① plasmid construction

P450 enzyme CYP107D1 (Olep), derived from Streptomyces antibioticus, was synthesized after codon optimization (BBa_K4759001) and subcloned to the plasmid pET-28a to obtain the recombinant plasmid pET-28a-oleP(BBa_K4759202). Since the CYP107 family was a typical three-component P450 enzyme that required a pair of redox partners to transfer electrons from the electron donor NAD(P)H to the P450 enzyme activity center, ferredoxin reductase CamA and ferredoxin CamB were selected as redox partners. The codon-optimized camA gene and camB gene were synthesized and subcloned to the plasmid pACYCDuet to obtain the recombinant plasmid pACYCDuet-camA/camB(BBa_K4759201).

② Expression of Olep in a recombinant strain

The above two recombinant plasmids were then transformed into the E. coli expression host C43 (DE3) to obtain one recombinant strain, named E. coli O1.

③ Enzyme purification and CO spectral assay

The purified Olep was obtained by Ni-NTA affinity chromatography. The expressed Olep presented a characteristic peak at 450 nm after the reaction with sodium dithionite and carbon monoxide. The results showed that Olep was an active P450 enzyme.

Fig. 1-1: Full wavelength scan of purified Olep in three states. The blue, orange, and red circles mark the characteristic absorbance of the purified Olep, Olep-CO complex and Olep-CO-sodium dithionite complex, respectively.

The recombinant strain E. coli O1 was cultivated using TB culture and grown at 37 °C until an OD₆₀₀ value of 0.6-0.8 was reached. 0.5 mM IPTG (isopropyl-β-D-thiogalactopyranoside) was added and the culture was incubated for another 20 h at 25 °C.

After the fermentation was completed, the fermentation broth was centrifuged at 4°C, 8000 rpm for 10 min. The bacteria were collected and washed with potassium phosphate buffer at pH 8.0. Then the bacteria were resuspended with potassium phosphate buffer (containing 500 mM sodium chloride, and 20 mM imidazole) at pH 8.0 and performed cell disruption with an ultrasonic disruptor on ice. The procedure of the ultrasonic crusher was as follows: POWER, 38%; 2 s on /3 s off; 5 min. The cell disruption solution was centrifuged at 4°C, 8000 rpm for 10 min, and the supernatant was collected as Olep crude enzyme. Olep pure enzyme was obtained by Ni-NTA affinity chromatography with a pH 8.0 potassium phosphate buffer (containing 500 mM sodium chloride and 200 mM imidazole). The purified Olep was desalted with Amicon Ultra 30 K ultrafiltration centrifuge tubes.
SDS-PAGE showed that the proportion of soluble Olep only can reach 32.4%.

Fig. 1-2: SDS-PAGE analysis of Olep. Lane 1, soluble expression of Olep; Lane 2, inclusion body of Olep; Lane 3, purified soluble Olep. M, marker

First, we selected high-copy plasmid pRSFDuet, medium-high copy plasmid pETDuet and pET28a, and low-copy plasmid pACYCDuet to express Olep in E. coli. The results showed that the recombinant strain constructed with high-copy plasmid pRSFDuet grew better and its conversion rate increased to 29.4%. In WT strain, Olep was expressed in (BBa_K4759202)and CamA/CamB was expressed in (BBa_K4759201); In the plasmid pETDuet, Olep was expressed in (BBa_K4759203) and CamA/CamB was expressed in (BBa_K4759201); In the plasmid pACYCDuet and pRSFDuet, Olep was expressed in (BBa_K4759203) and CamA/CamB was expressed in (BBa_K4759201).

Fig. 2-1: Selection of suitable plasmid. WT represents the E. coli O1 strain. The blue-filled triangle represents the biomass (OD₆₀₀). The red hollow triangle represents the conversion rate (%). Values and triangles represent the means and standard deviations of biological triplicates.

Subsequently, we selected Escherichia coli BL21 (DE3), C41(DE3), and C43(DE3) to express Olep. The results showed that C41(DE3) was the suitable expression host and the conversion rate of 6β-OH DCA was increased to 31.7%.

Fig. 2-2: Selection of proper host

Next, we fused six protein tags (SUMO, GST, MBP, TF, Trx, Nus) at the N-terminal of the Olep. Protein tags MBP and TF can significantly improve the solubility of Olep, but the catalytic efficiency of Olep was reduced.
The expression framework of elements is shown in the following parts:BBa_K4759204、 BBa_K4759205、BBa_K4759206、BBa_K4759207、BBa_K4759208、BBa_K4759209

Fig. 2-3: The application of protein tags

Then, we co-expressed commonly used commercial chaperone plasmids (pGro7, pKJE7, pGKJE8, PTf16, and pTf2) and plasmid pRSFDuet-camA-camB-oleP. However, these chaperones did not significantly increase the expression of Olep, nor did they improve the catalytic efficiency of substrates.

Fig. 2-4: The co-expression of molecular chaperones

Furthermore, the conditions of Olep expression were optimized. The results suggested that the lower temperature (25°C) and moderate concentrations of IPTG (0.5 mM) were crucial for Olep expression (67.1% solubility) and the highest conversion rate of 6β-OH DCA reached 34.8% at the optimal conditions.

Fig. 2-5: (A) The optimal induction temperature for Olep expression (°C).
(B) The proper concentration of IPTG for Olep expression (mM).

By screening different copy number plasmids, E. coli expression host, protein tags, molecular chaperones, and optimizing of induction conditions, the recombinant strain E. coli O2 was constructed. The expression system of E. coli O2 was high-copy recombinant plasmid pRSFDuet-camA-camB-oleP, expressing host E. coli C41(DE3), induction temperature 25°C, inducer IPTG concentration 0.5 mM.
The soluble expression of Olep reached 67.1%, and under the optimized conditions, 0.348 mg/mL 6β-OH deoxycholic acid (conversion rate 34.8%) could be produced using a final concentration of 1 mg/mL of deoxycholic acid. Compared with E. coli O1, the soluble expression of Olep was increased by 51.7%, and the catalytic efficiency was increased by 6.25 times.

In the part of the soluble expression of Olep, it was found that when the N-terminal of Olep was fused with the solubilizing label MBP or TF, although the soluble expression of Olep increased, the catalytic efficiency decreased .

After the end of induction, 1 mL of fermentation broth was collected and the cells were harvested by centrifugation at 14,000 rpm for 10 min at 4 °C. 0.2 mL of the supernatant was taken and mixed with 0.2 mL of ammonia and 1.6 mL of acetonitrile. Then centrifuged it at 14,000 rpm for 20 min, collected the supernatant, which was used to detect extracellular heme. Resuspended the pellet with 0.2 mL of ultrapure water, 0.2 mL of ammonia, and 1.6 mL of acetonitrile. The samples were then placed on ice and the cell suspension was sonicated with an ultrasonic disruptor for 10 min (power 30 W, 3 mm conical microtip probe with a pulse period of 3 s on/3 s off). The supernatant were collected by centrifugation at 14,000 rpm for 20 min, and it was used to detect intracellular heme.

When measuring the heme-binding rate, it was found that the heme binding rate of MBP-Olep and TF-Olep was only 21.1% and 16.8%, which was lower than that of the heme binding rate of the analytic label Olep (39.4%). Therefore, insufficient heme supply may be a key factor limiting the catalytic efficiency of Olep .
There were currently two main ways to enhance the heme supply of the P450 enzyme. One strategy was to exogenously add the precursor 5-aminolevulinic acid (ALA) of heme during P450 enzyme expression, and the other strategy was to enhance the synthetic pathway of heme in E. coli.

Fig. 3-1: The heme-binding ratio of wild-type Olep, MBP-Olep and TF-Olep The blue-filled triangle represents the heme-binding ratio (%). The red hollow triangle represents the conversion rate (%). Values and triangles represent the means and standard deviations of biological triplicates.

First, three different combinations of exogenous additions (direct addition of heme, addition of precursor ALA+FeSO₄, addition of precursor ALA+FeCl₃) were tested. The results showed that the addition of 0.64 mM ALA and 0.3 mM FeCl₃ during the induction of Olep increased the heme binding rate to 67.7% and the conversion rate of deoxycholic acid to 40.7%. Through the above experiments, we verified that the increase in heme content had a positive effect on P450 enzyme activity, but the exogenous addition of ALA increased the cost of whole-cell catalysis by 60%, so we tried to increase intracellular heme content by strengthening the heme synthesis pathway of E. coli itself.

Fig. 3-2: The effect of supplements on the Olep catalysis

We divided the synthesis pathway of heme into two parts, one was the synthesis pathway of heme premise ALA (upstream), and the other part was the synthesis pathway of ALA to heme (downstream).

For the upstream pathway, there were two pathways from glutamate to ALA, one was the C4 pathway and the other was the C5 pathway. The C4 pathway had been enhanced by the existing team, and the experimental results showed that the effect was not as good as the C5 pathway, so we chose the C5 pathway for modification. The C5 pathway had two key genes, one was hemA and the other was hemL. So we constructed plasmids containing hemA and hemL , and overexpressed them by transforming recombinant plasmids to E. coli.

For the downstream pathway, that was, the synthesis pathway of heme, it involved 7 genes, of which 4 genes (hemB, hemD, hemC, hemH) were more critical, according to the literature. Therefore, we also constructed and overexpressed plasmids containing hemB, hemD, hemC, and hemH.

The downstream pathway, that was, the synthesis pathway of heme, involved 7 genes, of which 4 genes (hemB, hemD, hemC, hemH) were more critical, according to the literature. Therefore, we also constructed and overexpressed plasmids containing hemB, hemD, hemC, and hemH.

Fig. 3-3: Heme biosynthetic pathways in E. coli. The purple arrow represents the C5 pathway and the pink arrow represents the downstream biosynthetic pathway of heme. The pCDFDuet-hemA-hemL plasmid was constructed to enhance the C5 pathway; the pETDuet-hemBDC-hemH plasmid was constructed to enhance the downstream biosynthetic pathway

Therefore, 3 recombinant strains were constructed, E. coli AL strain(The recombinant plasmid pCDFDuet-hemA-hemL was expressed in E. coli O2), E. coli BCDH strain (The recombinant plasmid pETDuet-hemB-hemC-hemD-hemH was expressed in E. coli O2), E. coli AL-BCDH strain(The recombinant plasmids pCDFDuet-hemA-hemL and pETDuet-hemB-hemC-hemD-hemH were expressed in E. coli O2).
hemA（ BBa_K4759020）and hemL（BBa_K4759025）were expressed in BBa_K4759210, hemH（BBa_K4759024）was expressed in BBa_K4759211. hemB（BBa_K4759021）/hemC（BBa_K4759022）/hemD（BBa_I716155）were expressed in BBa_K4759270

Fig. 3-4: The color of engineered strains and pure enzyme. 1: E. coli O1 strain; 2: E. coli O2 strain; 3: E. coli O2 strain cultivated with ALA and FeCl₃; 4: E. coli AL strain; 5: Olep enzyme purified from E. coli AL strain. The lower values represent the content of intracellular heme in different engineered strains

The experimental results showed that, without the addition of ALA and FeCl₃, the heme binding rate of Olep increased to 53.9%, and the conversion rate of deoxycholic acid increased to 41.4%.

Fig. 3-5: The effect of enhancing heme biosynthesis on the Olep catalysis

The P450 enzyme requires electrons (NADH, NADPH) to function. Electrons are first transferred to FDR (ferredoxin reductase), to FDX (ferro-reducing protein), and then to P450 enzyme under the action of redox partners, and finally to the heme center, after which the enzyme can react with the substrate and undergo hydroxylation.

Through extensive reading of the literature, we summarized 11 pairs of redox partners with good results. After that, we screened four pairs of redox partners with good effects by molecular docking and mathematical modeling and then verified them experimentally.

Generally, the method of determining whether the redox partners were suitable required tedious steps such as the construction of plasmids, heterologous expression, construction of catalytic systems, and detection of conversion rate after catalysis. Therefore, we wanted to find a convenient way to do a quick screening. We used the fluorescent protein sfGFP and successfully constructed a sensor to detect redox partners.

We divided sfGFP into N-terminal and C-terminal, and although these two parts were cut off, there was an interaction force between them. Thus, four iron redox proteins were fused to the N-terminal of sfGFP-1-10 and Olep to the C-terminal of sfGFP-11 , respectively, to obtain the recombinant plasmid pRSFDuet-BM3-GFP-1-10-GFP-11-oleP, pRSFDuet-camA-camB-GFP-1-10-GFP-11-oleP, pRSFDuet-FdR_0978-Fdx_1499-GFP-1-10-GFP-11-oleP, and pRSFDuet-petH-petF-GFP-1-10-GFP-11-oleP.

The above four recombinant plasmids were converted to BL21(DE3) to obtain recombinant strains G2 to G5.

The following table is to show parts used in this section.

Fig. 4-1: The scheme of constructing sfGFP sensor

Fig. 4-2: The self-assembly of Olep and Fdx based on the three-dimensional structure of sfGFP (PDB: 5BT0)

The recombinant strains G2 to G5 were subjected to shaker fermentation experiments. After the fermentation was completed, 200 uL bacteria were added to the 96-well plate with a microplate reader to determine biomass (wavelength 600 nm) and fluorescence value (excitation wavelength 488 nm, emission wavelength 520 nm). Then we calculated the fluorescence intensity (fluorescence value/biomass) of the strain. The fluorescence intensity of the recombinant strain G5 (containing recombinant plasmid pRSFDuet-petH-petF-GFP-1-10-GFP-11-olep) was the highest (1.2×10⁶) and 6 times higher than that of the control strain G2 (containing recombinant plasmid pRSFDuet-camA-camB-GFP-1-10-GFP-11-olep).

Fig. 4-3: BL21 morphology diagram seen under excitation light 488nm and emitted light 520nm, green is green fluorescence of sfGFP

We selected four conventional redox partners (BM3, CamA/CamB, SelFdR0978/SelFdx1499, PetH/PetF) in combination with the P450 enzyme. Four groups of redox partners were constructed on the high-copy plasmid pRSFDuet to obtain recombinant plasmids: pRSFDuet-BM3-olep, pRSFDuet-camA-camB-olep, pRSFDuet-FdR0978-Fdx1499-olep, and pRSFDuet-petH-petF-olep. Then they were transformed to C41 (DE3) to obtain the recombinant strain R2 to R5. The recombinant strains R2 to R5 were subjected to shaker fermentation experiments.
The recombinant strain R5 (containing recombinant plasmid pRSFDuet-petH-petF-olep) had the highest conversion rate , which significantly increased from 41.4% to 85.6%. Therefore, the redox partners PetH/PetF derived from Synechocystis was successfully screened as the most suitable redox partners for the P450 enzyme Olep, and the construction of the sfGFP sensor was verified , which could efficiently and accurately screen the redox partner adapted by the P450 enzyme.

The following table is to show parts used in this section.

Fig. 4-4: (A) Screening proper redox partners for Olep from different sources. The G1 strain that contains the empty pRSFDuet-1 plasmid was used as a control. The fluorescent intensities were calculated and the color of cells and fluorescent images were presented for G2-G5 strains that express different redox partners-sfGFP-1-10 and sfGFP-11-Olep, respectively.
(B) The conversion rates were calculated for R2-R5 strains that express different redox partners and Olep, respectively. The R1 strain that contains the empty pRSFDuet-1 plasmid was used as a control. The blue-filled triangle represents the fluorescent intensity/OD₆₀₀. The red hollow triangle represents the conversion rate (%). Values and triangles represent the means and standard deviations of biological triplicates

Fig. 4-5: The structures and interactions between Olep and Fdxs are presented. The key interacting residues in Olep-Fdx complexes are depicted as sticks and highlighted in yellow. Heme and substrates are displayed as sticks, colored in red and wheat, respectively. The Fe2S2 cluster is visualized as spheres. The distances (Å) between the iron–sulfur cluster and heme-iron are measured and indicated by dashed red lines. The interaction areas of Olep-Fdx are calculated by NovoPro (https://www.novopro.cn/). The numbers of hydrogen bonds and salt bridges are predicted by PDBePISA (https://www.ebi.ac.uk/pdbe/)

The fusion combination strategy of redox partners can help to improve the catalytic activity of Olep. To further screen the optimal redox partner, we decided to make a fusion combination of PetF\PetH.
We adopted the following strategies and all recombinant plasmid were transformed to C41 (DE3), respectively: recombinant strains R6 (PetH, PetF, and olep were fused by two linkers); recombinant strains R7 (petH and petF were fused by a linker); recombinant strains R8 (PetH and PetF was constructed on pACYCDuet, while oleP expression gene was constructed on pRSFDuet); recombinant strains R9 (petH and petF were fused by a linker, and were constructed on pACYCDuet, while the oleP expression gene was constructed on pRSFDuet).

Fig. 4-6: The fusion expression and different expressional ratio between Olep and PetH/PetF

The following table is to show parts used in this section.

Recombinant strains R6 to R9 were subjected to shake flask fermentation. HPLC assay for product generation. The transformation rate of recombinant strains R6 to R9 were lower than that of the recombinant strain R5.

Fig. 4-7: The conversion rate of DCA for R5-R9 strains. The blue-filled triangle represents the biomass (OD₆₀₀).

After obtaining the best redox partners PetH/PetF, we performed alanine scanning on petF to speculate which sites had a greater impact on its electron transport capacity. Finally, we found that after mutations in seven of them, the electron transport effect would change greatly, so we mutated the amino acids of these sites into other 19 amino acids by modeling, and selected 23 of them to get better results.
(Site-directed Mutagenesis parts can be search from BBa_K4759053 to BBa_K4759075, and constructed modeling screening for redox partners can be search from BBa_K4759076 to BBa_K4759099 )

Fig. 4-8: Fermentation of 23 mutants and control groups

We conducted control tests with the positive control group, negative control group, and wild-type strains, and finally selected 9 mutants with the highest fluorescence intensity for subsequent catalytic verification by detecting their green fluorescence intensity.

Fig. 4-9: 9 mutants + wild-type + negative control, 50 ml/250 ml system fermentation

By verifying the catalytic ability, we found that the substrate conversion of D68P was higher than that of the wild type in the nine strains with high fluorescence intensity, reaching 89.2%.

Fig. 4-10: A:Fluorescence intensity of wild type with 23 mutants
B: Conversion of the 9 mutants with the highest fluorescence intensity with wild type

First, we explored catalytic forms. We measured the substrate conversion rate of Whole-Cell, Permeabilized Cell, Crude enzyme, and fermentation, respectively, and found that the substrate conversion rate was the highest under whole-cell catalytic conditions.

Fig. 5-1: Effect of catalytic form.

The red hollow triangle represents the conversion rate (%). Values and triangles represent the means and standard deviations of biological triplicates.

In the next step, we explored the optimal biomass by changing the size of the cellular biomass added to the whole cell catalytic system. When the substrate concentration at the time of reaction was set to 1 mg/mL and the catalysis time was set to 12 h, the effect of biomass OD₆₀₀ (10, 15, 20, 25, 30, 35) on whole cell catalysis was investigated. The results showed that the conversion rate was the highest when OD₆₀₀ was selected as 30 .

Fig. 5-2: Effect of different biomass of resting whole-cell (OD₆₀₀)

Next, when the biomass OD₆₀₀ was set to 30 and the catalysis time was set to 12 h, the effect of substrate concentrations (0.5 mg/mL, 1.0 mg/mL, 1.5 mg/mL, 2.0 mg/mL, 2.5 mg/mL, 3.0 mg/mL, 3.5 mg/mL) on whole-cell catalysis was investigated and we found that the conversion efficiency was the highest when the substrate concentration was 0.5 mg/mL .

Fig. 5-3: Effect of concentration of DCA (mg/mL)

Subsequently, when the biomass OD₆₀₀ was set to 30 and the substrate concentration was set to 2.0 mg/mL, the effect of catalysis time (0 h, 2 h, 4 h, 6 h, 8 h, 10 h, 12 h, 14 h, 16 h, 18 h, 20 h, 22 h, 24 h) on whole cell catalysis was investigated and we found that the conversion rate of 20-24 h reached the highest and remained unchanged by controlling different catalytic time. Considering the time was too long, we chose the catalytic time of 12h, and the conversion rate could reach more than 90%.

Fig. 5-4: Effect of catalytic time (h)

P450 enzyme catalysis required redox partners to transfer NADPH electrons to heme to catalyze the substrate. The content of NADPH in E. coli cells was low, which can only meet its own growth needs. By constructing a cofactor circulatory system, the effect of NADPH on whole-cell activity was investigated. However, the cost of exogenous NADPH addition was higher and the catalytic efficiency was low. NAD+ kinase (NadK) and membrane-bound hydrogenase (PntAB) can be used to enhance the circulation of NADPH. Thus, 3 recombinant strains are constructed. The recombinant plasmid pACYCDuet-nadK is expressed in E. coli C1 (C2 strain). The recombinant plasmid pACYCDuet-pntAB (BBa_K4759269) was expressed in E. coli C1 (C3 strain). The recombinant plasmid pACYCDuet-pntAB-nadK(BBa_K4759271) was expressed in E. coli C1 (C4 strain). Without the addition of NADPH, the conversion rate of whole-cell catalysts prepared by the C4 strain was as high as 99.1% .

Fig. 5-5: Effect of NADPH addition and intracellular NADPH regeneration system