Transcriptomic responses to environmental temperature by turtles with temperature-dependent and genotypic sex determination assessed by RNAseq inform the genetic architecture of embryonic gonadal development

Abstract

Vertebrate sexual fate is decided primarily by the individual's genotype (GSD), by the environmental temperature during development (TSD), or both. Turtles exhibit TSD and GSD, making them ideal to study the evolution of sex determination. Here we analyze temperature-specific gonadal transcriptomes (RNA-sequencing validated by qPCR) of painted turtles (Chrysemys picta TSD) before and during the thermosensitive period, and at equivalent stages in soft-shell turtles (Apalone spinifera-GSD), to test whether TSD's and GSD's transcriptional circuitry is identical but deployed differently between mechanisms. Our data show that most elements of the mammalian urogenital network are active during turtle gonadogenesis, but their transcription is generally more thermoresponsive in TSD than GSD, and concordant with their sex-specific function in mammals [e.g., upregulation of Amh, Ar, Esr1, Fog2, Gata4, Igf1r, Insr, and Lhx9 at male-producing temperature, and of β-catenin, Foxl2, Aromatase (Cyp19a1), Fst, Nf-kb, Crabp2 at female-producing temperature in Chrysemys]. Notably, antagonistic elements in gonadogenesis (e.g., β-catenin and Insr) were thermosensitive only in TSD early-embryos. Cirbp showed warm-temperature upregulation in both turtles disputing its purported key TSD role. Genes that may convert thermal inputs into sex-specific development (e.g., signaling and hormonal pathways, RNA-binding and heat-shock) were differentially regulated. Jak-Stat, Nf-κB, retinoic-acid, Wnt, and Mapk-signaling (not Akt and Ras-signaling) potentially mediate TSD thermosensitivity. Numerous species-specific ncRNAs (including Xist) were differentially-expressed, mostly upregulated at colder temperatures, as were unannotated loci that constitute novel TSD candidates. Cirbp showed warm-temperature upregulation in both turtles. Consistent transcription between turtles and alligator revealed putatively-critical reptilian TSD elements for male (Sf1, Amh, Amhr2) and female (Crabp2 and Hspb1) gonadogenesis. In conclusion, while preliminary, our data helps illuminate the regulation and evolution of vertebrate sex determination, and contribute genomic resources to guide further research into this fundamental biological process.

Fig 1

Overlap of differentially-expressed genes by temperature (controlling FDR at 1e-10) for various categories: (a) overlap between male- and female-producing temperatures at Stage 22 of Chrysemys based on expression levels normalized by (1) upper-quartile (UQ100); (2) upper-quartile excluding the top 1 percentile of transcripts with the highest expression (UQ99); (3) upper-quartile and house-keeping genes (UQHK100); and (4) upper-quartile and house-keeping genes excluding the top 1 percentile of transcripts with highest expression (UQHK99). (b) and (c): across stages in Chrysemys picta and Apalone spinifera. (d): Overlap of genes of interest present in turtles with the mouse gonadal genes described in Jameson et al., (2012). DE = differentially-expressed, GoI = genes of interest described in Table 5. Differential expression was assessed between temperatures at each developmental stages for each species separately.

Fig 2. Average qPCR expression at 26°C (blue) and 31°C (red) across developmental stages from qPCR experiments (top panels; modified from Valenzuela et al., 2013) and RNAseq (bottom panels; this study).

Mean and one standard deviation are presented. Asterisks denote significant differential expression by temperature. Boxed stages fall within the thermosensitive period.

Fig 3. Differential expression in Chrysemys picta and Apalone spinifera turtles based on RNAseq of a subset of genes involved in the mammalian urogenital pathway and other genes of interest for turtle gonadogenesis.

The list of genes and activation or repression information derives from Valenzuela 2008b, Liu et al, 2009, Chassot et al., 2012, Eggers et al., 2014, Lai et al., 2014 plus other sources cited in the text, while information about the timing and pattern of expression correspond to those observed in turtles during the present study. Approximate equivalency is provided between mice and turtle developmental stages of gonadal development. [not sig. diff. exp. = Not significant differential expression].

Fig 4. P-values of differentially-expressed genes (after applying Benjamini-Hochberg correction) linked to mammalian urogenital pathways, showing higher expression at 26°C (blue) and 31°C (red).

Highly differentially-expressed genes (while controlling false discoveries at 1e-10) identified in dark blue and dark red. Light colored cells denote significance at a standard α = 0.05. GA = gene absent in transcriptome.

Fig 5

Eigengene networks and dendrograms illustrating the co-expression patterns in developing gonad of (a) Chrysemys at 26°C (b) Chrysemys at 31°C, (c) Apalone at 26°C and (d) Apalone at 31°C, of 981 genes of interest described in Table 5. Colors along the x and y-axes represent clusters of genes showing similar expression. Gene order varies by plot along the X and Y-axes.

Fig 6. Comparative transcriptional patterns in Chrysemys, Apalone and Trachemys for a subset of common elements in the regulatory network of sexual development across embryonic stages at male- (low) and female- (high) producing temperatures for the TSD taxa.