The strain Si 85 used in this study was isolated from the fruiting bodies of D. sinensis collected from a fallen angiosperm branch in Changbai Mountain, Antu County, Yanbian Prefecture, Jilin Province, China (altitude: 574 m; 42°31'31"N, 128°3'15"E). The specimen is currently deposited at the Herbarium of School of Ecology and Nature Conservation of Beijing Forestry University in China. Mycelia were harvested following incubation at 28 °C for 7–14 days on potato dextrose agar media (20.0 g/l agar, 20.0 g/l glucose, 200.0 g/l potato and 1.0 l distilled water). Genomic DNA was extracted from the mycelia using the cetyltrimethyl ammonium bromide method (Watanabe et al. 2010). Whole-genome sequencing was performed using the PacBio Sequel and MGISEQ2000 platforms (Nextomics Biosciences Co., Ltd.), according to the manufacturer’s instructions, using GrandOmics (https://www.grandomics.com/, Wuhan, China). The raw reads were subjected to quality control to acquire clean data and used to assemble the mitochondrial genome with GetOrganelle v.1.7.7.0 (Jin et al. 2020). The genome sequence was assembled on Bandage (Wick et al. 2015) to verify its ring structure. Subsequently, the sequence was automatically annotated with MITOS (Donath et al. 2019), after which it was manually corrected. The assembled and complete mitochondrial genome was deposited in GenBank under the accession number NC_085747.

Using published fungal mitochondrial sequences, the PCGs and ribosomal RNA (rRNA) genes of D. sinensis were annotated and the sequences of D. sinensis matched those of closely-related species. The transfer RNA (tRNA) genes were annotated using tRNAscan-SE (http://lowelab.ucsc.edu/tRNAscan-SE/) (Chan and Lowe 2019). A mitochondrial genome map of D. sinensis was drawn using the OGDRAW web server (https://chlorobox.mpimp-golm.mpg.de/OGDraw.html) (Greiner et al. 2019).

CodonW v.1.4.4 (Sharp and Li 1986) was applied to calculate the relative synonymous codon usage (RSCU) values and the data generated above were collated and visualised in R v.4.3.1. The mitochondrial repeats were divided into three main parts for analysis, from which MISA v.2.1 (Thiel et al. 2003; Beier et al. 2017), TRF (Benson 1999) and REPuter (Kurtz et al. 2001) were used to detect simple sequence repeats (SSRs), tandem repeats and interspersed repeats, respectively. The identified repeats were visualised using Circos v.0.69.9 (Krzywinski et al. 2009). The AT and GC skews of PCGs in Daedaleopsis were calculated using the following formulas: AT-skew = (A–T)/(A+T) and GC-skew = (G–C)/(G+C). R was used to visualise PCG length variation, GC content, AT skew and GC skew. The three mitochondrial genomes of Daedaleopsis species were aligned using MAFFT v.7.471 (Katoh and Standley 2013; Rozewicki et al. 2019) and the Ka and Ks values of the genes, representative of the mean numbers of non-synonymous and synonymous substitutions in each non-synonymous and synonymous sites, were determined using KaKs_Calculator v.2.0 (https://sourceforge.net/projects/kakscalculator2/) via the MLWL calculation method (Wang et al. 2010). The Kimura-2-parameter (K2P) genetic distances were calculated by MEGA v.11 (Tamura et al. 2021). The GenBank IDs for the mitochondrial genomes of the other two Daedaleopsis species were as follows: D. confragosa: NC_084114; D. nitida: NC_087776.

Subsequently, the correlation between intron number and mitochondrial genome size was analysed using the Pearson correlation coefficient. Further, the introns of cox1 genes in the mitochondrial genomes of 23 species in Polyporaceae were classified into different position classes (Pcls), according to the method described by Férandon et al. (2010). The cox1 genes in the mitochondrial genomes of 23 Polyporaceae fungi were aligned with those of Ganoderma calidophilum using Clustal W (Thompson et al. 1994; Li et al. 2019b). Each Pcl comprises introns inserted at the same location in the cox1 gene. The same Pcls usually have high sequence similarity.

The NCBI website was used to search for publicly available mitochondrial genome data for fungi, selecting suitable species according to fungal taxonomy categories. We selected mitochondrial data of fungi belonging to Basidiomycota for analysis; information on the 87 varieties is included in Suppl. material 1: table S1. The gene sequences of the PCGs were extracted and imported into MAFFT v.7.471 for multiple sequence alignment (Katoh and Standley 2013). MACSE v.2 (Ranwez et al. 2018) and Gblock (Talavera and Castresana 2007) were used to optimise and trim the MAFFT alignment results, respectively. Multiple genes were concatenated and PartitionFinder v.2.1.1 (Lanfear et al. 2017) was applied to identify optimal partitioning strategies and evolutionary models. The phylogenetic relationships of the mitochondrial genomes of 87 Basidiomycota species were determined using IQ-tree and MrBayes v.3.2, following the Bayesian Inference (BI) and Maximum Likelihood (ML) methods (Ronquist et al. 2012; Nguyen et al. 2015). All analyses were conducted within PhyloSuite (Zhang et al. 2020). The results were visualised and embellished on the iTOL website (https://itol.embl.de/) (Letunic and Bork 2007).

The Deepred-mt tool, which is based on a convolutional neural network model, was applied to predict cytidine-to-uridine RNA editing events in the mitochondrial genome of D. sinensis (Edera et al. 2021). Deepred-mt’s predictions were generally regarded as reliable and data with a probability value higher than 0.9 for further analysis were selected. The high threshold of 0.9 ensures confidence, thereby improving the accuracy and reliability of RNA editing analyses in mitochondrial genomes.

The mitochondrial genomes of nine closely-related species were analysed using BLAST (Chen et al. 2015) and a multiple synteny plot was generated and visualised using MCScanX software (Wang et al. 2012). Mauve v.2.4.0 was applied to analyse and visualise the collinearity and rearrangement between homologous regions of the three Daedaleopsis species (Darling et al. 2004).

BI Bayesian Inference

BPP Bayesian posterior probability

K2P Kimura-2-parameter

ML Maximum Likelihood

PCG Protein-coding gene

Pcl Position class

rRNA Ribosomal RNA

RSCU Relative synonymous codon usage

SSR Simple sequence repeat

tRNA Transfer RNA

Fig. 1A presents a representative image of the fruiting body of D. sinensis in the wild. We newly assembled the complete mitochondrial genome of D. sinensis and compared and analysed it with the publicly-available mitochondrial genomes of D. confragosa and D. nitida. The mitochondrial genome of D. sinensis is a closed circular DNA molecule 69,155 bp in length, with a GC content of 25.0%. The GC content of D. confragosa and D. nitida were similar to D. sinensis, at 25.2% and 25.3%, respectively. The mitochondrial genome of D. sinensis comprises 15 PCGs, two rRNA genes, 26 tRNA genes and one DNA polymerase gene (dpo) (Fig. 1B; Suppl. material 1: table S2). An entire set of PCGs in D. sinensis was detected, including atp6, atp8, atp9, cob, cox1, cox2, cox3, nad1, nad2, nad3, nad4, nad4L, nad5, nad6 and rps3 genes. The two rRNA genes identified were the small (rns) and large subunit ribosomal RNA (rnl), which encode the small and large RNA molecules of the mitochondrial ribosome, respectively. D. sinensis also contains a dpo gene, which is known to have a plasmid origin, encodes a DNA-directed DNA polymerase and comprises mitochondrial plasmid-related genes. The proportion of AT nucleotides in the mitochondrial genome of D. sinensis was greater than that of GC nucleotides.

Each mitochondrial genome of the three Daedaleopsis species contained two rRNA genes: rns and rnl (Suppl. material 1: table S1). The average lengths of rnl and rns in the three species were 2,293 and 1,265 bp, respectively. The 26 tRNAs contained in the D. sinensis were folded into classic clover structures (Suppl. material 2: fig. S1). The 26 tRNA genes identified in D. sinensis encode 20 standard amino acids and range in length from 71 to 88 bp. Owing to its sizeable extra arm, the trnS gene had the largest volume of all the detected tRNA genes. Due to the secondary structure folding process of tRNA, base mismatches occur. The tRNA genes shared by the three Daedaleopsis species varied in location between the three mitochondrial genomes. A total of 39 base mismatches were identified in tRNA genes in the mitochondrial genome of D. sinensis, amongst which 38 were G-U mismatches and one was an A-C mismatch. The mismatched base pairs were distributed in different parts of the tRNAs. Of the tRNA genes shared by the three Daedaleopsis species, there are five variable sites, shown in red in Suppl. material 2: fig. S1. The structure in blue depicts one tRNA in the D. nitida that is completely distinct from the other Daedaleopsis species.

The RSCU of the three Daedaleopsis species were nearly identical, with only minimal differences. The most used codons for each amino acid are depicted in Fig. 2. Arginine (Arg), leucine (Leu) and serine (Ser) were the most frequent amino acids, whereas methionine (Met) and tryptophan (Trp) were less common. Amongst the 64 codons in PCGs, an RSCU value equal to 1 indicates no preference for the amino acid, while an RSCU value greater than 1 indicates a preference for the amino acid. The RSCU values for Met (AUG) and Trp (UGG) were 1. Moreover, 29 codons had RSCU values greater than 1 and were classified as the high-frequency codons. The highest RSCU value was 2.67 for Arg (AGA), indicating a clear preference for AGA, followed by Leu (UUA) and the end codon (UAA), with RSCU values of 2.56 and 1.92, respectively (Suppl. material 1: table S3).

Figure 2. 

Bar chart of RSCU values, which represent the specific codon frequency compared with the expected frequency of uniform synonymous codon usage, with the x-axis representing the codon families. The box on the bottom represents all the codons encoding each amino acid.

The mitochondrial genome of D. sinensis contains 15 PCGs totalling 13,725 bp in length or 19.85% of the total length of the mitochondrial genome. The shortest (atp8) was 159 bp, while the longest (nad5) was 1,986 bp, indicating significant length disparities. Two of the 15 PCGs detected (cox3 and nad6) showed obvious variations in length amongst the three Daedaleopsis species analysed (Fig. 3A). The initiation and termination codons for the 15 PCGs were ATG and TAA, respectively. PCGs were classified into four categories according to their various functions. These genes included three ATP synthase-related genes (atp6, atp8 and atp9), four cytochrome-related genes (cox1, cox2, cox3 and cob), seven NADH dehydrogenase-related genes (nad1, nad2, nad3, nad4, nad4L, nad5 and nad6) and one ribosomal protein-encoding gene (rps3) (Suppl. material 1: table S2). The length of the cox3 gene varied the most amongst Daedaleopsis species, with D. nitida having the longest cox3 gene. With the exception of atp8, which was completely unchanged in all three Daedaleopsis species, and cox3, which was significantly different, the GC contents of the other PCGs did not differ remarkably (Fig. 3B). There are fewer noticeable differences in the AT and GC skews of 15 PCGs in the three Daedaleopsis species, with the majority showing negative AT skews and positive GC skews (Fig. 3C, D). The AT skew in 15 PCGs varied amongst the three Daedaleopsis species, indicating that A/T mutations occurred more frequently in PCGs. Most PCGs were positive for GC skew, suggesting that most PCGs tend to evolve towards being G-rich rather than C-rich in the leading strand of PCGs.

Since the mitochondrial genomes of the three Daedaleopsis species contain 14 core PCGs and one rps3 gene, here 15 PCGs were used to calculate the K2P genetic distances and substitution rates (Fig. 4). While nad6 presented the highest average non-synonymous and synonymous replacement rates, atp8 presented the lowest average Ka and Ks rates, both equal to zero. In addition, atp8 in the three species were identical in size and arrangement. Therefore, a comparison of the atp8 gene sequences revealed that atp8 was highly conserved across the three Daedaleopsis species. All the PCGs had Ka/Ks values of less than 1.0, implying that the genes of Daedaleopsis were under pressure from purifying selection. The K2P genetic distance amongst the nad6 genes was the greatest, followed by that amongst the nad4 genes, indicating significant divergence during evolution. Amongst the mitochondrial genomes of the three Daedaleopsis species, the atp8 and nad3 genes presented the lowest average K2P genetic distances, suggesting high levels of conservation.

Figure 3. 

Variations in the length and base composition of individual PCGs in the mitochondrial genomes of the three Daedaleopsis species A PCG length variation B GC content of the different PCGs C AT skew D GC skew.

Figure 4. 

Genetic analysis of the 15 PCGs conserved in the mitochondrial genomes of the three Daedaleopsis species. Ka, the mean number of non-synonymous substitutions per non-synonymous site; Ks, the mean number of synonymous substitutions per synonymous site; K2P, genetic distance according to the Kimura-2-parameter.

The introns are unevenly distributed in the host genes and show a clear preference for cox1 genes. A total of 374 introns were detected in the mitochondrial genomes of 23 Polyporaceae species, of which 199 were detected in the cox1 gene, accounting for 53.21% of total introns (Suppl. material 1: table S1). The cox1 gene is the largest host gene in mitochondrial introns and its intron dynamics could, therefore, significantly affect the organisation and size of mitochondrial genome. Pearson correlation analysis further revealed a high correlation between the number of introns and sizes of mitochondrial genomes of Polyporaceae (Fig. 5A). Introns in Polyporaceae have a significant effect on mitochondrial genome size. Based on the cox1 gene of Ganoderma calidophilum, the introns of the cox1 gene of 23 Polyporaceae species were divided into different Pcls, while the introns belonging to the same Pcls were considered homologous introns (Fig. 5B). The 199 introns in the cox1 genes of 23 mitochondrial genomes were classified into 30 Pcl types, indicating the rich diversity of intron types in Polyporaceae. A total of 199 introns were also detected in the cox1 gene of 23 Polyporaceae species, of which five introns belonged to group II, one was unknown and the rest belonged to group I. A total of 30 Pcls were detected in 23 Polyporaceae species, of which P706 was the most common intron type, detected in 16 Polyporaceae, followed by P612 and P731 in 15 Polyporaceae. Seven of 30 Pcls (S309, P394, P401, P703, P720, P864, P1125) were only detected in one of the 23 Polyporaceae species. Four intron types (P490, P612, P706, P1262) were relatively common, present in all three Daedaleopsis species. The differences in intron type and number amongst the three Daedaleopsis species further indicated that intron gain/loss occurred during the evolution of the Daedaleopsis mitochondrial genomes. Rarer Pcls, such as P900 and P1057, have also been found in Daedaleopsis, Ganoderma and Perenniporia, indicating that there may be potential gene transfer events between the relatively distant species in mitochondrial genome evolution. Compared to D. sinensis, the mitochondrial genome of D. confragosa contains five non-homologous introns, indicating greater differentiation of intron evolution.

Figure 5. 

A Pearson correlation analyses between the number of introns and sizes of mitochondrial genomes of 23 Polyporaceae species B Pcl information of cox1 genes. Introns in cox1 genes of 23 mitochondrial genomes were classified into different Pcls using the cox1 gene of Ganoderma calidophilum as the reference. The phylogenetic positions of the 23 species were established using the BI and ML methods, based on concatenated mitochondrial genes. The species indicated by the abbreviations are as follows: Dconf, Daedaleopsis confragosa; Dsine, D. sinensis; Dniti, D. nitida; Ffome, Fomes fomentarius; Pepim, Porogramme epimiltina; Dsqua, Dichomitus squalens; Gappl, G. applanatum; Gleuc, G. leucocontextum; Gtsug, G. tsugae; Gsine, G. sinensis; Gresi, G. resinaceum; Gsuba, G. subamboinense; Gwebe, G. weberianum; Gcali, G. calidophilum; Gmere, G. meredithae; Gflex, G. flexipes; Gling, G. lingzhi; Gluci, G. lucidum; Gsich, G. sichuanense; Gpseu, G. pseudoferreum; Gmult, G. multipileum; Pfrax, Perenniporia fraxinea; Psuba, P. subacida. Species and NCBI accession numbers for each of the mitochondrial genomes used are provided in Suppl. material 1: table S1.

The repeats of the mitochondrial genome of D. sinensis were analysed and 41 SSRs were identified. These included 38 (92.68%) monomeric, one (2.44%) dimeric, one (2.44%) trimeric and one (2.44%) hexameric SSR (Suppl. material 1: tables S4, S5). Overall, seven tandem repeats in the mitochondrial genome of D. sinensis were identified, with a similarity of at least 75% and lengths ranging from 6 to 24 bp (Suppl. material 1: table S6). Additionally, 50 pairs of interspersed repeats with a length of at least 30 bp were also discovered, including 30 pairs of forward repeats (F), 19 pairs of palindromic repeats (P) and one pair of complementary repeats (C) (Suppl. material 1: table S7). Suppl. material 2: fig. S2 depicts the distribution of all these repeats in the mitochondrial genome of D. sinensis.

Subsequently, the arrangements of the 15 PCGs and the two rRNA genes were compared in the mitochondrial genomes of the 23 Polyporaceae species (Fig. 6). The comparisons revealed that mitochondrial gene rearrangements were detected between species of different genera. Two gene pairs, nad2/nad3 and nad4L/nad5, were found in the mitochondrial gene arrangement of 23 Polyporaceae species. However, within the mitochondrial genes of Fomes fomentarius, inversion of the nad2 and nad3 gene pair was observed. Compared with those of D. confragosa and D. nitida, the mitochondrial genome of D. sinensis presented evidences of large-scale rearrangements, including gene migration, inversion and insertion. Collinearity analysis of the whole mitochondrial genome was further conducted on the three closely-related Daedaleopsis species. Five homologous regions (A to E) were detected in each of the mitochondrial genomes (Fig. 7). The homologous region C in the other two species was significantly smaller than that in the mitochondrial genome of D. nitida, indicating that the mitochondrial genome contracted during evolution. Collinear analysis of nine phylogenetic relatives revealed further gene rearrangement between D. sinensis and related species (Suppl. material 2: fig. S3). Taiwanofungus camphoratus, Trametes coccinea and Wolfiporia cocos have larger mitochondrial genomes than other species, which may contribute to their greater collinearity. The collinearity results revealed that the mitochondrial genomes of the nine Polyporaceae species were not statistically consistent, indicating that the existence of extensive gene rearrangements occurred between the mitochondrial genomes of D. sinensis and its close relatives.

Figure 6. 

Comparisons of the gene sequences of 23 Polyporaceae species. Starting with cox1, all genes, including PCGs and rRNA genes, are displayed in the order in which they occur in the mitochondrial genome.

Figure 7. 

Collinearity analysis of the mitochondrial genes of the three Daedaleopsis species. Rectangular blocks of the same colour represent homologous regions between different mitochondrial genomes.

Comparative analysis of the mitochondrial genomes of 23 Polyporaceae species revealed the significant variation in the size of 23 mitochondrial genomes, ranging from 40,719 to 124,588 bp, with an average size of 79,583 bp (Suppl. material 1: table S1). The GC content of the 23 mitochondrial genomes also varied, ranging from 24.1% to 28.7%, with an average GC content of 26.1%. The three Daedaleopsis species contained lower GC content than the average GC content of the 23 mitochondrial genomes. Positive AT skews were found amongst 13 of the 23 mitochondrial genomes, while 20 had positive GC skews. Each of the 23 mitochondrial genomes contains two rRNA genes, while the number of tRNA genes ranged from 22 to 29. The number of introns contained in the 23 mitochondrial genomes also varied widely.

A phylogenetic tree of 87 fungal species in the Basidiomycota division was constructed using the DNA sequences of conserved PCGs. Suppl. material 1: table S1 lists the species, NCBI accession numbers and characteristics of the mitochondrial genomes investigated in the phylogenetic study. The phylogenetic tree was constructed using the selected common PCGs, which included the atp6, atp8, cox1, cox2, cox3, cob, nad1, nad2, nad3, nad4, nad4L, nad5 and nad6 genes. It also showed that 74 out of the 84 branching nodes had bootstrap values greater than 90%, including 72 nodes with 100% bootstrap values. In comparison, 70 of the 84 branching nodes had Bayesian posterior probability (BPP) values higher than 90, including 54 nodes with 100 BPPs. D. sinensis is closely related to the two species within the Daedaleopsis genus and is classified on the same branch as Fomes fomentarius (Fig. 8).

Figure 8. 

Molecular phylogeny of the 87 investigated species based on BI and ML methods of the conserved mitochondrial PCGs. The BPP and bootstrap values are indicated next to each branch. The asterisks indicate that the BPP and bootstrap values are 1 and 100%, respectively. The GenBank accession numbers are displayed as suffixes following species names.

We identified 43 C-to-U RNA editing sites in 14 mitochondrial PCGs (Suppl. material 1: table S8). Suppl. material 2: fig. S4A presents the distribution of the predicted RNA editing sites in each gene. Amongst these genes, cox1, nad4 and nad5 had the most RNA editing sites, each with six sites, while atp9 had no RNA editing sites. This study further revealed that RNA editing events could introduce stop codons in the nad1 gene.

Suppl. material 2: fig. S4B depicts the effects of RNA editing events on amino acids. These findings indicate that 39 amino acids experienced non-synonymous changes, altering the types of encoded amino acids. The most frequent transition was the Ser to phenylalanine (Phe) transition, followed by the conversion of proline (Pro) to Leu and Ser to Leu. A smaller number of amino acids also underwent synonymous substitution, including isoleucine, Phe, Pro and valine, each with one synonymous substitution. The RNA editing sites involved a total of 18 codon modifications, of which 39.5% resulting in amino acid changes with no change in hydrophobicity, 2.3% were predicted to switch from hydrophobic to hydrophilic and 55.9% changed from hydrophilic to hydrophobic (Suppl. material 1: table S9).

In this study, the mitochondrial genome of D. sinensis was sequenced, assembled and analysed. Despite sharing the same genus, the three species all have distinct mitochondrial genome sizes. Fungal mitochondrial genomes range in size from tens to hundreds of kilobases, primarily due to introns, intergenic regions, repeats, horizontal gene transfer and plasmid-derived dynamic change regions (Burger et al. 2003; Wang et al. 2020a, b; Ma et al. 2022). In the Polyporaceae species involved in our study, the largest mitochondrial genome was 3.06 times larger than the smallest. The differences in gene content between the different species indicated that gene gain and loss occur in the mitochondrial genome of fungi during the long-term evolution process, which is very common in the mitochondrial genome. Although the mitochondrial genomes of the three Daedaleopsis species differed in size, their PCGs were similar in type and quantity, consisting primarily of atp6, atp8, atp9, cob, cox1, cox2, cox3, nad1, nad2, nad3, nad4, nad4L, nad5, nad6 and rps3 genes. Variations in gene order amongst different species may account for differences in mitochondrial genome size. In addition, the dpo gene found in the mitochondrial genome of D. sinensis may be involved in the horizontal gene transfer of DNA polymerase genes to the mitochondrial genome (Férandon et al. 2013). The three Daedaleopsis species contain two RNA genes and 26 tRNA genes, but their gene sizes and base composition showed variation. It has been reported that tRNA anticodon arm mutations may also lead to changes in mRNA codon specific recognition and could thus affect protein synthesis (Ding et al. 2019; Lin et al. 2021). As a result, it is necessary to further study the subsequent effects of tRNA mutations.

ATG and TAA are the typical start and stop codons used in the mitochondrial genomes. In the present study, codon usage analysis revealed that the third codon position of the mitochondrial genome of D. sinensis prefers adenine/thymine (A/T), which is more pronounced and prevalent in higher plants (Yang et al. 2021). The use of codons also plays an important role in gene expression levels and protein structure (Liu et al. 2021). The influence of codon bias on protein structure is reflected primarily in the stage from mRNA translation to protein. The encoding genes of proteins with different tertiary structures often affect the translation extension speed of different regions and regulate the translation efficiency and accuracy by using different codon biases, thereby affecting the folding of new-born peptide chains in the translation process, as well as the spatial conformation of the translated protein (Liu et al. 2021; Moss et al. 2024). Codon preference analysis further helps in understanding the molecular mechanism of gene transcription and translation, determining the optimal codon and predicting some genes with unknown functions (Jiang et al. 2008; Liang 2010).

The evolutionary rate of genes in the mitochondrial genome is generally regulated by purifying selection, mutation and directed selection. The ratio of the non-synonymous replacement rate to the synonymous replacement rate is also an important index to determine the evolutionary selection pressure on PCGs in the mitochondrial genome (Hurst and Jiggins 2005). The gene with the highest Ka/Ks ratio is the nad6 gene, while the Ka and Ks values of the nad6 gene are close to each other, indicating that the nad6 gene is primarily subjected to harmless neutral selection during genetic evolution and that it is under weak natural selection pressure. The Ka/Ks ratio of the atp6, atp9, cox1, cox3, nad1, nad4L and nad5 genes is relatively low, indicating that these PCGs are under strong natural selection pressure and may be bound by the functions of the proteins encoded by the genes to ensure the normal biological function of the proteins. These findings also indicate that these genes play important roles in the survival and evolution of species. Moreover, the evolution rate of these genes is slow, while the level of genetic variation is low, making them suitable for the study of the phylogenetic evolution of Daedaleopsis species, as well as the identification of molecular markers for the construction of biological barcodes between different species. Fungal mitochondrial genomes frequently include SSRs, tandem repeats and interspersed repeats. During evolution of the mitochondrial genome, recombination may have caused these abundant repeats to alter the size of the mitochondrial genome (Burger et al. 2003). The high A/T composition of SSRs may further contribute to the abundance of AT in the mitochondrial genome of D. sinensis. The interspersed repeats had a total length of 3,193 bp, accounting for 4.62% (3,193/69,155) of the mitochondrial genome length. The longest palindromic repeat (P) spanned 251 bp, while the longest forward repeat (F) extended to 159 bp. These repeats may also influence the size of the mitochondrial genome, as well as demonstrate the frequency of intermolecular recombination in the mitochondrial genome (Wynn and Christensen 2019).

Repeats are a part of the gene regulatory network, which can co-regulate gene expression with other signalling molecules or homeopathic expression elements and are rich in genetic information (Lu et al. 2020). Some repeats can also specifically bind to certain proteins, thereby facilitating the assembly of nucleic acids into various higher-level structures (Hoskins et al. 2002). As such, repeats greatly expand and enrich genetic information in the process of species evolution and generate the impetus of evolution (Eichler and Sankoff 2003; Lu et al. 2021). Understanding and studying these repeats will contribute to our knowledge of DNA maintenance and evolution in mitochondria (Wynn and Christensen 2019).

Introns are commonly detected in fungal mitochondrial genomes, while the size and gain/loss of introns may affect the organisation and size of fungal mitochondrial genomes (Sandor et al. 2018). Research has also been shown that introns are the primary factors affecting the size variation of fungal mitochondrial genomes (Friedrich et al. 2012). Even within the single fungal genus, the number and species of introns may vary (Jung et al. 2010; Joardar et al. 2012; Beaudet et al. 2013). Many fungi contain both group I and II introns, but group I introns are more common in fungi than II (Lang et al. 2007). In Basidiomycota, most introns belong to group I, while the cox1 gene is accepted to be the largest host gene of group I introns (Wang et al. 2020b). Some homologous introns were simultaneously detected in species with relatively distant phylogenetic relationships, which may indicate a potential gene transfer event. Simultaneously, the differences in introns between species in the same family or genus may indicate the variability and dynamics of intron distribution amongst species (Li et al. 2022). Further sequencing and analysis will be needed to reveal the origin of these introns, including transfer events and evolutionary mechanism.

The 43 RNA editing sites identified were distributed across 14 PCGs in the mitochondrial genome of D. sinensis. RNA editing events cause amino acid composition variations, resulting in differences in encoded information. RNA editing is essential for the synthesis of functional proteins in the mitochondrial system. These functional proteins exhibit closer sequence conservation after editing with homologues in other systems (Brennicke et al. 1999). Further, RNA editing events in D. sinensis were predicted. RNA editing may also alter the codon initiation and termination of PCGs. However, the initiation and termination codons triggered by RNA editing may result in more conserved proteins, thereby regulating mitochondrial gene expression (Galtier 2011). RNA editing also altered 55.9% of the amino acids in D. sinensis from hydrophilic to hydrophobic. This change in hydrophobicity has previously been shown to be closely related to protein conformation and function (Brenner et al. 2019). The hydrophobicity of amino acids reflects protein folding, while their hydrophilic/hydrophobic interactions are amongst the most important forces for maintaining protein tertiary structure (Lacey et al. 1992). The prediction of RNA editing events facilitates the study of mitochondrial gene expression mechanisms and is a valuable reference to predict gene function (Li et al. 2014; Hao et al. 2021).

Homology analysis is another analytical approach crucial for elucidating species evolution (Bonierbale et al. 1988; Tanksley et al. 1988; Lagercrantz et al. 1996; Wang et al. 2015). The present study explored the homologous collinear regions between D. sinensis and other eight fungi, while the more heterogeneous collinear maps revealed the presence of gene rearrangements in the mitochondrial genome of D. sinensis, which could function as potential drivers of D. sinensis evolution (Yang et al. 2016; Ye et al. 2018). Mitochondrial genome rearrangement is common in plants, animals and fungi, while the analysis of variations in mitochondrial gene sequences can reveal the phylogenetic status of a species and its phylogenetic relationships (Sankoff et al. 1992; Boore 1999; Li et al. 2018, 2019a). Our study revealed widespread gene inversion, insertion and migration amongst the three Daedaleopsis species, indicating that mitochondrial gene organisation is highly variable. Herein, several repeats were simultaneously discovered in the mitochondrial genome of D. sinensis. According to previous studies, the accumulation of repeats is one of the primary causes of fungal mitochondrial gene rearrangement (Li et al. 2020; Chen et al. 2021). Simultaneously, the arrangement of mitochondrial genes can provide important information regarding genetic variation and phylogenetic relationships between species (Zheng et al. 2018; Li et al. 2021). A large number of studies have shown that a more complex rearrangement mechanism may exist in fungal mitochondrial genomes than in animals (Song et al. 2024). The gene arrangement of 23 species in the family Polyporaceae also shows its complexity.

Herein, a phylogenetic tree involving 87 Basidiomycota species was constructed using both BI and ML analyses of the concatenated conserved mitochondrial PCG genes (Fig. 8). Four species, namely, D. confragosa, D. sinensis, D. nitida and F. fomentarius, were grouped in a closely-related cluster with high support values. Moreover, D. sinensis shares a tighter evolutionary link with D. confragosa than other relatives. Compared with phylogenetic trees constructed from whole genomes, phylogenetic trees constructed from mitochondrial PCGs have certain similarities (Ma et al. 2024). However, the mitochondrial phylogenetic tree had a relatively low support for several other branches. Hence, the evolutionary lineages on the basis of mitochondrial genomes may not adequately reflect actual evolutionary links. Mitochondria, which are semi-autonomous organelles with their own genetic expression system, may evolve in divergent directions from the nuclear genome (Yin et al. 2015; Liu et al. 2020, 2022). Mitochondrial DNA, which evolves at a faster rate than nuclear DNA, is easy to analyse and can be used to assess the interrelationships between different genera or species of fungi and between different strains of the same species (Ye et al. 2018).

The authors have declared that no competing interests exist.

No ethical statement was reported.

This study was supported by the National Natural Science Foundation of China (32070016, 32270016 and 32325001) and the Beijing Nova Program (20230484322).

JXM: Data curation, Formal analysis, Investigation, Methodology, Software, Visualisation, Writing-original draft, Writing-review and editing. HJL: Formal analysis, Methodology, Resources, Software, Visualisation, Writing-original draft. CJ: Formal analysis, Methodology, Software, Visualisation, Writing-original draft. HW: Formal analysis, Methodology, Writing-original draft. LXT: Formal analysis, Methodology, Writing-original draft. JS: Conceptualisation, Data curation, Formal analysis, Funding acquisition, Investigation, Project administration, Resources, Supervision, Validation, Writing-original draft, Writing-review and editing. BKC: Conceptualisation, Data curation, Funding acquisition, Project administration, Validation, Writing-review and editing.

Jin-Xin Ma https://orcid.org/0000-0003-2069-7493

Hai-Jiao Li https://orcid.org/0000-0002-7227-4479

Can Jin https://orcid.org/0000-0003-4573-6911

Hao Wang https://orcid.org/0000-0002-4880-7001

Lu-Xin Tang https://orcid.org/0000-0002-5753-6354

Jing Si https://orcid.org/0000-0001-9229-0727

Bao-Kai Cui https://orcid.org/0000-0003-3059-9344

The complete mitochondrial genome of D. sinensis has been deposited in the GenBank database under the accession number NC_085747. All data generated or analysed during this study are included in this article [and its Suppl. materials 1, 2].