Background

Due to the complicated physical structure, sequence composition, and abundance of repeat sequences, there are still certain challenges associated with the assembly and annotation of vascular plant mitochondrial genomes [1,2,3,4,5]. An increasing number of plant mitochondrial genomes are being reported as a result of the ongoing advancements in sequencing technology [5,6,7,8]. However, gymnosperms only make up a small portion of the published plant mitogenomes; the majority come from angiosperms. Only 33 gymnosperm mitogenomes (including draft mitogenomes) have been reported to date [4, 8,9,10,11,12,13]. Previous research has demonstrated that the gymnosperm mitogenome has undergone substantial changes in genomic size, physical structure, and gene composition [4, 8,9,10,11,12,13]. For example, the genome size of Larix sibirica Ledeb. was 33 times larger than that of Ginkgo biloba L. (G. biloba) [11, 12]. The related groups, Pinus taeda L. (Pi. taeda) and Welwitschia mirabilis Hook. f. (W. mirabilis), differ by 12 protein-coding genes (PCGs, 41 vs. 29) [11].

Gymnosperms can be classified into five main lineages, including Cycads, Ginkgo, Pinaceae, Gnetophytes, and Conifer II (non-Pinaceae conifers or Cupressophyta), according to recent studies [14]. Cupressaceae, as the most widely distributed and the second largest family among gymnosperms (the number of species is only second to Pinaceae), is an essential component of Conifer II. However, Cupressaceae has only reported six assembled mitogenomes to date, including five draft mitogenomes and one complete mitogenome [9, 10, 13]. This may greatly hinder a deep understanding of the evolution of mitogenomes in this large family, even gymnosperms.

Thuja sutchuenensis Franch. (Th. sutchuenensis), belonging to an astern Asian-North American disjunct genus, was first discovered by French missionary P. G. Farges in 1892 and officially published by Franche in 1899 [15, 16]. Since Farges collected it in 1900, this species has not been found in the wild for over 90 years [16]. As a result, it was listed as being "extinct in the wild" in some monographs and red lists [17,18,19]. Th. sutchuenensis was not rediscovered until 1999, during a comprehensive investigation of rare and endangered plants in Chengkou County [16]. However, recent studies have indicated that both the existing population and wild seedlings of Th. sutchuenensis were rare, and reproductive barriers were the main reason for its population decline [20,21,22]. Therefore, Th. sutchuenensis is now assessed as an endangered species and the I-class national key protected wild plant in China. Previous research has demonstrated that there is a close relationship between the mitogenome and cytoplasmic male sterility (CMS), and abortion occurred during the growth of ovulate strobilus, microstrobilus and seeds of Th. sutchuenensis [22,23,24]. Hence, studying the mitogenome of Th. sutchuenensis is helpful to understand the molecular mechanism of the reproductive barriers, and is of great significance for the protection of this endangered plant.

In this study, the complete mitogenome of Th. sutchuenensis was sequenced, assembled and validated. Then, we compared it with multiple reported mitogenomes of gymnosperms. Our research provides more evidence that the plant mitochondrial genome contains several sequence components with complex structures. The investigation of the Th. sutchuenensis' mitogenome also provides supporting data for the study of the reproductive obstacles of this endangered species and the genetic evolution of the gymnosperm mitogenome.

Results

Assembly, annotation and validation of Th. sutchuenensis mitochondrial genome

Assembly results indicated that the complete mitochondrial genome of Th. sutchuenensis is 2.46 Mb, including 4 contigs. Bandage [25] was used to visualize the complete mitogenome, which is composed of two components: a larger 2.2-Mb component (including 3 contigs, labelled 1–3) and a smaller 251-kb circular contig (labelled 4) (Fig. 1). Contigs 1, 2 and 3 had overlapping regions with each other, with lengths of 1,390,975 bp, 519,836 bp and 293,570 bp, respectively. These 3 contigs formed a multibranched structure and were treated as putative linear molecules here. The shortest segment, contig 4, was putatively circular with a length of 251,475 bp.

Fig. 1
figure 1

The assembly graph of the mitogenome of Thuja sutchuenensis

Full size image

To validate the accuracy of the assembly conformation of Th. sutchuenensis, PCR and Sanger sequencing were carried out. We designed four pairs of specific primers (AF + AR, BF + BR, CF + CR and DF + DR) for PCR amplification. The two connected regions of contig 1 and contig 3 should successfully generate PCR products by primer pair AF + AR and BF + BR, respectively, while the connected regions of contig 1 and contig 2, contig 2 and contig 3 should successfully generate PCR products by primer pairs CF + CR and DF + DR, respectively. The PCR amplification results showed that the lengths of bands were consistent with that expected (Additional file 1: Fig. S1), and the Sanger sequencing results confirmed the presence of this complex connected structure of the assembly conformation (Additional file 1: Fig. S2).

According to the annotation results, the mitogenome of Th. sutchuenensis had 32 different PCGs with known functions, including 24 core mitochondrial genes and 8 variable genes. The core genes were composed of nine NADH dehydrogenase genes, five ATP synthase genes, four ubiquinol cytochrome c reductase genes, three cytochrome c oxidase genes, 1 transport membrane protein gene, 1 maturases gene and 1 cytochrome c biogenesis gene. The variable genes included five small subunits of ribosome genes (rps-), two large subunits of ribosome genes (rpl-) and one succinate dehydrogenase gene (Table 1). Fourteen introns were discovered in 5 of the 32 PCGs (nad5 had four introns; nad 1, nad 2 and nad4 had three introns; and cox2 had one intron). In addition, three rRNA genes, five tRNA genes of mitochondrial origin (trnD-GUC, trnE-UUC, trnI-CAU, trnfM-CAU, trnY-GUA), and one tRNA gene of chloroplast origin (trnW-CCA) were also annotated in the mitogenome. The relative order and direction of these genes are shown in the mitogenome map (Fig. 2).

Table 1 Gene composition in the mitogenome of Thuja sutchuenensis
Full size table
Fig. 2
figure 2

Gene map of the complete mitogenome of Thuja sutchuenensis. For contigs 1, 2, and 3, genes above and below the horizontal line are transcribed in clockwise and counterclockwise directions, respectively. For contig 4, genes inside and outside the outer circle are transcribed in clockwise and counterclockwise directions, respectively

Full size image

Repeat sequences in the mitogenomes of gymnosperms

The analysis of repeat sequences indicated that 894, 293, 202 and 127 simple sequence repeats (SSRs) were found in contigs1-4 of Th. sutchuenensis, respectively, and the tetranucleotide polymers accounted for the largest proportion in each contig (Additional file 1: Fig. S3, Additional file 2: Table S2.1-S2.4). In addition, more than 2,000 dispersed repeats with a length greater than or equal to 30 bp were detected in each contig (Additional file 2: Table S2.5-S2.8). Among these dispersed repeats, only three pairs of complementary repeats were observed in contig 2 and contig 4, respectively, but not in contig 1 and contig 3 (Additional file 1: Fig. S4). The analysis also identified abundant tandem repeats in these 4 contigs, with 769, 270, 181, and 116 tandem repeats found in contigs 1–4, respectively (Additional file 2: Table S2.9-S2.12). In total, 1,516 SSRs, 15,317 pairs of dispersed repeats, and 1,336 tandem repeats were found in the mitogenome of Th. sutchuenensis. Different types of repeat sequences are shown in Fig. 3 in a circle diagram. The analysis of long repeats (> 500 bp) showed that 31 long repeats were found in the whole mitogenome of Th. sutchuenensis, which were main located among contig1, contig 3, and contig 4 (Additional file 1: Fig. S5; Additional file 2: Table S2.13).

Fig. 3
figure 3

The distribution of repeats in the mitogenome of Thuja sutchuenensis. The inner circle shows the dispersed repeats with a length greater than or equal to 50 bp, in which blue represents forward repeats, purple represents palindromic repeats, orange represents reverse repeats. The two outer circles show tandem repeats and simple sequence repeats as short bars, respectively. A, B, C and D represent the distribution of repeat sequences of contig 1–4, respectively

Full size image

We also identified the repeat sequences of mitogenomes of six other representative gymnosperms and compared them with Th. sutchuenensis. The results indicated that all species had the highest proportion of tetranucleotide polymers except for W. mirabilis and Taxus cuspidata Sieb. et Zucc. (Ta. cuspidata), which were dominated by pentanucleotide polymers and dinucleotide polymers, respectively (Additional file 1: Fig. S6A, Additional file 2: Table S2.14-S2.20). The total number of SSRs indicated that Cycas taitungensis C. F. Shen et al. (C. taitungensis) and G. biloba had the fewest SSRs (both less than 200), Pi. taeda and W. mirabilis had a moderate number of SSRs (500–700), while Ta. cuspidata, Platycladus orientalis (L.) Franco (Pl. orientalis) and Th. sutchuenensis harboured the most SSRs (both more than 1,000). Comparative analysis of dispersed repeats indicated that forward repeats and palindromic repeats accounted for the largest proportion in gymnosperms (Additional file 1: Fig. S6B, Additional file 2: Table S2.5-S2.8, S2.21-S2.27). Among the 7 gymnosperms, Pl. orientalis and Th. sutchuenensis harboured the largest number of extra-long repeats (more than 5,000 bp), far more than the other five gymnosperms (Additional file 1: Fig. S7; Additional file 2: Table S2.13, S2.21-S2.25, S2.28). And although the mitogenome of Th. sutchuenensis was slightly smaller than that of Pl. orientalis, it possessed the longest repeat sequence identified in seven gymnosperms, which was up to 47,016 bp long.

Identification of mitochondrial plastid DNAs (MTPTs) and synteny analysis

The blast analysis for the mitogenome and plastid genome indicated that 17 homologous fragments were identified, with a total length of 3,481 bp, accounting for 0.14% of the total mitogenome (Additional file 1: Fig. S8, Table 2). Among those fragments, the longest was 660 bp, and the shortest was 63 bp. Annotation analysis for these homologous fragments indicated that only two incomplete rRNA genes (rrn18 and rrn26) were found, which appeared in the mitochondrial and plastid genomes simultaneously (rrn18/rrn16 and rrn26/rrn23). Therefore, this could not represent the existence of sequence migration between the two genomes, as these rRNA genes may be of a common origin. In conclusion, no significant sequence migration was detected between the chloroplast and mitochondrial genomes of Th. sutchuenensis.

Table 2 Mitochondrial plastid DNAs (MTPTs) identified in mitogenome of Thuja sutchuenensis
Full size table

Synteny analysis indicated that there were a large number (10,496) of homologous collinear blocks among the seven gymnosperms, especially between Th. sutchuenensis and Pl. orientalis, C. taitungensis and G. biloba (Fig. 4). The analysis also found that collinear blocks with inconsistent arrangement orders accounted for a large proportion (average proportion: 52%), which may indicate that these seven representative gymnosperms have experienced a large number of genome rearrangement events, and the mitogenome structure of gymnosperms was not conserved. Previous studies have identified 29 conserved gene clusters in angiosperms [26]. However, only 3 conserved gene clusters were found in the mitogenome of Th. sutchuenensis, namely, cox3-sdh4, nad3-rps12 and rps3-rpl16 (Additional file 1: Fig. S9). The number and composition of gene clusters varied greatly among these seven species, except for two species of Cupressaceae (Th. sutchuenensis and Pl. orientalis), which had three identical gene clusters.

Fig. 4
figure 4

The collinear blocks among gymnosperm mitogenomes. Bars indicated the mitogenomes, and the ribbons showed the homologous sequences (> 500 bp) between the adjacent species. The red areas and gray areas indicate collinear blocks with inconsistent and consistent arrangement orders, respectively. The four lines of numbers beside the arrow represent the genome size of two species and the respective proportions of all homologous sequences between two species

Full size image

Variation in gene composition in the mitogenomes of gymnosperms

We compared the mitogenomes of Th. sutchuenensis with 6 other representative gymnosperm species to show the different gene compositions among gymnosperm mitogenomes. Among the 7 species representing 5 lineages of gymnosperms, C. taitungensis, G. biloba and Pi. taeda had the highest number of PCGs (41), while the fewest PCGs were found in W. mirabilis (29, Fig. 5A, Additional file 1: Table S3.1). In addition, three species from Conifer II, including Ta. cuspidata, Pl. orientalis and Th. sutchuenensis, harboured identical numbers and compositions of PCGs (32). A comprehensive reannotation of tRNA for the mitogenomes of seven gymnosperms showed that compared with the basal group of gymnosperms (C. taitungensis and G. biloba), numerous loss events of mitochondrial-derived tRNA occurred in the evolutionary history of gymnosperms. Only four tRNAs were found to be conserved in seven gymnosperms, namely, trnD-GUC, trnE-UUC, trnI-CAU and trnY-GUA (Fig. 5B). Additionally, the occurrence of chloroplast-derived tRNAs seemed to be random. Interestingly, we found that trnfM-CAU (mt origin) was conserved in all species except for W. mirabilis, whereas trnfM-CAU (cp origin) was found only in this species.

Fig. 5
figure 5

The composition of protein coding genes (A) and tRNA genes (B) in the 7 gymnosperm mitogenomes

Full size image

Variation in substitution rates of mitochondrial PCGs among gymnosperms

We extracted 28 shared conserved mitochondrial genes from 7 representative gymnosperms and calculated the nucleotide substitution rates at two levels (species and genes). The nucleotide substitution rate at the species level was calculated using a supermatrix concatenating 28 genes. The results showed that most dN/dS values were less than 1.0, with the exception of W. mirabilis vs. Pi. Taeda and Pl. orientalis vs. Th. sutchuenensis, indicating that almost all gymnosperms were negatively selected (Additional file 1: Fig. S10, Additional file 2: Table S2. 29). At the gene level, the paired nucleotide substitution rates of these 28 genes were calculated respectively. The results indicated that most genes had dN/dS values less than 1.0 (Fig. 6, Additional file 2: Table S2. 30), suggesting that they had undergone purifying selection, especially atp1, cox2, nad7 and rps12, which were less than 1.0 in all species. In contrast, a few genes, including atp8, ccmB, mttB and sdh4, had dN/dS values that were greater than 1.0 in most species, suggesting potential positive selection. The boxplot also revealed that atp1, nad4L and nad9 had unusually low dN/dS values, which may indicate that they have undergone intense purification and were well conserved in the mitogenome of gymnosperms (Fig. 7).

Fig. 6
figure 6

Heatmap of pairwise dN/dS ratios among 28 mitochondrial genes in the 7 gymnosperm mitogenomes

Full size image
Fig. 7
figure 7

Boxplots of pairwise dN, dS values and their ratio among 28 mitochondrial genes in the 7 gymnosperm mitogenomes

Full size image

Discussion

The National Center for Biotechnology Information (NCBI) released 7,943 chloroplast genomes and 470 mitochondrial genomes as of July 28, 2022. As an increasing number of mitogenomes are published, studies have found that the size of plant mitogenomes varies greatly, ranging from 66 kb to 11.7 Mb [12, 27]. Additionally, studies have revealed that plant mitogenomes differ from chloroplast genomes (always containing the conserved single loop structure) and should be presented in multiple sequence units to show the dynamic connection of the mitogenome [3, 5, 6, 28, 29]. The complete mitochondrial genome of Th. sutchuenensis was assembled and validated in this study, which consists of one circular contig and three linear contigs with overlapping regions among each other (Fig. 1, 2). The use of long reads is critical to achieving this complete assembly. The multipart mitogenome described in our study was also reported in other gymnosperms, such as Pl. orientalis, Larix sibirica and Picea sitchensis (Bong.) Carrière, which were composed of 2, 9 and 13 segments, respectively [4, 12, 13].

SSRs and dispersed repeats are important components of repeat sequences in the mitochondrial genome [5, 6, 8]. By comparing the repeat sequences of 7 gymnosperm mitogenomes, it is interesting to find that the number of SSRs in gymnosperms form a definite hierarchy and could well reflect the phylogenetic relationship. The number of SSRs could be divided into three groups, namely, C. taitungensis and G. biloba (both less than 200), Pi. taeda and W. mirabilis (500–700), and Ta. cuspidata, Pl. orientalis and Th. Sutchuenensis (more than 1,000), which may indicate that the size of the SSR sequence of gymnosperms has gradually increased during evolution (Additional file 1: Fig. S6A).

Previous studies have indicated that repeat sequences in the mitogenome are crucial for intermolecular recombination, especially long repeat sequences (more than 1 kb), which may cause high-frequency recombination, leading to isomerization of the genome into multiple major forms [11, 29, 30]. Our study revealed that both the number and length of long repeats of Th. sutchuenensis and Pl. orientalis were far more than those of the other 5 gymnosperms, and the longest repeat was also found in Th. sutchuenensis, with a length of more than 47 kb (Additional file 1: Fig. S7). This longest repeat was located in contig 3, whose two ends formed a complex connected structure with contig1 and contig2 (Fig. 1, 3; Additional file 1: Fig. S5). The accuracy of this connected structure has also been confirmed by our experiment. Therefore, it can be inferred that the recombination mediated by these abundant long repeats may help the mitogenome of Th. sutchuenensis to form a more complicated conformation than that of other gymnosperms [29,30,31].

The mitogenome of angiosperms has 29 conserved collinear gene clusters, according to earlier research [26]. Among the angiosperms selected in that study, the number of collinear gene clusters ranged from 7 to 23. However, compared to angiosperms, our study revealed that gymnosperms have a significantly lower number of collinear gene clusters, with G. biloba having the most, with 8 gene clusters, and W. mirabilis having the fewest, with just 2 (Additional file 1: Fig. S9) [8]. Only one collinear gene cluster, nad3-rps12, is shared by all of the selected gymnosperm species. Additionally, comparative research revealed that Th. sutchuenensis and Pl. orientalis have identical compositions of collinear gene clusters. These findings indicated that in gymnosperms, the orders of mitochondrial genes may only be conserved within families but vary greatly in different lineages of gymnosperms, which may be connected to frequent recombination caused by abundant repeat sequences in gymnosperm mitogenomes.

Previous studies on the tRNAs of gymnosperms were mostly based on software prediction, which may overestimate the number of tRNAs. BLAST based on sequences could more accurately reflect the composition of tRNAs. Thus, we integrated a tRNA database from previous studies and used it to identify potential tRNAs for 7 representative gymnosperms [26, 32, 33]. The analysis showed that extensive loss events of tRNA (mitochondrial origin) have occurred in the evolutionary history of gymnosperm mitogenomes, and only a few tRNAs were conserved in the evolutionary process (Fig. 5B). Surprisingly, our results indicated that tRNAs of chloroplast origin may have a potential functional complement. For example, W. mirabilis lost the trnfM-CAU of mitochondrial origin that was present in the other 6 gymnosperms, but it was found to harbour the only trnfM-CAU of chloroplast origin, which may play a complementary role in function.

The dN/dS analysis in our study increased our comprehension of the mitochondrial gene evolution of gymnosperms. The analysis indicated that most genes of these 7 gymnosperms had dN/dS values that were less than 1.0, which may imply that they were conservative and underwent purifying selection (Fig. 6). This inference was consistent with the conclusions of other studies on angiosperms [6, 34]. However, we also identified that four genes, atp8, ccmB, mttB and sdh4, may have undergone positive selection in the evolution of gymnosperms in most species. These four genes are involved in the biosynthesis of ATP synthase, ubiquinol cytochrome c reductase, transport membrane protein and succinate dehydrogenase, respectively [5, 7]. Based on previous studies, it is speculated that these four genes may have undergone positive selection under environmental stress to generate new functions to adapt to the new environment.

Conclusion

In this study, we successfully assembled the mitochondrial genome of Th. sutchuenensis by using Nanopore and DNBSEQ reads, which is the second most complete mitogenome of Cupressaceae. The complex conformation of the mitogenome of Th. sutchuenensis, whose presence was validated by experiments and sequencing, was composed of three linear contigs and one small circular contig. The results of this study further proved the authenticity of the mitogenome's complex structure in seed plants. By comparing the mitogenomes of five different lineages of gymnosperms, we discovered that the number of SSRs may have increased during the evolutionary history of gymnosperms and that these abundant long repeats may contribute to the generation of the complex conformation of the mitogenome of Th. sutchuenensis. The analysis of the collinear block and the gene cluster revealed that the orders of mitochondrial genes of Th. sutchuenensis was not conserved, which indirectly proved that this mitogenome might have undergone a high level of recombination. In addition, the analysis of tRNA genes indicated that only a few tRNAs were conserved in the evolutionary process, and tRNAs of chloroplast origin may have potential complementary functions. The investigation of Th. sutchuenensis mitogenome in this study also offers a theoretical foundation for removing the reproductive barriers faced by this threatened species.

Methods

Sampling and genome sequencing

We collected young needles of Th. sutchuenensis from the ecological park of Southwest University and extracted total genomic DNA using the CTAB method [35]. Voucher specimens for this plant were collected and deposited in the herbarium of Southwest University, Chongqing, China, with accession number: 20210425TS-1. A DNA library with an insert size of 350 bp was constructed using the kit and sequenced using the DNBSEQ sequencing platform. Soapnuke (V 1.6.5) [36] was used to filter the raw data, which could remove low-quality reads and some reads with adapter contamination or PCR duplication. Finally, a total of 40 G of clean data was obtained.

The total genomic DNA of the identical plant sample used for DNBSEQ short-read sequencing was also subsequently used for Oxford Nanopore sequencing, which followed the standard protocol provided by Oxford Nanopore Technologies (ONT) company and included sample quality detection, library construction, library quality detection, library sequencing and other processes. In total, 15 Gb of sequence reads were obtained, and 13.71 Gb remained after filtering and qualification. The average read length of filtered reads was 8.16 kb (N50 = 17.69 kb), and the longest read was 733.08 kb.

Mitogenome assembly and annotation

We first used Flye (version 2.9-b1774) [37] to assemble the draft mitogenome through Nanopore sequencing data. Here, all the raw data were used as input files, and the nuclear genome, mitochondrial genome and chloroplast genome were assembled. The minimum overlap was set to 2,000, and the other parameters were set to default. For the assembled contigs, we exported them into FASTA format and made a database by using makeblastdb. The conserved mitochondrial genes of Ta. cuspidata (MN593023) and Pl. orientalis (OL703044 and OL703045) were used as the query sequence, and the BLASTn program [38] was used to identify the assembly contigs containing mitochondrial genes. Finally, we imported the GFA format file generated by Flye in Bandage [25] software, the identified mitochondrial contigs were retained and the others were removed. These mitochondrial contigs had overlapping regions with each other, and formed a complex graph, which was considered the draft mitochondrial genome. Subsequently, we mapped these long reads to the draft genome to resolve repeat regions. In short, for each repeated region, the path supported by most long reads was the optimal path. The repeat resolved sequences represent the major conformation of the dynamically changing mitochondrial genome. Finally, four contigs were obtained and considered to be the hypothetical mitogenome of Th. sutchuenensis (Fig. 1). To obtain the final mitogenome sequence, we used both short-read data and long-read data to correct these four contigs with minimap2/miniasm [39, 40], racon (v1.4.20) and pilon (v1.23) [41, 42].

Validating the assembled conformation

PCR experiments with specific primer pairs were used to amplify the connected regions of three contigs, which could verify the presence of this complex conformation. Four pairs of specific primers were designed for 4 connecting regions using Primer blast of NCBI, which are listed in Table S3.2 (Additional file 1: Fig. S11, Additional file 3: Table S3.2). PCRs were performed in a 50 μl mixture, including 10 μl buffer, 4 μl dNTPs, 1 μl hyPerFUsion™ High Fidelity DNA Polymerase, 2 μl total DNA, 2 μl each of the forwards and reverse primers, and 29 μl ddH2O. After an initial denaturation at 94 °C for 2 min, PCRs were performed for 35 cycles. Each cycle consisted of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, and elongation at 72 °C for 1 min. The PCR products were sequenced by Sanger sequencing.

Genome annotation

Geseq [43] was used to annotate the mitogenome. For the annotation of transfer RNA genes, we first built a local tRNA database based on many previous studies and then used four contigs as queries to blast potential tRNA genes [26, 32, 33]. After blast, sequences with an identity less than 85% or a length difference (compared with the target sequence) greater than 10 bp were removed, which may be considered incomplete. Second, tRNAscan SE [44] was used to reconfirm the accuracy of the blasting results. Finally, each annotation error of the mitochondrial genome was manually modified and corrected by Apollo [45]. The mitogenome map was drawn with OGDRAW [46]. The annotated mitogenome of Th. sutchuenensis was submitted to NCBI under accession numbers ON603305-ON603308.

Identification of repeat elements

The repeat sequences of each contig were annotated separately. Then, the long repeats (> 500 bp) were annotated for the whole mitogenome of Th. sutchuenensis. The online website MISA (https://webblast.ipk-gatersleben.de/misa/) was used to identify SSRs, with minimum repetition numbers of mono-, di-, tri-, tetra-, penta-, and hexanucleotides was 10, 5, 4, 3, 3, and 3, respectively [47]. Dispersed repeats were identified by using REPuter (https://bibiserv.cebitec.uni-bielefeld.de/reputer), with the hamming distance = 3, maximum computed repeats = 5,000, and minimal repeat size = 30 [48]. We also used Tandem Repeats Finder (https://tandem.bu.edu/trf/trf.html) to identify tandem repeats with default settings [49]. Additionally, six mitochondrial genomes of representative gymnosperms were downloaded from NCBI, including C. taitungensis (AP009381), G. biloba (KM672373), W. mirabilis (KT313400), Pi. taeda (MF991879), Ta. cuspidata (MN593023) and Pl. orientalis (OL703044, OL703045). These six mitogenomes, together with Th. sutchuenensis, were analysed for repeat sequences.

Identification of MTPTs and synteny analysis

To identify the potential homologous sequences that may be transferred between the plastid (MH784400.1) and mitogenome of Th. sutchuenensis, we used BLASTn [38] to compare the two organelles, with the following parameter settings: E-value = 1e-5, NumofHits = 50,000, NumofAligns = 25,000. TBtools [50] was used for the visualization of the blast results. The identified homologous sequences were also extracted and then annotated with GeSeq. In addition, MCscan [51] was used to plot multiple synthetic plots of Th. sutchuenensis and six other gymnosperms based on sequence similarity. Moreover, we searched for gene clusters common to gymnosperms by simple visual inspection to evaluate the conservation of gene orders for the above seven gymnosperms.

Estimation of nucleotide substitution rates

We used Phylosuite (V1.2.1) [52] to identify and extract the homologous mitochondrial genes in these six representative gymnosperms plus Th. sutchuenensis. The corresponding nucleotide sequences were aligned and concatenated by using Mafft (v7.450) [53] and Phylosuite, respectively. The yn00 module in PAML (v4.9) [54] was used to estimate the pairwise nucleotide substitution rates, including the nonsynonymous substitution rate (dN), synonymous substitution rate (dS), and ratio of dN to dS. Single-gene matrix and concatenated matrix were used as input files to calculate the nucleotide substitution rates at the gene level and species level, respectively. TBtools and R-package (ggplot) were used to draw boxplots and heatmaps for pairwise dN/dS values [50, 55].