In this study, wild-type strains (WT) were isolated from oxytropis glabra in Gansu Province and analyzed by PCR detection (primers were ITS1: 5′TCCGTAGGTGAACCTGCGG 3′/ITS4: 5′TCCTCCGCTTATTGATATGC 3′, respectively), and through NCBI Blast, it was found that the sequence of this strain was 100% similar to Alternaria sp. ZC-2014 (a type of Alternaria oxytropis) (Woudenberg et al., 2013). This fungus needs to be routinely cultivated on PDA (potato dextrose agar) media at a temperature of 20°C.

Seeds previously collected from Oxytropis glabra plants (Jiuquan, Gansu, coordinates) were used for all experiments described below. The method of seed coat removal was employed to qualitatively reduce the amount of A. oxytropis inoculum. Additionally, seed embryos without seed coats were inoculated into fungal cultures. Plants derived from these seeds or embryos were cultivated in an artificial climate chamber under a 16-h light cycle, with day/night temperatures maintained at 20°C/18°C.

Alternaria oxytropis was inoculated onto PDA plates and cultured at 20°C. Mycelia was collected daily from the 6th to the 10th day, three repetitions were set for each group. After drying, the mycelia mass was measured, and it was promptly ground into a fine powder using a grinder, and then the mycelia is extracted with methanol (the method was consistent with Song et al., 2019). The mycelia was analyzed using Triple Quad 5500+QTRAP Ready and using the methods of (Noor et al., 2020). SW concentration was tested three times for each strain.

Alternaria oxytropis was inoculated onto PDA plates and cultured at 20°C. The culture media was collected daily from the 6th to the 10th day. The solid media was chopped into small pieces and soaked in 400 mL of methanol overnight. After soaking, ultrasonic extraction was carried out for 60 minutes, repeated three times. Combined with solvent evaporation and concentration, the crude extract was obtained and dissolved with methanol. Three parallel replicates were set up for each strain. The specific detection method is the same as that for the “Section 2.3” in the previous section.

The pSilent-swnR vector was constructed in two steps using the pSilent-1 plasmid. Based on the sequence of the swnR gene from A. oxytropis (KY365742.1), primers were designed to amplify two fragments, a 332 bp forward and reverse fragment of the swnR gene, each including homologous arms (primer sequences are listed in Supplementary Table S3). The pSilent-1 plasmid was first digested with Kpn I/Bgl II and then subjected to double digestion with Xho I/Hind III. The swnR gene fragments 1 and 2 were successfully amplified. These two swnR gene fragments were then ligated into the enzyme-digested pSilent-1 plasmid fragment. The ligation product was transformed into competent cells, and positive transformants were selected. Sequencing was performed on the transformants, and those with sequences completely matching the target swnR fragment were identified as containing the pSilent-swnR silencing vector. These positive transformants were used for further experimentation.

A. oxytropis was inoculated onto PDA plates and cultured at 20°C for 6–10 days. The mycelia was scraped off, washed with an osmotic stabilizer (1.2 M KCl), and centrifuged. The centrifuged A. oxytropis mycelia was weighed, and different enzymatic digestion solutions were prepared with varying enzyme combinations at a solid-to-liquid ratio of 1:10. The enzyme combinations tested included: 1% cellulase + 3% lysing enzyme, 1% snailase + 3% lysing enzyme, 1% cellulase + 1% snailase, and 1% snailase + 1% cellulase + 3% lysing enzyme. The 50 mL centrifuge tubes containing the enzymatic digestion solutions were wrapped in aluminum foil to protect them from light and placed on a shaker at different temperatures (20°C, 25°C, 30°C, and 35°C) with a shaking speed of 100 rpm. Digestion was carried out for various time intervals (2h, 3h, 4h, 5h, 6h). After digestion, 20 mL of 1.2 M KCl was added to each sample for further processing.

The enzymatic digestion mixture of A. oxytropis was filtered through fungal-specific filter cloth to separate the protoplasts from the remaining mycelia clumps. The protoplasts were then washed repeatedly with 1.2 M KCl to ensure complete collection. The resulting A. oxytropis protoplast suspension was centrifuged, and the supernatant was carefully removed. The protoplast pellet was resuspended in an appropriate volume of STC buffer and centrifuged again at 5,000 rpm for 15 min. After this step, the purified protoplasts were resuspended in STC buffer and counted using a hemocytometer. This process was repeated six times to accurately determine the protoplast concentration.

Approximately 400 μL of prepared protoplasts (2–5 × 106 protoplasts/mL) was aliquoted into two sterile 1.5 mL centrifuge tubes. To each tube, 5 μg of silent vector was added respectively, mixed gently, and left to stand for 20 min (gently inverting the tubes). Then, 1 mL of 40% PTC Buffer was added to each tube, and the tubes were incubated at room temperature for 20 min. After that, 5 mL of TB3 Buffer (containing 50 μg/mL ampicillin) was added, and the protoplasts were cultured overnight at 30°C, shaking at 100 rpm.

The overnight-cultured Alternaria oxytripis protoplasts were centrifuged (5,000 rpm), and the supernatant was carefully removed. The protoplasts were then resuspended in an appropriate amount of regeneration lower layer media, followed by the addition of 1.5 μg/mL hygromycin and ampicillin. The mixture was quickly spread onto plates and cultured at 20°C for 24 h. Afterward, regeneration upper layer media containing 2.0 μg/mL hygromycin was poured onto the plates, and the culture continued at 20°C. Single colony transformants were observed on the plates. These transformants were picked and transferred onto hygromycin-resistant media for further selection. Wild-type A. oxytropis was used as a control. The transformants growing on the selective media were expanded for further experiments.

WT and the selected transformants were inoculated onto PDA media and cultured for 8 days. Fresh mycelia from each strain grown on the PDA media was collected, frozen in liquid nitrogen, ground into powder, and dried. The SW content and swnR gene expression level of the transformants were measured using the previously described methods, and silent strains (pSilent-swnR) with significantly reduced SW content and swnR gene expression were screened, three repetitions were set for each group. WT and the selected silent strains were inoculated on PDA media and continuously cultured for five generations. The SW content in the mycelia of the fifth generation was measured to determine whether there was a significant difference in SW synthesis between the silent strain and WT, and to assess whether the difference remained stable.

On the 8th day of culturing the fifth-generation wild-type and silent strains of A. oxytropis, mycelia and media from both WT and pSilent-swnR strains were collected. Using the previously described methods, LC-MS was used to detect the content of swainsonine (SW), pipecolic acid (Pa), and L-lysine (LY). The differences in the levels of these three products in the mycelia between the WT and pSilent-swnR strains were compared to determine the regulatory role of the swnR gene in the swainsonine biosynthesis pathway in A. oxytropis.

At the same time, total RNA from both WT and pSilent-swnR strains was extracted and reverse transcribed. Using cDNA as the template, the expression level of the swnR gene was detected via qRT-PCR to determine the effect of this gene's expression on swainsonine production.

The wild-type and swnR gene silent strains of A. oxytropis were inoculated on PDA media and incubated upside down for 8 days in a fungal incubator at 20°C. The changes in colony morphology and the growth status of the strains were observed at various time points. And then mycelial slides were prepared; the mycelia were observed and recorded using an optical microscope. Finally, transmission electron microscopy (TEM) was used to observe and compare the ultrastructural differences of mycelia between WT and pSilent-swnR strains.

WT and the silent strains (pSilent-swnR) of A. oxytropis were inoculated on PDA media and cultured upside down in a fungal incubator at 20°C for 8 days. Mycelia from WT and two pSilent-swnR strains with the highest and lowest swainsonine production were collected. The mycelia of these three strains were divided into three groups, each with six samples. These samples were sent to Maiwei Company (Yangling, Shaanxi) for metabolomics analysis, with three replicates per group. Pairwise comparisons between the three groups were performed. Additionally, in-depth data mining was performed using the Maiwei Cloud platform (http://www.international.biocloud.net), enabling a series of personalized metabolomics operations, such as annotating differential metabolites using the KEGG database (Kanehisa and Goto, 2000) and enriching differential metabolic pathways.

RNA sequencing was performed using the mycelia of WT and two pSilent-swnR strains. RNA extraction, quality verification, library preparation, and sequencing were performed by BMKCloud (Beijing, China). After sequencing, data analysis was conducted using the bioinformatics analysis pipeline provided by BMKCloud (http://www.biocloud.net). The Clean Data was assembled to obtain the Unigene library for this species. Based on this, the following analyses were performed: library quality assessment, structural level analysis, differential expression analysis, gene function annotation, and functional enrichment analysis. Furthermore, BMKCloud (http://www.biocloud.net) also provided options for in-depth data mining and various personalized transcriptomics operations, such as gene search and plotting, unique and shared gene analysis, protein interaction network construction, gene set enrichment analysis (GSEA), and co-expression analysis of differentially expressed genes.

The technique for inoculating Oxytropis glabra plants with fungi (specific steps, based on Grum et al., 2012), and the cultivation conditions.

The stems and leaf tissues of the 14-day-old wild-type, OE, and co-cultivated plants are taken from the culture bottles. These tissues are prepared for scanning electron microscopy (SEM) to observe the colonization of fungal mycelia in the plants for each group.

After 14 days of growth, wild-type, OE, and co-cultivated plants are taken out of the culture bottles. The growth morphology of the plants is observed, and the total height, number of leaves, number of branches, and fresh weight of the seedlings are recorded.

After Oxytropis glabra seeds were cultured on PDA media for 5 days, they were transferred to culture bottles containing MS media and cultured for 14 days. The seedlings from each group were collected.

Swainsonine extraction method: weigh 25 mg of dried plant sample into a 2 mL centrifuge tube using an analytical balance. Grind the sample into a fine powder using a grinder. Add 1.5 mL of anhydrous ethanol and shake for 18 h to extract. After centrifugation, collect the supernatant. The sediment can be re-extracted using the same method to increase extraction efficiency. Combine the supernatants and filter through a microporous membrane into a 1 mL auto-sampler vial. The swainsonine content in the extract is detected using LC-MS, under the same conditions and calculation methods as used for swainsonine detection in fungal samples.

The swainsonine content in the extract is divided by the dry weight of the seedlings to obtain the swainsonine content per milligram of plant seedlings.

The analyses not explicitly stated in this study were all conducted using IBM SPSS Statistics software (version 26.0). A two-factor correlation analysis (Spearman) was employed to assess the relationship between the production of swainsonine (SW) and the expressions of genes swnH1, swnH2, swnN, swnK, and swnR. Subsequently, the correlations of the genes were analyzed using Origin software. One-way analysis of variance (ANOVA) was performed on the expression levels of genes within the SWN gene cluster, the contents of SW, Pa, and LY detected by LC-MS, as well as various statistical data such as plant fungal biomass, fresh weight, height, leaf number, and branch number. The significance threshold was as follows: ^****P < 0.0001 indicates an extremely significant difference between groups (except for the three silent transformant mycelia where the content of SW was not reached P < 0.0001 compared with the wild type, so the P values used were P < 0.05 and P < 0.01, while for the remaining experiments, P < 0.0001 was used).

In order to screen and identify key genes involved in swainsonine biosynthesis in A. oxytropis, we measured the swainsonine content in the fungal mycelia from days 6 to 10, along with the expression levels of five genes from the SWN gene cluster (Figures 2B–F). We found that the expression levels of swnH2, swnK, and swnR peaked on day 8 during the observation period, which closely mirrored the trend in swainsonine production (Figures 2A–C, F). Correlation analysis further demonstrated a strong positive correlation (R = 1.000) between swainsonine production and swnR gene expression in A. oxytropis (Figure 2G). These results indicate that the swnR gene is a key factor involved in the swainsonine biosynthesis pathway in A. oxytropis.

To perform gene editing on A. oxytropis, it was crucial to obtain a sufficient quantity of protoplasts for subsequent experimental procedures. Thus, we undertook a series of explorations and optimizations for protoplast preparation conditions. By treating 8-day-old mycelia with different enzyme mixtures and durations, we found that a mixture of three enzymes (3% CX+1% LYX+1% CX) produced the highest number of protoplasts, approximately 8.3 × 105 protoplasts/mL, after 5 h of treatment (Figures 2H, I, K). Beyond this point, protoplast lysis began to occur which reduced the number of usable protoplasts (Figures 2I, K). We also optimized the enzymatic digestion temperature and fungal age for protoplast preparation. Results showed that the enzymatic solution was most effective at 30°C, producing the highest yield of protoplasts. As the temperature increased, protoplast yield significantly decreased (Figure 2J). We found 8-day-old mycelia consistently yielded sufficient protoplasts therefore 8d was used for subsequent experiments (Figure 2H).

Before conducting transformation experiments on A. oxytropis, it was essential to construct a silencing vector targeting the swnR gene. As part of this process, antibiotic sensitivity testing was performed to determine the appropriate type and concentration of antibiotics for selecting transformants and preparing selective media for protoplast transformation. Previous laboratory research suggested that a specific concentration of hygromycin could inhibit the growth of A. oxytropis. To further verify the fungal sensitivity to hygromycin, A. oxytropis was inoculated onto media containing varying concentrations of hygromycin. Results indicated that a concentration of 2.5 μg/mL completely inhibited the growth of A. oxytropis (Supplementary Figure S1A).

Given the sensitivity of A. oxytropis to hygromycin, the pSilent-1 vector containing the hygromycin resistance gene was selected as the empty vector for gene silencing. The pSilent-swnR silencing construct had been successfully constructed as confirmed by complete plasmid construct sequencing (Supplementary Figure S1B). The pSilent-swnR vector was then introduced into A. oxytropis protoplasts via PEG-mediated transformation. After 10–14 days, colonies appeared on PDA plates containing hygromycin. Four hygromycin-resistant transformants were randomly selected, and of these, three were confirmed to harbor the pSilent-swnR vector (Supplementary Figure S1C). These confirmed colonies were subsequently transferred to PDA plates for further analysis.

The expression levels of the swnR gene in the wild-type (WT) strain and the WT strain transformed with the empty vector pSilent-1 were comparable as determined by qPCR quantification. The mRNA levels of swnR were significantly reduced in three of the transformants (3-1, 3-2, 7-1, 91.30%, 71.61%, 34.32%, respectively), with strain 3-1 exhibiting a 91.51% reduction in swnR gene expression relative to the wild-type (P < 0.0001) (Figure 3A). Microscopic and ultrastructural observations of colony mycelia from the wild-type A. oxytropis (WT) and the silenced strains 3-1, 3-2, and 7-1 revealed no notable differences (Figures 3E–G).

Silencing the swnR gene led to a significant reduction in swainsonine production in A. oxytropis. Since the culture media used for growing A. oxytropis is plant-based and already contains the precursor lysine and the intermediate product L-pipecolic acid, we were only able to assess the levels of swainsonine and related compounds in the mycelia. As illustrated in Figure 3B, swainsonine levels in the mycelia of all three transformants were significantly reduced (P < 0.05), with strains 3-1 and 3-2 showing highly significant reductions in swainsonine content (P < 0.01). Figures 3C, D further indicate that lysine and pipecolic acid levels in the mycelia of all three transformants were also significantly decreased (P < 0.0001). These findings suggest that the reduction in swnR gene expression directly limits the availability of precursors for the swainsonine biosynthesis pathway, thereby fundamentally decreasing swainsonine production.

Further investigation was conducted to assess the effects of varying degrees of swnR gene silencing on fungal growth, development, and the synthesis and secretion of secondary metabolites. Principal component analysis (PCA) score plots revealed distinct separations among the silenced strains 3-1, 7-1, and WT (Figure 4A), with the separation trends correlating with changes in swnR gene expression levels. Compared with WT, strain 3-1 exhibited 164 upregulated genes and 293 downregulated genes, while strain 7-1 exhibited 25 upregulated and 175 downregulated genes. Cluster analysis further highlighted significant differences between the wild-type and each of the silenced strains (Figures 4B, C). Gene Ontology (GO) enrichment analysis of the upregulated genes in strain 3-1 revealed that these genes were predominantly involved in multiple metabolic pathways, including the carbohydrate metabolic process, oxidoreductase activity, and transmembrane transporter activity, indicating roles in secondary metabolism networks, cellular structural adaptations, and metabolic homeostasis mechanisms (Figure 4F). KEGG pathway enrichment analysis demonstrated that downregulation of swnR significantly activated several primary metabolic pathways in Alternaria oxytropis, including phenylalanine metabolism, carbohydrate metabolism, and porphyrin and chlorophyll metabolism. Notably, there was a marked upregulation in pathways associated with the metabolism of aromatic compounds and other secondary metabolites (Figure 4G). Conversely, downregulated genes in strain 3-1 were mainly enriched in pathways related to oxidative stress responses, substance transport, membrane dynamics, and transmembrane transporter activity, as revealed by GO analysis (Figure 4D). KEGG analysis further indicated significant suppression of several key metabolic pathways, including glyoxylate and dicarboxylate metabolism, peroxisome activity, and lipid metabolism and degradation (Figure 4E). In contrast, upregulated genes in strain 7-1 were enriched in only three metabolic categories: phenylalanine metabolism, transmembrane transporter activity, and membrane components. The downregulated genes in strain 7-1 showed patterns similar to those observed in strain 3-1, involving oxidative stress responses, carbon metabolism, glyoxylate and dicarboxylate metabolism, and membrane dynamics, with only a few biological pathways—characterized by low enrichment factors and limited statistical significance—differing between the two groups (Supplementary Figures S2A–F).

Transcriptomic analysis of swainsonine biosynthesis-associated genes revealed significant differential expression. Comparative assessment of expression levels within the SWN gene cluster and the P5CR gene between the wild-type strain and the silenced strain demonstrated significantly reduced expression of swnR (TRINITY_DN2817_c1_g1), swnN (TRINITY_DN2817_c0_g1), swnK (TRINITY_DN7706_c0_g1), and P5CR (TRINITY_DN1901_c0_g1). Furthermore, both coding sequence (CDS) and amino acid sequence analyses indicated that swnH1 and swnH2 correspond to the same genetic locus (TRINITY_DN8537_c0_g1). This gene (swnH1/H2) similarly exhibited significantly reduced expression (Supplementary Table S4).

After identifying significant transcriptomic differences among the silenced strains 3-1, 7-1, and WT, we further explored the metabolic differences between the wild-type and the two silenced strains by integrating transcriptomic and metabolomic analyses. Non-targeted metabolomics was employed to detect metabolites in each group, aiming to investigate the metabolic response mechanisms triggered by changes in swnR gene expression levels. Across the three groups, a total of 6,841 metabolites were identified, with the most abundant classes including amino acids and their derivatives, benzene and substituted derivatives, heterocyclic compounds, and organic acids and their derivatives (Figures 5D, E). PCA score plots revealed that while there was some overlap between the metabolites of silenced strains 3-1, 7-1, and WT, clear separations were also evident (Figure 5A). PLS-DA analysis further confirmed distinct separations between the silenced strains and the wild-type, indicating that changes in swnR expression significantly impact the metabolic profiles of A. oxytropis (Figures 5B, C). Specifically, 830 differential metabolites were identified between the wild-type and strain 3-1, with 396 metabolites significantly upregulated and 434 downregulated (Figure 5D). These metabolites were mainly enriched in pathways related to carbon metabolism, lipid metabolism, antioxidative responses, autophagy and organelle renewal, and secondary metabolite biosynthesis, showing high consistency with transcriptomic results (Figure 5G, Supplementary Figure S3). Similarly, 893 differential metabolites were found between the wild-type and strain 7-1, with 400 significantly upregulated and 493 downregulated (Figure 5E). As in strain 3-1, these differential metabolites were enriched in pathways related to secondary metabolite biosynthesis, lipid metabolism, and amino acid metabolism. However, a notable distinction in the strain 7-1 group was the enrichment of differential metabolites in the pathway “microbial metabolism in diverse environments,” suggesting that microbial metabolic activities were strongly regulated (Figure 5H, Supplementary Figure S3B). These findings underscore the critical role of swnR in regulating secondary metabolism pathways, suggesting that modulation of swnR expression profoundly impacts the production of bioactive metabolites and the adaptive capacity of A. oxytropis under environmental stress.

Metabolomic analysis identified alterations in metabolites associated with swainsonine biosynthesis. Specifically, decreased abundance of L-pipecolic acid and L-Saccharopine was observed. These metabolites were co-enriched within the same metabolic pathway. In contrast, levels of L-lysine, L-glutamic acid, and L-proline showed no significant differences (Supplementary Table S5). Other putative biosynthetic intermediates were not detected in our metabolomics dataset.

To test whether changes in swnR gene expression affect material transport across the cell membrane in A. oxytropis, we used LC-MS to measure swainsonine levels in both the mycelia and culture media of WT and the 3-1 and 7-1 silenced strains on the 8th day. In WT, the ratio of swainsonine content in the mycelia to the culture media was 1:3, while in the 3-1 and 7-1 strains, the ratios were ~ 1:1 and 1:1.5, respectively (P < 0.0001) (Figure 5F).

To investigate whether the silenced A. oxytropis strains affect the growth of Oxytropis glabra seedlings, we co-cultivated the silenced strains with peeled seed embryos. For unpeeled O. glabra seeds (WTP), fungal mycelia appeared around the seedlings 5 days after germination on PDA media and continued to grow around the seedlings after they were transferred to MS media (Figures 6A, B). In contrast, no fungal mycelia was observed around the peeled seed embryos under the same conditions. For seedlings co-cultivated with the silenced strains 3-1 and 7-1, no fungal mycelia initially appeared on the PDA media, but 14 days after transferring to MS media, fungal growth became visible around the roots (Figure 6C). Further analysis using scanning electron microscopy revealed fungal mycelia from both the silenced and wild-type strains within longitudinal sections of petioles from O. glabra seedlings. The mycelia extended into the intercellular spaces between the cell walls, exhibiting branching. However, in the longitudinal sections of petioles from plants grown from peeled seed embryos (NC), no fungal mycelia was detected (Figure 6D).

We observed, measured, and analyzed the fresh weight, plant height, leaf number, and branching number of Oxytropis glabra seedlings co-cultivated with WTP as well as the silenced strains 3-1 and 7-1. The statistical analysis revealed no significant differences in these growth parameters between seedlings co-cultivated with the silenced strains and those co-cultivated with WTP (Figures 6E–H). Furthermore, comparisons of fresh weight and plant height between seedlings from peeled and unpeeled seeds showed that the presence or absence of fungal symbiosis had no significant effect on these growth parameters. Therefore, it can be inferred that the silenced strains 3-1 and 7-1 do not have a significant negative impact on the growth and development of O. glabra seedlings. There were subtle observable differences between WT and silenced strains, however, these differences were not statistically significant.

After measuring the swainsonine content in the fungal mycelia and culture media, we observed that swainsonine levels in the silenced strains 3-1 and 7-1 decreased to varying degrees compared to WT. To determine if co-cultivation with these silenced strains would result in similar reductions in Oxytropis glabra, we set up negative controls (NC) and positive controls (WTP) and used LC-MS to measure the swainsonine content in the seedlings. The results showed that seedlings from peeled seeds contained the lowest levels of swainsonine, while unpeeled wild-type seedlings had the highest levels. Seedlings co-cultivated with the silenced strains 3-1 and 7-1 exhibited a significant reduction in swainsonine content compared to the wild-type (P < 0.0001) (Figure 6I). Additionally, the swainsonine content in seedlings co-cultivated with strain 7-1 was slightly higher than that in seedlings co-cultivated with strain 3-1, consistent with the observed changes in swainsonine levels in the fungal mycelia and culture media.

The detection of swainsonine content in the silenced symbionts was consistent with our initial experimental expectations. To more precisely compare the impact of the silenced strains on swainsonine levels in the symbiotic plants, we also measured the fungal biomass within the symbionts. The results showed that the biomass of the 7-1 group was significantly higher than that of the 3-1 group, nearly double (Figure 6J). This suggests that the silenced strain 7-1 has a higher infection efficiency and a greater ability to establish symbiosis with the plant. However, based on the ratio of swainsonine production by the strains, the swainsonine content in plants co-cultivated with strain 7-1 was about three times that of strain 3-1 group. The actual results, however, did not align with this prediction, the reason for this discrepancy remains to be further investigated. This divergence could be attributed to unknown regulatory mechanisms that influence swainsonine production during the symbiotic interaction between the fungus and the plant.