Mitogenome organisation and nucleotide composition
The mitochondrial genome of S. fasciolata, a double-stranded, circular molecule measuring 16,570 base pairs, aligns with the size range typical for the genus Schistura (Supplementary Table S1, from S. incerta: 16,561 bp to S. geisleri: 16,819 bp). This mitogenome consists of a canonical set of 13 PCGs, two rRNAs (12S rRNA and 16S rRNA), 22 tRNAs, and a control region known as the D-loop. Positioned between the tRNA-Pro and tRNA-Phe genes, the D-loop region serves as a crucial regulatory element for initiating replication and transcription. This organization is typical of metazoan mitochondrial genomes [37, 38]. Notably, the mitochondrial genome analysis reveals the presence of 23 intergenic spacers of variable lengths, including a significant − 10 base pair overlap between the ATP8 and ATP6 genes, indicative of direct genetic contiguity [38, 39]. Analysis of the S. fasciolata mitogenome nucleotide composition shows 30.14% adenine (A), 16.69% guanine (G), 26.37% thymine (T), and 26.80% cytosine (C). The A + T content (56.51%) predominates over the G + C content (43.49%) (Table S2), highlighting the genome's nucleotide bias and providing insights into its evolutionary adaptations and constraints.
Additionally, the mitogenome comprises two ribosomal RNAs, 12S rRNA (952 bp) and 16S rRNA (1,657 bp) (Table 1). Positioned between tRNA-Phe and tRNA-Leu, and separated by tRNA-Val, these rRNAs have high A / T contents of 50.32% and 55.94%, respectively (Supplementary Table S2). Their positioning and nucleotide composition, indicated by a positive AT-skew and a negative GC-skew, are essential for the structural and functional integrity of ribosomal RNA, crucial for protein synthesis [18].
Transfer RNAs, ribosomal RNAs
The S. fasciolata mitochondrial genome contains 22 tRNA genes, ranging in size from 66 bp (tRNA-Cys) to 76 bp (tRNA-Lys), each displaying the classic cloverleaf secondary structure (Table 1, Fig. 1). This set highlights the structural integrity and functional importance of mitochondrial tRNAs, particularly the trnA gene's retention of its dihydrouridine (DHU) arm, demonstrating these molecules' conservation across mitochondrial genomes. Despite some tRNAs lacking the D-arm, notably in trnS for AGY/N codons, this trait is widespread, signifying evolutionary adaptations for mitochondrial efficiency [40]. Further analysis of tRNA base-pairing shows mostly classical Watson-Crick A-U and G-C matches, with deviations like unmatched pairs U-U, U-C, A-A, and C-C in various tRNAs, indicating specific evolutionary adaptations. Additionally, a rare C-A pairing in trnF, trnH, trnM, and trnR was noted (Fig. 1). These observations highlight the nuanced balance between conserving structural features essential for function and introducing variations that may confer advantages in the mitochondrial context.
Table 1
Analysis of mitochondrial genome characteristics of S. fasciolata.
Genes | Strand | Position (start-end) | Length (bp) | Initiation codon | Stop codon | Anticodon | Intergenic nucleotide |
trnF | N | 1–69 | 69 | | | GAA | |
rrnS | N | 70 − 1,021 | 952 | | | | 2 |
trnV | N | 1,024 − 1,095 | 72 | | | TAC | 19 |
rrnL | N | 1,115-2,771 | 1,657 | | | | |
trnL2 | N | 2,772-2,846 | 75 | | | TAA | |
nad1 | N | 2,847-3,821 | 975 | ATG | TAA | | 7 |
trnI | N | 3,829-3,900 | 72 | | | GAT | -2 |
trnQ | J | 3,899-3,969 | 71 | | | TTG | 1 |
trnM | N | 3,971-4,039 | 69 | | | CAT | |
nad2 | N | 4,040 − 5,084 | 1,045 | ATG | T(AA) | | |
trnW | N | 5,085 − 5,155 | 71 | | | TCA | 2 |
trnA | J | 5,158-5,226 | 69 | | | TGC | 1 |
trnN | J | 5,228-5,300 | 73 | | | GTT | 1 |
OL | N | 5,302-5,332 | 31 | | | | -2 |
trnC | J | 5,331-5,396 | 66 | | | GCA | |
trnY | J | 5,397-5,465 | 69 | | | GTA | 1 |
coI | N | 5,467-7,017 | 1,551 | GTG | TAA | | |
trnS2 | J | 7,018 − 7,088 | 71 | | | TGA | 2 |
trnD | N | 7,091 − 7,163 | 73 | | | GTC | 13 |
coII | N | 7,177-7,867 | 691 | ATG | T(AA) | | |
trnK | N | 7,868-7,943 | 76 | | | TTT | 1 |
atp8 | N | 7,945-8,112 | 168 | ATG | TAA | | -10 |
atp6 | N | 8,103-8,786 | 684 | ATG | TAA | | -1 |
coIII | N | 8,786-9,570 | 785 | ATG | TA(A) | | -1 |
trnG | N | 9,570-9,642 | 73 | | | TCC | |
nad3 | N | 9,643-9,992 | 350 | ATG | T(AA) | | -1 |
trnR | N | 9,992 − 10,061 | 70 | | | TCG | |
nad4l | N | 10,062 − 10,358 | 297 | ATG | TAA | | -7 |
nad4 | N | 10,352 − 11,733 | 1,382 | ATG | TA(A) | | |
trnH | N | 11,734 − 11,803 | 70 | | | GTG | |
trnS1 | N | 11,804 − 11,870 | 67 | | | GCT | 1 |
trnL1 | N | 11,872 − 11,944 | 73 | | | TAG | |
nad5 | N | 11,945 − 13,783 | 1,839 | ATG | TAA | | -4 |
nad6 | J | 13,780 − 14,301 | 522 | ATG | TAG | | |
trnE | J | 14,302 − 14,370 | 69 | | | TTC | 4 |
cob | N | 14,375 − 15,515 | 1,141 | ATG | T(AA) | | |
trnT | N | 15,516 − 15,586 | 71 | | | TGT | -2 |
trnP | J | 15,585 − 15,654 | 70 | | | TGG | 6 |
OH | N | 15,661 − 16,401 | 741 | | | | 168 |
Protein-coding genes and codon usage
The S. fasciolata mitochondrial genome contains 13 PCGs for mitochondrial function: seven NADH dehydrogenases (ND1-6, ND4L), three cytochrome c oxidases (COI, COII, COIII), two ATP synthase subunits (ATP6, ATP8), and one cytochrome b (CYTB), totaling 16,570 base pairs. All genes are similar in length and identical in gene arrangement to those in other Schistura species. This arrangement, with 12 PCGs on the major strand and ND6 on the minor strand, exemplifies the coding efficiency and conservation of the mitochondrial genome. The PCGs employ standard start codons, primarily ATN and TTG, and their stop codons exhibit specificity for precise gene expression: TAA for COI, ATP8, ATP6, ND1, ND4L, and ND5; TAG for ND6; and a unique TA dinucleotide or a single T for the others, suggesting post-transcriptional modifications to complete the standard stop codons (Fig. 2).
Analysis of Relative Synonymous Codon Usage (RSCU) values provides further insight into the mitogenome’s evolutionary dynamics, revealing codon usage biases that favor genes encoding Ala, Arg, Pro, Ser2, Thr, and Leu1 over those encoding Asn, Asp, and Cys (Fig. 3). This codon usage bias reflects adaptive evolutionary strategies aimed at enhancing translational efficiency and accuracy [42], underscoring the nuanced interplay between genetic code specificity and mitochondrial function optimization in S. fasciolata.
Phylogenetic relationships
In this study, we reconstructed phylogenetic relationships using ML methods and BI methods, yielding a phylogenetic tree with a single topology and high support rate (with Bootstrap Support (BS) values consistently above 75 and BI posterior probabilities all exceeding 0.85), significantly enhancing our understanding of their phylogenetic relationships (Fig. 4).
Phylogenetic analyses show that Acanthocobitis, Barbatula, Nemacheilus and Homatula form a distinct monophyletic lineage, respectively (Fig. 4), which was in line with previous studies [1, 9, 11, 13]. On the other hand, our study reveals a widespread paraphyly within the tribe Nemacheilini, which is also prevalent across the entire Nemacheilidae family [4, 10]. Specifically, Schistura, Troglonectes, and Heminoemacheilus form a paraphyletic lineage. Luo et al. uncovered paraphyly between Heminoemacheilus and Paranemachilus, recommending merging them to ensure the monophyly of Paranemachilus [41]. Conversely, our research challenges the classification of Heminoemacheilus, particularly highlighting H. zhengbaoshani placement within Troglonectes, which also questions Troglonectes previously accepted monophyly [7, 42]. The extensive paraphyly across tribe Nemacheilini, even family Nemacheilidae, might be related with the rapid evolution, extensive radiation, and remarkable adaptability of species, reflected in their diversity and wide ecological niche occupation [1, 2, 3], as well as geological event [43, 44], niche competition [45], hybridism [46], and rapid adaptive evolution in response to environmental pressures [47]. These comprehensive factors collectively contribute to the rapid evolution within the family Nemacheilidae. Additionally, the genus Schistura was clustered into two clades in this study (clade 1 and clade 2 in Fig. 4), among which the S. incerta (No. MK361215.1) was nested with S. fasciolata (clade 2 in Fig. 4), our results align with prior studies [12–14] which imply the paraphyletic origin for genus Schistura. Of cause, it possibly resulted from the complex gene flow [48] or need redefinition of species. Meanwhile, the same situation for Acanthocobitis botia in genus Acanthocobitis, and Barbatula toni in Barbatula.
Selection analyses
In this study, we examined the evolutionary dynamics of 13 mitochondrial PCGs of five genera in Nemacheilini: Acanthocobitis, Barbatula, Homatula, Schistura, and Troglonectes. By applying the M0 model, we found omega (ω) ratios significantly below 1, ranging from 0.00659 (COI) to 0.12080 (ATP8), across the tribe Nemacheilini. This result emphasizes the role of purifying selection in the oxidative phosphorylation (OXPHOS) pathway. Notably, the ATP8 gene showed the highest dN/dS value, indicating potential accelerated evolution, similar with prior research on species such as Glyptothorax macromaculatus [49], Orthoptera insects [50], Desis jiaxiangi [51], and Lamprologus [52]. In contrast, the COI gene displayed the lowest dN/dS value, suggests it may have faced much stronger evolutionary constraints, which consist with Singh et al. [53], Yang et al. [54], and Li et al. [55]. This evidence underlines the extensive use of the COI gene in phylogenetic reconstructions across species groups due to its low mutation rate. Furthermore, M1 model analysis revealed significant dN/dS ratio variability among mitochondrial PCGs, indicating a diversity of selective pressures in Nemacheilini (Table 2). This variability suggests that while some genes are highly conserved due to their critical roles in cellular metabolism and energy production, others may evolve more rapidly, possibly reflecting adaptations to different ecological niches or life-history strategies within the tribe.
In order to further evaluate the significance of evolutionary rates among different genus, the Kruskal-Wallis H test was implemented. The results uncovered significance of evolutionary rates for the ND4, ND5, ND6, ATP6, and CYTB genes (Table 3). Specifically, the ND4, ND5, and CYTB genes have undergone the accelerated evolution in cave-adapted genera Homatula and Troglonectes. It might be related to the environmental pressures of their cave-dwelling lifestyle, such as low temperatures, and low oxygen levels, where it needs to much more accurate energy consumption [55, 56]. The studies have confirmed that ND5 interacts with protein subunits, playing a crucial role in regulating respiration [57]. And, the transmembrane helices in ND4 and ND6, crucial for the proton transfer process in complex I, interact with at least one lipid molecule [57]. Consequently, the accelerated evolution of ND4 in Homatula and ND6 in Troglonectes may significantly influence mitochondrial regulation of lipid synthesis and metabolism, potentially conferring a survival advantage in environments with irregular resources. However, the ATP6, ND4, ND5, ND6, and CYTB genes exhibit lower evolutionary rates in Acanthocobitis. In summary, our findings underscore the complexity of adaptive evolution within the tribe Nemacheilini, with various taxa exhibiting unique evolutionary trajectories in mitochondrial genes, reflecting their diverse ecological adaptations (Fig. 5)
Table 2
The results of evolution rate of tribe Nemacheilini lineages detected by free ratio model.
Genes | Model | -lnL | 2ΔlnL | df | P-value | ω-value |
COI | M0 | -18431.620056 | 136.609028 | 112 | 0.056934002 | 0.00659 |
M1 | -18363.315542 | |
COII | M0 | -7382.775398 | 131.968276 | 112 | 0.09574604 | 0.01357 |
M1 | -7316.791260 | |
COIII | M0 | -8628.686080 | 139.28185 | 112 | 0.041230627 | 0.01570 |
M1 | -8559.045155 | |
ND1 | M0 | -14520.605528 | 199.603388 | 112 | 6.96069E-07 | 0.02220 |
M1 | -14420.803834 | |
ND2 | M0 | -17380.652518 | 174.248292 | 112 | 0.000151349 | 0.05310 |
M1 | -17293.528372 | |
ND3 | M0 | -5171.078323 | 157.301696 | 112 | 0.003095583 | 0.05282 |
M1 | -5092.427475 | |
ND4L | M0 | -3594.222622 | 92.17911 | 112 | 0.914096146 | 0.02056 |
M1 | -3548.133067 | |
ND4 | M0 | -20760.870248 | 174.96119 | 112 | 0.000131825 | 0.03231 |
M1 | -20673.389253 | |
ND5 | M0 | -28335.202604 | 249.026302 | 112 | 2.02416E-12 | 0.05280 |
M1 | -28210.689453 | |
ND6 | M0 | -7610.482435 | 239.374544 | 112 | 2.95995E-11 | 0.04300 |
M1 | -7490.795163 | |
ATP6 | M0 | -9989.832989 | 204.373632 | 112 | 2.28909E-07 | 0.03085 |
M1 | -9887.646173 | |
ATP8 | M0 | -1974.880635 | 99.76198 | 112 | 0.78945838 | 0.12080 |
M1 | -1924.999645 | |
CYTB | M0 | -15779.463851 | 304.378548 | 112 | 1.05877E-19 | 0.02112 |
M1 | -15627.274577 | |
Table 3
Statistical test of evolutionary rates in different genes.
Genes | Test of Homogeneity of Variances | Kruskal-Wallis Test |
Levene Statistic | Levene Sig. | Chi-Square | Asymp. Sig. |
ND1 | 2.9614 | 0.0394 | 6.7316 | 0.1508 |
ND2 | 1.7365 | 0.169 | 8.1148 | 0.0875 |
ND3 | 2.9061 | 0.0479 | 9.6162 | 0.0474 |
ND4L | 0.5202 | 0.723 | 4.1544 | 0.3855 |
ND4 | 1.3608 | 0.2674 | 14.7231 | 0.0053 |
ND5 | 4.7105 | 0.0041 | 15.0192 | 0.0047 |
ND6 | 0.9982 | 0.4286 | 0.7011 | 0.9512 |
COI | 1.0037 | 0.43 | 8.6007 | 0.0719 |
COII | 1.2243 | 0.342 | 8.3019 | 0.0811 |
COIII | 1.4934 | 0.2437 | 9.4803 | 0.0502 |
ATP6 | 1.8928 | 0.1459 | 12.5892 | 0.0135 |
ATP8 | 0.8693 | 0.5049 | 3.738 | 0.4426 |
CYTB | 1.5131 | 0.2275 | 16.8024 | 0.0021 |