Negative Selection in Oreochromis niloticus × O. aureus Hybrids Indicates Incompatible Oxidative Phosphorylation (OXPHOS) Proteins

Abstract

Crossing Oreochromis niloticus (On) females with O. aureus (Oa) males results in all-male progeny that are essential for effective tilapia aquaculture. However, a reproductive barrier between these species prevents commercial-scale yield. To achieve all-male progeny, the currently used practice is crossing admixed stocks and feeding fry with synthetic androgens. Hybrid tilapias escaping to the wild might impact natural populations. Hybrids competing with wild populations undergo selection for different stressors, e.g., oxygen levels, salinity, and low-temperature tolerance. Forming mitochondrial oxidative phosphorylation (OXPHOS) complexes, mitochondrial (mtDNA) and nuclear DNA (nDNA)-encoded proteins control energy production. Crossbred tilapia have been recorded over 60 years, providing an excellent model for assessing incompatibility between OXPHOS proteins, which are critical for the adaptation of these hybrids. Here, by comparing nonconserved amino acid substitutions, across 116 OXPHOS proteins, between On and Oa, we developed a panel of 13 species-specific probes. Screening 162 SRA experiments, we noted that 39.5% had a hybrid origin with mtDNA-nDNA allele mismatches. Observing that the frequency of interspecific mtDNA-nDNA allele combinations was significantly (p < 10^-4) lower than expected for three factors, UQCRC2, ATP5C1, and COX4B, we concluded that these findings likely indicated negative selection, cytonuclear incompatibility, and a reproductive barrier.

Keywords: cichlid fish; hybrid selection; introgression; mitonuclear incompatibility.

Figure 1

Variation between O. niloticus (On) and O. aureus (Oa). Two Oreochromis species important for the aquaculture industry were compared: (a) Schematic representation of the key differences between On and Oa, which are associated with a preference for a specific water depth, adaptation to different oxygen concentrations and temperatures, and feeding habits, with On and Oa having preferences for omnivorous and herbivorous diets, respectively; (b) nDNA-encoded OXPHOS genes with nonconserved variations (red font) between the On and Oa ortholog genes were used to determine peptide probes. These allele probes were applied for TBLASTN searches against 162 SRA libraries annotated as On and Oa to determine their frequencies. The mtDNA-encoded probes were used to verify the species annotation. The table shows the basic statistics of these analyses, ¹ Genomic position in On genome build (nucleotide accession GCF_013358895.1). ² Number of cases in which one or two foreign alleles were detected. ³ Frequencies of alien alleles among all the fish alleles. ⁴ A bold underlined font denotes a significantly (p < 10⁻⁴) lower frequency than expected. Three genes that presented a low frequency of alien alleles are denoted by bold font in the last column.

Figure 2

Diagram deciphering position of 13 non-conserved amino acid variations between O. niloticus (On) and O. aureus (Oa). Three-dimensional (3D) visualization of these variations was performed using Mol* 3D Viewer (RCSB PDB, Mol* 3D Viewer, version 4.11.0) and the 3D structure data of assembled respiratory protein complexes, which were deposited for Bos taurus in the protein data bank format (PDB). The protein complexes are displayed following their orientation on the inner mitochondrial membrane and colored by the Chain Id default option. The relevant gene symbols of the variable proteins are shown, and variations (1–10) are indicated next to the orthologous residues highlighted by the software tool. (a) Complex III (bovine PDB 1SQB). Each polypeptide within the complex has two instances (colored similarly) and these are as follows (from top left to bottom right): subunit UQCRH (grey,); UQCRQ (light brown); mtDNA-encoded cyc1 UQCR4 (bordeaux); UQCRFS1 (green I); UQCR10 (blue); mtDNA-encoded subunit CYTB (purple); UQCR11 (green II); UQCRB (yellow); UQCRC1 (green III); proteolytic procced polypeptide of UQCRFS1 (red); UQCRC2 (orange). With two instances each, four variable amino acids are denoted as follows (mutation positions are of the On and Oa orthologous protein, GenBank protein accessions: ADD71229, ADD71255, and XP_003438780, XP_031595481, respectively): S360F (1), F365L (2); T258A (3); P404T (4). (b) Complex IV (bovine PDB 5B1A). COX4I1 (dark yellow); COX6C (grey); COX5A (red); mtDNA-encoded COX2 (purple); mtDNA-encoded COX1 (green III); COX5B (green II); COX6B (brown); mtDNA-encoded COX3 (green I); COX6A2 (orange); COX7A1 (light brown); completely hidden behind are COX7B, COX7C, COX8B (pink, yellow, and grey, respectively). Three variable amino acids are denoted as follows (mutation positions are of the On orthologous protein GenBank protein accessions: XP_003457010, XP_031584659, ADD71219, ADD71232, ADD71171, and ADD71236, respectively): mutation on an illustration of the cleaved mitochondrial transit-peptide A17V (5), A516T, G521E and T524A are three mutations on the c-terminal, which is outside of the homologous region with cattle; thus, the c-terminal end is indicated (6); A107V (7). (c) Complex V (bovine PDB 7AJB). Each polypeptide within the complex has at least two instances (colored differently), and these are as follows (from top left to bottom right): ATP5O (brown, yellow); ATP5F1 (light brown, turquoise); ATP5J (bordeaux, purple); ATP5A1L with 6 instances (dark yellow, green II, light brown, bordeaux, grey, green I); ATP5B with 6 instances (purple, red, orange, blue, yellow, green II); ATP5H (pink, light blue), ATP5C1 (pink, brown); ATPIF (green I, pink); ATP5E (light blue, light brown); ATP5D (turquoise, grey); ATPG2 with 16 instances (yellow, dark yellow, beige, light brown, grey, grey II, green III, red, orange, blue, purple, green II, bordeaux, purple II, green I, orange); mtDNA-encoded ATP6 (grey, pink); mtDNA-encoded ATP8 (orange, green III); ATPMJ (dark yellow, green I); ATP5MG (orange, green II); ATP5ME (yellow, light green); ATP5MK (grey, light brown). With two instances each, four variable amino acids are denoted as follows (mutation positions are of the On and Oa orthologous protein, GenBank protein accessions: XP_003448120, XP_031602241, ADD71138, and ADD71248, respectively): A79T (8), A136T (9), adjacent A136T and P137L (10).