Introduction

Taxus species are widely distributed in the world [1, 2]. In China, most of the Taxus species have been included in the national protection scope for rare and endangered plants [3, 4]. Taxus plants adapt to a variety of changing environments, from cold zones to temperate zones across a latitudinal gradient in China [5]. The conservation of endangered Taxus trees and their biodiversity, including population size, geographical distribution, and habitat, is an urgent priority [6].

As medicinal plants, Taxus contain natural anticancer components, such as taxol (generic name paclitaxel) and its precursors, which have high medicinal value [7, 8]. Modern medical research confirms that taxol can be used to treat a variety of cancers, including ovarian, lung, cervical, and esophageal cancers [9]. Due to its low content in Taxus tissue, the significant medicinal value of paclitaxel keeps its price high [10].

Endophytic fungi, which are widely present in plant tissues, play a crucial role in plant growth, development, and secondary metabolism, typically without causing any damage to the plant itself [11, 12]. The relationship can vary from latent pathogens or saprotrophs to mutualism, and the fungi can be facultative endophytes. In most cases, endophytic fungi contribute to plant survival by promoting plant growth, inhibiting competitors, and enhancing stress resistance [13, 14]. In addition, some endophytic fungi produce compounds similar to those produced by their host plants. For instance, Taxol was identified and isolated from endophytic fungi found in T. brevifolium [15]. Over time, numerous taxol-producing fungi have been isolated from various Taxus species, such as Aspergillus fumigatus from T. wallichiana, TMS-26 from T. media, Cladosporium from Iranian T. baccata, and Paraconiothyrium sp. from English T. baccata [16,17,18]. However, recent studies have challenged the idea of taxol production by endophyte species, as the biosynthetic gene cluster has never been discovered in endophytes [19]. Throughout our research, we have not yet identified a fungus capable of producing taxol. Therefore, our study focuses on screening fungi that can influence the synthesis of taxol in Taxus, rather than those that can directly produce taxol. A comprehensive analysis of endophytic fungi in Taxus plants is essential for identifying novel taxol biosynthesis-related fungi.

The culture-dependent method is widely used to determine the diversity of endophytic fungi [20]. However, due to limitations in microbial culture technology, isolating and culturing all endophytic fungi in Taxus trees is challenging. High-throughput sequencing plays a crucial role in exploring the diversity of endophytic fungi, providing comprehensive information about these fungi in different Taxus species [21]. To date, the diversity of endophytic fungi has been investigated in various Taxus species. Molecular phylogenetic analysis identified 245 endophytes in T. globose [22]. Morphological and molecular identification isolated 81 endophytic fungi from T. media [23]. In the Himalayan yew (Taxus wallichiana), 13 distinct endophytic fungi were identified using the Internal Transcribed Spacer (ITS)-PCR method [24]. A total of 115 endophytic fungi isolates were isolated from the bark of T. chinensis by ITS1-5.8 S-ITS2 sequencing method [25]. In the Iranian yew (T. baccata), 125 fungal strains from 21 genera and 26 species were isolated and identified [26]. However, the variations in endophytic fungal communities among different Taxus species remain largely unknown.

A recent report revealed that fungal elicitors could significantly enhance taxol biosynthesis in Corylus avellana cell culture [27]. To date, several endophytic fungi from Taxus trees have been reported to be involved in Taxol biosynthesis. For example, Aspergillus niger was isolated from the inner bark of T. chinensis, and Pseudodidymocyrtis lobariellae was isolated from the needles of T. chinensis as fungal elicitors [28, 29]. Taxomyces andreanae, a fungal endophyte isolated from the Yew tree Taxus brevifolia, is another alleged taxol-producing fungus [30]. Screening for additional Taxol biosynthesis-related endophytes is an environmentally friendly approach to enhance the mass production of Taxol.

In the current investigation, three Taxus species were studied, including T. media, a natural hybrid known for its high taxol content, T. mairei, found in the Yangtze River Basin of China, and T. cuspidata, primarily located in Changbai Mountain. The study aimed analyze the differences in endophytic fungi communities [31, 32]. Through Pearson correlation analysis, several taxol biosynthesis-related fungal genera were identified. Furthermore, effect of inoculation with Alternaria isolates on the gene expression and metabolite accumulation of T. mairei was determined by transcriptomic and untargeted metabolomic studies. These findings offer crucial foundational information for the holistic utilization of Taxus and aid in the identification of taxol production-related endophytic fungi.

Materials and methods

Plant samples and DNA extractions

Barks were collected from five-year-old cultivated Taxus plants. T. media trees were obtained from Quzhou City, China, T. mairei plants were obtained from Yaan City, China, and T. cuspidata from Jilin City, China. All plants were kept in a growth chamber at Hangzhou Normal University, Hangzhou, China, at 28 ± 2 °C with a light/dark cycle of 14/8 h and a 65–70% relative humidity for one month. Total DNAs from the twigs of different Taxus trees were isolated using a Fast-DNA SPIN extraction kit (MP Biomedicals, Santa Ana, CA, USA) according to the kit’s instruction. Doble Distilled H2O (ddH2O) was used as a negative control. Total DNAs were eluted in 50 µL of ddH2O and stored at -80 °C until used. For each Taxus tree, fifteen biological replicates were performed for ITS sequencing.

PCR amplification and ITS sequencing

The ITS region of the eukaryotic small-subunit rRNA gene was amplified with primers ITS7 (5’-GTGARTCATCGAATCTTTG-3’) and ITS4 (5’-TCCTCCGCTTATTGATATGC-3’) [33]. Specific barcodes were connected to the 5’ end of ITS4/7 primers. PCR amplification was carried out in a reaction mixture solution containing 25 ng of template DNAs, 12.5 µL of PCR Premix, 2.5 µL of each primer, and ddH2O to adjust to a 25 µL final volume. The PCR condition to amplify the ITS fragments consisted of an initial denaturation at 98℃ for 30 s; 35 cycles of denaturation at 98 ℃ for 10 s, annealing at 55 ℃ for 30 s, and extension at 72 ℃ for 45 s; and final extension at 72℃ for 10 min. The PCR products were checked by 2% agarose gel electrophoresis. The purified PCR products were quantified using the Qubit kit (Invitrogen, USA). The DNA library for sequencing was prepared on an Agilent 2100 Bioanalyzer (Agilent, USA) with a Library Quantification Kit for Illumina (Kapa Biosciences, Woburn, MA, USA). The resulting libraries were sequenced using the 250PE MiSeq platform provided by LC-Bio (Hangzhou, China).

OTU and diversity analysis

Raw reads were isolated based on their unique barcodes. After removing the barcode and primer sequence, clean reads were quality filtered using FastQC software (v.0.10.1). Chimeric sequences with ≥ 97% similarity were produced and assigned to specific operational taxonomic units (OTUs) by Verseach (v.2.3.4). The representative sequence for each OTU was then assigned to each representative sequence using the Ribosomal Database Project classifier program.

OTUs abundance levels were normalized, and multiple sequence alignment analysis was performed using PyNAST software [34]. Alpha diversity parameters, such as Chao1, Shannon, Simpson, and Observed species, were utilized to analyze the complexity of species diversity. These were calculated using QIIME (v. 1.8.0). Beta diversity analysis was conducted to assess differences in species complexity among samples. This analysis was performed through principal co-ordinates analysis (PCoA) and QIIME software (Version 1.8.0), respectively.

Phylogenetic study

An unrooted phylogenetic tree was constructed using MEGA 6.1 (http://www.megasoftware.net/) employing the neighbor-joining method. The values marked on the branch of the phylogenetic tree mainly represent the support or reliability of the branch. To quantify the effectiveness of the branch, the values are calculated with bootstrap method.

Pearson correlation coefficient

The correlation between taxoids and OTUs was made with the Pearson correlation coefficient using the ‘corrplot’ package available in OmicStudio (https://www.omicstudio.cn/home). Pearson’s correlation coefficient cut-off of 0.6 and P value < 0.05 were stablished to consider significant and strong correlations.

Isolation and characterization of endophytic fungal isolates

Fresh leaves of T. mairei were surface washed and sterilized with 75% ethanol and 1.5% sodium hypochlorite washing solution twice. The clean samples were cut into pieces and placed on solid Potato Dextrose Agar (PDA) medium supplemented with 30 µg/mL streptomycin sulfate. The emerging fungal colonies were harvested and isolated by plate streaking. The genomic DNA of each fungal isolate was extracted using the FirePureTM Microbial DNA Extraction Kit (FG0308, FIREGEN). The ITS region was amplified using ITS4/7 primers. The PCR products were recovered, sequenced, and searched against the NCBI GenBank database to identify fungal species and genera (Table S1).

Inoculation of endophytic fungi with T. mairei leaves

The clean leaves of T. mairei were collected and sprayed with 1 mL of spore suspension containing the two endophytic fungal isolates (L7 and M14). The concentration of the spore suspension was adjusted to 1 × 106 spores/mL. ddH2O was used as a control. After 48 h incubation, the treated leaves were cleaned and harvested for transcriptomic and metabolic analyses.

RNA isolation and RNA-sequencing

Total RNAs of T. mairei leaves were extracted using an RNeasy Mini Kit (Qiagen, Hilden, Germany). The purified and fragmented mRNAs were used to construct cDNA libraries using a VAHTS Universal V6 RNA-seq Library Prep Kit according to the manufacturer’s instructions. RNA sequencing was conducted on an Illumina 6000 platform (OEbiotech, Shanghai, China) to produce raw sequences according to the manufacturer’s protocol.

Analysis of the transcriptomic data

Raw fastq format reads were processed to produce clean reads using fastp program. The clean reads for each sample group were mapped onto the T. mairei genome using HISAT2 software [8]. Fragment Per Kilobase of transcript per Million mapped reads (FPKM) method was applied to calculate the read counts of each gene by HTSeq-count. Differentially expressed genes were screened using the DESeq2 software based on a threshold of P value < 0.05 and fold change > 2 or fold change < 0.5. Most of the differentially expressed genes (DEGs) were classified into different Gene Ontology (GO) terms and/or Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways by the Blast2GO program (https://www.blast2go.com/). For the enrichment analysis, a GO or KEGG group with a corrected P value < 0.05 was considered significant.

Untargeted metabolomic analysis

The separation was carried out on a 100 × 2.1 mm, 1.7-µmparticle size Waters ACQUITY UPLC BEH C18 column using an UPLC system (Waters, Herts, UK). Ahigh resolution MS/MS Waters Xevo G2-XS Q-TOF (Waters, Herts, UK) was used to detect metabolites eluted from the column. Raw data of LC-MS/MS with a matrix of molecular features, including retention time (RT), and mass/change ratio (m/z), was processed using the XCMS program. The identification of compounds is based on multiple dimensions such as RT, precise mass number, secondary fragments, and isotopic distribution. KEGG, Lipidmaps (v2.3), METLIN database, and LuMet Plant3.0 local database were used for identification and analysis. The intensity of peak data was further processed by an in-house software MetaX. For quality control, the identifications of precursor ions of the expected positive ion adduct with less than a 5 ppm error were defined using high-resolution MS.

Results

Sample sequence characteristics

In total, 585,049 raw tags were obtained in T. cuspidata, 554,417 in T. mairei, and 559,155 in T. media. The quality of sequencing data was evaluated by various parameters, including sequence numbers, sequence lengths, effective ratios, Q20/Q30 quality values, and GC content in each sample (Table S2). The number of valid reads per sample ranged from 7,822 to 62,334, with the average read length for each sample ranging mainly from 200 to 400 bp (Fig. 1a). A total of 271 OTUs were detected in all libraries, including 230 OTUs in T. media, 115 in T. mairei, and 177 in T. cuspidata. In detail, 52 OTUs were common among the three Taxus species, while 1 OTUs, 27 OTUs and 44 OTUs were unique in T. cuspidata, T. mairei, and T. media, respectively (Fig. 1b and Table S3). The abundance of OTUs in the three Taxus species was shown by a heatmap, suggesting significant differences in fungal communities among the three Taxus species (Fig. 1c).

Fig. 1
figure 1

Overview of the sequencing data. (a) The length of all tags. (b) Venn diagram shows the OTUs shared by T. cuspidate, T. mairei, and T. media samples. (c) A heatmap of the abundance of OTU in the three Taxus species (N = 15). The heatmap scale ranges from 0 to 25 on a log2 scale

Full size image

Alpha and beta diversity analysis

The endophytic fungal communities in Taxus trees were significantly influenced by host plants. Analysis of the Chao1 dilution curve showed that the curve reached an asymptote, indicating that the sequencing depth was adequate (Fig. 2a). Based on the Chao1 and Observed_species indexes, the richness of endophytic fungal species in T. mairei is the highest, followed by T. cuspidate and T. media (Fig. 2b). Similarly, the analysis of Shannon index and Simpson parameters showed that T. mairei exhibited the highest diversity of endophytic fungi among the three Taxus species, followed by T. media and T. cuspidata (Fig. S1a and b). The UniFrac distance measurement and multivariate analysis were applied to calculate the differences in endophytic fungal communities among 45 Taxus samples. The UPGMA tree analysis showed that the endophytic fungal communities of T. cuspidata and T. media were clustered together, indicating a closer genetic relationship between T. media and T. cuspidata (Fig. 2c). PCoA revealed significant variations in fungal community composition among different Taxus trees (Fig. 2d). The highest variations in the microbiota of the sample groups were 40.31% (PCo1) and 28.66% (PCo2), highlighting the distinctions in endophytic fungi among the three Taxus species.

Fig. 2
figure 2

Alpha diversity analysis. Four parameters, including (a) Rarefaction: Chao1 index curve, (b) Rarefaction: Observed_species index curve, (c) Dendrogram analysis and (d) PCA, were analyzed

Full size image

Taxonomic analysis of endophytic fungal communities

After filtering out the rare OTUs, the remaining OTUs were grouped into 2 phyla, 6 classes, and 19 genera (Table S4). The two identified phyla were Ascomycota and Basidiomycota, and the relative abundance of these phyla varied among the three Taxus species. No Basidiomycota fungi were detected in T. cuspidate. Ascomycota fungi accounted for 95.5% in T. media and only 99.0% in T. mairei (Fig. 3a).

Fig. 3
figure 3

Taxonomic analysis of endophytic fungal communities. (a) Taxonomic analysis of endophytic fungi at phylum level. (b) Taxonomic analysis of endophytic fungi at class level. (c) Taxonomic analysis of endophytic fungi at genus level

Full size image

Among the eight classes identified, the dominant class was Dothideomycetes, which accounted for 93.1%, 76.9% and 70.0% in T. cuspidate, T. mairei, and T. media, respectively. The second dominant class was Eurotiomycetes, accounting for 11.7% in T. mairei and 10.8% in T. media. However, no Eurotiomycetes fungi were detected in T. cuspidate. Ascomycota, an unidentified fungus in the phylum Ascomycota, was the second dominant class in T. media, accounting for 11.6%, and in T. cuspidate, accounting for 6.9%. No Ascomycota was detected in T. mairei. The remaining enriched classes were Sordariomycetes, Microbotryomycetes, Basidiomycota, and Lecanoromycetes. All the above fungal classes were not detected in T. cuspidate, while Basidiomycota and Lecanoromycetes were not detected in T. mairei (Fig. 3b). At the genus level, the dominant genus in T. cuspidate was Lophiostoma (75.6%), the dominant genera in T. media were Capnodiales (24.1%) and Davidiella (15.8%), and the dominant genera in T. mairei were Ramichloridium (28.6%) and Alternaria (19.9%) (Fig. 3c). The detailed information of endophytic fungal communities is listed in Table S4.

Relationships between different fungal genera and the accumulation of taxoids

To reveal the relationship between the differential fungal genus and taxoid accumulation, a previously published untargeted metabolome was used for Pearson correlation analysis [32]. The detailed information regarding the relationship between the differential fungal genus and the accumulation of taxoids is presented in Table S5. Ramichloridium, Alternaria, Guignardia, Glomerella, and Phyllosticta were negatively correlated with the accumulation of Taxol (paclitaxel) (Rho < -0.99) (Fig. 4a). Capnodiales, Devriesia, Ascomycota, Paramycosphaerella, Basidiomycota, Ceramothyrium, and Sporobolomyces were positively correlated with the accumulation of 10-deacetyl-2-debenzoylbaccatin III (Rho < -0.99) (Fig. 4b). Most of the differential fungal genus were positively correlated with the accumulation of 3’-N-debenzoyl-2’-deoxytaxel and 3’-N-debenzoyltaxol (Fig. 4c and d). Most of the differential fungal genus, except for Lophiostoma, were negatively correlated with the accumulation of baccatin III and 10-deacetylbaccatin III (Fig. 4e and f).

Fig. 4
figure 4

Pearson correlation analysis of differential fungal genus and host’s taxoid accumulation levels. Analysis of paclitaxel (a), 3’-N-debenzoyltaxol (b), 3’-N-debenzoyl-2’-deoxytaxol (c), 10-deacetyl-2-debenzoylbaccatin III (d), baccatin III (e), and 10-deacetylbaccatin III (f)-related differential fungal genus

Full size image

Screening of taxol biosynthesis-related fungal isolates

ITS sequencing identified a T. mairei-specifically accumulated endophytic fungal genus, Alternaria (Fig. 5a). According to the Pearson correlation analysis, the Alternaria genus showed a negative correlation with the accumulation of Taxol (Rho = -0.93), suggesting its role in the negative regulation of the Taxol biosynthesis pathway (Fig. 5b). Twigs of T. mairei were used for screening the endophytic fungal isolates, and two Alternaria isolates, L7 and M14, were isolated from the twigs of T. mairei (Fig. 5c). The phylogenetic tree displayed similar taxonomic affinities among identifiable taxa and the two Alternaria isolates, L7 and M14 (Fig. 5d).

Fig. 5
figure 5

Screening of taxol biosynthesis-related fungal isolates. (a) The relative abundance level of Alternaria genus in three different Taxus species. (b) Pearson correlation analysis of differential fungal genus and host’s taxoid accumulation levels. (c) Photograph of two fungal isolates belonging to Alternaria genus. (d) Phylogenetic relationships of Alternaria isolates L7 and M14 with other Alternaria spp. based on the ITS sequences. Red stars indicated Alternaria isolates L07 and M14

Full size image

Effect of Alternaria isolate inoculation on the expression pattern of T. mairei

To investigate the role of the Alternaria genus in Taxol biosynthesis, transcriptomic analysis was performed. Transcriptome study revealed that 4298 DEGs, including 1832 up- and 2466 down-regulated genes, were detected in the L7/CK comparison, and 4784 DEGs, including 2082 up- and 2702 down-regulated genes, were detected in the M14/Ck comparison (Fig. 6a). A Venn diagram illustrated that 3348 DEGs were common to both comparisons (Fig. 6b). GO enrichment analysis revealed that a number of DEGs were enriched in Taxol biosynthesis pathway-related GO terms, such as the paclitaxel biosynthetic process (GO:0042617, P = 2.03E-11) and paclitaxel metabolic process (GO:004261, P = 5.86E-23). Additionally, L7 and M14 inoculations activated several hormone biosynthesis and signaling transduction pathways (Fig. 6c).

Fig. 6
figure 6

Effect of inoculation withAlternariaisolates L7 and M14 on the host Taxus trees. (a) The spores of Alternaria isolate L7 and M14 was sprayed onto the twigs of T. mairei. (b) The number of up-regulated and down-regulated genes after Alternaria isolate L7 and M14 spore infection. (c) GO enrichment analysis of the DEGs after Alternaria isolate L7 and M14 spore infection. (d) The expression level of taxol biosynthesis pathway genes after Alternaria isolate L7 and M14 spore infection. Red indicated up-regulated genes and green indicated down-regulated genes. The heatmap scale ranges from − 2 to + 2 on a log2 scale. (e) The number of phytohormone-related DEGs. Red indicated up-regulated genes and green indicated down-regulated genes

Full size image

A number of Taxol biosynthesis-related genes were analyzed by transcriptomic analysis. Interestingly, the expression levels of the most identified genes in the Taxol biosynthesis pathway were significantly inhibited by L7 and M14 inoculations (Fig. 6d). Most of the ethylene signaling pathway-related genes were significantly up-regulated by L7 and M14 inoculations. Meanwhile, most of the ABA signaling pathway-related genes were significantly down-regulated by L7 and M14 inoculations (Fig. 6e). Our data confirmed that Alternaria has a negative regulatory effect on the host’s Taxol biosynthesis by regulating the host’s hormone pathways.

Effect of Alternaria isolate inoculation on the metabolite accumulation of T. mairei

After 48 h incubation, the twigs of T. mairei were collected and cleaned for untargeted metabolomic analysis. In total, 7029 annotated metabolites were detected (Table S6). Quality control of the ion feature data was analyzed (Fig. S1), indicating good data repeatability. KEGG classification analysis identified four taxoids, including 10-deacetylbaccatin III, baccatin III, 10-deacetylytaxol, and taxol, and seven phytohormones, including (±) abscisic acid, (S)-(+)-abscisic acid, (+)-7-iso-jasmonate, (+)-7-iso-jasmonate, 24-epi-brassinolide, gibberellin A36, and gibberellin A37 (Table S6). Interestingly, the four detected taxoids in T. mairei were significantly inhibited by L7 inoculation. (±) Abscisic acid and (+)-7-iso-jasmonate were significantly up-regulated by the L7 inoculation, and gibberellin A36 and gibberellin A37 were significantly down-regulated by the L7 inoculation (Fig. 7).

Fig. 7
figure 7

Effect of inoculation with Alternaria isolate on the metabolite accumulation of T. mairei. Relative accumulation levels of four taxoids and seven hormones. Each value is the mean ± standard deviation of three biological repeats. *P < 0.05

Full size image

Discussion

Recently, the diversity of plant endophyte communities has gradually become a hot topic. Endophytic fungi colonizing plant tissues have beneficial effects on the host ecosystem [35]. Endophytic fungi form a close relationship with host plants during long-term co-evolution, providing nutrients for plant defense and enhancing plant resistance [36]. Furthermore, endophytic fungi influence the growth and production of key secondary metabolites in host plants [37]. Our study analyzed the diversity of endophytic fungi in three different Taxus species, providing a foundation for studying the impact of plant-endophytic microbial interaction on the host’s Taxol biosynthesis.

Flores-Cotera et al. screened the endophytic fungi of T. globosa and obtained 245 endophytes from 21 genera [22]. In 2013, 81 fungal isolates belonging to 8 genera were recovered in T. media based on morphological and molecular identification [23]. A total of 145 endophytic fungi, comprising 124 taxa, were isolated from the barks, branches, and leaves of T. mairei [38]. In the present study, the characteristics of endophytic fungi in Taxus were analyzed and compared. A total of 271 endophytic fungal OTUs were detected in three Taxus species, representing 19 genera, suggesting that our data provides sufficient coverage.

Altitude has a significant impact on the interaction between endophytic fungi and their hosts. Below an altitude of 2000 m, Ascomycota occupies a larger proportion of endophytic fungi than Basidiomycota, becoming the dominant phylum [39]. The distribution of endophytic fungi varied greatly among three Taxus species, primarily due to host genotype and growing environment [39]. In T. cuspidate and T. mairei, Ascomycota is the dominant phylum, suggesting that these species are more sensitive to altitude than the hybrid variety T. media. In T. globosa, the Mexican yew, Dothideomycetes was considered one of the dominant classes [22]. In our study, Dothideomycetes and Eurotiomycetes were the predominant classes found in all Taxus trees.

At the genus level, Alternaria and Davidiella fungi are adapted to extreme low-temperature environments, exhibiting good adaptability to continuous low temperatures and high ultraviolet radiation [40]. Previous studies have also found evidence of altitude-related competition between Alternaria and Davidiella [25]. Taxus trees grow in areas with high altitudes and low temperatures [41]. Therefore, Alternaria and Davidiella play an important role in the adaptability of Taxus trees to low temperatures and high altitudes. In addition to environmental adaptation, endophytic fungi contain many important plant-derived secondary metabolites [42, 43]. Newly identified 3,4-dihydroisocoumarin derivatives were detected in the endophytic fungus Lophiostoma sp. Sigrf10, which is isolated from Siraitia grosvenorii [44]. Several novel compounds, including scorpinone, 5-deoxybostrycoidin, and 4-methyl-5,6-dihydro-2 H-pyran-2-one, were extracted from the endophytic fungus Lophiostoma sp. Eef-7, which is isolated from Eucalyptus exserta [26]. The Lophiostoma genus occupied the largest proportion in T. cuspidata, suggesting an essential role of Lophiostoma in Taxus secondary metabolism.

To date, a large number of Taxol-producing fungi have been identified, offering a potential shortcut for the industrial production of Taxol [45]. For example, Alternaria alternata F3 is the first reported taxol-producing endophytic fungus isolated from Taxus fruits [46]. Endophytic fungi have a significant impact on the growth, development, and taxane accumulation of Taxus trees. Three promising strains, Kocuria TRI2-1, Micromonospora TSI4-1, and Sphingomonas MG-2, were also found to significantly enhance the accumulation of taxanes in T. yunnanensis stem cells [47]. The endophytic fungus Pseudodidymocyrtis lobariellae KL27-FB has been found to significantly enhance the expression of key genes involved in the Taxol biosynthesis pathway [28]. Our study suggests that two Alternaria isolates (L7 and M14) may have a negative regulatory effect on Taxol biosynthesis by down-regulating the expression of Taxol biosynthesis-related genes. In Taxus tissues, taxoids play a role in the growth inhibition of certain specific endophytic fungi. Various endophytic fungi played opposite roles in maintaining the balance of Taxol biosynthesis in Taxus trees. In our study, two fungi negatively correlated with Taxol biosynthesis were screened, which can be used as antagonistic fungi to screen endophytic fungi positively correlated with Taxol biosynthesis.

Phytohormones, such as jasmonic acid (JA), gibberellic acid (GA), abscisic acid (ABA), and ethylene, have been reported to be involved in the regulation of Taxol biosynthesis [48]. Our RNA-seq analysis showed that most of the ethylene signaling pathway-related genes were significantly up-regulated under L7 and M14 infections, suggesting that Alternaria isolates might affect Taxol biosynthesis by activating the ethylene signaling pathway. Moreover, the expression of ABA signaling pathway genes was significantly down-regulated by L7 and M14 infections, indicating that ABA has the opposite effect of ethylene in the regulation of Taxol biosynthesis. Several hormones, such as JA, ABA, GA, and ET, were reported to be involved in Taxol biosynthesis [49, 50]. The two Alternaria strains might inhibit Taxol biosynthesis by regulating various hormone signaling pathways.

In summary, our study initially examined the variances in abundance and community of endophytic fungi among three Taxus species using high-throughput sequencing. A total of 271 OTUs representing 19 genera of fungi were detected in three library groups. Taxonomic analysis of endophytic fungal communities suggests significant differences in endophytic fungi among different Taxus trees. A Pearson correlation analysis was conducted to predict the presence of taxol biosynthesis-related fungal genera and to screen two Alternaria isolates. The two Alternaria isolates (L7 and M14) may have a negative regulatory effect on taxol biosynthesis by modulating hormone signaling pathways. Our findings will serve as a foundation for advancing research in the utilization of Taxus and will assist in the identification of endophytic fungi associated with taxol production.