Mitochondrial genomes of five Hyphessobrycon tetras and their phylogenetic implications

Abstract To date, the taxonomic status and phylogenetic affinities within Hyphessobrycon, even among other genera in Characidae, remain unclear. Here, we determined five new mitochondrial genomes (mitogenomes) of Hyphessobrycon species (H. elachys, H. flammeus, H. pulchripinnis, H. roseus, and H. sweglesi). The mitogenomes were all classical circular structures, with lengths ranging from 16,008 to 17,224 bp. The type of constitutive genes and direction of the coding strand that appeared in the mitogenomes were identical to those of other species in Characidae. The highest value of the Ka/Ks ratio within 13 protein‐coding genes (PCGs) was found in ND2 with 0.83, suggesting that they were subject to purifying selection in the Hyphessobrycon genus. Comparison of the control region sequences among seven Hyphessobrycon fish revealed that repeat units differ in length and copy number across different species, which led to sharp differences in mitogenome sizes. Phylogenetic trees based on the 13 PCGs did not support taxonomic relationships, as the Hyphessobrycon fish mixed with those from other genera. These data were combined to explore higher level relationships within Characidae and could aid in the understanding of the evolution of this group.


| INTRODUC TI ON
Characidae is the most diverse family of tropical fish, with approximately 163 genera and more than 1,050 valid species, of which 231 have been described in the last 10 years. This family richness accounts for about 52% of all species in the order (Mirande, 2019;Paz et al., 2014;Veríssimo-Silveira et al., 2010). Most members of the Characidae are small-sized fish of <8 cm in standard length, reaching as much as 20 cm in some predatory genera. Fish of this family are characterized by a small adipose fin on the caudal peduncle, and most species have small, beautiful bodies and gentle temperaments (Mathubara & Toledo-Piza, 2020). Characidae is one of the most popular ornamental fish groups in the world, with great economic value (Mirande, 2019;Sun et al., 2021). Many fish of the family are known in the aquarium market under the popular name of "tetras" (Camacho et al., 2020;Leggatt & Devlin, 2020;Liu, Sun, et al., 2020;Liu, Xu, et al., 2020;Paz et al., 2014). Until now, the classification of tetra fish is mainly based on morphological characteristics. Indeed, molecular phylogeny might differ from morphological classification within tetra fish Liu, Xu, et al., 2020;Mirande, 2019).
The high diversity of Neotropical freshwater fish has been challenging to classify (DoNascimiento et al., 2017;Mirande, 2019;Silva et al., 2020). Although the complex biogeographic patterns of some of these taxa (extending over vast continental areas) have been the focus of much research recently, Characidae, which have small body size and relatively uniform morphology, are still poorly understood. Furthermore, new genera and species in this family are being validated and described (Albornoz-Garzon et al., 2019;DoNascimiento et al., 2017;Mathubara & Toledo-Piza, 2020). Hyphessobrycon, one of the richest genera of vertebrates with 109 species, is the most diverse fish genus that dominates vertebrate neotropical freshwater. Native to the Neotropics, Hyphessobrycon is widely distributed from southern Mexico to Argentina (Rio de la Plata), with the greatest species diversity found in the Amazon River basin (Faria, Bastos, et al., 2020;Paz et al., 2014). Classifying the genus and even the entire family of Characidae is currently challenging.
Considering the limited research using molecular data to infer taxonomic relationships, it is necessary to make comprehensive comparisons of morphological and genetic features of many species, to better understand the phylogenetic relationships with Characidae (Mirande, 2019). Our study on five new mitogenomes of Hyphessobrycon will help to improve the current classification of tetra fish by comparing the differences between mitogenomes of fish belonging to the same genus. Specifically, the mitochondrial characteristics of these five species, including gene order, genome size, nucleotide composition, codon usage, tRNA secondary structure, and noncoding control region (CR), were comparatively analyzed. This study provides new insights into the phylogeny and classification of tetra fish.

| Ethics statement
The collection and sampling of the specimens were reviewed and approved by Nanjing Forestry University. All specimens for this study were collected in accordance with Chinese laws. All the experiments were performed with animal welfare and care.

| PCR amplification and sequencing
According to the already published mitogenomes of Characidae spe-
The rate of nonsynonymous substitutions (Ka), rate of synonymous substitutions (Ks), and ratio of Ka/Ks were determined using DnaSP V.6.0 for these five Hyphessobrycon species (Rozas et al., 2017). tRNA genes were identified using tRNAscan-SE Search Server (available at http:// lowel ab.ucsc.edu/tRNAs can-SE/) (Chan & Lowe, 2019). Some tRNAs not detected by tRNAscan-SE were determined in the unannotated regions by sequence similarity to tRNAs of other fish.

| Phylogenetic analyses
In addition to five newly sequenced mitogenomes, 33 species from 26 genera of Characiformes, Cyprinus carpio from Cypriniformes, and Lateolabrax japonicus from Perciformes were used for phylogenetic analyses. Their accession numbers and information are listed in Table S1. After sequence alignment and model prediction using MAFFT v7.313 and ModelFinder, phylogenetic analyses were conducted for each dataset using Bayesian inference (BI) and maximum likelihood (ML) methods available in the PhyloSuite v1.2.1 (Kalyaanamoorthy et al., 2017;Katoh & Standley, 2013;Zhang, Gao, et al., 2020;Zhang, Sun, et al., 2020). BI analyses were performed with MrBayes 3.2.6 (Huelsenbeck & Ronquist, 2001) and run for a million generations, with tree sampling every 100 generations and a burn-in of 25% trees, while ML analyses were performed using the TIM2+F+R5 model in the IQ-TREE (Nguyen et al., 2015). Clade support was assessed using a nonparametric bootstrap with 1,000 replicates, and phylogenetic trees were viewed and edited in iTOL (available at https://itol.embl.de/) (Letunic & Bork, 2021).

| Genome organization and base composition
The complete mitogenomes of the five fish were typically circular,  Mitogenomes of the five fish encoded all 37 typical mitochondrial genes (13 PCGs, 22 tRNAs, and 2 rRNAs) and one noncoding CR.
Twenty-six genes were transcribed from the majority strand (J strand), and the remaining nine genes were from the minority strand (N strand) in these five mitogenomes. The gene orders of the five fish ( Figure 1; Table S2) were found to be identical to those of two other species of this genus that have been previously sequenced Liu, Xu, et al., 2020;Yan et al., 2020).
The skewness of the base composition in nucleotide sequences was used to measure the relative numbers of A to T (AT-skew) and G to C (GC-skew), and the nucleotide compositions of 19 complete or nearly complete mitogenomes in Characidae were investigated by calculating the percentages of AT-skew and GC-skew ( Figure 2).
The results of the nucleotide skew statistics showed that the ATskews in PCGs, tRNAs, and rRNAs of five whole mitogenomes were almost all positive, while the GC-skews were all obviously negative.
The low GC-skew values among the analyzed mitogenomes (−0.269 -−0.221) indicated the occurrence of more Cs than Gs, which was also observed in other announced Characidae fish mitogenomes.   Figure 1; Table S2). One PCG (ND6) was transcribed from the N strand, while the remaining 12 genes were from the J strand ( Figure 1 and  (Silva et al., 2016).

| PCGs and codon usage
Furthermore, four termination codons were found in the PCGs of the five mitogenomes, namely TAA, TAG, AGG, and T (Table S2).
In all mitogenomes, the occurrence frequency of the termination codon TAA was higher than those of the other three termination codons, while the termination codon AGG occurred the least.
Summaries of the relative synonymous codon usage (RSCU) and number of amino acids in 13 PCGs were calculated for the five mitogenomes, as shown in Figure 3. The amino acid compositions and RSCUs of these mitogenomes were found to be largely similar. Further, we calculated Ka/Ks ratios for each PCG of these mitogenomes, as shown in Figure 4, and Ka/Ks ratios of five species were compared in turn with each other. In evolutionary analysis, it is necessary to understand

| rRNA and tRNA genes
Two rRNA genes (12S and 16S rRNAs) were transcribed from the J strand in the five mitogenomes. The large rRNA (16S rRNA) was found between tRNA-Val and tRNA-Leu, while the small rRNA (12S rRNA) was located between tRNA-Phe and tRNA-Val. Lengths ranged from 945 to 952 bp in 12S rRNA and from 1,669 to 1,679 bp in 16S rRNA in the mitogenomes.  (Table S2). The tRNA regions of these five mitogenomes were 1,556, 1,555, 1,561, 1,558, and 1,559 bp, accounting for 9. 0%, 9.7%, 9.2%, 9.1%, and 9.7% of the whole mitogenomes, respectively. These five mitogenomes have 22 typical tRNA genes, with eight transcribed from the N strand and 14 from the J strand. The sizes of these tRNAs ranged from 68 to 74 bp.
Except for tRNA-Phe of H. sweglesi, all the tRNAs could be folded into secondary structures. The peculiar structures of tRNAs have also been reported in previous studies (Yuan et al., 2015). The ten most diverse tRNAs of all the five genomes are shown in Figure 5.

| CR
CR is located between the genes tRNA-Pro and tRNA-Phe. This region is responsible for regulating transcription and replication.
A + T contents in the CRs of the five mitogenomes were 67.7%, 76.2%, 73.2%, 66.4%, and 78.4%, respectively. According to previous reports, the CRs of fish vary significantly between different species and even within the same species (Buroker et al., 1990;Cui et al., 2010;Gong et al., 2015;Padhi, 2014  H. roseus were classified with other species of different genera and slightly away from the other fish of the same genus, which showed similarity to previous reports (Guimarães et al., 2019(Guimarães et al., , 2020. Although the species in Paracheirodon were classified together, Astyanax and

| Phylogenetic relationships
Hyphessobrycon species were not classified together, indicating that there may be some problems with the basis of classification. The result was indeed quite different from the existing classification system, although it only involved 27 genera in Characiformes. Additionally, we found that many species of the same genus have large differences in morphological characteristics, while species from different genera have similar morphological characteristics, which we hope to verify in future studies.

ACK N OWLED G M ENTS
We thank Nan Xu and Wentao Ye from Nanjing Forestry University for their assistance in data analysis and image rendering. This study was supported by the National College Students Innovation

CO N FLI C T O F I NTE R E S T
All the authors declared no potential interest.