Genome organization and nucleotide composition
The complete mitochondrial genomes of
C.p. japonicus (MF578940) and
H. rondeletii (MF578941) were 16532 bp and 16525 bp in length, respectively. Their circular gene map was shown in Fig 1 and the characteristics of the two complete mitochondrial genome were shown in Table 2. The overall A+T content was same in these flying fishes (56.7%), which was similar to that in other flying fish reported
(Chou et al., 2016). Eight and nine overlaps were found in
C.p. japonicus and
H. rondeletii, respectively. The length of the longest overlaps in two fishes were both 7 bp. Twelve and fourteen intergenic spacers were detected in
C.p. japonicus and
H. rondeletii, respectively. The biggest one located between tRNA-Asn and tRNA-Cys which was 38 bp in length.
Protein-coding genes and codon usage
The total lengths of 13 protein-coding genes in
C.p. japonicus and
H. rondeletii were 11428 bp and 11432 bp, respectively. The A+T content was 56.40% in
C.p. japonicus and 56.30% in
H. rondeletii. There were typical location and arrangements of protein-coding genes in these two fishes and only two protein-coding genes (ND5 and ND6) encoded by L-strand (Table 2).
12 of 13 protein-coding genes started with ATG, while the COI gene used GTG as a start codon. Five protein-coding genes ended with typical complete codon TAA or TAG. Eight used an incomplete stop codon (T or TA). There was no mutation in the conserved ATP8 gene and the most variation was found in the ND2 gene in these two fishes.
Transfer and Ribosomal RNA genes
There were 22 transfer RNA genes interspersed in two fishes’ mitochondrial genomes and ranged from 66 bp (tRNA-Cys) to 75 bp (tRNA-Leu). The tRNA-Leu and tRNA-Ser genes both showed two forms (Table 2). The tRNA-Ser (AGY) was lack of DHU arm and the rest 21 tRNA genes could fold into the typical clover-leaf structures.
The two ribosomal RNA genes (12SrRNA and16SrRNA) located between tRNA-Phe and tRNA-Leu and were separated by tRNA-Val. Their length in
C.p. japonicus were 950 bp and 1703 bp and those in
H. rondeletii were 944 bp and 1688 bp. The composition of A+T (55.6% and 55.5%) was higher than that of G+C (44.4% and 44.5%).
Non-coding region
There were two non-coding regions in the mitochondrial genome of two flying fishes (D-Loop L-strand origin (OL). The putative OL could fold a sTable stem-loop secondary structure, with 12 nucleotides in the loop and 10 paired nucleotides in the stem. The conserved 5'-GCCGG- 3' motif, which was thought being related to the transition from RNA synthesis to DNA synthesis
(Hixson et al., 1986), was not found in the OL region of two flying fishes. Although C-rich was detected in the loop of many reported teleost fish
(Johansen et al., 1990), it was not observed in this study. According to the previous researches, C-rich region might be the initiation of primer synthesis
(Taanman, 1999) which suggested that the OL might have no function in
C.p. japonicus and
H. rondeletii. The result was similar to the study of
Cheng et al., (2010).
The lengths of D-Loop in
C.p. japonicus and
H. rondeletii were 871bp and 862bp
, respectively. The A+T content of D-Loop in these two flying fishes were 64.6% and 65.3%, which was higher than the average of the whole mitogenome (56.1% and 56.1%) and that of 13 protein coding genes (56.4% and 56.3%). The feature had been reported in other fishes
(Dong et al., 2017). As described in other fishes
(Lee et al., 1995; Hurst et al., 1999), the D-Loop possessed the typical tripartite regions. The genetic distance between two D-Loop regions was smaller than that of five protein coding genes (Fig 2), which is the same as that in other four flying fishes (
C. arcticeps, C. atrisignis, C. doederleinii and
C. unicolor). It indicated that the D-loop region in flying fishes was conservative.
Phylogenetic analysis
The ML tree was built with significant bootstrap supports (Fig 3). The tree topologies indicated that
C.p. japonicus and
H. rondeletii clustered in a group
. Three species from
Cheilopogon genus formed a single group which was considered to be the sister group of the first one. This might result from the lack of the data of flying fishes in this topology. Another reason might be the large variations in some protein coding genes (ND2, ND4 and ND5). The genus
Hirundichthys showed closest relationship with the genus
Cheilopogon under the family of Exocoetidae. This study would advance our understanding of the population genetic structure, phylogenetic classification and evolution of flying fishes.