Introduction

Synodontis (Cuvier, 1816) (Teleostei, Mochokidae) is a popular fish in aquariums. It is important to correctly identify these fish, especially in fishing competitions where species identification can affect records. However, due to the unreliable taxonomic nature of prominent morphological features, certain species are extremely difficult to identify1. The genus Synodontis belongs to the family Mochokidae, which is abundant in all the three East African Great Lakes. Therefore, it is an ideal system for comparing data from phylogenetic and geographic analyses with existing data on the family Cichlidae. Among the nine non-cichlid families shared by the three East African Great Lakes, Mochokidae shows the closest overlap in habitat with cichlids, as it mainly occurs in coastal and sub-coastal regions2. In East African mochokids, Synodontis is more diverse in species than Chiloglanis. Moreover, most Synodontis species are omnivorous, which enables them to cope with seasonal changes in food richness and habitat3,4. Compared to many more specialized fish species, such as cichlids, which adapt to lakes, their feeding habits broaden their food niche and enable them to better settle in different habitats.

In recent years, DNA barcode technology has gradually developed as a molecular identification method5. In the process of species identification in animals and plants, DNA barcode technology mainly uses a COI gene sequence that is approximately 650 bp long to distinguish species, gradually realizing automation and standardization of the identification process, thereby reducing the dependence of species identification on traditional, empirical morphological classification methods. COI, an important mitochondrial gene, is characterized by maternal inheritance, simple structure, moderate evolutionary rate, high polymorphism, and easy amplification using universal primers6. It is commonly used for species identification and analysis of phylogenetic relationships between closely related species, subspecies, and geographic populations. Currently, DNA barcode technology is widely used for species identification of freshwater fish7 and marine fish8, with a recognition rate of over 90%9. In addition, DNA barcode technology has shown significant advantages in the identification of species and evaluation of their genetic diversity10.

The mitochondrial genome of animals has typical genetic characteristics, such as maternal inheritance patterns, conservative coding regions, fast evolution of control regions, high mutation rates, and independent replication units, making it a very useful material for molecular evolutionary research. The mitochondrial genome of the vast majority of metazoans is a double-stranded closed circular DNA molecule, with a size of 14–20 kb, encoding a total of 37 genes, including 22 transfer RNA (tRNA) genes, 13 protein-coding genes (PCGs), and two ribosomal RNA (rRNA) genes11. Additionally, there is a long non-coding region in the mitochondrial genome called the control region (D-loop)12.

Currently, mitochondrial genome data are widely used to study phylogenetic relationships at different taxonomic levels in bony fish13. However, there are few comparative studies based on mitochondrial genome data from Mochokidae. Therefore, this study focused on Synodontis eupterus and Synodontis polli and measured their mitochondrial genome sequences. The structural characteristics of the mitochondrial genome sequences were also analyzed, and using the mitochondrial genome sequences of 12 Mochokidae species published in the GenBank database, a phylogenetic tree was constructed to provide a basis for further research on their genetic evolution and classification.

Materials and methods

Sample collection and DNA extraction

All methods were performed in accordance with the relevant guidelines. All specimens in this study were collected in accordance with Chinese laws. The collection and sampling of the specimens were reviewed and approved by the Animal Ethics Committee of Nanjing Forestry University. All experiments were conducted with respect to animal welfare and care. The study complied with CBD and Nagoya protocols and with the ARRIVE guidelines (https://arriveguidelines.org). Both fish samples were sourced from flower and bird markets in Fuzimiao, Qiqiao Weng, Nanjing, Jiangsu Province, China. After morphological identification, muscle tissue samples from both were taken and stored at − 80 °C for genomic DNA extraction. Approximately 50 mg of S. eupterus and S. polli muscle tissues was taken and used for genomic DNA extraction using a TIANamp Genomic DNA Kit blood/cell/tissue genomic DNA extraction kit (DP201101X, TIANGEN). The purity and concentration of the DNA were determined using an ultraviolet spectrophotometer, while the integrity was determined by 1% agarose gel electrophoresis. To ensure accuracy in sample identification, universal primers were used to amplify the Cytb, COI, and 16S rRNA genes.

Genomic sequencing

DNA was fragmented by mechanical interruption (ultrasonic), purified, and end-repaired, followed by the addition of A at the 3′ end and connection to a sequencing connector. Agarose gel electrophoresis was used to select the size of the fragments, followed by PCR and enrichment analysis, and a sequencing library was constructed. The constructed library was purified to remove connector pollution, and a library quality inspection was carried out. The qualified libraries were sequenced using the Illumina NovaSeq platform.

Sequence assembly and analysis

To reduce the complexity of sequence assembly, bowtie2 v2.2.4 software14, which is a very sensitive local software, was used for alignment with the mitochondrial genome database, and the aligned sequence was used as the mitochondrial genome sequence (mtDNA sequence). The mitochondrial genome was assembled using the SPAdes15 software, without relying on the reference genome. Quality control was performed after the assembly was completed using the reference sequence of Synodontis petricola MZ930090 (https://www.ncbi.nlm.nih.gov/nuccore/MZ930090). Using Mitos216 (http://mitos2.bioinf.uni-leipzig.de), we annotated the assembled sequences, compared the Mitos2 annotation results with those of related species, corrected the standards, and obtained the final annotation results. A mitochondrial genome map was created using MitoFish (https://mitofish.aori.u-tokyo.ac.jp/). Comparative analysis of mitochondrial genome structure for close-source species was performed using the PhyloSuite v1.2.1 software17.

Systematic evolution analysis

The complete mitochondrial genome sequences of Mochokidae were downloaded from NCBI, with Corydoras aeneus MZ571336 and Austroglanis sclateri MZ930070 as outgroups. We performed evolutionary tree analysis using tandem sequences of 13 PCGs, performed multiple sequence alignment using MAFFT v7.313 software, optimized the results through MACSE v2.03 comparison, and then performed tandem analysis after block pruning. Using ModelFinder for partition model prediction, maximum likelihood phylogenies were inferred using IQ-TREE18, and Bayesian inference phylogenies were inferred using MrBayes 3.2.619.

Results

Basic characteristics of mitochondrial genome

The mitochondrial genome sizes of S. eupterus and S. polli were 16,579 and 16,544 bp (Supplementary Information 1 and 2), respectively, with a total of 37 genes, including 13 PCGs, 22 tRNA genes, and 2 rRNA genes (Figs. 1 and 2). Analysis of the nucleotide composition of S. eupterus and S. polli showed that the A + T content of S. eupterus accounted for 56.0% of the entire mitochondrial genome, whereas the A + T contents of PCGs, tRNA, and rRNA accounted for 55.8, 56.5, and 54.3% of the entire mitochondrial genome, respectively (Table 1). S. polli A + T content accounted for 55.0% of the entire mitochondrial genome, whereas the A + T content of PCGs, tRNAs, and rRNAs accounted for 54.6, 56.7, and 53.9% of the entire mitochondrial genome, respectively (Table 2).

Figure 1
figure 1

Mitochondrial genome map of Synodontis eupterus.

Full size image
Figure 2
figure 2

Mitochondrial genome map of Synodontis polli.

Full size image
Table 1 Nucleotide composition of Synodontis eupterus mitochondrial genome.
Full size table
Table 2 Nucleotide composition of Synodontis polli mitochondrial genome.
Full size table

Previous studies have shown that base composition skewness plays an important role in transcription and replication. The AT skewness of the S. eupterus mitochondrial genome (0.121) was similar to that of the S. polli (0.109) mitochondrial genome, indicating that the content of adenine (As) was higher than that of thymine (Ts). The negative skewness (− 0.309) of the mitochondrial genome GC of S. eupterus was lower than that of S. polli GC (− 0.296), indicating that the cytosine (Cs) content was higher than that of guanine (Gs) (Tables 1 and 2).

Two overlapping gene regions, with sizes of 7 bp and 10 bp, were found in the mitochondrial genome of S. eupterus. Overlapping fragments in genome are generally only 7–10 bp in fish, whereas in mammals they can generally reach 40–46 bp20. The mitochondrial genome of S. eupterus has 11 gene intervals, with a length distribution of 1–32 bp. The interval between tRNA-Asn and tRNA-Cys was the longest at 32 bp (Table 3).

Table 3 Mitochondrial genome characteristics of Synodontis eupterus.
Full size table

In the mitochondrial genome of S. polli, two gene regions overlapped with neighboring genes, with overlapping region lengths of 7 and 10 bp. The mitochondrial genome of S. polli consists of 11 gene-spacer regions, with nucleotide lengths ranging from 1 to 32 bp. The interval between tRNA-Asn and tRNA-Cys was the longest at 32 bp (Table 4).

Table 4 Mitochondrial genome characteristics of Synodontis polli.
Full size table

Although the genome was arranged compactly as a whole, 11 gene intervals were found in both the S. eupterus and S. polli mitochondrial genomes, with a total length of 64 bp each. The longest interval was 32 bp, and the shortest interval was 1 bp, which was found in multiple locations.

Thirteen PCGs with a total length of 11,397 bp were obtained from the entire mitochondrial genome of S. eupterus, with an A + T content of 55.8%. Among the 13 PCGs, 12 (NDl, ND2, COI, COII, ATP6, ATP8, COIII, ND3, ND4L, ND4, ND5, and Cytb) are located on the heavy chain, and one (ND6) is located on the light chain. Except for COI, which uses GTG as the starting codon, all the other PCGs use ATG as the starting codon. This is common in the mitochondrial genomes of other vertebrates21. COI, ATP8, ND4L, and ND5 use TAA as the termination codon; ND1 and ND6 use TAG as the termination codon; ATP6 is encoded by the incomplete termination codon TA-; and COII, COIII, Cytb, ND2, ND3, and ND4 are encoded by the incomplete termination codon T– (Table 3).

The mitochondrial genome of S. polli contains 13 PCGs, with a length of 11,397 bp and an A + T content of 54.6%. Except for COI, which uses GTG as the starting codon, all other PCGs use ATG as the starting codon, which is consistent with the characteristics of S. eupterus. COI, ATP8, ND1, ND4L, and ND5 use TAA as the termination codon; ND6 uses TAG as the termination codon; ATP6 is encoded by the incomplete termination codon TA-; and COII, COIII, Cytb, ND2, ND3, and ND4 are encoded by the incomplete termination codon T (Table 4). The similarity between the genes of S. eupterus and S. polli is as high as 91%, indicating a close genetic relationship.

tRNA and rRNA

tRNAs play a crucial role in the adaptation of molecules during protein synthesis. The total length of tRNAs in S. eupterus was 1561 bp, with an A + T content of 56.5%, AT skewness of 0.019, and GC skewness of 0.053. The rRNA length of this species is 2643 bp, with an A + T content of 54.3%. The percentage of As (33.9%) was higher than that of Ts (20.4%), resulting in an AT skewness of 0.249. Similarly, the percentage of Gs (20.0%) was lower than that of Cs (25.7%), resulting in a negative GC bias of 0.125 (Table 1). It was predicted that there were 22 tRNAs in S. eupterus, with a length range of 66–75 bp.

The tRNA length of S. polli was 1562 bp, with an A + T content of 56.7%, AT bias of 0.019, and GC bias of 0.062. The rRNA length of S. polli is 2640 bp, with an A + T content of 53.9%. The percentage of As (33.6%) in S. polli is higher than that of Ts (20.3%), resulting in an AT bias of 0.247, whereas the percentage of Gs (20.2%) was lower than that of Cs (25.8%), resulting in a negative GC bias of 0.122 (Table 2). It was predicted that there are 22 tRNAs in the S. polli mitochondrial genome, with a length range of 66–75 bp. Post-transcriptional processing and modification of tRNA molecules can decrease their stability22, resulting in changes in their structure.

The interactions between codons and anticodons are significantly influenced by the modification position, which is usually near the swinging position. This characteristic is well-preserved in eukaryotes and directly affects the regulation of translation efficiency, transfer, and maintenance23. The stability of tRNAs depends on the changes in their central tRNA structure, which may lead to tRNA degradation and differentiation.

Control region and codon usage bias

Sequencing revealed that the mitochondrial genomes of S. eupterus and S. polli each contain control regions. The D-Loop has the fastest evolution rate in the mitochondrial genome, and its sequence variation is also the highest24. This study compared and analyzed the control region sequences of the two species and found that the control region lengths of S. eupterus and S. polli were 931 bp and 898 bp, respectively, with an A + T content of 60.9% for the control regions of both species. Previous studies have shown that mitochondrial DNA is important for studying various fish species, including bony fishes. They used control regions as markers for studying intra-species variation, which varies in many vertebrates, such as bony fish25, humans26, and birds27. The relative synonymous codon usage (RSCU) of amino acid utilization in the mitochondrial genomes of S. eupterus and S. polli is shown in Fig. 3.

Figure 3
figure 3

Codon usage of protein-coding genes of Synodontis eupterus and Synodontis polli.

Full size image

The RSCU analysis showed that the mitochondrial genomes of S. eupterus and S. polli had the highest frequency of the amino acid codons for Leu, Ala, Thr, Ile, and Gly, with fewer Cys amino acid codons. The number of hydrophobic amino acid codons in the vertebrate mitochondrial genome is higher than the number of hydrophilic amino acid codons28. This indicates that the genomic regions close to the D-loop are extensively utilized and exhibit high translation efficiency, and can be effectively translated into the vertebrate mitochondrial genome.

Phylogenetic analysis

We constructed phylogenetic trees of 14 Mochokidae fish species, including S. eupterus and S. polli, and two outgroup species based on PCG data. The phylogenetic trees constructed using the two methods had the same topological structure (Fig. 4). The results showed that the target species S. eupterus and Synodontis clarias clustered together, which confirmed the results of Dayet al.29. The other target species, S. polli and S. petricola, clustered into one branch, which is consistent with previous research results30,31,32. It is worth noting that Mochokiella paynei, Mochokus brevis, and 9 species of Synodontis genus converged into one branch, and M. paynei clustered into the genus Synodontis; a similar phenomenon was observed by Schedel et al.33.

Figure 4
figure 4

Bayesian inference (BI) (A) and maximum likelihood (ML) (B) phylogenetic trees based on the nucleotide datasets for 13 protein-coding genes from the mitogenomes of 14 Mochokidae fishes and two outgroups. The numbers along the branches indicate the Bayesian posterior probability values and ML bootstrap values, respectively.

Full size image

Discussion

This study systematically analyzed the structural characteristics, base composition, codon preferences, and PCGs of the mitochondrial genomes of S. eupterus and S. polli. The results indicated that S. eupterus and S. polli have a significant AT preference, which is similar to the base composition of vertebrate mitochondrial genomes34. The G-base content in S. eupterus and S. polli is similar to that in other bony fish, such as Sillago aeolus (18.75%)35 and Oryzias celebensis (17.60%)36,37, indicating significant anti-guanine effects20. In the mitochondrial genomes of S. eupterus and S. polli, all 13 PCGs, except ND6, were located on the heavy chain. In the codon preference analysis, the ATG codon usage frequency was the highest. The results of our analysis of start and end codons, lengths, and PCGs of the 37 genes in the mitochondrial genomes of S. eupterus and S. polli were consistent with the conclusions drawn by previous researchers38.

During mitochondrial whole-genome sequencing, mtDNA is prone to mutations and is difficult to repair, and the method for indirectly obtaining mtDNA information from high-throughput sequencing data has an important problem of sequence contamination39, which leads to erroneous research and inference of biodiversity, population genetics, species phylogenetic relationships, and mitochondrial diseases. To avoid contamination, we used various methods, such as single fragment extension and direct mapping of near source species, to assemble the mitochondrial genome. In this study, we sequenced the entire mitochondrial genomes of S. eupterus and S. polli, revealing the genetic characteristics and differentiation of the related species. By combining morphological1,2,3,4 and bioinformatic analyses, S. eupterus and S. polli were accurately distinguished.

mtDNA has a fast evolution rate and is a good source of genetic material. Phylogenetic studies of mtDNA have been widely used to outline relationships between species40. However, when studying small domains, there are limited data and information obtained from mtDNA; therefore, taxonomy is considered the foundation for understanding biodiversity and evolutionary behavior. Phylogenetic analysis was used to compare and study the similarities and differences within a family41. At present, phylogenetic trees constructed by the maximum likelihood and Bayesian methods are widely accepted, and this study combined these two methods to predict phylogeny. We found that S. eupterus and Synodontis clarias clustered together, while S. polli and S. petricola, clustered into one branch, consistent with previous studies29,30,31. Moreover, Mochokiella paynei, Mochokus brevis, and 9 species of Synodontis genus converged into one branch, and M. paynei clustered into the genus Synodontis; consistent with previous observations33.

Wong et al.42 used DNA barcode technology to detect three types of cod ingredients in 96 fish and seafood products extracted from markets and restaurants in northeastern North America, including Gadus morhua, Theragra chalcogramma, and Merluccius paradoxus. Lakra et al.43 analyzed the phylogenetic evolution of 115 fish species in the Indian Ocean using COI genes and found that the traditional taxonomic characteristics of the groups formed by COI genes in the NJ evolutionary tree were consistent, and the phylogenetic relationships between these groups were well revealed. Mat et al.44 sequenced a 36-bp long COI fragment from 723 individuals of 652 hypothesized species of the family Carangidae distributed in the waters of the Malay Archipelago in India and compared the variability of mitochondrial DNA COI fragments within and between species to evaluate the applicability of COI fragments for species identification. They found that all species formed monophyletic clusters in the phylogenetic tree, indicating that DNA barcode technology has high application value in fish species identification. This study was based on the mitochondrial genome, and it identified base differences in the COI genes of S. eupterus and S. polli. Specific primers will thus be designed based on this and the two will be identified using DNA barcode technology.

The evolution rate of protein coding genes is moderate, and different protein coding genes exhibit different evolutionary characteristics. They can be grouped based on the phylogenetic relationships of genes, but the results obtained from different groups are not the same. Therefore, a systematic analysis of each group is necessary. The DNA sequence of mitochondrial coding genes is the preferred gene for phylogenetic analysis, and the effective population size is one-quarter of that of nuclear autosomal genes. Therefore, gene trees constructed based on mitochondria have a higher probability of consistency with species trees than those constructed on nuclear autosomes. Therefore, they are often used to estimate the development history of recent evolutionary groups45. However, the functional differences of different genes may lead to different intensities of natural selection throughout history, resulting in the use of different genes in molecular phylogenetic analysis to obtain completely different gene trees. Therefore, compared to phylogenetic trees constructed with DNA barcodes or other single genes, phylogenetic trees constructed with mitochondrial whole genome exhibit more optimized stability and accuracy.

Conclusion

This study involved phylogenetic analysis of mitochondrial genomes to accurately distinguish S. eupterus and S. polli, laying a foundation for the establishment of a clearer classification system for Mochokidae fish and providing new directions for further classification research. The mitochondrial genome sizes of S. eupterus and S. polli were 16,579 and 16,544 bp, respectively, with a total of 37 genes, including 13 PCGs, 22 tRNA genes, and 2 rRNA genes. Phylogenetic trees constructed using the maximum likelihood and Bayesian methods are widely accepted, and were combined in the present study to predict phylogeny. We found that S. eupterus and Synodontis clarias clustered together, while S. polli and S. petricola, clustered into a single branch.