Thyroid hormone membrane receptor binding and transcriptional regulation in the sea urchin Strongylocentrotus purpuratus

Abstract

Thyroid hormones (THs) are small amino acid derived signaling molecules with broad physiological and developmental functions in animals. Specifically, their function in metamorphic development, ion regulation, angiogenesis and many others have been studied in detail in mammals and some other vertebrates. Despite extensive reports showing pharmacological responses of invertebrate species to THs, little is known about TH signaling mechanisms outside of vertebrates. Previous work in sea urchins suggests that non-genomic mechanisms are activated by TH ligands. Here we show that several THs bind to sea urchin (Strongylocentrotus purpuratus) cell membrane extracts and are displaced by ligands of RGD-binding integrins. A transcriptional analysis across sea urchin developmental stages shows activation of genomic and non-genomic pathways in response to TH exposure, suggesting that both pathways are activated by THs in sea urchin embryos and larvae. We also provide evidence associating TH regulation of gene expression with TH response elements in the genome. In ontogeny, we found more differentially expressed genes in older larvae compared to gastrula stages. In contrast to gastrula stages, the acceleration of skeletogenesis by thyroxine in older larvae is not fully inhibited by competitive ligands or inhibitors of the integrin membrane receptor pathway, suggesting that THs likely activate multiple pathways. Our data confirms a signaling function of THs in sea urchin development and suggests that both genomic and non-genomic mechanisms play a role, with genomic signaling being more prominent during later stages of larval development.

Keywords: echinoderm; genomic; nongenomic; sea urchin; skeletogenesis; thyroid; thyroid hormone response element; transcriptome.

Figure 1

Thyroid hormones bind to membrane protein extracts from Strongylocentrotus purpuratus gastrulae and can be displaced by integrin ligands. (A–G) In fluorescence anisotropy competitive binding assays, T4, T3, T2, Tetrac, and RGD peptide bind to membrane protein extract. Normalized mP: Millipolarization normalized on a scale of 0-100. We were not able to detect binding of rT3. Out of the tested ligands, T4 showed the highest affinity for membrane proteins, with a calculated Ki of approximately 1.1 x 10^-9 M. (H) In a saturation binding assay, thyroxine labeled with rhodamine (RHT4) binds to membrane protein extract with a Kd of approximately 9.5 x 10^-8 M, roughly two orders of magnitude lower in affinity when compared to unlabeled T4.

Figure 2

Thyroid hormone exposure accelerates skeletogenesis in Strongylocentrotus purpuratus larvae. (A) Representative images of TH exposure effects on skeletogenesis in the gastrula and late-stage larvae. In gastrulae, TH exposure accelerates initiation of skeletogenesis with initial spicules for larval skeleton appearing several hours earlier compared to the control (no TH). In the rudiment, THs accelerate development of skeletal features, including tube feet, spines, and test. ls: Larval skeleton, r: Rudiment, s: Skeleton, sp: Juvenile spines. (B) THs accelerate skeletogenesis in Gastrula in a MAPK-dependent manner. Data reproduced from . All THs and RGD peptide are shown at a concentration of 10^-7M, PD98059 at 5 x 10^-6M. * indicates a rate of skeletogenesis statistically different from the control (Binary logistic regression with Bonferroni corrected p-values). THs, including T4, T3, and T2, accelerate skeletogenesis.†: While we have displayed here all THs at 10^-7M, a higher concentration of T4 is able to outcompete the inhibitory effect of RGD peptide, as discussed in . (C) THs accelerate skeletogenesis in late-stage larvae in a partially MAPK-dependent manner, including T3 and T4 (One-way ANOVA with Bonferroni-corrected t tests; F(10,121) = [20.88], p = 4.96e-22). The effect of T2 was not significant after correcting for multiple comparisons (adjusted p = 0.063). Unlike in gastrulae, inhibiting MAPK with PD98059 is insufficient to fully prevent the effect of T4 on skeletogenesis. * indicates a rate of skeletogenesis statistically different from the control (Bonferroni-corrected t-test, p<0.05).

Figure 3

Summary of transcriptome data structure. Gene expression clusters by age and hormone treatment T4-exposed samples are dissimilar from control and T3-exposed samples within each age group. (A) UMAP analysis of individual replicates reveals clustering primarily by age, with highly divergent gene expression patterns between the gastrulae and older larvae groups. Control and T3-exposed groups cluster more tightly together than with T4-exposed groups. (B) Heatmap of gene expression with genes separated into 4 clusters by k-means using the ComplexHeatmap R package. Gene counts are normalized to z-scores. The heatmap reveals that genes are differentially regulated in gastrulae and older larvae (9613 DEGs between 27-day old and gastrulae control groups), as well as in T4-exposed groups relative to control and T3-exposed groups in older larvae (1730 DEGs in T4 27d relative to control).

Figure 4

Summary of top ten significantly enriched GO slim categories of upregulated and downregulated genes. Gene ontology (GO) enrichment was determined by Fisher’s exact test in T3 and T4-exposed gastrulae (G), 23-day old larvae (23d), and 27-day old larvae (27d). More GO categories were enriched in the upregulated subset of DEGs. The most differentially enriched groups between (A) upregulated and (B) downregulated DEGs in each major GO category were cytoskeleton, and plasma membrane (CC), signaling, vesicle-mediated transport, and microtubule-based movement (BP), and cytoskeletal protein binding and cytoskeletal motor activity (MF). Cells are coloured and GO slim categories are sorted by number of DEGs assigned to each GO slim category. Every displayed GO slim category was significantly enriched in at least one sample. Fisher’s exact test was used to determine significantly enriched GO categories relative to control groups. The top ten groups sorted by lowest p values (<0.05) are displayed for each major GO category (Cellular Component, Biological Process, and Molecular Function). GO annotations were sourced from Ensembl Metazoa and mapped to GO slim annotations.

Figure 5

Summary of upregulated and downregulated genes by manually annotated functional group. DEGs were determined using DESeq2 comparisons between TH-exposed groups and the control group of the same age. T4 resulted in regulation of more genes than T3, with dramatically more DEGs in older larvae compared to gastrulae. More DEGs are upregulated than downregulated. Individual categories are explored in more detail in Figures 6 – 12 .

Figure 6

T4 regulates gene expression of skeletogenesis-related genes in older larvae. (A) Heatmap of top 30 skeletogenesis-related genes, sorted by p-value (low to high, determined by DESeq2) and clustered by expression pattern. Colours are scaled to log2(foldchange) and capped at 2-fold, while numbered cells display foldchange. Spicule proteome was updated from . PMC-expressed genes were obtained from . The Skeletogenic GRN list was compiled from the Davidson Lab Gene Regulatory Network model hosted on BioTapestry. (B) Volcano plots of all skeletogenesis related genes.

Figure 7

T4 regulates gene expression of thyroid hormone-related genes in older larvae. Deiodinases and some putative TH transporters are strongly upregulated. Tg-like cholinesterases (>40% identity to Tg with >4% tyrosine content) and sulfotransferases are generally downregulated in 27-day-old post-rudiment development larvae (27d). (A) Heatmap of top 30 TH-related genes, sorted by p-value (low to high, determined by DESeq2) and clustered by expression pattern. Colours are scaled to log2(foldchange) and capped at 2-fold, while numbered cells display foldchange. Genes were compiled from annotations in the S. purpuratus 5.0 genome release and verified with manual BLAST searches against chordate genomes. (B) Volcano plots of all TH-related genes.

Figure 8

T4 regulates gene expression of nuclear hormone receptors in older larvae. Notably, all the upregulated nuclear receptors excepting NR5A2 are in the NR1 family. The putative nuclear TH receptor is among those genes found to be upregulated in larvae with rudiments (27d), but not in younger larvae without rudiments (23d). (A) Heatmap of top 30 nuclear hormone receptor genes, sorted by p-value (low to high, determined by DESeq2) and clustered by expression pattern. Colours are scaled to log2(foldchange) and capped at 2-fold, while numbered cells display foldchange. Genes were compiled from annotations in the S. purpuratus 5.0 genome release and verified with manual BLAST searches against chordate genomes. (B) Volcano plots of all nuclear hormone receptor genes.

Figure 9

T4 regulates gene expression of apoptosis-related genes in older larvae. The most highly upregulated and downregulated genes are putative orthologs of cytochrome C (Cyt C), an inhibitor of apoptosis. (A) Heatmap of top 30 apoptosis-related genes, sorted by p-value (low to high, determined by DESeq2) and clustered by expression pattern. Colours are scaled to log2(foldchange) and capped at 2-fold, while numbered cells display foldchange. Apoptosis-related genes were compiled from the Reactome pathway database and matched to S. purpuratus via annotations available on Echinobase or by best BLAST match to the S.p. 5.0 genome. (B) Volcano plots of all apoptosis related genes.

Figure 10

T4 regulates gene expression of adhesome-related genes in older larvae, however not to the extent of other categories we examine. Integrin alpha-8-like is dramatically downregulated, while integrin alpha-PS1 and several cadherins are moderately upregulated, along with elements of the Cadhesome and Adhesome. Transglutaminases are also notably downregulated in older larvae. (A) Heatmap of top 30 adhesome-related genes, sorted by p-value (low to high, determined by DESeq2) and clustered by expression pattern. Colours are scaled to log2(foldchange) and capped at 2-fold, while numbered cells display foldchange. Genes were compiled from annotations in the S. purpuratus 5.0 genome release and verified with manual BLAST searches against chordate genomes. (B) Volcano plots of all adhesome related genes.

Figure 11

T4 regulates gene expression of neuronal genes in older larvae. Aromatic L-amino acid decarboxylase (AADC), a key component of dopamine and serotonin synthesis, is dramatically upregulated by T4 in older larvae. As well, histamine receptors and several acetylcholine receptors are also upregulated. Several serotonin and acetylcholine receptors are downregulated. (A) Heatmap of top 30 neuronal genes, sorted by p-value (low to high, determined by DESeq2) and clustered by expression pattern. Colours are scaled to log2(foldchange) and capped at 2-fold, while numbered cells display foldchange. Genes were compiled from annotations in the S. purpuratus 5.0 genome release and verified with manual BLAST searches against chordate genomes. (B) Volcano plots of all neuronal genes.

Figure 12

Some immune genes are strongly upregulated in our T4-exposed groups (5.7% of total annotated genes), including a variety of SRCR-domain containing proteins. Several clotting and coagulation-related genes were strongly downregulated by T4. (A) Heatmap of top 30 immune genes, sorted by p-value and clustered by expression pattern. Colours are scaled to log2(foldchange) and capped at 2-fold, while numbered cells display foldchange. Immune gene list was compiled and updated from . (B) Volcano plots of all immune genes.

Figure 13

Genes with thyroid hormone response elements (TREs) <500 bp upstream are more likely to be regulated by THs. (A) Genes with >10 nearby TRE halfsites were regulated an average of 2.6-fold and 2.0-fold more strongly by T4 and T3 respectively when compared with genes that have no nearby TRE halfsites. Similarly, genes with nearby DR4 (B) and DR0-6 (C) sites were more heavily regulated by THs than genes with no nearby TRE sites. Nearby TRE sites were the strongest predictor of regulation in 27-day-old larvae, compared to younger larvae and gastrulae. * represents p<0.05 by two tailed T test comparison to sites with no nearby TRE.

Figure 14

(A) Average enrichment of predicted TRE sites within <500 bp upstream of >2-fold DEG transcription initiation sites, relative to unregulated genes. Genes which are at least 2-fold upregulated by THs are more likely to have nearby (<500bp) TRE sites. In 23 and 27-day old larvae, T3 requires more nearby TRE sites to regulate genes than T4. In particular, genes sensitive to T3 regulation in 27-day old larvae had over 2-fold enrichment of nearby TRE halfsites, DR4 and DR0-6 predicted sites. * represents p<0.05 by two tailed T test comparison to unregulated genes. (B) Proportion of DEGs with nearby predicted TRE sites. Significantly more TRE halfsites and DR4 sites were present in genes regulated by T4 in Gastrulae, and T3 in 27-day-old larvae, with a trend of increased TRE sites near DEGs in 23-day-old larvae as well. * represents p<0.05 by two-tailed T test comparison to unregulated genes.