C. paspali MJXA-WT is an industrial ergometrine-producing strain frozen at −80°C in our lab at Qingdao Institute of Bioenergy and Bioprocess Technology, Chinese Academic of Science, China. The strains were maintained on PDA (39 g·L⁻¹ PDA dry powder, BD company) at 25°C for 6 days, and then 3 cm² fresh mycelium from the PDA plates was transferred to a 250-mL flask containing 25 mL of seed medium (20 g·L⁻¹ mannitol, 10 g·L⁻¹ succinic acid, 2 g·L⁻¹ soybean cake powder, 1 g·L⁻¹ KH₂PO₄, 0.3 g·L⁻¹ MgSO₄·7H₂O, and pH 5.0) and cultured for 4 days at 25°C, 220 rpm. The seed liquid was inoculated into 250-mL flask containing 25 mL of fermentation medium (100 g·L⁻¹ sorbitol, 35 g·L⁻¹ succinic acid, 20 g·L⁻¹ corn steep powder, 0.5 g·L⁻¹ yeast extract powder, 0.022 g·L⁻¹ FeSO₄·7H₂O, 0.01 g·L⁻¹ ZnSO₄·7H₂O, 0.7 g·L⁻¹ MgSO₄·7H₂O, pH 5.5). The flasks were kept in the dark for 12 days at 25°C, 220 rpm. An inoculum ratio of 15% (v/v) was used in all bioprocess. All the experiments were repeated three times in this study.

Primers used in this study are listed in Supplementary Table S1. To knock out gene lpsB, primer pairs Up-lpsB-F/Up-lpsB-hph-R and Down-lpsB-hph-F/Down-lpsB-R were used to amplify the flanking 5' and 3' DNA of the lpsB gene from the genome of C. paspali MJXA-WT. The resulting DNA fragments were fused with the hph marker by fusion PCR. Primers N-lpsB-F/N-lpsB-R were used to amplify the gene-targeting cassette.

About 3 cm² of fresh mycelium from the PDA plates was crushed and inoculated into 100 mL of PDB medium and cultured for 3–4 days at 25°C, 220 rpm. The mycelium was collected by filtering with sterile 300 mesh nylon cloth, and washed twice with 0.7 M KCl solution. The mycelium was immersed in the enzymatic solution (0.1% of lywallzyme), and incubated at 30°C for 1 h to prepare protoplasts. The culture solution was filtered with 500 mesh nylon cloth and centrifuged for 12 min at 4,000 rpm to obtain protoplasts. The protoplasts were resuspended with 10 mL of STC solution (0.85 M sorbitol, 10 Tris-HCl pH 8.0, 50 mM CaCl₂), and centrifuged for 10 min at 3,400 rpm. The precipitation of protoplasts was diluted to 10⁷ cells/mL. Then 2–4 μg of DNA fragments, 100 μL protoplast solution and 50 μL of ice-cold PSTC (25% PEG 6000, 1 M D-sorbitol, 10 mM Tris-HCl pH 8.0, and 50 mM CaCl₂) were mixed, and incubated on ice for 25 min. After that, 1 mL of PSTC was added and incubated for 20 min. Afterwards, 20 mL of liquid top agar (PDB with 0.5% agarose and 0.4 M D-sorbitol) was added and the mixture was spread on PDAS (PDA with 0.4 M D-sorbitol). The plates were cultured in the dark for 7 days at 30 C.

LA and ILA were analyzed by high-performance liquid chromatography (HPLC) equipped with a C18 reversed-phase column (Agilent, 4.6 × 150 mm, 5 μm). 1 mL of fermentation supernatant was extracted with 4 mL of mixed solvent containing acetonitrile and water (1:1, v/v) for HPLC and LC-MS measurement. The mobile phases were as follows: water with 0.1% (NH₄)₂CO₃ (solvent A); and 75% acetonitrile (solvent B). Chromatography was carried out over a flow rate of 1.2 mL/min, with a stepped gradient as follows: 13%–43% B from 0–12 min, 43% B for 1 min, 13% B for 1 min. LA and ILA were quantified by comparing peak areas to standard curves under 254 nm. The titers of LA and ILA were quantified by HPLC according to the calibration curves respectively. The calibration curves were presented as Supplementary Figure S1.

To determine the significant factors influencing the LA and ILA titer in submerged fermentation, six factors including inoculum amount (A), succinic acid (B), corn steep powder (C), yeast extract (D), sorbitol (E), and MgSO₄·7H₂O (F) were handpicked for PBD. The final titer of total lysergic acid (TLA), including LA and ILA, was chosen as the response. For each independent factor, there were high (+1) and low (−1) levels, which were determined based on the single factor experiment, and used to assess the influence of each factor. The coded level and actual level of each factor are listed in Table 1. All experiments were repeated three times.

The most monumental factors were chosen to identify their optimal level with the LA and ILA titer as the response. Three significant factors identified by PBD were corn steep powder (A), succinic acid (B) and yeast extract (C). For each variable, there were low (−1), middle (0) and high (+1) levels. The coded level and actual level of three factors is listed in Table 2. All the experiments were repeated three times.

Inefficient genetic manipulation system restricted metabolic engineering of C. paspali. Protoplast-mediated transformation is widely and efficiently used in ascomycetes transformation. Therefore, we choose this method to establish the genetic manipulation system in C. paspali. As an industrial strain, C. paspali MJXA-WT lost the sporulation ability, the mycelium from PDA plates was used as seeds for cultivation of mycelium. Poor mycelium leads to difficulties in protoplast preparation and low efficiency of cell wall regeneration. Homologous recombination usually occurs in logarithmic growth period. The regeneration efficiency was mainly depended on the culture time of mycelia. We tried to apply the mycelia of non-sporulating C. paspali MJXA-WT cultured in original medium for different time to transformation. However, it showed very low regeneration efficiency probably due to the slow growth state. Luckily, we found that the mycelia of C. paspali MJXA-WT grew very fast only in the PDB medium, and the hyphae of C. paspali MJXA-WT cultivated in PDB medium for 3-days was applied to the enzymatic digestion in next step. Enzyme solution of 0.1% lywallzyme was used to digest cell wall, which produced 1.0 × 10⁶ cells/mL at 30°C in an hour or so. STC solution was used to wash and resuspend the protoplasts at 1.0 × 10⁷ cells/mL. In the PEG-CaCl₂ mediated transformation, 2–4 μg of DNA fragments, a cassette containing PgpdAt (constitutive promoter of glyceraldehyde-3-phosphate dehydrogenase from Aspergillus terreus), sgfp (synthetic green fluorescent protein), TtrpC (terminator of tryptophan synthetase from Aspergillus nidulans) and hph (hygromycin B phosphotransferase gene), was added in 100 μL protoplasts solution. The D-sorbitol concentration of regeneration medium was optimized, and the optimal concentration was 0.4 M (Supplementary Table S2). Beginning with 1.0 × 10⁷ protoplasts, approximately 120–150 positive clones were obtained in a reaction (Supplementary Figure S2). Transformants were verified by genomic PCR. The representative mutant strains displayed significant green fluorescence under the fluorescence microscope, while the parental strain did not show any visible fluorescence (Figure 2). Consequently, an efficient genetic manipulation system, which was based on the protoplast-mediated transformation, was established in the industrial strain C. paspali MJXA-WT.

The gene lpsB was involved in the modification of LA in the biosynthetic pathway of ergometrine (Figure 1). We tried to knock out lpsB to prevent the biosynthesis of ergometrine, resulting in the accumulation of LA. The targeting element was constructed using hph as a marker. About 1 kb region in lpsB gene was replaced by the hph (2.2 kb) (Figure 3A). As shown in Figure 3B, the lpsB gene was successfully disrupted in mutant strains (Figure 3B). Knocking out lpsB eliminated the production of ergometrine but accumulated two new products with identical molecular ion peaks ([M + H]⁺ 269.1303) (Figure 3C; Supplementary Figure S3). The new products were identified as LA and ILA by further NMR analyses (Supplementary Figures S4, 5) (Liu and Jia, 2011). The ¹H and ¹³C NMR data of ILA showed quite similar signals to those of LA. The noticeable differences were the occurrence of C-8 (δC 42.7, δH 3.54–3.50) in ILA, rather than C-8 (δC 44.4, δH 4.09–4.04) carbons in LA. Supplementary Table S3 indicated the attribution of ¹H and ¹³C NMR data for LA and ILA. Due to its instability, LA could be partially converted to its isomer ILA (Stoll et al., 1949; Himmelsbach et al., 2014; Yao et al., 2022). The percentages of LA were stabilized at around 45%, while the percentages of ILA were stabilized at around 55%. Base on this, we used the titer of TLA to evaluate the change of production in the next step.

Since various factors in the fermentation medium affect the titer of TLA, an effective experimental approach, such as Plackett-Burman and Box-Behnken design, is essential to achieve the optimal production. First, single-factor experiments were selected to assess the influence of inoculum amount and five key components of the medium (succinic acid, corn steep powder, yeast extract, sorbitol, and MgSO₄·7H₂O) on the titer of TLA. The influence of different inoculum amount on the titer of TLA was shown in Figure 4A. With the increase of inoculum amount, the titer of TLA improved. The highest titre was obtained when the inoculum amount was 4 mL, reaching 0.83 g·L⁻¹. The production curve of the titer of TLA tended to flatten when the inoculum amount was more than 4 mL. The effect of succinic acid with the addition between 25 g·L⁻¹ to 45 g·L⁻¹ was not significant (Figure 4B), the production was highest when the concentration was 35 g·L⁻¹. A high titer of TLA was obtained with 20 g·L⁻¹ corn steep powder added, while the titer of TLA decreased when the corn steep powder concentration was further increased (Figure 4C). When the concentration of yeast extract was 0.25 g·L⁻¹, the titer of TLA reached the maximum, and then decreased significantly with the increase of yeast extract content (Figure 4D). Increasing sorbitol and MgSO₄·7H₂O concentrations enhanced the titer of TLA. A maximum titer of TLA of 1.09 g·L⁻¹ and 1.23 g·L⁻¹ was achieved when sorbitol and MgSO₄·7H₂O were added in the medium with the concentration of 120 g·L⁻¹ and 1.5 g L⁻¹, respectively (Figures 4E, F).

PBD, a two-level experimental design (Abdel-Fattah et al., 2005; Abuhena et al., 2022), was further used to screen out the crucial factors which have a significant impact on the titer of TLA. The experimental data were shown in Table 3. And analysis of experimental results was listed in Table 4 using Design Expert 12.0 software. As the F-value was 27.85 and the p-value of 0.0011 was less than 0.05, the model was significant. Two factors, succinic acid and yeast extract were identified as the significant factors influencing the titer of TLA, with corresponding p-values less than 0.05. The corresponding p-values of corn steep powder was less than 0.001, implying that it was highly significant. Others factors, inoculum amount, sorbitol and MgSO₄·7H₂O, with corresponding p-values more than 0.1, implied that they were not significant. Another important evaluation indicator of the model quality is R². The values of R², predicated R² and adjusted R² were 0.9709, 0.9361, and 0.8327, respectively, indicating that the predicted model was highly fit and effective. The final equation in terms of actual factors was derived as follows: Y = 1.903 + 0.009 A + 0.256 B − 0.363 C + 0.218 D + 0.051 E − 0.006 F

According to the regression coefficient, four factors (inoculum amount, succinic acid, yeast extract and sorbitol) showed positive effects on the titer of TLA. Besides, succinic acid and MgSO₄·7H₂O showed negative effects. Succinic acid, corn steep powder, and yeast extract were chosen as key factors for further optimization. Succinic acid in the tricarboxylic acid cycle was usually used as a carbon source, which may promote the high level of oxidative metabolism in the Claviceps cells, promoting the biosynthesis of secondary metabolites (Hulvová et al., 2013). Fermentation nitrogen (corn steep powder and yeast extract) are particularly rich in protein and amino acid, such as tryptophan and methionine, which may promote the synthesis of ergot alkaloids (Hulvová et al., 2013; Chen et al., 2017).

BBD is a reliable method which is usually used to identify the optimum response region for significant factors (Hegazy et al., 2022). In this study, three significant factors (corn steep powder, (A); succinic acid, (B); yeast extract, (C)) were further explored at three levels. The experimental data were shown in Table 5. And analysis of experimental results was listed in Table 6 using Design Expert 12.0 software. As the F-value was 271.56 and the p-value was less than 0.0001, the model was highly significant. The corresponding p-values of A, B, and C were both less than 0.0001, suggesting that these factors were highly significant. In the same light, the corresponding p-values of A², B², and C² were both less than 0.0001. The values of R², predicated R² and adjusted R² were 0.9971, 0.9935, and 0.9731, respectively, indicating that the model was effective and reliable. To predict the optimal point, the final equation fitted to the experimental response results was represented below: Y = 3.38 + 0.36 A + 0.301 B + 0.414 C + 0.055 A B + 0.135 A C + 0.113 B C − 0.703 A 2 − 0.615 B 2 − 0.525 C 2

According to the equation, the predicted optimum of the cultural medium consisted of 37.2 g·L⁻¹ succinic acid, 23.5 g·L⁻¹ corn steep liquor powder, and 0.4 g·L⁻¹ yeast extract powder, and the predicated titer was 3.6 g·L⁻¹. The value of R² was 0.997, indicating a great degree of correlation between the experimental and the predicted values.

To better evaluate the interaction effect between three key factors, three-dimensional response surfaces and contour plots were created in Figure 5. Based on the surface drawing and contour drawing of response surface model, Figures 5A, B showed the interaction effect of corn steep powder and succinic acid on the titer of TLA with the concentration of yeast extract was 0.4 g·L^-1. Increasing the corn steep powder and succinic acid to the optimal point improved the titer of TLA. However, the increased amount of both factors leaded to the decline of the titer. Figures 5C, D showed the interaction effect of corn steep powder and yeast extract on the titer of TLA with the concentration of succinic acid was 37.2 g·L⁻¹. Increasing the addition of corn steep powder and yeast extract to the optimal point improved the titer of TLA. However, improving concentration of both factors leaded to the decline of the titer. Figures 5E, F showed the interaction effect of succinic acid and yeast extract on the titer of TLA with the concentration of corn steep powder was 23.5 g·L⁻¹. Increasing the concentration of succinic acid and yeast extract to the optimal point improved the titer of TLA. However, improving concentration of both factors leaded to the decline of the titer.

To verify the developed model, the time profile of TLA production under the optimal fermentation condition was analyzed. After 16 days culture, the final titer of TLA reached 3.7 g·L⁻¹ using the optimal fermentation medium (Figure 6), which was close to the predicted value 3.6 g·L⁻¹. Hence, the validity of this model was verified. Whereas, the titer of TLA reached 0.8 g·L⁻¹ using the original fermentation medium. It indicates that an efficient microbial fermentation technology for LA production has been developed using an engineered C. paspali strain.