a complex multi-enzyme catalytic process involving numerous redox reactions. In fact, the electron transfer in S. cerevisiae is a highly complex process from metabolism perspective^{26, 27}. Consistent with the findings of Chen et al.³, introducing GDH, FPK, and PTA enzymes into S. cerevisiae to drive the PPP for efficient NADPH regeneration in our study effectively enhanced Cho and Prn production. More innovatively, the ETE strategy used in this study introduced the electron transfer system from the species producing the target product, thereby directing more carbon flux towards the target product. This is because using the electron transfer elements from the species producing the target product could pair with ERG11 and ERG25^28–30, which promoting the enrichment of upstream metabolic products near the subcellular organelles preferred by downstream heterologous NADPH-dependent enzymes^21–24.

In conclusion, our study elucidated the electron transfer mechanisms of various types of NADPH-dependent enzymes involved in NPs biosynthesis and systematically designed the electron transfer process from regeneration to the final delivery to the active centers for the high-level production of natural products. Thus, the ETE strategy offered a novel and advanced strategy for developing cell factories to synthesize NPs efficiently.

Supplementary Information description

Supplementary Information for this work can be found online at the nature communications website. Materials and methods: Materials and Methods; Chemicals; Gene synthesis and plasmid construction; Culturing S. cerevisiae for cholesterol and pregnenolone production; Extraction of cholesterol and pregnenolone from S. cerevisiae; Phylogenetic and deep learning–based k_cat prediction for analyzing CYP11A1; Molecular docking, molecular dynamics simulation, and QM calculations; Gas chromatography and HPLC-MS analysis for fermentation products; Determining the NADPH/NADP⁺ ratio; Codon optimized sequences of genes used in this study ; NMR data; Cartesian coordinates of stationary points; Supplementary Tables (Table S1-S4); Supplementary Figures (Fig. S1-S27).

Acknowledgements

This work has been funded by the National Key Research and Development Program of China (2019YFA0905300).

Author Contributions

Q.C, W.W, and J.Z designed the experiments. Q.C, Z.C, R.Q, J.H, H.C, and W.W performed the molecular biology experiments. Q.C, X.W, and K.W performed the molecular simulation experiment. Q.C, Y.R, and J.Z wrote and revised the manuscript.

Competing interests

This article does not pose any conflict of interest.

Ethical statement

This article does not contain any studies with human participants or animals performed by any authors.

Park, SY; Eun, H; Lee, MH; Lee, SY. Metabolic engineering of Escherichia coli with electron channelling for the production of natural products. Nat Catal. 2022 5(8):726-737.
Li, X; Roberti, R; Blobel, G. Structure of an integral membrane sterol reductase from Methylomicrobium alcaliphilum. Nature. 2015 517(7532):104-107.
Chen, R; Gao, J; Yu, W; Chen, X; Zhai, X; Chen, Y; Zhang, L; Zhou, YJ. Engineering cofactor supply and recycling to drive phenolic acid biosynthesis in yeast. Nat Chem Biol. 2022 18(5):520-529.
Zhou, Z; Zhang, X; Wu, J; Li, X; Li, W; Sun, X; Wang, J; Yan, Y; Shen, X; Yuan, Q. Targeting cofactors regeneration in methylation and hydroxylation for high level production of Ferulic acid. Metab Eng. 2022 73:247-255.
Rashida, Z; Laxman, S. The pentose phosphate pathway and organization of metabolic networks enabling growth programs. Curr Opin Syst Biol. 2021 28:100390.
Tyburski, R; Liu, T; Glover, SD; Hammarström, L. Proton-Coupled Electron Transfer Guidelines, Fair and Square. J Am Chem Soc. 2021 143(2):560-576.
Cai, X; Li, Y; Zhang, Y; Lin, W. π–π Interaction-Driven Charge Separation and Interlayer Transfer in Polymeric Carbon Nitride. ACS Catal. 2023 13(24):15877-15885.
Greene, BL. Progress and Opportunities in Photochemical Enzymology of Oxidoreductases. ACS Catal. 2021 11(23):14635-14650.
Zhang, R; Zhang, Y; Wang, Y; Yao, M; Zhang, J; Liu, H; Zhou, X; Xiao, W; Yuan, Y. Pregnenolone Overproduction in Yarrowia lipolytica by Integrative Components Pairing of the Cytochrome P450_scc System. ACS Synth Biol. 2019 8(12):2666-2678.
Ding, J; You, S; Zhang, J; Zhang, H; Wang, H; Zhang, W; Qi, W; Su, R; He, Z. Rational design of 17β-hydroxysteroid dehydrogenase type3 for improving testosterone production with an engineered Pichia pastoris. Bioresour Technol. 2021 341:125833.
Lin, YC; Cheung, G; Porter, E; Papadopoulos, V. The neurosteroid pregnenolone is synthesized by a mitochondrial P450 enzyme other than CYP11A1 in human glial cells. J Biol Chem. 2022 298(7):102110.
Qin, Z; Zhang, Y; Liu, S; Zeng, W; Zhou, J; Xu, S. Combining Metabolic Engineering and Lipid Droplet Assembly to Achieve Campesterol Overproduction in Saccharomyces cerevisiae. J Agric Food Chem. 2024 72(9):4814-4824.
Zöllner, A; Pasquinelli, MA; Bernhardt, R; Beratan, DN. Protein Phosphorylation and Intermolecular Electron Transfer: A Joint Experimental and Computational Study of a Hormone Biosynthesis Pathway. J Am Chem Soc. 2007 129(14):4206-4216.
Su, H; Wang, B; Shaik, S. Quantum-Mechanical/Molecular-Mechanical Studies of CYP11A1-Catalyzed Biosynthesis of Pregnenolone from Cholesterol Reveal a C–C Bond Cleavage Reaction That Occurs by a Compound I-Mediated Electron Transfer. J Am Chem Soc. 2019 141(51):20079-20088.
Wei, W; Gao, S; Yi, Q; Liu, A; Yu, S; Zhou, J. Reengineering of 7-dehydrocholesterol biosynthesis in Saccharomyces cerevisiae using combined pathway and organelle strategies. 2022 13.
Wang, Z; Shaik, S; Wang, B. Conformational Motion of Ferredoxin Enables Efficient Electron Transfer to Heme in the Full-Length P450_TT. J Am Chem Soc. 2021 143(2):1005-1016.
Carroll, E; Ravi Gopal, B; Raghavan, I; Mukherjee, M; Wang, ZQ. A cytochrome P450 CYP87A4 imparts sterol side-chain cleavage in digoxin biosynthesis. Nat Commun. 2023 14(1):4042.
Li, F; Yuan, L; Lu, H; Li, G; Chen, Y; Engqvist, MKM; Kerkhoven, EJ; Nielsen, J. Deep learning-based k_cat prediction enables improved enzyme-constrained model reconstruction. Nat Catal. 2022 5(8):662-672.
Zhang, R; Zhao, Y; Sun, Y; Lu, X; Yang, X. Isolation, Characterization, and Hepatoprotective Effects of the Raffinose Family Oligosaccharides from Rehmannia glutinosa Libosch. J Agric Food Chem. 2013 61(32):7786-7793.
Chen, Q; Chao, Z; Wang, K; Wang, X; Meng, H; Liu, X; Shan, X; Zhou, J. Simplification of Corticosteroids Biosynthetic Pathway by Engineering P450BM3. ACS Catal. 2024:4117-4129.
Ly, TTB; Schifrin, A; Nguyen, BD; Bernhardt, R. Improvement of a P450-Based Recombinant Escherichia coli Whole-Cell System for the Production of Oxygenated Sesquiterpene Derivatives. J Agric Food Chem. 2017 65(19):3891-3899.
Qin, Z; Zhang, Y; Liu, S; Zeng, W; Zhou, J; Xu, S. Combining Metabolic Engineering and Lipid Droplet Assembly to Achieve Campesterol Overproduction in Saccharomyces cerevisiae. J Agric Food Chem. 2024 72(9):4814-4824.
Wang, S; Wang, L; Xu, J; Wang, Y; Xiang, L; He, X. Synergistic Combination of the Total Steroidal Saponins from the Berries of Black Nightshade and Adriamycin to Overcome Leukemia Multidrug Resistance. J Agric Food Chem. 2023 71(7):3315-3324.
Zhu, M; Wang, C; Sun, W; Zhou, A; Wang, Y; Zhang, G; Zhou, X; Huo, Y; Li, C. Boosting 11-oxo-β-amyrin and glycyrrhetinic acid synthesis in Saccharomyces cerevisiae via pairing novel oxidation and reduction system from legume plants. Metab Eng. 2018 45:43-50.
Zhao, X; Wei, W; Li, S; Gao, S; Liu, S; Zeng, W; Zhou, J. Elucidation of the Biosynthesis of Asiaticoside and Its Reconstitution in Yeast. ACS Sustainable Chem Eng. 2024 12(10):4028-4040.
King, E; Cui, Y; Aspacio, D; Nicklen, F; Zhang, L; Maxel, S; Luo, R; Siegel, JB; Aitchison, E; Li, H. Engineering Embden–Meyerhof–Parnas Glycolysis to Generate Noncanonical Reducing Power. ACS Catal. 2022 12(14):8582-8592.
Li, Y; Mao, J; Liu, Q; Song, X; Wu, Y; Cai, M; Xu, H; Qiao, M. De Novo Biosynthesis of Caffeic Acid from Glucose by Engineered Saccharomyces cerevisiae. ACS Synth Biol. 2020 9(4):756-765.
Liu, Y; Denisov, IG; Sligar, SG; Kincaid, JR. Substrate-Specific Allosteric Effects on the Enhancement of CYP17A1 Lyase Efficiency by Cytochrome b5. J Am Chem Soc. 2021 143(10):3729-3733.
Chen, J; Fan, F; Qu, G; Tang, J; Xi, Y; Bi, C; Sun, Z; Zhang, X. Identification of Absidia orchidis steroid 11beta-hydroxylation system and its application in engineering Saccharomyces cerevisiae for one-step biotransformation to produce hydrocortisone. Metab Eng. 2020 57:31-42.
Pourcel, L; Buron, F; Garcia, F; Delaloix, M-S; Le Fourn, V; Girod, P-A; Mermod, N. Transient vitamin B5 starving improves mammalian cell homeostasis and protein production. Metab Eng. 2020 60:77-86.

There is NO Competing Interest.

SupportingDates1.zip
Dataset 1
SupportingDates2.zip
Dataset 2
DatasFiguresmanuscript.xlsx
Dataset 3
manuscriptfigures.pptx
Dataset 4
ChenQHpregnenolonesupplementaryInformation.docx
rs.pdf
Reporting Summary
GraphicalAbstract.png

Download PDF

Version 1

posted

You are reading this latest preprint version

Electron transfer engineering of artificially designed cell factory for complete biosynthesis of natural products

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1