RNA-seq analysis of the gonadal transcriptome during Alligator mississippiensis temperature-dependent sex determination and differentiation

Abstract

Background: The American alligator (Alligator mississippiensis) displays temperature-dependent sex determination (TSD), in which incubation temperature during embryonic development determines the sexual fate of the individual. However, the molecular mechanisms governing this process remain a mystery, including the influence of initial environmental temperature on the comprehensive gonadal gene expression patterns occurring during TSD.

Results: Our characterization of transcriptomes during alligator TSD allowed us to identify novel candidate genes involved in TSD initiation. High-throughput RNA sequencing (RNA-seq) was performed on gonads collected from A. mississippiensis embryos incubated at both a male and a female producing temperature (33.5 °C and 30 °C, respectively) in a time series during sexual development. RNA-seq yielded 375.2 million paired-end reads, which were mapped and assembled, and used to characterize differential gene expression. Changes in the transcriptome occurring as a function of both development and sexual differentiation were extensively profiled. Forty-one differentially expressed genes were detected in response to incubation at male producing temperature, and included genes such as Wnt signaling factor WNT11, histone demethylase KDM6B, and transcription factor C/EBPA. Furthermore, comparative analysis of development- and sex-dependent differential gene expression revealed 230 candidate genes involved in alligator sex determination and differentiation, and early details of the suspected male-fate commitment were profiled. We also discovered sexually dimorphic expression of uncharacterized ncRNAs and other novel elements, such as unique expression patterns of HEMGN and ARX. Twenty-five of the differentially expressed genes identified in our analysis were putative transcriptional regulators, among which were MYBL2, MYCL, and HOXC10, in addition to conventional sex differentiation genes such as SOX9, and FOXL2. Inferred gene regulatory network was constructed, and the gene-gene and temperature-gene interactions were predicted.

Conclusions: Gonadal global gene expression kinetics during sex determination has been extensively profiled for the first time in a TSD species. These findings provide insights into the genetic framework underlying TSD, and expand our current understanding of the developmental fate pathways during vertebrate sex determination.

Fig. 1

Experimental design. Experimental design of the RNA-seq analysis is illustrated. Bipotential, sex fate commitment and sex differentiation period are indicated, with temperature sensitive period (TSP; indicated in light brown). The dotted line represents the end of the TSP. Eggs were first incubated under female producing temperature (FPT; indicated in red) until just prior to the onset of sexual differentiation (stage 19; Day 0), which were then either shifted to male producing temperature (MPT; indicated in blue) or kept at FPT. Gonadal regions were sampled from individuals at several subsequent time points (Day 3, 6, 12) with corresponding approximate developmental stage (Ferguson) displayed in the bottom table. Day 0–12 represents the timing of sexual differentiation, and three individuals per temperature condition per time points are used

Fig. 2

Overlap of development-dependent and sex-dependent differentially expressed genes. Venn diagram of differentially expressed genes (DEGs) in (a) Day 0 to Day 12 MPT (indicated in blue) conditions and (b) Day 0 to Day 12 FPT (indicated in red) conditions based on genome mappings of Tophat. c Venn diagram for number of DEGs between Day 0 to Day 12 MPT and FPT conditions at respective time points, based on genome mapping using Tophat. Number values in blue indicate the number of genes with MPT-biased expression, while values in red indicate the number of genes with FPT-biased expression. All DEGs were determined based on statistical significance (FDR < 0.01) using Cuffdiff software. Further details are available in Additional files 3 and 4

Fig. 3

Candidate temperature-responsive differential gene expression. a Venn diagrams show the number of DEGs between Day 0 and Day 3 incubated either under MPT (indicated in blue) or FPT (indicated in red). 131 genes (indicated in bold) were found to have MPT-specific gene expression movement, possibly in response to changes in incubation temperature. b 41 out of 131 genes were found to have been up- or down-regulated significantly enough to be sexually dimorphic. 17 genes displayed MPT-biased expression and 24 genes displayed FPT-biased expression. Red, blue, and grey indicate FPT-, MPT- or non-bias, respectively. c Top 10 biological process gene ontology terms mapped to the candidate temperature-responsive genes with highest node score, based on Blast2GO program

Fig. 4

Candidate potential critical genes central for sex determination. a Cross comparison between development- and sex-dependent DEGs (FDR < 0.01; Fig. 2) was performed among Day 0 to Day 12, and a total of 230 genes were identified as candidate genes for A. mississippiensis sex determination. Green and orange arrows indicate up- and down-regulation, respectively, with number of corresponding genes indicated. The majority of gene expression dynamics associated with development was observed to be MPT-specific. Further information is available in Additional file 6. b Hierarchical clustering analysis of sexual dimorphism (M-F log₂FC) in the candidate genes at Day 3, 6, and 12. Red color indicates high z-score (female biased expression) while blue indicates a low z-score (male biased expression)

Fig. 5

NetGenerator derived model prediction of MPT and FPT cascade. a-b Interpolated expression (dotted line; dots indicate actual recorded fold change) and inferred patterns derived from the predicted NetGenerator model (solid line) of putative transcriptional regulatory genes over the course of sex differentiation at both (a) MPT and (b) FPT. c-d Inferred network model at both (c) MPT and (d) FPT. Robust (solid line) and non-robust (dotted line) predicted interactions are displayed as gene-gene interactions (black), and temperature-gene interactions (MPT:blue; FPT:red). Thickened lines indicate interaction with high robustness (80 % or above). Following genes names were modified for inference network: LOC102562106 (JARID2-1), LOC102561337 (JARID2-2), LOC102569158 (WNT11), LOC102560544 (EEF1A1), LOC102577358 (DMRT3), LOC102573123 (SOX9), LOC102576325 (PAK1), LOC102559361 (EDNRB), LOC102563625 (PDE2A), and LOC102577040 (FOXL2)