Functional Characterization of Thyrostimulin in Amphioxus Suggests an Ancestral Origin of the TH Signaling Pathway

Abstract

Thyrostimulin, consisting of GpA2 and GpB5 subunits, has been identified in amphioxus, but to date, little is known about the roles of GPA2/GPB5‒type hormone in this evolutionarily important animal. We showed here that amphioxus GpA2, GpB5, and TSH receptor (TSHR) represent the archetypes of vertebrate TSHα, TSHβ, and TSHR, respectively, and both gpa2 and gpb5 were coexpressed in the Hatschek pit, a homolog of the vertebrate pituitary, in amphioxus. We also showed that recombinant amphioxus GpA2 and GpB5, like zebrafish TSHα and TSHβ, bound to both amphioxus and zebrafish TSHR and that tethered amphioxus thyrostimulin activated both protein kinase A and protein kinase C pathways in the cells expressing amphioxus TSHR. Moreover, we demonstrated that recombinant amphioxus thyrostimulin induced the production of thyroid hormone (TH) T4. Because genuine TSH is absent in amphioxus and thyrostimulin is the only and sole glycoprotein hormone, our data likely provide evidence that amphioxus thyrostimulin is a functional glycoprotein hormone that plays a role as TSH does in vertebrates. The data also suggest that the TH signaling pathway evolved in the basal chordate more than 500 million years ago.

Glycoprotein hormones (GpHs), including FSH, LH, and TSH are the pituitary hormones that have carbohydrate side chains. They are heterodimers of two noncovalently associated α and β subunits. The α subunits (FSHα, LHα, and TSHα) of these hormones are common; however, their β subunits (FSHβ, LHβ, and TSHβ) are unique and grant specific bioactivity. In general, FSH and LH regulate gonadal activity, and TSH regulates thyroidal activity. More than a decade ago, a novel GpH was identified and named thyrostimulin because it activated the TSH receptor (TSHR) (1). Compared with other GpHs, thyrostimulin is believed to consist of two distinct subunits known as GpA2 (α subunit) and GpB5 (β subunit), forming a heterodimer based on the colocalization of its subunits in the pituitary in vertebrates (2–5). It is thought to act mostly as a regulator with autocrine and paracrine functions rather than as a systemic hormone. However, whether thyrostimulin can heterodimerize in circulation and whether its individual subunits can function independently is still debated.

In addition to thyrostimulin, all jawed vertebrates have three heterodimeric pituitary GpHs: FSHα/FSHβ, LHα/LHβ, and TSHα/TSHβ. These classical FSH, LH, and TSH subunits in jawed vertebrates are not found in jawless vertebrates, such as the lamprey, and nonvertebrate chordates, such as amphioxus. The lamprey possesses one heterodimeric pituitary GpH consisting of GpA2/GpHβ as well as thyrostimulin (6–8), whereas amphioxus has only thyrostimulin consisting of GpA2/GpB5 as its sole GpH (7, 9–12). Previous studies have shown that in amphioxus, the gpa2 gene is expressed mainly in the anterior part of the nerve cord and the left side of the central canal (11), whereas the gphβ gene is expressed predominantly in the dorsal part of the nerve cord, the atrial cells of the gill, and the previtellogenic oocytes (12). Both phylogenetic and synteny analyses have shown that the thyrostimulin subunits are ancestral to the vertebrate GpH subunits (4, 9, 12), but their function remains poorly established to date (7, 9, 11, 12).

GpHs act through binding to specific receptors (13). For example, TSH binds to its G protein‒coupled receptor TSHR on thyroid epithelial cells, where it stimulates the production and secretion of the thyroid hormones (THs) T4 and T3, which regulate development, growth, and metabolism (14–16). Amphioxus is the basal extant chordate lineage, having a genome uncomplicated by extensive genomic duplication (17). It is hence an important reference for insights into the origin and evolution of the vertebrate neuroendocrine system (15, 18–21). Several lines of evidence suggest there is an active TH metabolism similar to that of vertebrates in amphioxus (22). First, both T3 and T4 have been detected in the amphioxus endostyle (23) and fixation of iodine reported in the same tissue (24). Second, peroxidases and deiodinases involved in the activation or deactivation of THs have been present in amphioxus (15, 25). Importantly, there is an active TH receptor in amphioxus (26, 27). Despite this progress, little is known about the regulation of TH production in this evolutionarily important animal. Notably, however, a gene encoding TSHR has been identified in amphioxus (15, 28).

In the absence of any genuine TSH in amphioxus, it was thus reasonable to hypothesize that a GpA2/GpB5 heterodimer may interact with TSHR in amphioxus, thereby regulating TH production similarly to the vertebrate TSH/TSHR system. This study was therefore undertaken to test the hypothesis. We showed that amphioxus GpA2, GpB5, and TSHR represent the archetypes of vertebrate TSHα, TSHβ, and TSHR, respectively, and that both gpa2 and gpb5 were coexpressed in the Hatschek pit, a homologue of the vertebrate pituitary, in amphioxus. We also showed that recombinant amphioxus GpA2 and GpB5, like zebrafish TSHα and TSHβ, was bound to both amphioxus and zebrafish TSHR and that a recombinant, single-chain amphioxus thyrostimulin activated both protein kinase A (PKA) and protein kinase C (PKC) pathways in the cells expressing amphioxus TSHR. Moreover, we demonstrated that recombinant amphioxus thyrostimulin induced the production of TH T4. These data demonstrate that amphioxus thyrostimulin is a functional GpH that plays a role similar to that of TSH in vertebrates.

Materials and Methods

Animal and cell

All animal experiments described in this paper were approved by the ethics committee of the Laboratory Animal Administration of Shandong province (permit no. SD2007695). Adult amphioxus Branchiostoma japonicum (formerly known as Branchiostoma belcheri tsingtauense) were collected from the sandy bottom of the sea near Shazikou, Qingdao, China, and maintained in glass containers with continuous aeration at room temperature. They were fed twice a day with single-celled algae. Zebrafish (Danio rerio) that were ∼5 months old were purchased from the local commercial market, maintained in well-aerated tap water at 27°C ± 1°C and fed live bloodworm and TetraMin Tropical Flakes fish food (Tetra, Fairfax, VA) twice a day. They were acclimatized for 1 week before the experiments.

HEK293T cells were gifts from Jianfeng Zhou, Laboratory of Molecular Medicine, School of Medicine and Pharmacy, Ocean University of China. The cells were cultured at 37°C with 5% CO₂ in DMEM (HyClone, Logan, UT) supplemented with 10% fetal bovine serum (Gibco, Carlsbad, CA), 10 U/mL of penicillin (Beyotime, Shanghai, China), and 100 mg/mL of streptomycin (Beyotime) as described previously (20).

Cloning, sequencing, and sequence analysis of Bjtshr, Bjgpa2, and Bjgpb5

Total RNAs were extracted from B. japonicum with RNAiso plus (TaKaRa, Dalian, China) according to the manufacturer’s instructions. First-strand cDNAs were synthesized with a reverse transcription kit (TaKaRa) using oligo (dT) primer after digestion with recombinant DNase I (RNase-free) (TaKaRa) to eliminate genomic contamination. Partial cDNA fragments of B. japonicum tshr, gpa2, and gpb5 genes, named Bjtshr, Bjgpa2, and Bjgpb5, respectively, were amplified by PCR with the primer pairs P1 and P2, P3 and P4, and P5 and P6 (29), which were designed on the basis of tshr, gpa2, and gpb5 gene sequences found in the Branchiostoma floridae genome database (http://genome.jgi.doe.gov/Brafl1/Brafl1.home.html). To obtain full-length cDNA sequences of Bjtshr, Bjgpa2, and Bjgpb5, 3′ and 5′ RACE was performed using the BD SMART^TM RACE cDNA amplification kit (Clontech, Beijing, China) according to the manufacturer’s instructions, with the Bjtshr gene-specific primer pairs P7 and P8 and P9 and P10, Bjgpa2 gene-specific primer pairs P11 and P12 and P13 and P14, and Bjgpb5 gene-specific primer pairs P15 and P16 and P17 and P18 (29), respectively. The clones obtained were sequenced, and the overlapping sequences were assembled. On the basis of the cDNA sequences assembled, the full-length opening reading frames (ORFs) of Bjtshr, Bjgpa2, and Bjgpb5 were obtained by PCR with the primer pairs P19 and P20, P21 and P22, and P23 and P24 (29), respectively, and verified by sequencing.

Sequence comparison against the GenBank protein database was performed using the BLAST network server at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/BLAST/). Protein domains were analyzed using the SMART program (http://smart.embl-heidelberg.de/). The molecular mass and isoelectronic point of the proteins were determined using ProtParam (http://web.expasy.org/protparam/). The three-dimensional (3D) structures of BjTSHR, BjGpA2, and BjGpB5 as well as human TSHR, TSHα, and THSβ were established using the SWISS-MODEL online software at the Expert Protein Analysis System (https://swissmodel.expasy.org/interactive). Multiple alignments of the protein sequences were generated using the ClustalW program. Phylogenetic trees were constructed using the amino acid sequences of available TSHR, TSHα, and THSβ proteins, including BjTSHR, BjGpA2, and BjGpB5, by MEGA (version 6.0) according to the neighbor-joining method, and the reliability of each node was estimated by bootstrapping with 1000 replication. The putative transcription factor binding sites were predicted using the JASPAR program (http://jaspar.genereg.net/). Information on exon-intron organization was obtained from the National Center for Biotechnology Information database (http://www.ncbi.nlm.nih.gov/) and Ensembl database (http://www.ensembl.org).

Semiquantitative real-time PCR

Semiquantitative real-time PCR (qRT-PCR) was used to determine the expression profiles of Bjgpa2 and Bjgpb5 in the different tissues of B. japonicum, including the Hatschek pit, gill, hepatic cecum, hindgut, muscle, notochord, testis, and ovary. Total RNAs were extracted from the tissues with RNAiso plus (TaKaRa) and digested with recombinant DNase I (RNase-free). The cDNAs were synthesized as described previously and used as a template for qRT-PCR. The gene elongation factor-like α (ef1-α) was chosen as the reference for internal standardization. The primers P25 and P26 specific to Bjgpa2, P27 and P28 specific to Bjgpb5, and P29 and P30 specific to ef1-α (29) were designed to amplify the specific fragments of the genes. qRT-PCR was performed on the ABI 7500 real-time PCR system (Applied Biosystems, Carlsbad, CA) as described previously (18). The reaction of each sample was performed in triplicate. At the end of each PCR reaction, dissociation analysis was performed to confirm the amplification specificity. Data were analyzed with 2^−∆∆CT based on CT values for Bjgpa2 (or Bjgpb5) and ef1-α to calculate the relative mRNA expression level.

In situ hybridization

In situ hybridization was carried out as described by Fan et al. (30). Briefly, B. japonicum were cut into three pieces and fixed in freshly prepared 4% paraformaldehyde in 100 mM of PBS (pH, 7.4) at 42°C for 8 hours. The samples were dehydrated, embedded in paraffin, and sectioned at 7 μm. The sections were mounted onto poly-l-lysine‒coated slides, dried at 42°C for 48 hours, and deparaffinized in xylene for 20 minutes (two changes for 10 minutes each), followed by immersion in absolute ethanol for 10 minutes (two changes for 5 minutes each). They were then rehydrated and brought to double distilled water treated with 0.1% diethypyrocarbonate. After prehybridization in a hybridization buffer containing 50% deionized formamide (v/v), 100 μg/mL of heparin, 5 × SSC, 0.1% Tween-20, 5 mM EDTA, 1× Denhardt solution, and 100 μg/mL of salmon sperm RNA at 50°C for 3 hours, they were hybridized in the same hybridization buffer with 1 μg/mL of digoxigenin (DIG)-labeled Bjtshr antisense RNA probe, DIG-labeled Bjgpa2 antisense RNA probe, or DIG-labeled Bjgpb5 antisense RNA probe at 50°C for 14 hours in a moist chamber. Subsequently, the sections were preincubated in 1% blocking reagent (Roche, Basel, Switzerland) in 100 mM of Tris-HCl (pH, 7.4) with 150 mM of NaCl for 2 hours at room temperature and incubated with anti-DIG alkaline phosphatase conjugated antibody (Roche) diluted 1:1000 in 1% blocking reagent for 2 hours at room temperature. After washing and staining, the sections were photographed under a BX51 Olympus microscope (Olympus, Tokyo, Japan).

Construction of expression vector

The complete cDNA region encoding the extracellular domain of BjTSHR (edBjTSHR) was amplified by PCR using the primers P31 and P32 containing EcoR V and EcoR I sites (29), respectively, and the complete cDNA region encoding the extracellular domain of D. rerio TSHR (edDrTSHR) amplified by PCR using the primers P33 and P34 (designed on the basis of the D. rerio TSHR cDNA sequence in GenBank ID: NM_001145763.2) with Sal I and Xho I sites (29), respectively. Both cDNAs were subcloned into the plasmid expression vector pET-32a (Novagen, Madison, WI) previously cut with the same restriction enzymes. The identity of the inserts was verified by sequencing, and the plasmids were designated pET-32a/edBjtshr and pET-32a/edDrtshr, respectively.

Similarly, the complete cDNA region encoding the mature BjGpA2 with the signal peptide deleted was amplified by PCR using the primers P35 and P36 with EcoR I and Xho I sites; the complete cDNA region encoding D. rerio mature TSHα (DrTSHα) with the signal peptide deleted was amplified by PCR using the primers P37 and P38 (designed on the basis of the D. rerio TSHα cDNA sequence in GenBank ID: NM_205687.2) with EcoR I and Xho I sites; the complete cDNA region encoding the mature BjGpB5 with the signal peptide deleted was amplified by PCR using the primers P39 and P40 with Nde I and EcoR I sites; and the complete cDNA region encoding D. rerio mature TSHβ (DrTSHβ) with the signal peptide deleted was amplified by PCR using the primers P41 and P42 (designed on the basis of the D. rerio TSHβ cDNA sequence in GenBank ID: NM_181494.2) with Nde I and EcoR I sites. The cDNAs were subcloned into the plasmid expression vector pET-28a (Novagen) previously cut with the same restriction enzymes. The identity of inserts was verified by sequencing, and the plasmids were designated pET-28a/Bjgpa2, pET-28a/Drtshα, pET-28a/Bjgpb5, and pET-28a/Drtshβ, respectively.

Expression, purification, and refolding of recombinant proteins

The plasmids pET-32a/edBjtshr, pET-32a/edDrtshr, pET-28a/Bjgpa2, pET-28a/Drtshα, pET-28a/Bjgpb5, and pET-28a/Drtshβ were transformed into Escherichia coli Transetta (DE3), and the cells were cultured overnight in lysogeny broth containing 100 μg/mL of ampicillin (Beyotime) (for pET-32a/edBjtshr and pET-32a/edDrtshr) or 50 μg/mL of kanamycin (Beyotime) (for pET-28a/Bjgpa2, pET-28a/Drtshα, pET-28a/Bjgpb5, and pET-28a/Drtshβ). When OD₆₀₀ reached ∼1.0, isopropyl β-d-thiogalactoside (IPTG) (Amresco, West Grove, PA) was added to the cultures at a final concentration of 1 mM, and the cultures were allowed to shake for 12 hours at 28°C. The recombinant proteins that were expressed in inclusion bodies were purified and refolded according to the method of Xu and Zhang (31) with modification of dialysis pH value. The refolded proteins were analyzed by SDS-PAGE on 12% gel and immunostained using mouse anti‒His-tag antibody (32) as the primary antibody, followed by incubation with horseradish peroxidase‒conjugated goat anti-mouse IgG (33) according to the method described previously by Wang et al. (34). Protein concentrations were determined by the bicinchoninic acid methods using BSA as a standard (CWBIO, Beijing, China).

ELISA

ELISA was used to assay the interaction of BjGpA2, DrTSHα, BjGpB5, or DrTSHβ with BjTSHR or with D. rerio TSHR (DrTSHR) according to the method of Sun et al. (35). Briefly, aliquots of 50 μL of recombinant edBjTSHR (redBjTSHR) or edDrTSHR (redDrTSHR) solution (40 μg/mL) were applied to a 96-well microplate and air dried at 25°C overnight. The plates were incubated at 60°C for 30 minutes to fix the proteins, and the wells were each blocked with 200 μL of 1 mg/mL BSA in 10 mM of PBS (pH, 7.4) at 37°C for 2 hours. After four washes with 200 μL of 10 mM of PBS supplemented with 1% Tween-20 (PBST), 50 μL of biotin-labeled recombinant BjGpA2 (rBjGpA2), DrTSHα (rDrTSHα), BjGpB5 (rBjGpB5), or DrTSHβ (rDrTSHβ) solution with different concentrations (0, 0.5, 1, 2, 4, 8, 16, and 32 μg/mL) were added to the wells. After incubation at room temperature for 3 hours, the wells were each rinsed four times with 200 μL of PBST, added with 100 μL of streptavidin–horseradish peroxidase (CWBIO) diluted to 1:5000 with 10 mM of PBS (pH, 7.4), and incubated at room temperature for 1 hour. The wells were washed as described previously and reacted with 75 μL of 0.4 mg/mL O-phenylenediamine (Amresco) in 51.4 mM of Na₂HPO₄, 24.3 mM of citric acid, and 0.045% H₂O₂ (pH, 5.0) at 37°C for 10 minutes. Subsequently, 25 μL of 2 mM H₂SO₄ was added to each well to terminate the reaction, and absorbance at 492 nm was monitored by a microplate reader (GENios Plus; Tecan, Männedorf, Switzerland). For control, the pET-32a His-tag protein (rTRX) was processed similarly. The experiments were performed in triplicate and repeated three times.

Construction of eukaryotic expression vectors

Recombinant single-chain amphioxus thyrostimulin (combination of GpA2 and GpB5) was constructed using synthetic linkers (5-residue Ser-Gly repeat) to join the α (GpA2) and β (GpB5) subunits. The construct has the β chain in the proximal postal position and the α chain in the distal postal position. The signal sequence of the α subunit was deleted from the construct, and the signal sequence of the β subunit was used to direct secretion of the tethered molecule. In brief, the complete coding region of Bjgpb5 and the region encoding mature Bjgpa2 with the signal peptide deleted were amplified by PCR using the sense primer P43 and the antisense primer P44 with the BamH I site, as well as the sense primer P45 with the BamH I site and the antisense primer P46 (29), which were designed on the basis of Bjgpb5 and Bjgpa2 gene sequences. The PCR productions of Bjgpb5 and Bjgpa2 were both digested with BamH I and ligated together using T4 DNA ligase (TaKaRa). After incubation at 4°C overnight, the linker fragment was used as a template to amplify the thyrostimulin of B. japonicum (Bjthyrostimulin) by PCR using the sense primer P47 with the Hind III site and the antisense primer P48 with the Age I site (29). The PCR products were digested with Hind III and Age I and subcloned into the plasmid expression vector pcDNA3.1/V5-His A (Invitrogen, Carlsbad, CA) previously cut with the same restriction enzymes. The recombinant plasmid was verified by sequencing and named pcDNA3.1/Bjthyrostimulin. Similarly, the full-length ORF of the Bjtshr gene was amplified by PCR using the primers P49 and P50 with the Hind III and Xho I sites, respectively (29). The PCR product was digested with Hind III and Xho I and subcloned into the plasmid expression vector pcDNA3.1/V5-His A (Invitrogen) previously cut with the same restriction enzymes. The identity of the insert was verified by sequencing, and the plasmid was designated pcDNA3.1/Bjtshr.

Recombinant Bjthyrostimulin synthesis in HEK293T cells

HEK293T cells were seeded in a cell culture flask (Eppendorf, Hamburg, Germany) and cultured at 37°C with 5% CO₂ in DMEM (HyClone) containing 10% fetal bovine serum (Gibco), 10 U/mL of penicillin (Beyotime), and 100 mg/mL of streptomycin (Beyotime). At 90% to 95% confluency, the cells were transfected with the plasmid pcDNA3.1/Bjthyrostimulin using Lipofectamine 2000 Reagent (Invitrogen) according to the instructions of the manufacturer. Briefly, pcDNA3.1/Bjthyrostimulin plasmid was mixed with Lipofectamine 2000 (Invitrogen), and the mixture was added into cells and incubated at 37°C for 6 hours. Subsequently, the transfection medium was replaced with serum-free DMEM medium (HyClone). After 72 hours, the medium containing recombinant Bjthyrostimulin (rBjthyrostimulin) protein was collected, filtered through a 0.45-μm filter membrane (Millipore, Billerica, MA), and purified by Ni-nitrilotriacetic acid (Ni-NTA) resin column (GE Healthcare, Beijing, China). The purified rBjthyrostimulin was concentrated at 4°C in centrifugal filter units (Millipore) according to the manufacturer’s instructions. The purity of rBjthyrostimulin was analyzed by SDS-PAGE on a 12% gel and stained with Coomassie Brilliant Blue R-250 (Beyotime). In parallel, the purified rBjthyrostimulin was verified by MALDI/TOF MS analysis and Western blotting as described previously. Finally, the concentrated rBjthyrostimulin was sterilized with 0.22-μm filter membrane (Millipore), aliquoted, and stored at −20°C until use. Protein concentration was determined by the bicinchoninic acid method using BSA as a standard (CWBIO).

Luciferase reporter gene assays

For luciferase assays, HEK293T cells were seeded in 24-well plates and cultured until reaching a confluency of 90% to 95% before transfection with Lipofectamine 2000 Reagent (Invitrogen). For each transfection, 400 ng of the pCRE-Luc or pSRE-Luc reporter plasmid (Biovector, Beijing, China), 400 ng of pcDNA3.1/Bjtshr, and 40 ng of pRL-TK (Promega, Madison, WI) were cotransfected into the cells as described previously. Twenty-four hours after transfection, the cells were serum starved for 18 hours and then stimulated with saline (control) or various concentrations of rBjthyrostimulin, rBjGpA2, rBjGpB5, or rTRX for an additional 6 hours. Luciferase activity in the cell extracts was determined using a luciferase assay system (GeneCopoeia, Rockville, MD), according to the user manual, in a luminometer (Promega) as described previously by Wang et al. (20).

Assay for in vivo bioactivity of rBjthyrostimulin

Because amphioxi are small and have an open circulation system, zebrafish were used to test in vivo bioactivity of Bjthyrostimulin. A total of 150 D. rerio with an average body weight of ∼400 mg were divided into three groups of D. rerio (50 fish per group; male/female, 1:1). Each fish of group 1 was injected IP with 20 μL of various doses of rBjthyrostimulin (0.1 and 1 μg/g body weight), whereas each fish of group 2 or group 3 was injected IP with 20 μL of rTRX (0.1 and 1 μg/g body weight) or 20 μL of saline. At 4 hours postinjection, blood samples were collected from the caudal vessels of the fishes with a heparinized syringe and needle and centrifuged at 3000g for 10 minutes. The sera were pooled, aliquoted, and stored at −20°C until use. ELISA was used to assay the concentrations of T4 in the sera using the double-antibody sandwich ELISA kit (MLBIO, Shanghai, China) according to the manufacturer’s instructions.

Statistical analysis

All the experiments except gene cloning and recombinant expression were conducted at least three times. Statistical analyses were performed using GraphPad Prism 5 software. The data obtained were subjected to statistical analysis by one-way ANOVA, and difference at P < 0.05 was considered significant. All data were expressed as mean ± SD.

Results

Sequence and phylogeny of Bjgpa2, Bjgpb5, and Bjtshr

The full-length cDNA of Bjgpa2 obtained (GenBank accession no. MH329280) was 1115 bp long and contained an ORF of 381 bp, a 5′–untranslated region (UTR) of 309 bp, and a 3′-UTR of 425 bp. The ORF of the cDNA encoded a deduced protein of 126 amino acids with a calculated molecular weight of ∼13.6 kDa and an isoelectric point of 8.20. BjGpA2 possessed an N-terminal signal peptide and a C-terminal GpH α domain that are characteristic of human TSHα (Fig. 1A). Molecular modeling revealed that the 3D structure of BjGpA2 consists of one α-helix and seven β-sheets, which is similar to that of human TSHα containing one α-helix and six β-sheets (Fig. 1B). Moreover, the phylogenetic tree showed that BjGpA2 was located at the base of vertebrate TSHα proteins (29), reflecting the phylogeny of chosen organisms. These findings suggest that BjGpA2 may represent the archetype of vertebrate TSHα.

Figure 1.

The structures of amphioxus GpA2, GpB5, and TSHR. (A) The domain structure of amphioxus GpA2 was predicted by the SMART program. (B) The 3D structure of human TSHα and amphioxus GpA2 was generated by SWISS-MODEL online software. (C) The domain structure of amphioxus GpB5 was predicted by the SMART program. (D) The 3D structure of human TSHβ and amphioxus GpB5 was generated by SWISS-MODEL online software. (E) The domain structure of amphioxus TSHR was predicted by the SMART program. (F) The 3D structure of human TSHR and amphioxus TSHR was generated by SWISS-MODEL online software. (G) Diagram of the genomic structures of rat TSHβ and amphioxus GpB5 is shown. The gray and black rectangles represent noncoding exons and coding exons, respectively. The horizontal lines between two rectangles represent introns. The numbers in the rectangles and above the lines represent the nucleotide numbers of the exons and introns, respectively. The numbers in parentheses represent the phase of the introns. The identical residues between Rat TSHβ negative TH response element and its similar sequence in amphioxus are marked with asterisks. The arrows indicate the genomic position of the sequences. Genomic sequences are from GenBank accession nos. NM_013116 (rat) and XM_019770501 (amphioxus). (H) Predicted potential transcription factor binding sites are shown. Various binding sites were marked with different colors. The transcription start site (+1) is depicted by an arrow.

The full-length cDNA of Bjgpb5 obtained (GenBank accession no.: MH329281) was 1012 bp long and had an ORF of 402 bp, a 5′-UTR of 61 bp, and a 3′-UTR of 549 bp. The ORF of the cDNA coded for a deduced protein of 128 amino acids with a calculated molecular weight of ∼14.3 kDa and an isoelectric point of 6.30. BjGpB5 possessed an N-terminal signal peptide and a C-terminal GpH β domain that are typical of human TSHβ (Fig. 1C). Molecular modeling revealed that the 3D structure of BjGpB5 consists of one α-helix and five β-sheets, which is similar to that of human TSHβ containing one α-helix and five β-sheets (Fig. 1D). Moreover, the phylogenetic tree demonstrated that BjGpB5 was located at the base of vertebrate TSHβ proteins (29), well reflecting the phylogeny of chosen organisms. These findings suggest that BjGpB5 may represent the archetype of vertebrate TSHβ.

The full-length cDNA of Bjtshr obtained (GenBank accession no.: MH329282) was 2778 bp long and possessed an ORF of 2208 bp, a 5′-UTR of 94 bp, and a 3′-UTR of 476 bp. The ORF of the cDNA encoded a deduced protein of 735 amino acids with a calculated molecular weight of ∼79.2 kDa and an isoelectric point of 7.93. BjTSHR contained an N-terminal signal peptide, a leucine rich repeat N-terminal domain, three leucine-rich repeats domains, and a C-terminal seven-transmembrane domain that are typical of human TSHR (Fig. 1E). Molecular modeling revealed that the 3D structure of BjTSHR consists of 16 α-helices and 20 β-sheets, which is similar to that of human TSHR containing 20 α-helices and 15 β-sheets (Fig. 1F). Moreover, the phylogenetic tree demonstrated that BjTSHR was located at the base of vertebrate TSHR proteins (29), again reflecting the phylogeny of chosen organisms. These findings suggest that BjTSHR may represent the archetype of vertebrate TSHR.

Genomic structure and functional binding sites of BjGpB5

Analysis of the genomic structures revealed that BjGpB5 consisted of three exons interspaced by two introns (Fig. 1G). The introns begin with GT and end with an AG dinucleotide, a sequence thought to be necessary for correct RNA splicing of various eukaryotic genes. Of note, the rat TSHβ gene also consists of three exons interspaced by two introns. It is apparent that BjGpB5 shares a genomic structure similar to that of vertebrate TSHβ. Analysis of potential transcription factor binding sites showed that like rat tshβ, Bjgpb5 has important pituitary transcription factor binding sites, such as MED-1, POU1F1, PROP1, Pitx1, Pitx2, and GATA2, in its upstream region (Fig. 1H).

The TSHβ gene was shown to be negatively regulated by TH at the transcriptional level (36) through the negative TH response element (nTRE). In the rat TSHβ gene, a 17-bp DNA fragment (CGCCAGTGCAAAGTAAG) located at the end of the first exon has a proven association with the repressive effects (37, 38). Notably, a similar sequence (CTGGATTGTAAAGTAAG) with 71% identity to nTRE in rat tshβ was found at the front of the first intron in Bjgbp5 (Fig. 1G). Collectively, these data provide additional evidence that Bjgpb5 is homologous to vertebrate tshβ and suggest that the transcription of Bjgpb5 may be regulated similarly to that of rat tshβ.

Tissue-specific expression of Bjgpa2, Bjgpb5, and Bjtshr

qRT-PCR assay was carried out to examine the expression patterns of gpa2 and gpb5 in the different tissues of B. japonicum (Fig. 2A). It was found that Bjgpa2 was expressed primarily in the hindgut, hepatic cecum, and ovary and at lower levels in the Hatschek pit, gill, and testis (Fig. 2B), and Bjgpb5 was expressed mainly in the hindgut, gill, Hatschek pit, hepatic cecum, and ovary and at lower levels in the testis, muscle, and notochord (Fig. 2C).

Figure 2.

Tissue-specific expression of Bjgpa2 and Bjgpb5 genes. (A) Diagram shows various tissues of amphioxus. (B and C) Relative expression of (B) Bjgpa2 and (C) Bjgpb5 genes in the different tissues, including the Hatschek pit, gill, hepatic cecum, hindgut, muscle, notochord, testis, and ovary (relative to whole body). The ef1-α gene was chosen as an internal control for normalization. A total of 10 amphioxi were used per group. Vertical bars represent the mean ± SD (n = 3). *P < 0.05.

In situ hybridization was also performed to examine the mRNA distribution of Bjgpa2, Bjgpb5, and Bjtshr in the different tissues of B. japonicum. As shown in Fig. 3, positive signals of Bjgpa2 (Fig. 3A) and Bjgpb5 (Fig. 3B) mRNAs were observed in the Hatschek pit, cerebral vesicle, endostyle, gill, hepatic cecum, hindgut, ovary, and testis, basically consistent with the qRT-PCR results. Notably, we clearly showed by both PCR analysis and in situ hybridization that Bjgpa2 and Bjgpb5 were coexpressed in the Hatschek pit, which is in line with the coexpressions of gpa2 and gpb2 detected in the presumptive area of the Hatschek pit in some developing embryos (9) but in contrast to the nonexpression of gpb5 in the pit of mature adults (12). Bjtshr was expressed predominantly in the endostyle, gill, hepatic cecum, hindgut, ovary, and testis (Fig. 3C). These findings indicate that both Bjgpa2 and Bjgpb5 as well as Bjtshr are expressed in a tissue-specific fashion.

Figure 3.

In situ hybridization of (A) Bjgpa2, (B) Bjgpb5, and (C) Bjtshr transcripts in the different tissues of amphioxus are shown. (a‒d) Relative expression of Bjgpa2 gene in the different tissues detected by in situ hybridization using Bjgpa2 antisense RNA probes. (e‒h) Control. In situ hybridization was performed using Bjgpa2 sense RNA probes. (i‒l) Relative expression of Bjgpb5 gene in the different tissues detected by in situ hybridization using Bjgpb5 antisense RNA probes. (m‒p) Control. In situ hybridization was performed using Bjgpb5 sense RNA probes. (q‒t) Relative expression of Bjtshr gene in the different tissues detected by in situ hybridization using Bjtshr antisense RNA probes. (u‒x) Control. In situ hybridization was performed using Bjtshr sense RNA probes. Scale bars, 150 μm. Cv, cerebral vesicle; E, endostyle; Gi, gill; H, Hatschek pit; Hc, hepatic cecum; Hg, hindgut; Mu, muscle; No, notochord. Ov, ovary; Te, testis.

Binding of rBjGpA2 and rBjGpB5 to rBjTSHR and rDrTSHR

The proteins rBjGpA2, rDrTSHα, rBjGpB5, rDrTSHβ, redBjTSHR, and redDrTSHR expressed in E. coli were purified by chromatography on Ni-NTA resin columns. The purified rBjGpA2, rDrTSHα, rBjGpB5, rDrTSHβ, redBjTSHR, and redDrTSHR were analyzed by SDS-PAGE, and each yielded a single band of ~14.8 kDa, ~14.2 kDa, ~13.6 kDa, ~18.4 kDa, ~54.4 kDa, and ~66.9 kDa, respectively, well matching the expected sizes (Fig. 4A‒4F). Western blotting showed that rBjGpA2, rDrTSHα, rBjGpB5, rDrTSHβ, redBjTSHR, and redDrTSHR reacted with the anti‒His-tag antibody, indicating that rBjGpA2, rDrTSHα, rBjGpB5, rDrTSHβ, redBjTSHR, and redDrTSHR were all correctly expressed.

Figure 4.

Interaction of BjGpA2, DrTSHα, BjGpB5, or DrTSHβ with BjTSHR or with DrTSHR. Purification and identification of recombinant (A) BjGpA2, (B) BjGpB5, (C) edBjTSHR, (D) DrTSHα, (E) DrTSHβ, and (F) edDrTSHR. The recombinant proteins were expressed in E. coli, and the purified proteins were analyzed by 12% SDS-PAGE and identified by Western blotting. Lane M, marker; lane 1, total cellular extracts from E. coli Transetta (DE3) without the expression vector; lane 2, total cellular extracts from E. coli Transetta (DE3) containing the expression vector before induction; lane 3, total cellular extracts from IPTG-induced E. coli Transetta (DE3) containing the expression vector; lane 4, recombinant proteins purified on Ni-NTA resin column; lane 5, Western blotting of corresponding purified recombinant proteins immunostained with anti‒His-tag antibody. (G and H) Binding of rBjGpA2, rDrTSHα, rBjGpB5, or rDrTSHβ to (G) redBjTSHR or (H) redDrTSHR analyzed by ELISA. Wells coated with redBjTSHR or redDrTSHR were individually incubated with different concentrations of biotin-labeled rBjGpA2, rDrTSHα, rBjGpB5, or rDrTSHβ. rTRX, instead of the recombinant proteins, was used as a negative control. The binding was detected using streptavidin–horseradish peroxidase. OD, optical density; WB, Western blotting.

An ELISA assay was conducted to determine the interaction of rBjGpA2 or rBjGpB5 with redBjTSHR or redDrTSHR and the interaction of rDrTSHα or rDrTSHβ with redBjTSHR or redDrTSHR. It showed that not only were rBjGpA2 and rBjGpB5 specifically bound to redBjTSHR and redDrTSHR but rDrTSHα and rDrTSHβ were also specifically bound to redBjTSHR and redDrTSHR (Fig. 4G and 4H). These findings indicate that BjGpA2 and BjGpB5, like zebrafish TSHα and TSHβ, could interact with TSHR.

Functional interaction of rBjthyrostimulin with BjTSHR

The recombinant protein rBjthyrostimulin (combination of GpA2 and GpB5) expressed in HEK293T cells was purified by chromatography on an Ni-NTA resin column. SDS-PAGE analysis showed that purified rBjthyrostimulin yielded a single band of ∼25 kDa, matching its expected size (Fig. 5A). Western blotting showed that rBjthyrostimulin reacted with the anti‒His-tag antibody, indicating that it was correctly expressed (Fig. 5A). MALDI/TOF MS analysis further verified the accuracy of the sequence of Bjthyrostimulin (Fig. 5B).

Figure 5.

rBjthyrostimulin functionally interacted with BjTSHR expressed on cultured eukaryotic cells. (A) The recombinant protein rBjthyrostimulin was expressed in HEK293T cells, and the purified proteins were analyzed by 12% SDS-PAGE and identified by Western blotting. Lane M, marker; lane 1, purified recombinant protein rBjthyrostimulin; lane 2, Western blotting of corresponding purified recombinant proteins immunostained with anti‒His-tag antibody. (B) Peptides detected in MALDI/TOF MS analysis. (C) Induction of CRE-driven luciferase activities in HEK293T cells transfected with BjTSHR by rBjthyrostimulin, rBjGpA2, rBjGpB5, and rTRX. (D) Induction of SRE-driven luciferase activities in HEK293T cells transfected with BjTSHR by rBjthyrostimulin, rBjGpA2, rBjGpB5, and rTRX. The data are expressed as the change in luciferase activity over basal activity. Vertical bars represent the mean ± SD (n = 3).

To examine the functional interaction of Bjthyrostimulin, BjGpA2, and BjGpB5 with BjTSHR, CRE and SRE reporter gene assays were performed. As reporter genes, CRE and SRE are used to activate PKA and PKC pathways, respectively. Graded concentrations of rBjthyrostimulin, rBjGpA2, rBjGpB5, or rTRX were applied to HEK293T cells, which were transfected with Bjtshr. The cells transfected with the empty vector exhibited no response to rBjGpA2, rBjGpB5, rBjthyrostimulin, or rTRX treatment (data not shown). By contrast, rBjthyrostimulin apparently triggered the receptor signaling pathways in a dose-dependent manner in the cells expressing BjTSHR through both CRE-Luc and SRE-Luc reporter systems (Fig. 5C and 5D). However, neither rBjGpA2 nor rBjGpB5 showed obvious effect. These findings show that only the GpA2/GpB5 recombinant rBjthyrostimulin, but not its individual subunits rBjGpA2 or rBjGpB5, was able to interact with the receptor BjTSHR in HEK293T cells and activate the downstream signal transduction.

Induction of TH by rBjthyrostimulin

Because rBjthyrostimulin showed functional interaction with BjTSHR, it was used for the following experiments. Adult zebrafish were given an IP injection with rBjthyrostimulin at doses of 0.1 μg/g or 1 μg/g, and plasma T4 level was measured 4 hours later. The survival rate of D. rerio was 100% at 4 hours after injection. As shown in Fig. 6, a significant increase (P < 0.05) in plasma T4 level was observed in the fish treated with different doses of rBjthyrostimulin compared with rTRX- or saline-injected fish. This indicates that Bjthyrostimulin could induce the production of T4 in zebrafish.

Figure 6.

Effects of administration of rBjthyrostimulin on TH secretion in zebrafish. Adult zebrafish were injected IP with various doses of rBjthyrostimulin. rTRX and saline-injected groups served as controls. Blood samples were collected, and plasma T4 levels were determined. Vertical bars represent the mean ± SD (n = 3). *P < 0.05.

Discussion

Amphioxus has been shown to have an organ homologous to the vertebrate pituitary, named the Hatschek pit (18, 39–41), as well as an organ homologous to the thyroid gland, named the endostyle (23, 24, 26, 27, 42, 43), suggesting the presence of a vertebratelike pituitary-thyroid axis at the histological level. Biochemical and molecular evidence also suggests that there is an active TH signaling pathway in amphioxus (22), such as the production of THs in the endostyle (23), deactivation of THs by the deiodinases (15, 25), and functional interaction of TH with the TH receptor (27). However, not much is known about the regulation of TH production in this evolutionarily important animal (44). We report the cloning and characterization of the amphioxus GpH Bjthyrostimulin, which consists of BjGpA2 (9, 12) and BjGpB5 (9, 11) subunits and the GpH receptor BjTSHR (28). Both the phylogenetic and modeling analyses presented here support the ancestry of BjGpA2, BjGpB5, and BjTSHR to the vertebrate TSHα, TSHβ, and TSHR, respectively. In particular, Bjgpb5, like rat tshβ, contains pituitary transcription factor binding sites such as MED-1, POU1F1, PROP1, Pitx1, Pitx2, and GATA2 in its upstream region and the nTRE at the front of the first intron, suggesting that the transcription of Bjgpb5 is possibly regulated similarly to rat tshβ. We also showed that Bjgpa2 and Bjgpb5 are coexpressed in the Hatschek pit of the adult amphioxus. Moreover, we demonstrated that recombinant GpA2 and BjGpB5 bound to BjTSHR and DrTSHR and that tethered rBjthyrostimulin activated both PKA and PKC pathways in HEK293T cells transfected with Bjtshr. In addition, rBjthyrostimulin significantly stimulated the production of T4 in zebrafish. Because genuine TSH is absent and thyrostimulin is the sole and only GpH in amphioxus, our data here in all possibility provide evidence that amphioxus thyrostimulin is a functional GpH that is capable of binding to TSHR and regulating TH production the same way as in the vertebrate TSH/TSHR system (Fig. 7), as initially proposed by Hsu et al. (45). The data also suggest an ancient origin of the TH signaling pathway within the chordate lineage.

Figure 7.

Diagram showing the hypothetical endocrine control system of THs in amphioxi and vertebrates. Based on the hypothesis that the amphioxus cerebral vesicle, Hatschek pit, and endostyle are homologous to the vertebrate brain, pituitary, and thyroid, respectively, the endocrine system in amphioxus is composed of the cerebral vesicle, Hatschek pit, and endostyle. In contrast, in vertebrates the hypothalamus-pituitary-thyroid axis controls TH functions.

Generally speaking, GpA2 and GpB5 require a compulsory heterodimeric association to play a role. This is clearly supported by our results showing that neither rBjGpA2 nor rBjGpB5 could activate the receptor TSHR in HEK293T cells, though they both could bind to TSHR in vitro. For heterodimerization, coexpression of GpA2 and GpB5 was necessary. We have shown in this study that both Bjgpa2 and Bjgpb5 are coexpressed in the Hatschek pit of amphioxus, meeting the requirement for heterodimerization. In addition, we also observed coexpression of Bjgpa2 and Bjgpb5 in other tissues, such as the hepatic cecum, hindgut, gill, ovary, and testis in amphioxus. Similarly, apart from the pituitary, coexpression of these subunits in vertebrates, including the lamprey, is too widespread in many other tissues (1, 9, 11, 12, 45–48). This widespread tissue distribution suggests that the role of thyrostimulin is pleiotropic, but whether these subunits form a heterodimer in these other tissues is unknown. In amphioxus, the dynamic expression patterns of gpa2 and gpb5, particularly in the cerebral vesicle, club-shaped gland, or Hatschek pit precursor, strongly suggest that they might be involved in the development of these structures in a heterodimerization-independent manner (9). Despite previous research, it is clear that our knowledge is rather limited regarding the roles and their mode of action of GpA2 and GpB5 in the extrapituitary tissues, which remains to be explored.

In summary, our study highlights the identification and characterization of the functional GpH thyrostimulin in amphioxus via showing coexpression of gpa2 and gpb5 in the Hatschek pit, interaction of recombinant GpA2 and GpB5 with TSHR, activation of TSHR by tethered thyrostimulin, and stimulation of T4 production by tethered recombinant thyrostimulin. These findings expand our understanding of the ancient origin of the TH signaling pathway within the chordate lineage.

Abbreviations:

Abbreviations:
 
• 3D

  three-dimensional

 
• Bjthyrostimulin

  thyrostimulin of Branchiostoma japonicum

 
• DIG

  digoxigenin

 
• DrTSHR

  Danio rerio TSH receptor

 
• DrTSHα

  Danio rerio mature TSHα

 
• DrTSHβ

  Danio rerio mature TSHβ

 
• edBjTSHR

  extracellular domain of BjTSHR

 
• edDrTSHR

  extracellular domain of Danio rerio TSH receptor

 
• ef1-α

  elongation factor-like α

 
• GpH

  glycoprotein hormone

 
• nTRE

  negative thyroid hormone response element

 
• IPTG

  isopropyl β-D-1-thiogalactopyranoside

 
• Ni-NTA

  Ni-nitrilotriacetic acid

 
• PBST

  10 nM of PBS supplemented with 1% Tween-20

 
• PKA

  protein kinase A

 
• PKC

  protein kinase C

 
• qRT-PCR

  semiquantitative real-time PCR

 
• rBjGpA2

  recombinant BjGpA2

 
• rBjGpB5

  recombinant BjGpB5

 
• rBjthyrostimulin

  recombinant Bjthyrostimulin

 
• rDrTSHα

  recombinant Danio rerio mature TSHα

 
• rDrTSHβ

  recombinant Danio rerio mature TSHβ

 
• redBjTSHR

  recombinant extracellular domain of BjTSHR

 
• redDrTSHR

  recombinant extracellular domain of Danio rerio TSH receptor

 
• TH

  thyroid hormone

 
• TSHR

  TSH receptor

 
• UTR

  untranslated region

Acknowledgments

Financial Support: This work was supported by grants from the Natural Science Foundation of China (31772442 to S.Z.) and the Blue Life Breakthrough Program (MS2017NO02) of the Laboratory for Marine Biology and Biotechnology, Qingdao National Laboratory for Marine Science and Technology, China (to S.Z.).

Disclosure Summary: The authors have nothing to disclose.

Author notes

These authors contributed equally.