Identification of a Parathyroid Hormone in the Fish Fugu rubripes


  • Dr Martin serves as a consultant. All other authors have no conflict of interest


A PTH gene has been isolated from the fish Fugu rubripes. The encoded protein of 80 amino acid has the lowest homology with any of the PTH family members. Fugu PTH(1–34) had 5-fold lower potency than human PTH(1–34) in a mammalian cell system.

Introduction: Parathyroid hormone (PTH) is the major hypercalcemic hormone in higher vertebrates. Fish lack parathyroid glands, but there have numerous attempts to identify and isolate PTH from fish.

Materials and Methods: Polymerase chain reaction (PCR) was performed with primers based on preliminary data from the Joint Genome Institute database. PCR amplification was performed on genomic DNA isolated from Fugu rubripes. PCR products were purified and DNA was sequenced. All sequence was confirmed from more than one independently amplified PCR product. Multiple sequence alignments were carried out, and the percentage of identities and similarities were calculated. An unrooted phylogenetic tree, using all the known PTH and PTH-related protein (PTHrP) amino acid sequences, was determined. Synthetic peptides were tested in a biological assay that measured cyclic adenosine 3′,5′-monophosphate formation in UMR106.1 cells. Rabbit polyclonal antisera specific for N-terminal human PTHrP and one rabbit polyclonal antiserum specific for N terminus hPTH were used to test the cross-reactivity with fPTH(1–34) in immunoblots.


Amajor change in calcium control between lower and tetrapod vertebrates is the evolution of the parathyroid gland and the hormone it secretes, parathyroid hormone (PTH). Amphibians were the first animals to have a distinct parathyroid gland. Despite the absence of a parathyroid gland, fish respond to changes in ambient calcium levels and regulate plasma calcium accordingly. Since the amino acid sequence of mammalian PTH was determined in 1970,(1) efforts have been made to find homologs of PTH in lower vertebrates. There have been a number of reports of the presence of immunoreactive PTH-like proteins in fish detected by antisera to mammalian PTH,(2) but a fish PTH homolog has not been identified.

Another calcium-regulating factor of higher vertebrates, parathyroid hormone-related protein (PTHrP) has been identified in lower vertebrates, in the circulation and tissues of a bony fish.(3) Further studies identified PTHrP in the tissues from cartilaginous fish,(4,5) as well as the earliest vertebrates, namely the lamprey.(6) The genes for PTHrP from Fugu rubripes(7) and Sparus aurata(8) were isolated and cloned. Although PTHrP is a hypercalcemic factor in mammals, as well as exerting a number of other actions,(9) its role(s) in fish has not been fully defined, with only one study(10) suggesting that PTHrP might be a hypercalcemic factor in fish.

In mammals, two G-protein-coupled, seven transmembrane receptors bind PTH and PTHrP. One of these, PTH1R, binds both PTH and PTHrP, whereas PTH2R favors PTH.(11,12) The fish PTH1R and PTH2R have been isolated and cloned from zebrafish,(11,12) as well as a novel fish PTH3R.(11) Zebrafish PTH2R was activated by hPTH but not by hPTHrP or fPTHrP, thus suggesting that PTH might occur in fish.(12) Recent studies indicate that 39 amino acids of tuberoinfundibular peptide (TIP39) may be an endogenous ligand for the PTH2 receptor(13) and that TIP 39 could be a distant relative of PTH and PTHrP.(14)

The structure of mammalian PTH and PTHrP, along with their receptor interactions, suggested that these two molecules share a recent evolutionary history. This makes fish, the oldest group of vertebrates, a pivotal group for the study of these two hormones. In the course of studies of the molecular evolution of the PTH family, we have isolated the DNA that encodes a fish homolog of PTH and demonstrated its activity in a mammalian cell-based assay.


Polymerase chain reaction and automated sequencing of Fugu PTH(1–80)

Primers for polymerase chain reaction (PCR) were designed from some preliminary data obtained from Joint Genome Institute ( using known PTH amino acid sequences. PCR primers were designed: forward primer, 5′-CAGTGAGTGAAGTCCAGCTCA-3′; and reverse primer, 5′-CTTCACTCCTGTGATTTGAGCA-3′. PCR amplification was performed on approximately 100 ng genomic DNA isolated from F.rubripes. PCR products were purified using a commercially available kit (UltraClean PCR Clean-Up DNA Purification Kit; Geneworks, Adelaide, Australia), and DNA was sequenced using the ABI Prism BigDye Terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer, Boston, MA, USA). All sequences were confirmed from more than one independently amplified PCR product.

Multiple sequence alignments were carried out using ClustalW(15) and displayed in Boxshade ( Percentage identities and similarities were calculated using Gap.(16) An unrooted phylogenetic tree, using all the PTH and PTHrP amino acid sequences currently lodged in SwissProt/TrEMBL, was determined.(17)

Synthetic peptides

The N terminus region of the protein fPTH(1–34) was synthesized using an Applied Biosystems 433A peptide synthesizer (Foster City, CA, USA) using Rink resin and Fmoc chemistry with a Fastmoc 0.1 Dry Conditions monitor. The completed peptides were simultaneously deprotected and cleaved from the resin (cleavage was carried out in 82.5% trifluoroacetic acid with Reagent K; Auspep, Parkville, Australia; consisting of 5% phenol, 5% water, 5% thioanisole, and 2.5% ethandithiol). The peptides were extracted from the resin in 20% (vol/vol) acetonitrile and 0.1% (vol/vol) trifluoroacetic acid and dried down. The peptides were purified by sequential ion-exchange chromatography (MacS) with 20% (vol/vol) acetonitrile (Mallinkrodt HPLC grade; Sigma, St Louis, MO, USA) and 0.1% (vol/vol) trifluroacetic acid using a gradient of 0–1 M guanidine hydrochloride. The fractions were checked by mass spectrometry and pooled. The pooled samples were purified by preparative low pressure reversed phase chromatography (25 × 800 column, C18, 250 Å, Vydac resin) with an acetonitrile gradient in the presence of 0.1% (vol/vol) trifluoroacetic acid. Mass spectrometry verified purity of the synthetic peptide (PerSeptive Biosystems Voyager DE with Data Explorer Software Version 4.0). The synthetic fPTH(1–34) and fPTHrP(1–34) were analyzed by nanospray mass spectrometry (Applied Biosystems QSTAR Pulsar).

Biological activity

The PTH-like bioactivity was assayed by measuring cyclic adenosine 3′,5′-monophosphate (cAMP) production in UMR106.01 cells, which were grown to 90% confluence. Before assaying, cells were washed once with PBS and equilibrated for 20 minutes in medium containing 0.1% bovine serum albumin (BSA) and 1 mM-isobutylmethylxanthine (Sigma). Cells were subsequently stimulated at 37°C for 10 minutes in the absence and presence of increasing hormone concentrations. The cells were washed once with PBS, and cAMP was extracted with 1.5 ml acidified ethanol. Samples were evaporated to dryness, reconstituted in assay buffer, and assayed by a specific cAMP radioimmunoassay.(18)


Eight rabbit polyclonal antisera specific for N-terminal hPTHrP and one rabbit polyclonal antiserum specific for N terminus hPTH were used to test the cross-reactivity with fPTH(1–34). Seven antisera (R88, R1904, R1942, R87, R1348, R196, R212) were raised against hPTHrP(1–14), and one antiserum (R190) was raised against hPTHrP(1–141). The anti-PTH polyclonal was raised against hPTH(1–34) (BioGenex, San Ramon, CA, USA). The anti-hPTHrP antisera have been used successfully in immunohistochemistry and Western blotting with tissues taken from bony and cartilaginous fishes.(3,4) The anti-PTH antiserum has been used in immunohistochemistry of human parathyroid material and Western blotting.(3) Ten, 25, and 50 μg of Fugu PTH were spotted along side a positive control [i.e., hPTHrP(1–34)] on nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany).

After an initial incubation period with Tris-buffered saline, 0.2% Tween, and 5% skim milk powder to block nonspecific binding sites, the nitrocellulose membrane was incubated with primary polyclonal antibody, followed by secondary antibody (anti-rabbit serum conjugated to horseradish peroxidase; Dako, Carpenteria, CA, USA) incubation. Three washes on a shaker table were done between incubations. A BM chemiluminescence blotting system (Roche Applied Sciences, Mannheim, Germany) was used to detect specific dot blots.


Using primers designed from the database sequence, PCR amplification of fugu genomic DNA yielded a 244-bp product (Fig. 1B). The nucleic acid sequence of fPTH is shown with forward and reverse primers, with the stop codon identified in Fig. 1A. This sequence, when translated, predicts an 80 amino acid sequence, suggesting that this is the entire coding region of the fPTH gene (Fig. 2). This has very low overall homology to either cPTH or hPTH (identity: 36% and 32%, respectively). The similarity between cPTH and hPTH is 49%, whereas the similarity between fPTH and hPTH is 44%. The sequence identity between fPTHrP and hPTHrP is 53%, and the similarity is 64%. If only the N-terminal 34 amino acid region is considered, fPTH(1–34) has 56% identity to cPTH and 53% to hPTH, but the similarity is 68% to cPTH and 65% to hPTH. By comparison, sequence identity in the N-terminal regions of fPTHrP and hPTHrP is 59%, and similarity is 77%. When the amino acid sequences of the PTHs and PTHrPs are represented in an unrooted phylogenetic tree (Fig. 3), the sequences segregate into two distinct groups. Tetrapod PTHrPs form a tight cluster and reflect the very high level of homology among the amino acid sequences, and the majority of tetrapod PTHs form another tight cluster. Rat and mouse PTH group together on a separate branch from other tetrapod PTHs. fPTH and fPTHrP have a lower number of amino acid substitutions between them than do any of the other PTHs and PTHrPs from the same species.

Figure FIG. 1..

(A) Nucleic acid sequence of fugu PTH gene with the positions of the forward and reverse PCR primers highlighted (bold and boxed) and the stop codon underlined. (B) Agarose gel electrophoresis of 244-bp fugu PTH PCR products. M, pGEM DNA molecular weight marker; F, fugu genomic DNA template; N, negative control template (water).

Figure FIG. 2..

Multiple sequence alignment of the deduced amino acid sequence of fugu PTH(1–80) chicken PTH, equine PTH, canine PTH, bovine PTH, cat PTH, porcine PTH, rat PTH, mouse PTH, and human PTH above fugu PTHrP, chicken PTHrP, and human PTHrP. Identical residues are shaded in black, and similar residues are shaded gray within PTH and PTHrP.

Figure FIG. 3..

The relationship between PTH and PTHrP proteins from higher vertebrates and fish in an unrooted tree inferred from the amino acid sequences. The length of each branch is proportional to the number of substitutions between the two endpoints, as indicated by the scales.

The purified synthetic fPTH(1–34) had a mass of 4154.75 Da, and fPTHrP(1–34) had a mass 4126.45 Da. The fully automated synthesis of fPTHrP(1–34) resulted in the quantitative transamidation of Asp10 by piperidine, adding 67 Da ( = 67&AvgMass =67&Margin=0), but this was overcome by the manual addition of Asp10. The purity of synthetic peptides of fPTH(1–34) and fPTHrP(1–34) was checked with mass spectrometry, and these traces show a single dominant peak, indicating that the peptides are full length and are between 90% and 95% pure (data not shown).

fPTH(1–34) stimulated cAMP formation in UMR 106.01 cells (ID50 = 17 ± 3.6 nM, n = 4) in a dose-dependent manner, with a potency that was consistently less than that of fPTHrP(1–34) (ID50 = 1.4 ± 0.48 nM, n = 4; Fig. 4). fPTH(1–34) is also less potent than hPTH(1–34) and hPTHrP(1–34) (Fig. 4). However, the maximum amplitude of response to fPTH(1–34) was significantly greater than that achieved with the highest concentrations of hPTH, hPTHrP, or fPTHrP (Fig. 4). When UMR 106.01 cells were co-incubated with 100 nM fPTH(1–34) and a maximal dose of 10 nM of fPTHrP(1–34), no further increase in cyclase activity was observed, consistent with the hypothesis that the two peptides act through the same receptor, namely PTH1R (data not shown).

Figure FIG. 4..

The cAMP response to hPTHrP(1–34), hPTH(1–34), fPTHrP(1–34), and fPTH(1–34) in UMR106.01 cells. Results shown as the mean ± SE of triplicate determinations.

When increasing concentrations of hPTHrP(7–34) were co-incubated with either10 nM fPTH(1–34), 1 nM hPTH(1–34), 0.5 nM fPTHrP(1–34), or 0.5 nM hPTHrP(1–34), partial inhibition of the cAMP response was evident with all peptides (Fig. 5),(19) providing further evidence that fPTH(1–34) acts through the PTH1R.

Figure FIG. 5..

Antagonist activity of hPTHrP(7–34) against fPTHrP, fPTH, hPTHrP, and fPTH was measured and shown as mean ± SE of triplicate determinations. All incubations contained increasing concentrations of hPTHrP(7–34) either in the absence of other treatment or with 1 nM hPTH(1–34), 0.5 nM hPTHrP(1–34), 0.5 nM fPTHrP (1–34), or 10 nM fPTH(1–34) for 10 minutes, and the cAMP response was determined.

The immunoblots showed that fPTH(1–34) does not cross-react with any of the rabbit polyclonal antisera raised against human PTHrP(1–14) or the rabbit polyclonal antiserum that is raised against human PTH(1–34) (data not shown).


The amino-terminal region of fPTH identified in this work is homologous with the N terminus of tetrapod PTH and with PTHrP from both mammals and fish. Eighteen of the first 34 amino acids of fPTH are identical to those in hPTH, whereas 20 of the first 34 amino acids of fPTHrP and hPTHrP are identical and only 14 of the first 34 amino acids of fPTH are identical to those in fPTHrP. This suggests that the fish sequence that has been isolated is more like PTH than PTHrP. In the fPTH amino acid sequence, after the first 34 amino acids, there is no significant homology to either hPTH or cPTH, indicating weak evolutionary pressure to conserve the C terminus of the PTH molecule. This is in contrast to the homology of the fish PTHrP amino acid sequences with hPTHrP, as well as that seen among the tetrapod PTH members. This is consistent with an evolutionary pressure to conserve PTH within tetrapods and also with the view that PTH fulfills a fundamental role in higher vertebrates.

Phylogenetic analyses of the amino acid sequences supports the proposal that the PTH and PTHrP genes may have been duplicated in an event that predates the radiation of fishes. Certainly the distance between the two fugu proteins indicate that they are closer to the divergence of these two genes than in any of the higher vertebrates.

The biological assay data are consistent with an action of fPTH through the PTH1R, as is the case with hPTH and hPTHrP. The consistent finding that fPTH can stimulate cAMP formation to a greater maximum level than any of the other PTH peptides might indicate activation of another receptor, a possibility that will require further investigation. Structural analyses using nuclear magnetic resonance and X-ray crystallography, together with extensive studies of cross-linking of PTH and PTHrP analogs to the PTH1R, are all in accord with a model of PTH and PTHrP binding through participation of residues within the sequence between residue 15 and 31. There are a number of structural aspects of the N-terminal portion of the fPTH molecule that fit comfortably with what is known of PTH1R interactions. Photoaffinity cross-linking studies have identified certain residues that are crucial for binding of PTH and PTHrP to PTH1R. Residues Phe 23, Leu24, and Ile28 are close to the receptor, and photolabeling of Leu24 caused a 10-fold reduction in binding.(20) When this is considered together with the fact that residues Phe23, Leu24, and Ile28 are intolerant to substitution by polar residues,(21,22) the fPTH sequence is in keeping with that of other PTH and PTHrP homologs. Furthermore, the Arg20 and Leu24 of fPTH are in accord with the strict conservation of these residues throughout all known PTH and PTHrP sequences. On the other hand, residues Lys26, Gln29, and Asp30 can be mutated without effect on receptor binding.(21)

The potency of fPTH(1–34) in the present work has consistently been about one-tenth that of human PTH or PTHrP. This might be because of subtle conformational changes resulting from the different sequence in the C-terminal portion of fPTH(1–34), for example, the fact that all of residues 26, 27, 29, and 30 are variations from those positions in the other PTH/PTHrP homologs. While the reduced potency of fPTH in activating adenylate cyclase in a mammalian target cell is interesting, it remains to be discovered what is the true target in the fish and what is the potency of the peptide. It may be relevant to note that in the PTH3R discovered in zebrafish,(11) activation by hPTH was consistently 20-fold less potent than that either by hPTHrP or fPTHrP.

Expression studies of fPTH(1–80) have not yet been undertaken, but it is unlikely that fPTH(1–80) reported here is a pseudogene, because F.rubripes is a model compact vertebrate genome. Unlike the human genome, there is a very low abundance of repetitive DNA in the F.rubripes genome.(23)

Immunologically, fPTH(1–34) is not recognized by any of the hPTH or hPTHrP antisera, even at high concentrations of peptide and antisera. It is very unlikely that any antisera raised to hPTHrP, hPTH, and bovine PTH could localize the fish PTH homolog in fish tissues, because the N terminus of fPTH is the portion of the protein that is the most highly conserved. With the development of homologous antisera and probes, the investigation of the expression of PTH and its roles in fish and other lower vertebrates can begin.


This study was supported by the National Health and Medical Research Council of Australia.