Comparative mitochondrial genome analysis reveals intron dynamics and gene rearrangements in two Trametes species

Trametes species are efficient wood decomposers that are widespread throughout the world. Mitogenomes have been widely used to understand the phylogeny and evolution of fungi. Up to now, two mitogenomes from the Trametes genus have been revealed. In the present study, the complete mitogenomes of two novel Trametes species, Trametes versicolor and T. coccinea, were assembled and compared with other Polyporales mitogenomes. Both species contained circular DNA molecules, with sizes of 67,318 bp and 99,976 bp, respectively. Comparative mitogenomic analysis indicated that the gene number, length and base composition varied between the four Trametes mitogenomes we tested. In addition, all of the core protein coding genes in Trametes species were identified and subjected to purifying selection. The mitogenome of T. coccinea contained the largest number of introns among the four Trametes species tested, and introns were considered the main factors contributing to size variations of Polyporales. Several novel introns were detected in the Trametes species we assembled, and introns identified in Polyporales were found to undergo frequent loss/gain events. Large-scale gene rearrangements were detected between closely related Trametes species, including gene inversions, insertions, and migrations. A well-supported phylogenetic tree for 77 Basidiomycetes was obtained based on the combined mitochondrial gene set using 2 phylogenetic inference methods. The results showed that mitochondrial genes are effective molecular markers for understanding the phylogeny of Basidiomycetes. This study is the first to report the mitogenome rearrangement and intron dynamics of Trametes species, which shed light on the evolution of Trametes and other related species.

Trametes species, belonging to Polyporales, Basidiomycota, are widely distributed throughout the world 1 . Trametes species are often distributed on standing dead or fallen hardwood trees 2 . Like other species from the order Polyporales, Trametes species are efficient wood decomposers that cause white rot in colonized wood. Their efficient lignin decomposition ability makes Trametes important decomposers in ecosystems and plays an important role in the natural cycle of earth forest ecosystems. In addition, Trametes species have also been used in ex situ and in situ biosynthesis and bioremediation studies [3][4][5][6] . A series of lignin-degrading enzyme systems and genes have been found to be closely related to the excellent degradation ability of Trametes 7-9 , including lignin peroxidases, manganese peroxidases, and laccases. To date, only two mitochondrial genomes (mitogenomes) of Trametes species have been reported 10,11 , including T. cingulata and T. hirsute, and the characterizations and differentiation of mitogenomes in Trametes have not been fully analyzed.

Results
Features and PCGs of Trametes mitogenomes. The complete mitogenomes of T. versicolor and T. coccinea were circular with sizes of 67,318 bp and 99,976 bp, respectively (Fig. 1). The GC contents of the T. versicolor and T. coccinea mitogenomes were 25.44% and 24.68%, respectively (Table S1). Both the AT skew and GC skew in the T. versicolor mitogenome were positive. The mitogenome of T. coccinea contained negative AT skews and positive GC skews. We detected an entire set of core PCGs in the two Trametes mitogenomes we assembled, including atp6, atp8, atp9, cob, cox1, cox2, cox3, nad1, nad2, nad3, nad4, nad4L, nad5, nad6, and rps3 (Table S2). In addition to these core PCGs, ten and five free-standing PCGs (non-intronic ORFs) were detected in the T. versicolor and T. coccinea mitogenomes, respectively. Several genes encoding DNA polymerase and RNA polymerase were detected in the two Trametes mitogenomes. In addition, the T. versicolor and T. coccinea mitogenomes contained seven and five PCGs with unknown functions, respectively. We also detected 14 and 38 introns in the T. versicolor and T. coccinea mitogenomes, respectively, which harbored 13 and 34 intronic ORFs. Intronic ORFs in the two Trametes mitogenomes mainly encoded LAGLIDADG endonucleases, GIY-YIG endonucleases, and intron-encoded RNA maturase bI4; however, we found two and three intronic ORFs with unknown functions in the T. versicolor and T. coccinea mitogenomes, respectively. rRNA genes and tRNA genes. Both the T. versicolor and T. coccinea mitogenomes contained two rRNA genes: the small subunit ribosomal RNA (rns) and large subunit ribosomal RNA gene (rnl) (Table S2). We Colored blocks outside each ring indicate that the genes are on the direct strand, while colored blocks within the ring indicates that the genes are located on the reverse strand. www.nature.com/scientificreports/ detected two and eight introns in the rnl genes of T. versicolor and T. coccinea, respectively. The average size of the rnl gene was 3,013 bp, and the average length of rns genes in the two Trametes mitogenomes was 1,517 bp (excluding intron regions). The T. versicolor mitogenome contained longer rnl and shorter rns genes than the T. coccinea mitogenome.
The two Trametes mitogenomes we assembled each contained 27 tRNA genes (Table S2) that encoded 20 standard amino acids and ranged in length from 71 to 88 bp. The trnS gene contained the largest sizes among all tRNAs detected, mainly because of the large extra arm. Both mitogenomes of T. versicolor and T. coccinea contained two tRNAs that code for leucine and arginine with different anticodons, as well as three tRNAs encoding serine with different anticodons. The two mitogenomes also contained three tRNAs encoding methionine with the same anticodons. In addition, the T. versicolor mitogenome contained two trnE codons encoding glutamate, while the mitogenome of T. coccinea contained two trnC codons encoding cysteine.
Repetitive elements in Trametes mitogenomes. We detected 38 and 45 intragenomic duplications in the T. versicolor and T. coccinea mitogenomes, respectively (Table S3), through BlastN searches of the two Trametes mitogenomes against themselves. The sizes of these intragenomic duplications ranged from 33 to 974 bp in the two Trametes mitogenomes and the pair-wise nucleotide similarities of these duplications were between 66.40 and 100%. The largest repeats were found in the coding regions of orf425 and orf424 in the mitogenome of T. coccinea, while the largest repeats in the T. versicolor mitogenome were observed in the intergenic regions between nad5 and orf270, as well as between atp8 and trnR, with each repeating sequence being 240 bp long. Repeat sequences identified by BlastN searches accounted for 4.45% and 8.29% of the T. versicolor and T. coccinea mitogenomes, respectively. We also identified 33 and 7 tandem repeats in the mitogenomes of T. versicolor and T. coccinea, respectively (Table S4), using Tandem Repeat Finder. The longest tandem repeat sequence, which had a size of 121 bp, was observed in the intergenic region between orf110 and orf112 in the mitogenome of T. versicolor. Tandem repeats accounted for 2.39% and 0.27% of the T. versicolor and T. coccinea mitogenomes, respectively.
We also conducted BlastN searches of the two Trametes mitogenomes against their published nuclear genomes to determine whether any gene fragments were naturally transferred between the mitochondrial and nuclear genomes (Table S5). We detected one and four aligned fragments between the mitochondrial and nuclear genomes of T. versicolor and T. coccinea, respectively. These aligned fragments ranged from 45 to 173 bp, with pair-wise nucleotide similarities ranging from 90.75% to 100%. The largest aligned fragment was observed in the intergenic region between trnC and orf242 in the T. coccinea mitogenome. A total of 45 bp and 504 bp aligned fragments were detected in the T. versicolor and T. coccinea mitogenomes, respectively, indicating potential gene segment transfering events between mitochondrial and nuclear genomes may have occurred in the evolution of Trametes.
Mitochondrial gene rearrangement in Trametes species. We compared the arrangements of 15 core PCGs and 2 rRNA genes in the 4 Trametes mitogenomes reported in the present study (Fig. 2). Trametes cingulata and T. hirsuta were found to have an identical gene arrangement. We detected large-scale gene rearrangements in mitogenomes of T. versicolor and T. coccinea compared with the identical gene arrangement in T. cingulata and T. hirsute. These gene rearrangements included gene migrations, inversions, and insertions.
Mauve 25 revealed the presence of 10 homologous regions between the 4 Trametes species (Fig. 3), with different Trametes species found to contain various types and numbers of homologous regions. The T. hirsuta and T. versicolor mitogenomes contained additional homologous regions I and J, which involved the intergenic regions between neighboring genes atp9 and orf576 and between atp8 and trnE, respectively. Collinearity analysis  www.nature.com/scientificreports/ indicated that the mitogenomes of T. cingulata and T. hirsuta were highly collinear, and that the T. versicolor and T. coccinea mitogenomes rearranged in homologous regions C, F, and D. The results indicated that the mitogenomes of Trametes underwent gene rearrangements in evolution.
Variation, genetic distance, and evolutionary rates of core PCGs. Seven of the 15 core PCGs detected (cox2, nad2, nad3, nad4, nad5, nad6, and rps3) varied in length between the 4 Trametes mitogenomes tested (Fig. 4). The nad5 gene had the largest length variations between Trametes species, while the largest nad5 gene was observed in T. hirsute. All of the 15 core PCGs except atp6, cox2, and nad4L had GC contents that varied between Trametes species. Among the 15 core PCGs detected, the atp9 gene contained the highest GC content and the atp8 gene contained the lowest. GC skews of the 15 core PCGs varied, with most showing positive GC skews. The largest K2P genetic distance between Trametes species was observed in the rps3 gene, followed by nad3, indicating the two genes differentiated greatly during evolution (Fig. 5). The atp8 and nad4L genes exhibited the lowest mean K2P genetic distance between the four Trametes mitogenomes, indicating that they were highly conserved. The rps3 gene exhibited the highest mean non-synonymous substitution (Ka) rate, whereas the nad4L gene had the lowest Ka mean value among the 15 core PCGs detected. We also found that the nad3 gene had the highest mean synonymous substitution rate (Ks), whereas the atp8 gene exhibited the lowest Ks mean value among the 15 core PCGs detected. The Ka/Ks values for all core PCGs were less than 1, indicating that the core genes in Trametes species were subjected to pressure of purifying selection.
Intron dynamics of cox1 gene in Polyporales species. Pearson correlation analysis showed that there was a high correlation coefficient between the number of introns and the size of the mitogenome in the order Polyporales (Fig. 6). The dynamics of introns in Polyporales had a significant effect on Polyporales mitogenome size. The cox1 gene was found to be the largest host gene of introns in Basidiomycota 26,27 . In the present study, introns of Polyporales cox1 genes were classified into different position classes (Pcls) using the cox1 gene of Ganoderma calidophilum as a reference 22 , and introns belonging to the same Pcls were considered homologous introns. Overall, 153 introns were detected in the 14 cox1 genes of the order Polyporales, and the number of introns in each cox1 gene varied between 6 and 15, with an average of 11. Five of the 153 introns belonged to the group II, and the other introns belonged to the group I.
Overall, 32 Pcls were detected in the 14 Polyporales species (Fig. 7), with P706, P807, and P1305 being the most common intron types, occurring in 11 of the 14 Polyporales species. These commonly distributed introns may have been obtained from the common ancestor of Polyporales species. Seven of the 32 Pcls (P218, P309, P480, P726, P894, P941, and P1114) were only detected in one of the 14 Polyporales species. The rare introns in Polyporales, such as the P480 and P941, were also detected in distantly related species, such as Armillaria sinapina 28 and Agaricus bisporus 29 from the order Agaricales. These results indicated potential horizontal gene transfer events occurred during evolution. However, P218, P309, P276, P894, and P1114 were only distributed in Polyporales, and no homologous introns were found in other Basidiomycota species; accordingly, the origins of these rare introns need to be further investigated. P1107 was the most commonly distributed intron type in the four Trametes species we tested and was present in all four Trametes species. We detected two novel introns in our newly sequenced T. coccinea species, indicating diverse intron types in Trametes species. The varied Comparative mitogenome analysis and phylogenetic analysis. The 99.976 kb mitogenome of T.
coccinea was the largest mitogenome among the four Trametes species reported 10,11 (Table S1). When compared with other Polyporales mitogenomes reported, the Trametes mitogenomes had medium sizes. The GC contents of the Polyporales species varied between 24.0 and 36.3%, with an average of 26.8%. Most Polyporales we detected contained positive AT and GC skews, indicating preferences for A and G bases in the leading strand of Polyporales mitogenomes. All Polyporales mitogenomes tested contained an entire set of core PCGs and a variety of free-standing ORFs. The mitogenome of T. coccinea contained the most introns among the 14 Polyporales species tested. Additionally, the highest number of intronic ORFs was detected in the mitogenome of Phlebia radiata 30 , which contributed to the P. radiata mitogenome becoming the largest mitogenome among the 14 Polyporales species we tested. All 14 Polyporales mitogenomes contained two rRNA genes, and the number of tRNA genes in the 14 Polyporales species ranged from 25 to 29. An identical and well-supported phylogenetic tree was obtained using the maximum likelihood (ML) and Bayesian inference (BI) methods based on the combined mitochondrial gene set (15 core PCGs) (Fig. 8). All major clades within the phylogenetic tree were found to have good support values (BS ≥ 99; BPP ≥ 0.96). Overall, 77 Basidiomycota species were included in the phylogenetic analysis, and these were divided into 13 major clades corresponding to the orders Agaricales, Boletales, Russulales, Polyporales, Hymenochaetales, Gomphales, Cantharellales, Pucciniales, Tremellales, Trichosporonales, Microbotryales, Sporidiobolales, Microstromatales, Ustilaginales, and Tilletiales (Table S6). T. coccinea was a sister species to T. cingulata and T. versicolor was closely related to T. hirsute. www.nature.com/scientificreports/

Discussion
Size and content variations between Trametes mitogenomes. In this study, two mitogenomes of the genus Trametes, T. coccinea and T. versicolor, were newly assembled. Comparative mitogenomic analysis showed that the novel mitogenome of T. coccinea was largest among the four Trametes mitogenomes reported 10,11 . Introns were considered the main factors contributing to the size expansion of the T. coccinea mitogenome, and   www.nature.com/scientificreports/ the T. coccinea mitogenome was found to have the most introns among Trametes species. We also found a high correlation coefficient between the mitogenome size and the number of introns in Polyporales 22,30-35 . These findings further suggested that the dynamic change in introns is one of the main factors leading to the size variation of mitogenomes from the order Polyporales and other fungal species, which is consistent with the results of previous studies 29, 36,37 . We also found that the mitogenome content of Trametes species varied. The four Trametes mitogenomes varied in AT and GC skews. According to the second parity rule, as long as there is no mutation or selection bias, each base in the complementary DNA strand exists at approximately equal frequencies 38 . The presence of AT or GC skews on the same DNA strand from different species indicated that mitogenomes of different Trametes species underwent different mutations or environmental selection 39 . In the present study, the core PCGs of Trametes species were found to vary frequently in length and base composition. Different core PCGs showed varied evolution rates, and some core PCGs were relatively conserved between Trametes species, whereas others showed large differentiation. All core PCGs were found to have been subjected to the pressure of purifing selection. In addition to these core PCGs, some non-conserved PCGs were also detected in Trametes species, including intronic ORFs, DNA polymerases, and RNA polymerases. Introns encoding endonucleases mainly mediate intron homing in Trametes mitogenomes 40 , while DNA and RNA polymerases are thought to be obtained from plasmids [41][42][43] . Several PCGs with unknown functions were also detected in Trametes species, indicating that there are still genes with unknown functions in Trametes mitogenomes that need to be revealed.

Mitochondrial gene rearrangement in Trametes species. Mitochondrial gene rearrangement was
frequently detected in mitogenomes of animals, plants, and fungi 18,44,45 . Variations in the mitochondrial gene order can be used to infer the phylogenetic status and phylogenetic relationship of eukaryotic species 46,47 . The rearrangement of the animal mitogenome has been widely studied, and several models have been proposed to reveal the rearrangement of animal mitogenomes 48,49 . When compared with the mitogenomes of animals, the mitogenomes of fungi show greater variation in gene order 50 . In the present study, we detected large-scale gene rearrangements between Trametes species, including gene inversions, insertions, and migrations. Gene rearrangement events have even been observed between closely related Trametes species within the same clades, indicating that the mitochondrial gene arrangement of Trametes species is highly variable. However, the mechanism of mitochondrial gene rearrangement in fungi has not been elucidated to date. Previous studies showed that the accumulation of repetitive sequences was the main factor contributing to mitochondrial gene rearrangement in fungi 50 . We found a large number of repeat sequences in Trametes species, which may result in gene recombination and rearrangement in these species. The introns belonging to the same Pcl were considered homologous 29 . We found that the number and class of introns varied greatly between different Polyporales species, indicating that the loss/gain of introns occurred in the evolution of Polyporales. Several Pcls from common ancestors of Polyporales, such as P706, P807, and  www.nature.com/scientificreports/ P1305, were found to be widely distributed in Polyporales species. However, several introns were found only in one of the 14 Polyporales species, such as P480 and P941, whereas homologous introns were detected in distant species 28,29 , indicating potential intron transfer events. We also found some novel introns in Polyporales that have never been detected in other Basidiomycetes, including P218, P309, P726, P894, and P1114. The origin, evolution, and function of these novel introns need to be further analyzed.

Phylogeny of Basidiomycota species based on mitochondrial genes.
Owing to the limited morphological characteristics, it is difficult to classify and identify Basidiomycetes accurately only by morphology.
To date, only rRNA internal transcribed spacers (ITS) 57 , RNA polymerase II subunits (RPB), and elongation factor 1-ɑ (EF1ɑ) genes 58 have been widely used to evaluate the phylogeny and reconstruct the early evolution of fungi. Mitochondrial genes have been widely used as molecular markers to analyze phylogenetic relationships of animals because of their unique advantages 16,44 . When compared with the mitogenomes of animals, the mitogenomes of fungi have been less studied, especially those of Basidiomycetes 59 . Currently, less than 130 complete mitogenomes of Basidiomycetes are available in the NCBI database (https ://www.ncbi.nlm.nih.gov/genom e/brows e#!/overv iew/), which has limited comprehensive analysis of the origin and evolution of fungi. In this study, 77 Basidiomycete species were included in the phylogenic study. We have also obtained a phylogenetic tree with high support values based on the combined mitochondrial gene set, indicating that mitochondrial genes are a powerful tool for analyzing phylogenetic relationships of Basidiomycota species. With the rapid development of second and third generation sequencing technology, additional mitogenomes of fungi are needed to understand the origin and evolution of Basidiomycetes and other fungal lineages.

Materials and methods
Assembly and annotation of Trametes mitogenomes. The raw sequencing data used for T. versicolor and T. coccinea mitogenomes assembly were downloaded from the Sequence Read Archive (SRA) database under the accession numbers SRR3927404 and SRR1588030, respectively. We conducted a series of quality control steps to generate clean reads from the raw sequencing reads, including removing adapter reads using Adap-terRemoval v2 60 and filtering low-quality sequences using ngsShoRT 2.2 61 . We used the clean reads to assemble mitogenomes of T. versicolor and T. coccinea using SPAdes 3.9.0 62 . Several contigs were obtained using SPAdes 3.9.0 62 , and gaps between the contigs were filled using MITObim V1.9 63 . NOVOPlasty 64 was also used to assemble or test mitogenomes of the two Trametes species. After these steps, we obtained the complete mitogenomes of T. versicolor and T. coccinea, and then annotated them according to previously described methods 33 . We initially annotated the protein-coding genes (PCGs), open reading frames (ORFs), introns, rRNA genes, and tRNA genes in the two Trametes mitogenomes using MFannot 65

Repetitive element analysis.
We conducted BlastN searches of the two Trametes mitogenomes against themselves 76 to determine whether there were intragenomic duplications of large fragments or interspersed repeats in the two Trametes mitogenomes. We also detected tandem repeats (> 10 bp in length) in the two Trametes mitogenomes using a Tandem Repeats Finder 77 with the default parameters. In addition, BlastN searches of the two Trametes mitogenomes were conducted against their published nuclear genomes to determine whether there were gene segments that naturally transferred between mitochondrial and nuclear genomes 78,79 .
Comparative mitogenomic analysis and intron analysis. The genome sizes, GC contents, base compositions, start and stop codons, gene numbers, and intron numbers of Polyporales mitogenomes were compared to identify variations and similarities between them. Introns of cox1 genes in the 15 Polyporales mitogenomes that have been published to date were classified into different position classes (Pcls) according to the method described by Ferandon et al. 80 . The cox1 genes of 15 Polyporales mitogenomes were first aligned with the cox1 gene of the medical fungus Ganoderma calidophilum 22 using Clustal W 81 . Each Pcl was composed of introns inserted at the same position of the cox1 reference gene based on the insert sites (nt) in the corresponding reference gene. Identical Pcls were considered orthologous introns with high sequence similarity.

Phylogenetic analysis.
A phylogenetic tree of 77 Basidiomycota species was constructed to investigate the status of Trametes species in the phylum Basidiomycota using previously described methods 33,82 based on the combined mitochondrial gene set (15 core PCGs). We used Annulohypoxylon stygium from the phylum Ascomycota as the outgroup 83 . Individual mitochondrial genes were first aligned using the MAFFT v7.037 software 84  www.nature.com/scientificreports/ after which we concatenated the aligned mitochondrial genes into a combined mitochondrial gene set using the SequenceMatrix v1.7.8 software 85 . A preliminary partition homogeneity test was then conducted to detect potential phylogenetic conflicts between different genes. Best-fit models of partitioning schemes and evolution for the combined mitochondrial gene set were determined based on the PartitionFinder 2.1.1 software 86 . We constructed phylogenetic trees using both Bayesian inference (BI) and maximum likelihood (ML) methods, with MrBayes v3.2.6 87 used for the BI analysis and RAxML v 8.0.0 88 for the ML analysis.