Discovery of the First Germline-Restricted Gene by Subtractive Transcriptomic Analysis in the Zebra Finch, Taeniopygia guttata

Abstract

Developmentally programmed genome rearrangements are rare in vertebrates, but have been reported in scattered lineages including the bandicoot, hagfish, lamprey, and zebra finch (Taeniopygia guttata) [1]. In the finch, a well-studied animal model for neuroendocrinology and vocal learning [2], one such programmed genome rearrangement involves a germline-restricted chromosome, or GRC, which is found in germlines of both sexes but eliminated from mature sperm [3, 4]. Transmitted only through the oocyte, it displays uniparental female-driven inheritance, and early in embryonic development is apparently eliminated from all somatic tissue in both sexes [3, 4]. The GRC comprises the longest finch chromosome at over 120 million base pairs [3], and previously the only known GRC-derived sequence was repetitive and non-coding [5]. Because the zebra finch genome project was sourced from male muscle (somatic) tissue [6], the remaining genomic sequence and protein-coding content of the GRC remain unknown. Here we report the first protein-coding gene from the GRC: a member of the α-soluble N-ethylmaleimide sensitive fusion protein (NSF) attachment protein (α-SNAP) family hitherto missing from zebra finch gene annotations. In addition to the GRC-encoded α-SNAP, we find an additional paralogous α-SNAP residing in the somatic genome (a somatolog)-making the zebra finch the first example in which α-SNAP is not a single-copy gene. We show divergent, sex-biased expression for the paralogs and also that positive selection is detectable across the bird α-SNAP lineage, including the GRC-encoded α-SNAP. This study presents the identification and evolutionary characterization of the first protein-coding GRC gene in any organism.

Keywords: Germline-restricted chromosome; RNA-seq; genomics; next-generation sequencing and assembly; phylogenomics; soluble NSF attachment; zebra finch.

Figure 1. Discovery of a paralogous α-SNAP gene pair

A. Subtractive transcriptomic analysis used in this study. B. Overview of sequence comparison between assembled GRC (green) and somatolog (orange) α-SNAP sequences along with confirmation by cloning. C. Genomic DNA qPCR analysis confirming GRC α-SNAP is only detected in testis or ovary (germline) tissue (Primers F2+R2, see panel B). Error bars represent standard error of the mean. Two-way ANOVA identified testis signal (**** = p < 0.0001) as highly statistically significant, with n = 3 individuals of each sex tested. D. RT-qPCR analysis of expression of GRC α-SNAP showing strong ovary expression. Statistical significance calculated with Students 2-tailed T-test. E. RT-qPCR analysis of somatolog α-SNAP showing strong testis expression. Statistical significance calculated with Students 2-tailed T-test. See also Figure S1.

Figure 2. Avian dN/dS analysis of α-SNAP genes

A. Bayesian tree of birds and dN/dS analysis. Red boxes represent β-SNAP; blue dots represent α-SNAP proteins. Branch numbers indicate posterior probabilities and scale bar represents substitutions per site. Branch letters A–I correspond to Table 1 for ω-ratio (dN/dS) estimation of selection pressure. B. Analysis of ω for branch G using aBSREL [24] showing two selective regimes. Positive selection on this branch was statistically significant (p-value 0.0045, Table 1). C. Analysis of ω for branch A using aBSREL [24] showing two selective regimes, with positive selection affecting 25% of sites but at a lower overall level than branch G. See also Figure S3.

Figure 3. Multi-species Bayesian tree, confirming that both SNAP genes in finch are from the α-SNAP family

Red boxes, β-SNAP; blue dots, α-SNAP proteins. Branch numbers indicate posterior probabilities and scale bar represents substitutions per site. Related to Figure S2.