The sample of F. aubertii was collected from the campus of Qinghai Minzu University in Qinghai, China (36.59 ${}^{\circ }$N, 101.82 ${}^{\circ }$E) and identified by Professor Yong-chang Lu. The fresh young leaves were dried in silica gel and stored at –20 °C. Genomic DNA was extracted from the whole sample of F. aubertii using the modified CTAB method and was subsequently evaluated for purity and integrity using agarose gel electrophoresis and a NanoDrop 2000c spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The classified DNA samples were sent to Shanghai Origingene Biotechnology Pharmaceutical Technology Co., Ltd., where the genome DNA was sequenced using the Illumina NovaSeq 6000 platform (BIOZERON Co., Ltd., Shanghai, China) and Nanopore platform (Oxford Nanopore Technologies, Oxford, UK).

We used GetOrganelle v1.7.1 (Max Planck Institute of Molecular Plant Physiology, Munich, Germany) [21] to perform de novo assembly of the mitochondrial genome. The mitochondrial genome of Bougainvillea spectabilis (GenBank Accession Number: MW167296), a closely related species, was used as the reference sequence. To select the potential mitochondrial reads, we used BLAST searches against the mitogenome of B. spectabilis and the GetOrganelle results from a pool of Illumina reads. The mitochondrial Illumina reads were assembled into $\sim$50 mitogenome contigs using SPAdes v3.14.1 (Algorithmic Biology Lab, St. Petersburg Academic University of the Russian Academy of Sciences, St Petersburg, Russia) [22]. Nanopore reads were aligned against the GetOrganelle and SPAdes assembled scaffolds using BWA v0.7.1 (Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Cambridge, UK) [23]. The aligned Nanopore reads were extracted to perform de novo assembly of the mitochondrial genome using Canu v2.0, and then used Pilon v1.23 (https://github.com/broadinstitute/pilon/; Broad Institute of MIT and Harvard, Cambridge, MA, USA) [24] for error correction. The genes were predicted based on the method of homology alignment prediction. The coding genes, tRNAs, rRNAs, and possible pseudogenes were annotated by Blast+ 2.7.1 and tRNAscan-SE v2.0.7 (Todd Lowe Lab, Dept. of Biomolecular Engineering, School of Engineering University of California, Santa Cruz, CA, USA) [25]. tRNAscan-SE was also used to predict the secondary structure of tRNAs in F. aubertii. The mitochondrial genome map was drawn using OGDRAW (https://chlorobox.mpimp-golm.mpg.de/OGDraw.html) [26].

Tandem repeat sequences in the genome were identified by Tandem Repeats Finder (TRF) [27] (https://tandem.bu.edu/trf/trf.html) with the following parameters: match +2, mismatch –7, indel –7. We reported only the repeats with scores greater than 50. Identification of short tandem repeat (STR) or simple sequence repeats (SSR) was performed using MISA v2.1 MISA (Leibniz Institute of Plant Genetics and Crop Plant Research, Stadt Seeland, Germany) [28] with the following parameters:1–10, 2–5, 3–4, 4–3, 5–3, 6–3. The REPuter online version [29] (https://bibiserv.cebitec.uni-bielefeld.de/reputer/) was used to search scattered repeats in the mitochondrial genome with the following parameters: hamming distance-3, maximum computed repeats-50, and minimal repeat size-8.

To predict the RNA-editing sites within the mitochondrial genes of F. aubertii the online PmtREP (http://www.genepioneer.com/) was used, with a threshold value of 0.2 for the mitochondrial sequences.

We applied the cloud platform (http://www.genepioneer.com/) to analyze the GC content of the F. aubertii mitochondrial coding sequence (CDS). To visualize the GC skew, we uploaded the CDS of F. aubertii to the online Proksee [30] tool and selected the “GC Content” and “GC Skew” tabs.

The ratio of nonsynonymous substitutions (Ka) to synonymous substitutions (Ks) of protein-coding genes (PCGs) between F. aubertii and three closely related species (Nepenthes $\times$ventrata (MH798871.1), Agrostemma githago (MW553037.1), and Tetragonia tetragonoides (MW971440.1)) was calculated. The homologous protein-coding sequences were obtained by comparing protein-coding sequences among mitochondrial genomes of F. aubertii and its closely related species to find the best match using BLAST [31]. Mafft v7.427 (Research Institute for Microbial Diseases, Suita, Osaka, Japan) [32] (https://mafft.cbrc.jp/alignment/server/) was used for homologous protein sequence alignment, whereas a Perl script was utilized to map the aligned protein sequences to the coding sequence to obtain an aligned coding sequence. Finally, Ka/Ks Calculator 2.0 [33] 96–97 with MLWL model was applied to calculate Ka and Ks values.

A phylogenetic tree was constructed using the mitochondrial genome sequences of F. aubertii, Malania oleifera (Olacaceae, Santalales), and 15 other species from the order Caryophyllales, with M. oleifera used as an outgroup. We extracted the shared coding sequences from the 17 genomes mentioned above and used BLAST to identify homologous sequences. The resulting sequences were aligned using Mafft, and then concatenated. A trimming percentage of 0.7 was set using the trim function to remove any poorly aligned bases. The maximum likelihood phylogenetic tree was constructed, and the substitution model GTR+F+R3 was selected for the analysis using MEGA 7 [34]. The phylogenetic tree’s step values were determined for each branch by performing 1000 spontaneous replicate analyses.

A BLAST search was conducted between the mitochondrial and chloroplast genomes of F. aubertii to identify homologous regions. The resulting homologous segments were then compared and analyzed for sequence similarity, number of protein-coding genes, length, and composition of intergenic regions.

The second-generation sequencing platform yielded a data set of 4645.7 Mb of raw reads and 4432 Mb of clean reads, whereas the third-generation sequencing platform generated 21,993 reads with a total size of 67 Mb.

The genome (NCBI accession number: MW664926.1) obtained in this study is a circular DNA molecule with a length of 350,156 bp (Fig. 1), and a GC content of 44.73%. The base composition of the mitochondrial genome is asymmetric, and the coding and non-coding regions’ base composition, size, and proportion are shown in Table 1. The GC content in tRNA and rRNA is relatively high, accounting for 0.59% and 1.47% of the mitochondrial genome sequence, respectively.

Fig. 1.

A circular map of the F. aubertii mt genome. The counterclockwise genes are located on the outer side of the loop, whereas the clockwise genes are located on the inner side of the circle. The diagram shows the names of the contained genes. Colors are used to represent the different functional groups of genes. The grey circles inside represent GC content. The asterisks (*) represent intron-containing genes.

The F. aubertii mitochondrial genome contains 64 genes, including 34 PCGs, three rRNAs, and 27 tRNAs (Table 2). Out of the 34 protein-coding genes, only five (ccmFC, rps3, nad4, nad5, and nad7) contain introns and were categorized as cis-splice genes (Fig. 2). We examined 27 tRNA genes and determined that they can transport 17 standard amino acids. The length of tRNA ranged from 71 bp to 88 bp, with a length of 1816 bp in total. Among these genes, multicopy genes were found for trnE-UUC, trnH-GUG, and trnM-CAU. Additionally, we predicted the secondary structures of tRNAs (Fig. 3). All genes except for tRNA-Leu, tRNA-Ser, and tRNA-Tyr were predicted to have typical cloverleaf structures, and the base mismatches were mostly G-U. The mitochondrial genome contains three different tRNAs, each with different anticodons that specifically transport serine (Ser). These results are of great significance for further study of the function and stability of tRNA.

Fig. 2.

Schematic map of the cis-splicing genes. Black blocks represent exons, and white blocks represent introns. The numbers beneath the protein-coding genes indicate the locations of introns and exons. One intron is present in both ccmFC and rps3. Three introns are in nad4 and four in nad5 and nad7.

Fig. 3.

tRNA secondary structure prediction. Twenty-four tRNA secondary structures were identified in the mitochondrial genome of F. aubertii. The top left of the structure is the amino acid abbreviation letter. Base mismatches exist, usually U-C base mismatches. The mismatched bases are marked in red boxes.

Repetitive sequences can serve as genome-specific genetic markers for phylogenetic relationships among species. STRs show a high degree of polymorphism and are commonly used as molecular markers for genetic diversity studies and germplasm characteristics, identification, and selection. In the analysis of STR in the mitochondrial genome of F. aubertii, 77 STR loci were detected. There are 28 single nucleotide motifs, 16 dinucleotide motifs, seven trinucleotide motifs, 22 tetranucleotide motifs, and four pentanucleotide motifs. No hexanucleotide motifs were detected (Table 3). Additionally, five long tandem repeat sequences were detected by TRF software in the mitochondrial genome of F. aubertii (Table 4).

Scattered repetitive sequences are crucial in the study of gene mutation, genome origin and evolution, and species formation [35]. In this study, a total of 50 scattered repeats were identified in the mitochondrial genome of F. aubertii, consisting of 22 forward repeats and 28 palindromic repeats. No reverse repeats, or complementary repeats were detected. The longest identified repeat sequence was 5272 bp, whereas most of the repeats were 50–200 bp in length (Fig. 4).

Fig. 4.

Scattered repeats of F. aubertii mt genome. Different types of repeats are represented by different colors, blue represents forward repeats and orange represents palindromic repeats. The height of the colored bar represents the number of sequences. No scattered repeats in the length range of 2000–2999 bp were found.

The prediction of RNA editing sites in the mitochondrial genome coding genes of F. aubertii resulted in a total of 367 sites in 30 genes (Fig. 5). Among them, ccmB and nad4 contained the majority of RNA editing sites, with ccmB containing most of them (33 sites), accounting for 8.99% of the total RNA editing sites. The atp8 had the fewest RNA editing sites, accounting for only 0.27% of the total number of sites. The results showed that 7.9% of hydrophobic amino acids were converted to hydrophilic amino acids, 49.86% of hydrophilic amino acids were converted into hydrophobic amino acids, and 42.24% of hydrophobic amino acids remained unchanged (Table 5). RNA editing sites were C to T conversion, and the first base of the codon accounted for 32.0% of the editing sites, whereas the second base accounted for 68.0%. Some RNA editing sites may result in the formation of stop codons, but we did not identify such phenomena in the mitochondrial genome of F. aubertii. Furthermore, the proportion of amino acid conversion to leucine after RNA editing was the largest, accounting for 46.59%.

Fig. 5.

Distribution of RNA editing sites in F. aubertii mt protein-coding genes (PCGs). RNA editing sites in 30 coding genes in F. aubertii were predicted. There are no RNA editing sites in cox1, ccmFC, ccmFN, or atp9. The horizontal axis is the protein-coding gene, and the number on the top of the color bar refers to the number of predicted RNA editing sites.

The CDS of F. aubertii were analyzed to determine the GC content of each gene and calculate the average GC content (Table 6). The nad5 gene was the longest, possibly due to its introns. The GC content varied across PCGs, with matR gene having the highest GC content (51.5%) and the GC content of the nad4L being the lowest (34.65%). The AT skew average value was negative, and 13 genes had positive AT skew. Visual analysis of GC skew showed that 22 genes with positive GC skew had higher G-base content than C-base content, and 12 genes with negative GC skew and high C-base content (Fig. 6). Caryophyllales plants, including F. aubertii, had a mitochondrial genome size range of 247 Kb–509 Kb, with an average GC content of about 44% (Table 7).

Fig. 6.

GC skew of F. aubertii mt CDS. The outermost circle represents the coding area of a gene, the innermost circle represents the variation in GC content across the genome, and the innermost circle represents the GC skew, where green represents skew+ and purple represents skew-.

Comparing the Ka and Ks in homologous genes during evolution is essential for the study of gene functions and evolutionary relationships across species, as well as for exploring issues such as adaptation and genetic diversity of plant communities. The Ka/Ks ratios of 34 PCGs in the mitochondrial genomes of F. aubertii, N. ventrata, A. githago, and T. tetragonoides were analyzed (Fig. 7). Among these 34 genes, the rsp16 gene had a zero Ka/Ks value. Overall, four of the 34 genes had Ka/Ks values greater than one, indicating that these four genes were under positive selective pressure during evolution, which included nad4. rps13, rps1 and ccmFN. The remaining genes had values of Ka/Ks less than 1, which implies purifying selection and relative conservation.

Fig. 7.

Ka/Ks values of the PCGs. The mitochondrial protein-encoding genes of F. aubertii were compared with N. ventrata (MH798871.1), A. githago (MW553037.1), and T. tetragonoides (MW971440.1), respectively, for Ka/Ks analysis. The three closely related species contain protein-coding genes different from those in F. aubertii, so the Ka/Ks values of some genes in the figure are 0.

To better understand the phylogenetic position of F. aubertii, 16 plant mitogenomes from the NCBI database were downloaded (Table 7). The phylogenetic tree showed that most branch nodes had high support values above 99%, and species from the same family were grouped together, indicating high result reliability (Fig. 8). The phylogenetic tree also strongly demonstrated that F. aubertii and the N. ventrata clustered into one clade with a 92% bootstrap value (Fig. 7). Caryophyllales plants were further divided into two subgroups [36, 37]. The F. aubertii and the N. ventrata were non-core taxa of Caryophyllales, and other species belonged to the core group of Caryophyllales.

Fig. 8.

A phylogenetic tree of 17 species. Reconstruction of phylogenetic relationships was based on the shared CDS using the maximum likelihood model with M. oleifera (Olacaceae) as an outgroup. F. aubertii is indicated using the star icon. The plants grouped together are marked according to the family.

By analyzing the homology of chloroplast and mitochondrial genomes, we can deeply explore their evolutionary history and their roles in evolution. Fifty-eight homologous fragments were observed in the organellar genome of F. aubertii, with a total length of 47,757 bp, accounting for 29.2% and 13.6% of the chloroplast and mitochondrial genomes, respectively. There were 14 homologous genes in the mitochondrial genome, including trnW-CCA, trnV-GAC, trnS-GGA, trnR-ACG, trnN-GUU, trnM-CAU, trnL-CAA, trnI-GAU, trnH-GUG, trnD-GUC, trnA-UGC, rrn26, rrn18, and ccmC(Table 8).