De Novo Biosynthesis of Antidepressant Psilocybin in Escherichia coli

Genes and primers were synthesised by Tsingke Biotechnologies (Beijing, China). Phanta Super‐Fidelity DNA Polymerase (Vazyme) was used for PCR amplification. DNA restriction enzymes were purchased from New England Biolabs (USA). SE seamless cloning and assembly kit (Zoman) was used for in vitro DNA assembly and plasmid construction. Acetonitrile was HPLC grade and was purchased from Sigma‐Aldrich (St. Louis, MO, USA). All other chemicals and solvents were of analytical grade. All buffers and solutions were prepared in Milli‐Q water.

To obtain the active P450 enzyme PsiH in E. coli, the genes encoding PsiH, cytochrome P450 reductase (PcCPR) and cytochrome b5 (PcCYB5) from P. cubensis were codon optimised and synthesised. To improve the expression of psiH, the small ubiquitin‐like modifier (SUMO) coding sequence was also codon optimised and synthesised. Then, the wild‐type psiH and various psiH variants, including the genes encoding the truncated PsiH (trPsiH), SUMO‐PsiH, SUMO‐trPsiH, 5144‐trPsiH and SUMO‐5144‐trPsiH, were constructed and cloned into the NdeI and XhoI digested plasmid pET28a, respectively, resulting in plasmids pHZR01–6 (Table S1). The PcCPR gene was cloned into plasmid pGro7 and polycistronically expressed with the chaperone protein encoding genes groES and groEL under the control of the araBAD promoter, resulting in plasmid pHZR07. To further enhance the activity of PsiH, the PcCYB5 gene, which encodes the second ET protein of fungal P450s (Durairaj et al. 2016), was further cloned into plasmid pHZR07, resulting in plasmid pHZR08.

To optimise the first biosynthetic step of psilocybin, three L‐tryptophan decarboxylase genes, including psiD from P. cubensis, BaTDC from Bacillus atrophaeus and CrTDC from Catharanthus roseus, were codon optimised and synthesised. They were cloned into the NdeI and XhoI digested plasmid pRSFDuet‐GmR, which was derived from pRSFDuet‐1 by changing the kanamycin resistance gene to the gentamicin resistance gene, resulting in plasmids pHZR09–11, respectively.

To reconstitute the biosynthetic pathway of psilocybin in E. coli, the psiK and psiM genes from P. cubensis were codon optimised and synthesised. They were ligated to the two cloning sites of plasmid pRSFDuet‐GmR, resulting in plasmid pHZR12. The psiD gene was then cloned into pHZR12, resulting in plasmid pHZR13.

To improve the supply of cofactor NADPH for the PsiH‐catalysed hydroxylation reaction, the nadK gene from E. coli was amplified and cloned into the plasmid pCDFDuet‐Amp, which was derived from pCDFDuet‐1 by changing the streptomycin resistance gene to the ampicillin‐resistant gene, resulting in plasmid pHZR14. To improve the supply of substrate S‐adenosyl‐L‐methionine (SAM) for the PsiM‐catalysed methylation reaction, the metK gene was amplified from E. coli, and the SAM2 gene from Saccharomyces cerevisiae was codon optimised and synthesised. They were then cloned into plasmid pHZR14, resulting in plasmids pHZR15 and pHZR16, respectively. The constructed plasmids were transformed into E. coli strain DH5α for propagation and sequenced for further expression in E. coli strain BL21 (DE3). All plasmids constructed in this work are listed in Table S1, and the primers used in this study are listed in Table S2; the codon‐optimised sequences expressed in this work are shown in Table S4.

The CRISPR/Cas9‐mediated gene editing strategy was used to knock out the trpR and tnaA genes in E. coli BL21. The plasmid pCas‐prha‐NEW for expressing cas9 was transformed into E. coli BL21. Then, 10 mM of L‐arabinose was added for λ‐Red induction, and E. coli BL21 electroporation‐competent cells harbouring pCas‐prha‐NEW were prepared (Sharan et al. 2009). For electro‐transformation, 50 μL of cells were mixed with 100 ng of pTarget series and 400 ng of donor DNA (fragment containing 500 bp up‐ and downstream of the target genes). Electroporation was performed in a 1 mm Gene Pulser cuvette (Bio‐Rad) at 2.5 kV, and the product was immediately suspended in 1 mL of ice‐cold LB medium. Cells were recovered at 30°C for 1 h, and transformants were selected on LB agar containing kanamycin (50 μg/mL) and spectinomycin (50 μg/mL) overnight at 30°C. The knockout strains were further verified by colony PCR and DNA sequencing.

To cure the pTarget series, the knockout strain carrying both pCas‐prha‐NEW and the pTarget series was inoculated in 2 mL of LB medium containing kanamycin (50 μg/mL) and L‐rhamnose monohydrate (10 mM) for 16 h. The cells were then diluted and selected on LB plates containing kanamycin (50 μg/mL). Colonies were confirmed as cured by determining their sensitivity to spectinomycin (50 μg/mL). The cured colonies from the pTarget series were used in a second round of genome editing. Finally, pCas‐prha‐NEW was cured by growing the colonies in LB medium without any antibiotics overnight at 37°C.

To verify the expression level of different psiH variants, plasmids pGro7 and each of pHZR01–6 were co‐transformed into E. coli BL21 (DE3). Transformants were selected on the LB agar plates supplemented with 50 μg/mL kanamycin and 25 μg/mL chloramphenicol at 37°C overnight. The colony was then picked and cultured in 3 mL of LB medium at 37°C overnight. The incubated cells were then transferred to 100 mL of fresh LB medium (1% v/v) supplemented with 50 μg/mL kanamycin and 25 μg/mL chloramphenicol. When the OD₆₀₀ of the culture reached approximately 0.4 ~ 0.6, 0.1 mM IPTG and 500 μg/mL L‐arabinose were added to induce protein expression for 14 h at 20°C. 3 mL of cells were then collected and resuspended in 1 mL Tris–HCl buffer (50 mM, pH 7.5). Total cell lysis was used to detect protein expression by SDS‐PAGE analysis.

To test the catalytic ability of different PsiH variants, plasmids pHZR07 and each of pHZR01–6 were co‐transformed into E. coli BL21 (DE3), resulting in strains H1–6. The strains were cultured, and proteins were expressed as described above. 50 mL of cells were harvested by centrifugation. Then, 1 mL of the fresh LB medium supplemented with 500 μg/mL tryptamine was added, and the cells were further cultured for 1 day at 20°C. To verify the catalytic ability of different L‐tryptophan decarboxylases, plasmids pHZR09–11 were transformed into E. coli BL21 (DE3), resulting in strains D1–3. After the protein was expressed, 50 mL of cells were harvested by centrifugation. Then, 1 mL of the fresh LB medium supplemented with 500 μg/mL of L‐tryptophan was added, and the cells were further cultured for 1 day at 20°C. To verify the catalytic ability of PsiK and PsiM, the plasmid pHZR12 was transformed into E. coli BL21 (DE3), resulting in strain KM. After the protein was expressed, 50 mL of cells were harvested by centrifugation. Then, 800 μL of PsiH‐transformed products and 200 μL of the fresh LB medium were added, and the cells were further cultured for 1 day at 20°C. After the in vivo transformation, 500 μL methanol was added to 500 μL of the cell mixture. The mixture was vortexed and centrifuged at 12,000 rpm for 10 min. The supernatant was then used for HPLC analysis.

To achieve de novo biosynthesis of psilocybin in E. coli, the plasmids pHZR02, pHZR07 and pHZR13 were co‐transformed into E. coli BL21 (DE3), resulting in strain P03. Transformants were selected on the LB agar plate supplemented with 50 μg/mL kanamycin, 25 μg/mL chloramphenicol and 40 μg/mL gentamicin at 37°C overnight. The colony was then picked and cultured in 3 mL of LB medium at 37°C overnight. Subsequently, the incubated cells were transferred to 100 mL of the fresh LB medium (1% v/v) supplemented with antibiotics as described above. When the OD₆₀₀ of the culture reached approximately 0.4 ~ 0.6, 0.1 mM IPTG and 500 μg/mL L‐arabinose were added. The cells were further cultured at 20°C, 220 rpm for 24 h.

To detect the product, 500 μL methanol was added to 500 μL of the cell mixture. The mixture was vortexed and centrifuged at 12,000 rpm for 10 min. The supernatant was then used for HPLC analysis. To further improve psilocybin production, additional strains were constructed and tested, and they are listed in Table 1. All the proteins expressed in this study are listed in Table S3.

To test psilocybin could be de novo synthesised in strain P13, a modified M9 medium was used for fermentation. The modified M9 medium contained 17.4 g/L Na₂HPO₄·7 H₂O, 3 g/L KH₂PO₄, 1 g/L NH₄Cl, 0.5 g/L NaCl, 0.49 g/L MgSO₄·7H₂O, 0.11 g/L CaCl₂ and 20 g/L glucose. The colony of strain P13 on the LB plate was selected and cultured in 3 mL of M9 medium at 37°C overnight. Subsequently, the incubated cells were transferred to 100 mL of M9 medium in a baffled flask (500 mL) supplemented with the corresponding antibiotics. When the OD₆₀₀ of the culture reached approximately 0.4 ~ 0.6, 0.1 mM IPTG and 500 μg/mL L‐arabinose were added. The cells were further cultured at 20°C, 220 rpm for 120 h. At every 12‐h interval, 0.5 mL of the broth was collected and psilocybin production was determined. To further improve psilocybin production, the optimal fermentation mediums, including fresh LB, LBG (LB with 2% glucose) and TB medium, were investigated.

High performance liquid chromatography (HPLC) analysis was performed on a Waters 2695 system (USA) using a C18 analytical column (Gemini 250 × 4.6 mm, particle size 5 μm; Phenomenex). The mobile phase flow rate was 1 mL/min, and the column temperature was 25°C. The analytical method is a linear gradient of 5%–19% acetonitrile (MeCN)‐H₂O (v/v, 0.1% formic acid) for 12 min, followed by 100% MeCN for 4 min and rebalanced by 5% MeCN (v/v, 0.1% formic acid) for 6 min. Using this method, the following retention times were observed: norbaeocystin (5.92 min), baeocystin (6.41 min), psilocybin (6.75 min), 4‐hydroxytryptamine (6.95 min), norpsilocin (7.25 min), psilocin (7.93 min), tryptamine (8.96 min) and tryptophan (10.00 min).

For the liquid chromatograph‐high resolution mass spectrometer (LC‐HRMS) analysis, liquid chromatography (LC) separation was performed on an Agilent 1260 series system (Agilent, USA) with a C18 analytical column (Gemini 250 × 4.6 mm, particle size 5 μm; Phenomenex). The LC analytical method is the same as described above. Mass spectra were performed on an Agilent Accurate‐Mass‐Q‐TOF MS 6520 system equipped with an electrospray ionisation (ESI) source in positive mode. The LC‐HRMS data were recorded and analysed using MassHunter Workstation software (version B.04.00, Agilent, Santa Clara, CA, USA).

The chemical compounds psilocybin (Cayman), tryptamine (Macklin) and tryptophan (Macklin) were purchased as authentic standards, and their standard curves were obtained by HPLC analysis. The peak area of the HPLC trace was recorded under 280 nm for all compounds. The concentration of the authentic standards was set as x, and the peak area of the HPLC trace was set as y, resulting in a regression equation (Figure S10). Due to limited commercial availability and extremely high cost, baeocystin, norbaeocystin and 4‐hydroxytryptamine were quantified using the standard curve of their analogues (Adams et al. 2019). Baeocystin and norbaeocystin were quantified using the standard curve of psilocybin. 4‐hydroxytryptamine was quantified from the standard curve of 5‐hydroxytryptamine.

The hydroxylation of tryptamine catalysed by the P450 enzyme PsiH is the rate‐limiting step for the biosynthesis of psilocybin in E. coli. Due to the membrane‐bound nature of PsiH and the need for enzymatic redox partners for catalysis, de novo synthesis of psilocybin in E. coli has been hindered (Adams et al. 2019; Flower et al. 2023). Most eukaryotic P450 enzymes have N‐terminal transmembrane regions that are natively located in the membranes of the endoplasmic reticulum or mitochondrion (Durairaj and Li 2022). Therefore, the expression of eukaryotic P450s in E. coli often leads to the aggregation of insoluble proteins (Hausjell et al. 2018). Previous studies have shown that N‐terminal modifications could improve the expression of eukaryotic P450s in E. coli, such as deleting the hydrophobic domain (Park et al. 2014) or replacing the hydrophobic domain with the N‐terminal sequence of other P450s (Ichinose et al. 2015; Ichinose and Wariishi 2013). PsiH has an N‐terminal transmembrane region of 19 amino acids (Figure S1), and its C‐terminal catalytic domain is separated from the N‐terminal hydrophobic domain by a proline‐rich region (PPGPP) (Figure 1A). Using this region as a boundary, the N‐terminal hydrophobic region and the C‐terminal catalytic functional domain of PsiH were separated, and several chimeric PsiH proteins were constructed (Figure 1A). We deleted the N‐terminal domain (shown in red) of the wild‐type PsiH to generate the truncated PsiH protein (trPsiH). We also replaced the N‐terminal domain of the wild‐type PsiH with the corresponding domain of CYP5144C1, which has been reported to enhance the expression of several fungal P450s in E. coli (Ichinose et al. 2015), to form the chimeric protein 5144C1NTD‐trPsiH. In addition, we also constructed chimeric proteins SUMO‐trPsiH, SUMO‐5144C1NTD‐trPsiH and SUMO‐PsiH (Figure 1A) because the small ubiquitin‐like modifier (SUMO) is a widely used tag sequence to enhance protein expression in E. coli (Peroutka Iii et al. 2011). The chimeric proteins were expressed respectively based on the medium copy number plasmid pET28a in E. coli BL21 along with the molecular chaperones GroES/GroEL, which could facilitate the proper folding of eukaryotic P450s to obtain the active enzymes (Ahn et al. 2004; Ichinose et al. 2015). The expressed proteins in total cell lysate were then detected by SDS‐PAGE analysis. The results showed that all the PsiH variants were successfully expressed (Figure 1B), but they cannot be purified due to their insolubility. In terms of expression level, all the chimeric proteins have a higher expression level compared to the wild‐type PsiH (Figure 1B, line 2). Amongst them, the expression level of SUMO‐5144C1NTD‐trPsiH, SUMO‐trPsiH and trPsiH was significantly improved (Figure 1B, lines 5, 7 and 8). These results demonstrated that both deletion of the N‐terminal hydrophobic domain and addition of the SUMO sequence promoted the expression of the PsiH protein.

To test the catalytic ability of the chimeric PsiH proteins, the PcCPR gene from P. cubensis was cloned into the low copy number plasmid pGro7 and was polycistronically expressed with groEL and groES in E. coli. After the substrate tryptamine was fed to the engineered E. coli strains (Table 1), the bio‐transformed product 4‐hydroxytryptamine was detected by HRMS analysis (Figure S2B) and further quantified by HPLC analysis (Figure S2A). As shown in Figure 1C, all the strains produced 4‐hydroxytryptamine. Amongst them, the strain H2 expressing trPsiH produced the highest titre of 4‐hydroxytryptamine at 107.82 mg/L, which was 10.53‐fold higher than the production of 9.35 mg/L from the strain H1 expressing the wild‐type psiH. In addition, the strain H4 expressing SUMO‐trPsiH and strain H6 expressing SUMO‐5144C1NTD‐trPsiH produced 4‐hydroxytryptamine at 97.96 mg/L and 88.97 mg/L, respectively, which were 9.48‐ and 8.52‐fold higher than that of the strain H1. These results indicated that the production of 4‐hydroxytryptamine was positively related to the expression level of the chimeric PsiH proteins. Since the strain H2 expressing trPsiH produced the highest titre of 4‐hydroxytryptamine, trPsiH was chosen to reconstitute the psilocybin biosynthetic pathway in E. coli.

We also determined the effect of culture temperature on the catalytic ability of trPsiH. Expression of trPsiH was induced at 18°C, 20°C, 22°C, 25°C and 28°C, respectively. By feeding tryptamine, the production of 4‐hydroxytryptamine was quantified by HPLC analysis. The results showed that the strain H2 cultured at 18°C, 20°C and 22°C gave a comparable amount of 4‐hydroxytryptamine (Figure ), while the production of 4‐hydroxytryptamine decreased sharply when the strain H2 was cultured at 25°C, indicating that trPsiH might have a good expression level and catalytic efficiency in the range of 18 to 22°C. S2C

Having obtained the chimeric PsiH with high catalytic performance, we sought to reconstruct the entire biosynthetic pathway of psilocybin in E. coli. Decarboxylation of tryptophan catalysed by PsiD is the first step of psilocybin biosynthesis, and sufficient precursor supply of tryptamine is crucial to improve psilocybin production. Since the L‐tryptophan decarboxylases BaTDC from B. atrophaeus (Choi et al. 2021) and CrTDC from C. roseus (Noé et al. 1984) showed relatively low Km values towards L‐tryptophan and high turnover numbers, their coding genes were optimised and synthesised together with the psiD gene from P. cubensis. They were cloned into the plasmid pRSFDuet‐GmR and were expressed in E. coli BL21. By feeding tryptophan, tryptamine was produced in all three strains (Figure S3A), confirming that all three decarboxylases could convert tryptophan to tryptamine. We further performed the in vivo conversion using the enzymes expressed at different temperatures in the range of 18°C–37°C. We found that the titre of tryptamine from the strain expressing CrTDC decreased dramatically with increasing temperature, but the strain expressing psiD gave a relatively stable tryptamine production (Figure S3B). Meanwhile, the strain expressing BaTDC gave the highest titre of tryptamine at 20°C, which was improved by 45.8% compared with that of the strain expressing psiD. Therefore, BaTDC could be a preferred candidate for the reconstruction of the psilocybin biosynthetic pathway in E. coli.

Since the expression levels of psiK and psiM are crucial for the efficiency of psilocybin synthesis, we expressed the psiM and psiK genes based on the high copy number plasmid pRSFDuet‐GmR in E. coli BL21. By adding 4‐hydroxytryptamine‐containing supernatant, which was prepared from the tryptamine‐fed strain H2, the final product psilocybin was detected compared to the negative control and authentic standard (Figure S4A), demonstrating that PsiM and PsiK could convert 4‐hydroxytryptamine to psilocybin. We further expressed psiM and psiK at different temperatures ranging from 18°C to 37°C. We found that the titre of psilocybin increased with the increase in culture temperature and reached the highest titre of psilocybin at 37°C (Figure S4B).

Culture temperature is critical for the enzymatic activities in the de novo biosynthetic pathway of psilocybin. According to the effect of culture temperature on enzymatic performance, we found that low temperature was beneficial for PsiH and BaTDC (Figure S2C and 3B), and relatively high temperature was beneficial for PsiK and PsiM (Figure S3D). Considering that the decarboxylation of tryptophan catalysed by BaTDC is the first step and the hydroxylation of tryptamine catalysed by PsiH is the rate‐limiting step to obtain de novo biosynthesis of psilocybin in E. coli, we expressed the whole biosynthetic enzymes at 20°C, which was the appropriate temperature for BaTDC and PsiH expression, to ensure the abundant precursors for psilocybin biosynthesis. To obtain the de novo biosynthesis of psilocybin in E. coli, BaTDC, psiK and psiM were cloned into the high copy number plasmid pRSFDuet‐GmR, and the resulting plasmid pHZR13 was co‐transformed into E. coli BL21 together with plasmids pHZR02 and pHZR07 (Figure 2A), resulting in the initial strain P03. The strain P03 was cultured at 20°C for 24 h. The products were detected by HPLC analysis. As shown in Figure 2B, psilocybin was produced compared to the negative control and authentic standard. HRMS analysis further showed that the incomplete methylation intermediates norbaeocystin and baeocystin remained (Figure 2B,C, Figure S5). In addition, both norpsilocin and psilocin, the spontaneous or enzyme‐catalysed dephosphorylation products of baeocystin and psilocybin (Milne et al. 2020), were also detected (Figure 2B, Figure S5).

The successful biosynthesis of psilocybin indicated that E. coli could be developed into a robust platform for eukaryotic P450 characterisation and engineering (Park et al. 2022). The biosynthesis of 4‐hydroxytryptamine catalysed by PsiH requires a two‐ET from NADPH. The low titre of psilocybin and a large amount of residual tryptamine in strain P03 prompted us to improve the activity of PsiH by engineering its electron transfer (ET) pathway. PsiH belongs to the fungal class II P450 system (Durairaj et al. 2016), and its ET pathway consists of a membrane‐bound protein CPR containing the cofactors FAD and FMN, which transfer two electrons from NAD(P)H to the heme moiety (Figure 3A). Alternatively, it may contain a third small protein component, cytochrome b5, which transfers a second electron to the heme moiety, and overexpression of CYB5 to enhance the ET pathway has gained much success in enhancing production (Paddon et al. 2013; Sun et al. 2022; Yao et al. 2022). To improve the ET efficiency, the PcCYB5 gene from P. cubensis was polycistronically expressed with PcCPR in P03. The resulting strain P04 did give a significantly increased titre of norbaeocystin, baeocystin and psilocybin at 20.4, 8.1 and 4.1 mg/L, respectively, which was 5.7, 6.5 and 4.7‐fold higher than that of the parental strain P03 (Figure 3B). In addition, the titre of tryptamine in strain P04 decreased to 18.6 mg/L, which was 57.6% lower than that of strain P03 (Figure 3B). These results indicated that the catalytic activity of PsiH was enhanced by improving the ET pathway.

Redox cofactor NAD(P)H is the important electron source in cellular primary and secondary metabolism and is indispensable for the catalytic cycle of fungal P450s. Lack of NAD(P)H could reduce P450 activity due to inefficient electron transfer. Improving cellular NADPH levels could enhance the production of P450‐involved biosynthesis (Liu et al. 2021). Thus, we rerouted redox metabolism to improve the activity of PsiH. Previous studies showed that overexpression of the NAD+ kinase gene (nadK) in E. coli improved NADPH supply, which further led to bacterial isobutanol production (Shi et al. 2013). Therefore, we overexpressed the native nadK gene on the plasmid pCDFDuet‐Amp in the strain P04. The resulted strain P05 gave increased titers of norbaeocystin, baeocystin and psilocybin at 48.6, 11.5 and 9.5 mg/L, which were increased by 134%, 42% and 132% compared to that of its parental strain P04, respectively (Figure 3B), demonstrating that the catalytic activity of PsiH was improved by enhancing NADPH supply.

Adequate precursor supply is critical for the biosynthesis of target compounds (Sang et al. 2023). Since tryptophan is the starting precursor for psilocybin biosynthesis, increasing tryptophan accumulation would enhance psilocybin production. The biosynthetic pathway of tryptophan has been extensively studied, and there are a number of reports describing modifications for increasing the supply of tryptophan (Ren et al. 2023). Tryptophanase (TnaA) catalyses the conversion of tryptophan to indole (Figure 4A), so knocking out the tnaA gene would block the metabolic flow to indole and thus increase tryptophan accumulation (Lee et al. 2020). We obtained the tnaA disruption mutant of E. coli BL21 by using the CRISPR/Cas9‐mediated gene editing tools (Figure S6) as described previously (Jiang et al. 2015) and named it strain P01. Then, PcCYB5, nadK and the psilocybin biosynthetic genes were expressed in strain P01, resulting in strain P06. The titers of tryptamine, norbaeocystin, baeocystin and psilocybin were 30.4, 95.4, 11.6 and 11.1 mg/L, which were improved by 27%, 96%, 0% and 17%, respectively, compared with those of strain P05(Figure 3B), showing that blocking the degradation of tryptophan could significantly improve the production of norbaeocystin. But the production of baeocystin and psilocybin was hardly improved, indicating that methylation was the rate‐limiting step. We further relieved the feedback inhibition of tryptophan biosynthesis in strain P01 by knocking out the transcriptional repressor encoding gene trpR (Figure 4A, Figure S6), resulting in strain P02. By co‐expressing PcCYB5, nadK and the psilocybin biosynthetic genes in strain P02, the resulting strain P07 produced tryptamine, norbaeocystin, baeocystin and psilocybin at the titers of 32.3, 105.3, 13.9 and 14.0 mg/L, representing an increase of 6.3%, 10.1%, 19.8% and 26.1%, respectively, over strain P06.

Psilocybin is a dimethylated compound, and S‐adenosyl‐L‐methionine (SAM) acts as the major methyl donor (Figure 5A). In the strain P07, the total titre of norbaeocystin and baeocystin, which were the major components of the products, was 119.2 mg/L. The two possible reasons were that: (i) the lack of SAM supplementation for PsiM‐involved methylation; (ii) the insufficient catalytic ability of methyltransferase PsiM. To test whether the supply of methyl donor was a limiting factor for psilocybin synthesis (Kunjapur et al. 2016; Shen et al. 2022), 0.5 mg/mL methionine (Met) was fed into the culture of strain P07. The results showed that strain P07 produced 11.9 mg/L psilocybin (Figure S7), which was lower than that of the control group. We further activated the SAM cycle by overexpressing the methionine adenosyltransferase gene to regenerate SAM. We first overexpressed the native metK gene (Liu et al. 2022) in strain P07 on plasmid pCDFDuet‐Amp, but the resulting strain P08 showed decreased production of baeocystin and psilocybin (Figure 5B). When metK was replaced by the SAM2 gene (Liu et al. 2022) from S. cerevisiae (strain P09), we found a similar result (Figure 5B). These results indicated that SAM supply was not the limiting factor for the catalytic efficiency of PsiM.

Since PsiM is the only enzyme that catalyses the successive methylation reaction of converting norbaeocystin to the final product, its catalytic activity is crucial for psilocybin production. To obtain a PsiM enzyme with higher catalytic activity, the cDNA of PacPsiM from Panaelus cyanescens, GdPsiM from Gymnopilus dilepis and PscPsiM from Psilocybe cyanescens (Figure S8) was codon optimised and synthesised to replace PcPsiM in strain P07 to generate strains P10–12 (Table 1). However, no strain produced a higher titre of psilocybin than that of strain P07 (Figure S9), suggesting that PcPsiM from P. cubensis was the optimal allele in E. coli.

We further increased an additional copy of PcPsiM on plasmid pHZR02 using an independent T7 expression frame after the psiH gene. The resulting plasmid pHZR20 was co‐transformed into strain P02 with plasmids pHZR08, pHZR13 and pHZR14 to generate the strain P13. Strain P13 showed an obviously increased psilocybin titre of 33.8 mg/L in flask fermentation, which was 1.4 times higher than that of strain P07. The titre of unmethylated norbaeocystin was 8.1 mg/L, which was 92.3% lower than that of strain P07, demonstrating that increasing the copy number of the psiM gene could greatly enhance the first step of methylation. However, excessive amounts of one‐step methylation product baeocystin (36.0 mg/L) were accumulated, indicating that the second round of methyl transfer was still inefficient, which may be attributed to the higher catalytic efficiency of PsiM towards norbaeocystin than that of baeocystin (Hudspeth and Rogge 2024).

To test the ability of the engineered strain in de novo synthesising psilocybin, a modified M9 medium was used for fermentation. The result showed that psilocybin reached its maximum titre of 9.4 mg/L at 72 h, demonstrating that psilocybin was de novo synthesised in strain P13. However, the production was lower compared to the previous E. coli co‐culture strategy (Flower et al. 2023), which may be attributed to the poor growth of strain P13 in M9 medium. To optimise the culture condition for improving psilocybin production, strain P13 was cultured in different fermentation media. Strain P13 was subjected to 100 mL of LB, LBG and TB medium, respectively, and was cultured in 500 mL baffled fermentation flasks. The results showed that psilocybin reached its maximum titre of 29.9 mg/L at 24 h in LB medium, 79.4 mg/L at 48 h in TB medium and 75.7 mg/L at 72 h in LBG medium, respectively (Figure 5C). The highest psilocybin production in TB medium increased by 232.2% compared to LB medium and was a 100‐fold improvement over the starting strain P03. LBG medium also produced a high titre of psilocybin, but it required a longer culture time than that of TB medium (Figure 5C), which may be attributed to the fact that E. coli consumed glucose for rapid growth when LBG medium was used, and the accumulation of acetic acid inhibited product synthesis until glucose was depleted (Lee 1996). Whereas the carbon source of TB is glycerol, E. coli consumed glycerol at a slow rate, resulting in a decrease in the carbon flux entering the fermentation and less accumulation of acetic acid (Lee and Chang 1990). Therefore, our results suggest that TB medium is suitable for psilocybin production in flask fermentation.