Medaka Bncra, but not Bncrb, is required on the medaka egg for fertilization.

A Genomic regions and resulting transcripts and proteins of the bncr locus in zebrafish (blue; GRCz11/danRer11) and medaka (yellow; Ensembl 93: Jul 2018 (GRCh38.p12)). Zebrafish Bncr is encoded by a single-exon gene (NM_001365726.1). The medaka bncr locus (ENSORLG00000004579) comprises three exons that are alternatively spliced to generate Bncra (exons 1 and 3; ENSORLT00000005754) and Bncrb (exons 1 and 2; ENSORLT00000005758). The location of the CRISPR-induced genomic deletions for medaka bncra (5-nt deletion in exon 3) and bncrb (38-nt deletion in exon 2) are indicated by asterisks. The gene structures are depicted with untranslated regions (thin rectangles) and coding sequences (thick rectangles). B Protein sequence alignment of the LU domains of Bncra and Bncrb from selected fish species (Supplementary Data File 2). Note that seahorse Bncrb is a predicted translation product from a genomic region. Purple shading indicates amino acids with at least 30% conservation. The percent amino acid sequence identity (% ID) within the mature domains is indicated. Disulfide bonds are indicated by orange brackets. C Expression values of bncra and bncrb transcripts in medaka ovary and testis based on RNA-seq³⁶ from three biological replicates for each tissue. The Y-axis is plotted in log₁₀ scale. TPM transcripts per million. D Quantification of in vivo fertilization rates from wild type and medaka bncra and bncrb mutants; tg[res], transgenic rescue. (Kruskal–Wallis test with Dunn’s multiple comparisons test; ns not significant). Means ± SD are indicated in (C) and (D).

Medaka and zebrafish sperm are compatible with multiple Bncr orthologs.

A Experimental setup for performing comparative medaka/zebrafish IVF with transgenic zebrafish bncr^−/− eggs expressing different fish Bncr orthologs. B IVF data obtained from transgenic zebrafish bncr^−/− lines expressing either zebrafish, carp, seahorse, fugu, or medaka Bncr with medaka vs. zebrafish sperm. N-glycans are depicted on each Bncr ortholog as sugar chain symbols on each glycosylated finger. (Two-tailed Wilcoxon matched-pairs signed rank test with the method of Pratt; ns not significant). C Plot of the bias index values derived from the IVF data in (A). The formula for the bias index is shown. (Two-tailed Wilcoxon signed rank test vs. theoretical median of 0 with the method of Pratt). Means ± SD are indicated in (B) and (C).

Medaka/zebrafish Bncr chimeras reveal specificity determinants in fingers 2 and 3.

A Zebrafish (blue) and medaka (dark yellow) mature Bncr protein sequence alignment and schematic of the Bncr protein fold. Fingers are labeled 1, 2, and 3 and correspond to the amino acids in boxes in the protein sequence alignment. Amino acid sites are numbered starting with the first site in the mature protein. Note that each finger is bounded by cysteine residues that keep disulfide bridges intact. “Top” and “base” are indicated. B Comparative medaka/zebrafish IVF data with Bncr chimera lines. N-glycosylation pattern of each chimera is depicted with sugar chain symbols. Means ± SD are indicated. (Two-tailed Wilcoxon matched-pairs signed rank test with the method of Pratt; ns not significant). C Plot of the bias index values derived from the IVF data in (B). Bias could not be calculated for data pairs for which the fertilization rates with both sperm were equal to 0. Means ± SD are indicated. (Two-tailed Wilcoxon signed rank test vs. theoretical median of 0 with the method of Pratt).

A positively selected, Oryzias-specific amino acid change hampers zebrafish sperm compatibility.

A Protein sequence alignment of fish Bncr orthologs and predicted ancestral states of Bncr. Fingers 1, 2, and 3 are indicated. Amino acid sites are numbered starting with the first site in the mature protein. Zebrafish-compatible sequences are blue, medaka-compatible sequences are yellow, and dually compatible sequences are green. Red rectangles demarcate sites 15, 45, and 63 which were tested individually and in combination for their role in species specificity, while N-glycosylation sites are marked with a black rectangle (see Fig. 5). The two positively selected sites are highlighted with asterisks. B Phylogenetic tree of the predicted Bncr ancestral states according to fish phylogeny. Tested nodes (A–G) are marked with a closed circle and colored according to compatibility as in (A). Nodes A–D were predicted to have the same sequence and are therefore equivalent. C Comparative medaka/zebrafish IVF data from the tested Bncr ancestral states. Means ± SD are indicated. (Two-tailed Wilcoxon matched-pairs signed rank test with the method of Pratt; ns not significant. D Plot of bias index values derived from the IVF data pairs from nodes E and F in (C). Bias was not calculated for nodes A–D and G for which the average fertilization rates with both sperm were <10%. Median (dashed line) and quartiles (dotted lines) are shown. (Two-tailed Wilcoxon signed rank test vs. theoretical median of 0 with the method of Pratt).

Zebrafish sperm favor a positively charged Bncr surface, while medaka sperm require finger 2 N-glycosylation for compatibility.

AlphaFold-predicted models of medaka Bncr (A) and zebrafish Bncr (B) (cartoon, left; surface representation depicting electrostatics, right). Amino acids that were mutated are indicated in the model as sticks and color-coded: hydrophobic (orange), positively charged (blue), polar (green). C Medaka/zebrafish IVF with medaka Bncr constructs, in which individual amino acids or combinations thereof were substituted for the corresponding amino acid(s) in zebrafish Bncr. D Plot of bias index values derived from the IVF data in (C). Bias could not be calculated for data pairs for which the fertilization rate with both sperm was equal to 0. Medians (dashed lines) and quartiles (dotted lines) are shown. E Medaka/zebrafish IVF with zebrafish Bncr constructs, in which individual amino acids or combinations thereof were substituted for the corresponding amino acid(s) in medaka Bncr. F Medaka/zebrafish IVF experiments to assess the importance of N-glycosylation in Bncr’s species specificity. IVF with zebrafish Bncr N-glycosylation site variants (left); IVF with medaka Bncr N-glycosylation site variants (right). G Medaka/zebrafish IVF experiments testing sufficiency of N-glycosylation pattern combined with specific amino acid changes for determining Bncr’s species specificity. IVF with zebrafish Bncr variants (left); IVF with medaka Bncr variants (right). H In vivo zebrafish fertilization rates of combined N-glycosylation and amino acid substitution Bncr variants. Means ± SD are indicated in (C, E–H). C, E–G: two-tailed Wilcoxon matched-pairs signed rank test with the method of Pratt; p values could not be calculated for samples in which all data points were 0. D: two-tailed Wilcoxon signed rank test vs. theoretical median of 0 with the method of Pratt.