In a previous study, strains H331 and H263 were isolated as dark septate endophytes (DSEs) from the healthy roots of Gymnocarpos przewalskii in the Minqin National Reserve (38°59'48"N, 103°2'33"E), Gansu Province, China, in September 2019 and July 2020, respectively (Li et al. 2021a). The fungal isolates were stored at - 80 °C, then transferred on to PDA, OA and MEA media and incubated in the dark at 27 °C for at least 28 days prior to morphological observations of colonies and microscopic features. The colours, texture and pigmentation of the colonies were recorded and their diameters were measured on the 7^th and 28^th days. Fresh mycelia were mounted in sterile double-distilled water and observed under a microscope (Nexcope NE900, Zhejiang, China). When necessary, frozen sections of microscopic structures were cut with a cryostat (CM1860, Leica, Germany). The sizes of the conidiogenous cells and conidia were measured at least 20 times. Specimens of the oven-dried cultures were preserved at the Herbarium of Mycology, Chinese Academy of Sciences (HMAS). Living cultures of all strains were deposited in the China General Microbiological Culture Collection Center (CGMCC), Beijing, China.

DNA was extracted using an alkaline thermolysis method. A small amount of fresh mycelium, visible to the naked eye, was picked from the margins of each colony and placed in a well of a 96-well PCR plate. To each well, 50 µl of sodium hydroxide (NaOH) solution (10 µmol/l) was added and the plate was then placed in a thermal cycler and incubated at 98 °C for 20 min. Sequence data for the ITS, LSU, SSU and TEF regions were generated in this study. The ITS region was amplified using the primer pair ITS4 and ITS5 (White et al. 1990), the LSU gene was amplified using LR0R and LR5 (Vilgalys and Hester 1990), the SSU region was amplified using NS1 and NS4 (White et al. 1990) and the TEF region was amplified using EF1-983F and EF1-2218R (Yang et al. 2019). All PCR reactions were performed in an EASTWIN™ Thermal Cycler (ETC811, Suzhou, China) with a 25 µl reaction system containing 12.5 µl of 2× buffer, 3 µl of template DNA, 1 µl of each primer and 7.5 µl of sterilised double-distilled H2O. The thermal cycling conditions for each locus are provided in Table 1. The PCR products were examined by electrophoresis at 120 V for 30 min in a 1.2% (w/v) agarose gel in 1× TAE buffer, then visualized under ultraviolet light after staining with ethidium bromide. The PCR products were sent to General Biology Company (Anhui, China) for Sanger sequencing. The sequences were reviewed and manually modified with Chromas v.1.0.1.1 to remove low-quality base calls from both ends. The modified sequences were subsequently deposited in GenBank.

Preliminary alignments and BLAST suggested that both H331 and H263 were highly similar in sequence and might represent a novel clade within Pleosporales. Representative sequences from Pleosporales members, particularly those from type materials, were selected for subsequent phylogenetic analysis to determine the taxonomic placement of H331 and H263 placement (Table 2, Please refer to Suppl. material 1 for detailed information and alignment matrices for all datasets). Multiple sequence alignment was performed using MAFFT v. 7.3.13. The alignments were reviewed in BioEdit v.7.0.9 to trim both ends and manually correct any obvious misalignments (Suppl. material 2: Align_combine_60). In this study, a 355 bp intron was identified in the SSU of strain H263 and was manually excised. The combined LSU, SSU, ITS and TEF sequence datasets were then analysed using Maximum Parsimony (MP), Maximum Likelihood (ML) and Bayesian Inference (BI) methods using PAUP v. 4.0, RAxML v. 8.2.12 and MrBayes v. 3.2.6, respectively (Suppl. material 2: comb60_PAUP.nxs, comb60_MrBayes.nxs). Dothidea sambuci (DAOM 231303) and Phaeosclera dematioides (CBS 157.81) in Dothideales were selected as outgroups, based on Ahmed et al. (2014).

The optimal models for phylogenetic analysis were predicted using MrModelTest v. 2.4, based on the Akaike Information Criterion (AIC). MP and BI analyses were performed using the GTR+I+G model and ML was calculated using the GTRGAMMA model. Both MP and ML were conducted with 1,000 bootstraps. In the BI analysis, eight Markov chains were run for 1,000,000 generations, with trees sampled every 100 generations. The first 25% of the trees were discarded as burn-in and the remaining trees were used to compute the Bayesian Posterior Probabilities (BPPs). The phylogenetic trees were visualised and edited with FigTree, TreeView and Adobe Illustrator.

Reverse transcription PCR (rtPCR) was conducted to acquire the cDNA of the 18s partial gene in this study. RNA was extracted from fresh mycelia using the RNAprep Pure Cell/Bacteria Kit (cat. DP430, TIANGEN BIOTECH, Beijing, China) following the manufacturer’s instructions. The template RNA-primer mixture was prepared by mixing 1 μg of total RNA, 1.0 μl of random hexamer primer and adjusting the final volume to 13.5 μl with RNase free ddH2O on ice. The mixture was incubated at 70 °C for 2 min and then placed immediately on ice. RNA was reversely transcribed using the PrimeScriptTM 1^st strand cDNA Synthesis Kit (Takara Bio, Japan). The following components were added to the 13.5 μl of the RNA-primer mixture on ice: 4.0 μl of 5×Reaction Buffer, 0.5 μl of Recombinant RNase Inhibitor, 1.0 µl of MMLV Reverse Transcriptase and 1.0 µl of dNTP Mix (10 mM each). The mixture was gently shaken, incubated at 42 °C for 60 min, then heated at 94 °C for 5 min to stop the reaction and subsequently placed on ice for further processing.

The cDNA templates were amplified using specific primer pairs designed for this study (F1, 5’-CGATACGGGGAGGTAGTGAC-3’; R1, 5’-TTCTCCAGGAAAGAAGGCCC-3’). The amplifications were performed in an ABI thermal cycler (ABI-2720, Applied Biosystems, USA) with a 25 µl reaction system containing 2.0 μl of cDNA template, 1.25 μl of forward and reverse primer (10 μM each), 12.5 μl of Q5 High-Fidelity 2× Master Mix and 8.0 μl of nuclease-free water. The PCR products were examined as described above. The PCR products were sent for sequencing using the Sanger method at Personalbio Technology Company (Shanghai, China). The sequences were reviewed and modified as described above and deposited in GenBank.

The 355 bp intron present in the 18S gene strain H263 was subjected to BLAST in GenBank, using the “Core nucleotide database (core_nt)” within the “Standard databases (nr etc.)” for all organisms, with uncultured/environmental sample sequences excluded and optimised for highly-similar sequences using the megablast algorithm. The searching identified 464 references. The 18S sequences of H263 and 464 references were aligned, manually modified and the 18S gene regions were excised using MAFFT and BioEdit, producing an alignment dataset consisting only of introns from 465 18S genes (Suppl. material 2: Align_onlyintron_465, Info_onlyintron_465). In addition, 287 references with alignment scores higher than 200 were selected for subsequent intron phylogenetic analysis. The 18S sequences from H263 and 253 references were aligned, manually modified and trimmed using MAFFT and BioEdit, as described above, for phylogenetic analysis. Identical sequences were removed and the modified alignments of 254 sequences (Suppl. material 2: Align_all_254) were divided into two datasets: one containing only intron sequences from the 18S gene of H263 and 253 references (Suppl. material 2: Align_onlyintron_254) and another containing the 18S exon sequences with introns removed (Suppl. material 2: Align_nointron_254).

The variation rates of nucleotide bases in the 18S genes carrying introns were calculated using the dataset nointron_254. Sequence lacking either the left or right flank of the intron were excluded, resulting in an alignment dataset with 229 sequences (Suppl. material 2: Align_nointron_229, Info_nointron_229).

The ratios of intron carriers to non-carriers within specific species were calculated, based on available data deposited in GenBank. References without species information (denoted as “sp.”) were excluded from the 254 references, which encompassed sequences from 126 fungal species. For each species, all its 18S sequences were retrieved from GenBank using the following search strategy: “SPECIES NAME” [Organism] AND (18s [All Fields] OR “small subunit ribosomal” [All Fields]) NOT 5.8s [All Fields]. The carrier:non-carrier ratios were calculated for species with more than two 18S gene records (Suppl. material 2: Info_species_ratio).

To investigate the spread of introns within Pleosporales, we compiled a dataset for ancestral character state reconstructions (ASRs). This dataset included introns from 125 fungal species (with Colletotrichum gloeosporioides excluded) within Pleosporales (Suppl. material 2: Align_ASR_onlyintron_125, Info_ASR_onlyintron_125).

PCoA ordination was performed, based on the dissimilarity matrix of the alignment dataset of onlyintron_465 with the cmdscale function and plotted in R (Suppl. material 2: R_codes_used_in_present_manuscript.r). To assess the evolutionary trajectory of introns within Pleosporales, the datasets onlyintron_254, nointron_254, and ASR_onlyintron_125 were subjected to ML analysis as described above for co-phylogeny analysis and ancestral character reconstruction. The best ML trees generated from the onlyintron_254 and nointron_254 datasets were used to perform PACo and parafit analyses to test the co-evolution between introns and the 18S gene that carries them (Suppl. material 2: RAxML_bipartitions.254_onlyintron_ML, RAxML_bipartitions.254_nointron_ML, Info_Co-phylogeny-254). The Procrustes Application to Co-phylogenetic analysis (PACo) was carried out using the PACo function from the R package “paco” (Balbuena et al. 2020) and the parafit analysis was performed using the parafit function from the R package “ape” (Paradis and Schliep 2019). The 125 species carrying introns within Pleosporales were coded according to their family placements (Suppl. material 2: Align_ASR_onlyintron_125, Info_ASR_onlyintron_125). Ancestral character estimation was performed using the ace function from the R package “ape”. The phylogenetic trees were visualised using tools available in the R packages “ggtree” and “phytools” (Yu et al. 2017; Revell 2024).

In our preliminary study, BLAST results indicated that strains H331 and H263 likely belong to Pleosporales. To determine their precise taxonomic placement, 54 species of 25 families in Pleosporales were included in the molecular phylogenetic analysis as reference sequences, with Dothidea sambuci and Phaeosclera dematioides used as outgroups (Table 2). The nucleotide sequences of ITS, LSU, SSU and TEF from various species were retrieved from GenBank, along with those generated in our study. The aligned and trimmed datasets for the four loci, comprising sequences from 58 species across 25 families in Pleosporales and two outgroups, were concatenated into a combined dataset of 3,499 bp. This dataset included 947 bp for SSU, 869 bp for ITS, 779 bp for LSU and 904 bp for TEF. Bayesian analysis produced a phylogeny tree with a topology similar to those derived from ML and MP analysis. The Bayesian consensus tree indicated that members of Trematosphaeriaceae formed a monophyletic clade with two main branches. Strains H331 and H263 formed a highly-supported clade (ML: 100%; MP: 100%; BI: 1.00; Fig. 1) within the Trematosphaeriaceae family (ML: 98%; MP: 83%; BI: 1.00), distinct from its sister clade, Fuscosphaeria, which exhibited long branches (ML: 100%; MP: 99%; BI: 1.00).

Figure 1. 

Phylogram of the combined genes of SSU, ITS, LSU and TEF, based on Bayesian analysis. The topology of the Bayesian tree was similar to those generated in the ML and MP analyses (Suppl. material 2: RAxML_bipartitions.comb60_ML, comb60_MPtree.nex, comb60_MrBayes.nxs.con). The numbers at each node represent the percentages of bootstrap values in ML (left) and MP (middle) calculated from 1,000 replications and the BPP (right). - indicates lack of support or support less than 50% (or 0.5 for BPP) for a particular clade. Scale bar indicates 0.1 expected changes per site.

In the present study, an inserted DNA fragment of 355 bp in length was detected in the 18S region of H263. In order to validate whether the fragment is an intron of 18S RNA gene, the rtPCR was conducted to determine whether the fragment is cut out during 18S RNA gene transcription. With total RNA extracted from fresh mycelia of H331and H263, the 18S RNA and 28S RNA were visualised with gel electrophoresis (Fig. 3A) and the cDNA of 18S RNA were successfully amplified (Fig. 3B). Sequencing indicated the 355 bp DNA insertion in the 18S region of H263 was cut off in matured rRNA and, thus, was validated as an intron (Fig. 3C).

BLAST was employed to query 464 references with intron hits, using a maximum of 5,000 target sequences (Suppl. material 1). Derived from 259 taxa, these reference sequences were exclusively 18S genes, the majority of which belong to the Kingdom of Fungi, with the exception of Oomycota (Fig. 4A). Interestingly, when the BLAST hits were filtered to include only those with alignment scores of at least 200, 287 references remained, representing 152 taxa, 143 of which were Pleosporales (Fig. 4B, Suppl. material 1). One exception was a reference identified as Colletotrichum gloeosporioides (MH051114, Glomerellales). However, BLAST results suggested the MH051114 was misidentified (data not shown) and we excluded it from our analysis. The alignment of the nointron_229 dataset indicated that intron insertion occurred at highly-conserved sites in the 18S gene of all 229 references (Fig. 4C). Furthermore, the alignments of the 465 intron sequences suggested that the 287 high-scoring introns were more aggregated in PCoA plots than the low-scoring introns (Fig. 4D). This suggests that these introns are phylogenetically affiliated descendants from common ancestors and are specific to the 18S gene of Pleosporales species. Based on the principles proposed by Zhang and Zhang (2019), we propose the name Nta.18SS453 for the intron found in H263, referring to its presence at site 453 in the 18S gene of N. tarda for convenience of reference in future research. Moreover, we define the Pcl as Ple.18S1, which includes Nta.18SS453 and similar introns on the 18S gene of other Pleosporales members.

Figure 3. 

The amplification of intron Nta.18SS453 with rtPCR and its location on 18S rRNA gene. A The RNA of 28S (smaller than 2000 bp) and 18S (larger than 750 bp) rRNA were detected with gel electrophoresis (1.2% agarose, 120 V, 23 min); B the PCR product of 18S rRNA cDNA in H263 and H331; C cDNA alignment indicated the intron started following a “GT” and ended with an “AG” and was split off during transcription.

Figure 4. 

Brief information on intron Pcl Ple.18S1. A Taxonomic distribution of Nta.18SS453 and similar introns (dataset onlyintron_465); B taxonomic distribution of Nta.18SS453 and 287 references with alignment scores higher than 200; C the maximum number of positional nucleotides in the alignment of 229 sequences (dataset nointron_229), indicating conservation at each nucleotide. The consensus sequence at both flanks of the insertion is shown at the top; D PCoA plot showing dissimilarities in the alignment of 465 intron sequences (dataset onlyintron_465).

To investigate whether Ple.18S1 spreads amongst Pleosporales via horizontal (random jumping) or vertical (inheritance) transmission, we examined the distribution of Ple.18S1 amongst various species within the order. Our results indicated that Ple.18S1 introns frequently increased and, subsequently, decreased at the species level. Amongst the 126 introns detected, 80 species were found to have more than one 18S sequence deposited in GenBank, of which 60 species simultaneously possessed isolates with and without the intron (Fig. 5A). Despite the observed intraspecies incongruous distribution, parafit and PACo analyses revealed that the phylogeny of introns was significantly correlated with the phylogeny of the 18S gene (ParaFitGlobal = 226.94, p = 0.001 in parafit and p = 0 in PACo). In the comparison between ML trees, no obvious co-phylogeny pattern was observed at the families’ level amongst the host species of the introns (Fig. 5B). Nevertheless, frequent co-phylogeny was noted at lower taxonomical hierarchies, indicated by parallel connections.

ASR provided further insight into the transmission dynamics of Ple.18S1 amongst host 18S genes. Fig. 6 illustrates the reconstruction of the host families of Ple.18S1 introns on an unrooted phylogenetic tree. The results revealed that the number of introns increased across most Pleosporales families. Introns hosted by Phaeosphaeriaceae formed polyphyletic clades, while, in families such as Cucurbitariaceae, Dictyosporiaceae, Lentitheciaceae and Thyridariaceae, most introns formed monophyletic clades, with a few exceptions, suggesting occasional interfamily host transitions. Some monophyletic clades were observed within genera, including Astrosphaeriella (Astrosphaeriellaceae), Longipedicellata (Bambusicolaceae), Periconia (Periconiaceae), Sulcatispora (Sulcatisporaceae), Cryptocoryneum and Digitodesmium (Pleosporales incertae sedis). These findings suggest that Ple.18S1 introns have undergone vertical transmission along the host phylogeny to a certain degree, remaining conserved at the genus level.

Figure 5. 

The Ple.18S1 intron distributions amongst fungal species. A Complete list based on all public sequences showing ratios between Ple.18S1 intron carriers and non-carriers within a certain species. The numbers in the right column indicate the numbers of 18S references deposited in the GenBank database; B co-phylogenetic comparison between ML trees in which 254 intron sequences and 254 18S sequences carried introns (left, dataset onlyintron_254; right, dataset nointron_254). Lines of different colours connect an intron and an 18S exon from the same reference in the GenBank database.

Figure 6. 

The ancestral state reconstruction of Ple.18S1 introns, showing their host alternation within Pleosporales, based on an unrooted ML tree (Suppl. material 2: RAxML_bipartitions.125_onlyintron_ML). The ancestral states are categorised into different families of intron carriers. Colours at each terminal node indicate the family of the intron host. Pie charts at each node indicate the relative likelihoods of ancestral states. Stacked colour chart approximately demonstrate the ecological traits of the habitats from which the fungi were recovered, according to the sequence information deposited in GenBank.

Nigromargarita tarda were determined to Pleosporales in a preliminary sequence search. Although this taxon lacks teleomorphs, its pycnidia and aseptate, hyaline, simple conidia resemble the Phoma-like taxa in Pleosporales (Gruyter et al. 2013). Nevertheless, the most distinguished morphological feature of N. tarda is its exceptionally small pycnidia, typically ranging from 31 to 39 μm in diameter and never exceeding 50 μm in our study. This is in stark contrast to most other species within Pleosporales, which possess pycnidia or other types of conidiomata with diameters exceeding 100 μm (Zhang et al. 2012; Li et al. 2020). Even amongst Phoma-like genera, known for their pycnidia of varying sizes and hyaline, aseptate conidia, it is unusual to find pycnidia as small as those observed in N. tarda (Aveskamp et al. 2010; Chen et al. 2017). Several Phoma-like species with relatively small pycnidia include Aposphaeria pulviscula ((30–)54–100 μm diam., (56–)84–133 μm high), Ascochyta neopisi (80–120 µm diam., 80–100 µm high) etc. (Li et al. 2020). However, their pycnidia are still considerably larger than those of N. tarda. Chaetophoma quercifolia which possesses pycnidial conidiomata up to 30 µm in diameter according to Sutton (1980), is somewhat comparable to N. tarda in pycnidial size. However, the phialidic conidiogenous cells of C. quercifolia are ampulliform, whereas those of N. tarda are straight or curved cylindrical, presenting a morphological distinction. C. quercifolia is classified as Pleosporales incertae sedis according to Li et al. (2020). Unfortunately, no sequence data for C. quercifolia have been published, precluding its inclusion in the phylogenetic analysis of N. tarda.

Members in Trematosphaeriaceae family can be divided to two ecological groups, viz. human dermatophytes, plant-associated fungi in marine habitats and root- associated fungi in terrestrial habitats. Nevertheless, this ecological divergence does not attribute to their phylogeny, as revealed in our analysis. In this study, multiple-locus phylogenetic analysis revealed that N. tarda represented an independent clade within the Trematosphaeriaceae under Pleosporales. This result also refuted N. tarda as any known Phoma-like taxa. Three genera in Trematosphaeriaceae have been established, based on sterile specimens, namely Emarellia, Fuscosphaeria and Meanderella (Borman et al. 2016; Pintye and Knapp 2021; Ahmed et al. 2022), while other members of the family produce septate ascospores. Unfortunately, N. tarda did not produce teleomorphic features in our study, preventing a direct comparison with other Trematosphaeriaceae members. Despite this, we still observed both similarities and distinct differences between N. tarda and other members of Trematosphaeriaceae. Several Trematosphaeriaceae members produced specialised hyphae structures, such as hyaline terminal chlamydospores in Fuscosphaeria hungarica, tuberculate in Meanderella rijsii and hyphopodia-like structures in Trematosphaeria pertusa (Suetrong et al. 2011; Pintye and Knapp 2021; Ahmed et al. 2022). N. tarda producespigmented, verrucous, thick-walled hyphae, a feature similar to M. rijsii. However, despite both being recovered from plant roots in arid ecosystems and sharing phylogenetic affinity, N. tarda is distinguished from F. hungarica with its pigmentation and unique morphological features of hypha, as well as its well-developed conidiogenous structures. Therefore, based on both morphological and molecular phylogenetic evidence, N. tarda forms an independent clade within the family Trematosphaeriaceae and represents a new fungal genus.

Introns are prevalent in both fungal nuclear and mitochondrial genomes and have become important markers in phylogenetic and genome evolution studies of fungi (Matheny et al. 2006; Wang et al. 2020; Lücking et al. 2021). They are widely, though sporadically, distributed across fungal nuclear genomes and amongst different species, indicating frequent horizontal transfers and recurrent cycles of invasion, degeneration, loss and re-invasion, which highlight the dynamic nature of intron transmission (Reeb et al. 2007). For example, group I introns in the nuclear small-subunit ribosomal DNA (nuc-ssu-rDNA) have been gained independently on two separate occasions via horizontal transmission in mushroom-forming genera: once in the lineage leading to Panellus and once in the lineage leading to Lentinellus and Clavicorona (Hibbett 1996). Similarly, incongruent branching patterns between intron-based and rDNA-based phylogenies in Sclerotiniaceae suggested independent horizontal intron transfers (Holst-Jensen et al. 1999). The mechanisms for intra- and interspecies intron transfer include intron homing and reverse splicing. Intron homing occurs when introns containing homing endonuclease genes (HEGs) spread by recognising specific DNA sequences, thereby facilitating their insertion into intron-less alleles (Haugen et al. 2004). In addition, Bhattacharya et al. (2005) demonstrated that group I introns spread by reverse splicing in Symbiotaphrina buchneri. In the present study, we identified a 355-bp intron, Nta.18SS453, in the 18S gene of N. tarda. Further analysis revealed that this intron belongs to the Pcl Ple.18S1 intron family, which is restricted to conserved insertion site within the 18S gene of Pleosporales species. It is, therefore, intriguing to explore the spread patterns and underlying mechanisms of this intron family within this fungal order.

The present study supports horizontal transmission as the primary mechanism for Ple.18S1 introns. The patterns in Fig. 5B and Fig. 6 indicated that the distribution of Ple.18S1 introns across Pleosporales families is phylogenetically uncorrelated, suggesting sporadic insertions in ancestral taxa of intron-rich families, such as Cucurbitariaceae, Dictyosporiaceae, Lentitheciaceae, Phaeosphaeriaceae and Thyridariaceae. In the ASR analysis, the ancestral nodes showing multiple potential states suggested that introns shifted hosts at the family level through horizontal transmission. Despite ecological data in GenBank references often being incongruous and incomplete, our analysis infers that this horizontal transmission is potentially related to specific habitats. For example, Ple.18S1 introns in Trematosphaeriaceae originated from two distinct sources: one transferred amongst fungi associated with healthy root (e.g. Fuscosphaeria hungarica from Festuca vaginata, Laburnicola radiciphila from Triticum aestivum according to GenBank and Nigermargarita tardus from Gymnocarpos przewalskii) and the other amongst fungi related to aquatic habitats (e.g. Clypeoloculus akitaensis from submerged woody plant, Trematosphaeria pertusa from a stump of Fraxinus excelsior in a swamp and Halomassarina thalassiae from mangrove (Zhang et al. 2008; Suetrong et al. 2009). Accordingly, our analysis reveals that Ple.18S1 introns exhibit phylogeny affinity within similar habitats rather than within same family (Trematosphaeriaceae). In addition, inter-family intron horizontal transmission is demonstrated by multiple examples in Fig. 6, such as from Cucurbitariaceae to Phaeosphaeriaceae and Leptosphaeriaceae and from Phaeosphaeriaceae to Massariaceae, Periconiaceae and Pleosporaceae. Notably, Nta.18SS453 was detected in N. tarda H263, but absent in H331, which was collected from the same location and substrate. This observation suggests that H263 acquired the intron from an undetermined source in the root habitat, whereas H331 did not.

Considering BLAST screening failed to identify Ple.18S1 introns in organisms other than fungi and that the habitats where horizontal transmission occurred are ecologically distinct, non-fungal organisms, such as bacteria, may not mediate this horizontal transmission. Ple.18S1 introns lack endonuclease and occur exclusively at conserved loci within the 18S RNA gene rather than spreading across the genome. Therefore, they are unlikely to follow spreading mechanisms such as intron homing. One possible spread mechanism of Ple.18S1 introns is reverse splicing, as proposed by Bhattacharya et al. (2005), which plays a significant role in the intron spread in Pezizomycotina nuclear rDNA. Another minor possibility is RNA mediated transmission, similar to that observed in plants (Adams et al. 2000; Adams and Palmer 2003; Keeling and Palmer 2008).

Beyond reverse splicing as a possible transmission mechanism, the interspecies transmission of Ple.18S1 introns within Pleosporales exhibited a phylogenetic pattern at low taxonomic hierarchies, such as the genus level. However, vertical transmission of Ple.18S1 introns from ancestral to descendant taxa during speciation is accompanied by frequent intron loss in certain species or strains (Fig. 5A). It has been suggested that fungi share a common ancestor with intron-rich genomes, yet massive intron reduction through intron loss has occurred in multiple fungal clades throughout their evolutionary history (Stajich et al. 2007; Lim et al. 2021). In the case of Ple.18S1 introns, a similar scenario is observed in Darksidea, a genus established in 2015, based on DSEs from semi-arid areas, which currently comprises nine species (Knapp et al. 2015; Romero-Jiménez et al. 2022; Tan and Steinrucken 2024a). Our results detected Ple.18S1 introns in six species of Darksidea with close phylogeny affinity, which suggested a vertical transmission from a common ancestor possessing the intron. However, while this observation may shift as more data accumulate, currently, no intron was detected in D. epsilon, D. eta and D. theta, which inhabit the same root habitat as other intron-carrying Darksidea species (Knapp et al. 2015; Tan and Steinrucken 2024a, 2024b). In addition, all six Darksidea species simultaneously possess strains that either carry or lack introns (Fig. 5A). This inter- and intra-species inconsistency of intron presence is likely attributed to intron loss. Therefore, frequent intron loss and gain events have been observed in our research, as in various fungal clades (Bolotin-Fukuhara and Fairhead 2014; Wu et al. 2021). The complex history of intron inheritance and lateral transfer within fungal rDNA implies that introns spread amongst genomes through various mechanisms (Bhattacharya et al. 2005).

It is well established that certain fungal introns are restricted to specific taxonomic or phylogenetic groups. For example, Holst-Jensen et al. (1999) reported that introns 788SSU and 798LSU were found exclusively within members of Helotiales. In a survey of introns in 39 mushroom-forming species, introns were confined to the genera Punellus, Cluvicoronn and Lentinellus, suggesting a pattern of horizontal transmission followed by vertical inheritance within these lineages (Hibbett 1996). One compelling question that arises is why Pleosporales specially accommodates Ple.18S1 members.

Beyond the genetic and biochemical mechanisms underlying horizontal transmission, the biological and ecological properties responsible for bringing the two taxa close enough are also essential (Holst-Jensen et al. 1999). Pleosporales is a fungal order closely associated with plant materials. Most of the Pleosporales reference sequences in our analysis were sourced from plant matter, including living or dead leaves, twigs, stems and roots; decaying litter; submerged wood; and soils under vegetation, many of which belong to DSEs. These habitats, characterised by temporal or spatial proximity, may provide ample opportunities for horizontal intron transfer amongst Pleosporales members.

The authors have declared that no competing interests exist.

No ethical statement was reported.

All the fungal strains used in this study have been legally obtained, respecting the Convention on Biological Diversity (Rio Convention).

This work was supported by the Advanced Talents Incubation Program of Hebei University (521100221030), the Introducing Overseas Talents Funding Project of Hebei Province (C20220512) and the Hebei Natural Science Foundation (C2023201009).

MYL, performed laboratory assays, analysed the data, prepared figures, wrote the manuscript; XS, conceived the study, analysed the data, prepared figures, edited the manuscript; YQL, performed laboratory assays; SHQ, performed laboratory assays; ML, collected samples, performed laboratory assays; XLH, edited the manuscript and contributed resources.

Meng-Yuan Li https://orcid.org/0009-0005-7352-7991

Xiang Sun https://orcid.org/0000-0002-6875-4971

Yu-Qing Liu https://orcid.org/0009-0005-2544-9394

Sheng-Hui Qin https://orcid.org/0009-0005-3095-8046

Min Li https://orcid.org/0009-0001-8215-8906

Xue-Li He https://orcid.org/0000-0002-2783-3390

The sequence data generated in current research have been deposited in GenBank and all datasets are provided in supplementary files.