Cell Transfection.

We found that rxLeptin activated the mouse LRb with a potency similar to recombinant human leptin (ANOVA, F = 25.71, P < 0.0001; EC₅₀ = ≈5 nM; Fig. 2C Left). Similarly, we found that both rxLeptin and recombinant human leptin activated the frog LR expressed in COS-7 cells with similar potencies in the low nanomolar range (ANOVA, F = 6.742, P = 0.014; EC₅₀ = ≈5 nM; Fig. 2C Right). There were no statistically significant differences in the responses between replicate experiments.

Tissue Distribution of Frog Leptin and LR mRNAs.

Using quantitative RT-PCR (RTqPCR), we found leptin mRNA in diverse frog tissues, including fat and liver, where leptin is primarily expressed in mammals and chicken (1, 21), but also in brain, pituitary gland, heart, gastrointestinal tract, lungs, kidney, and gonads (Fig. 3A); we obtained similar results with semiquantitative RT-PCR (data not shown). Expression levels varied among tissues, with the highest in the brain and heart. We detected leptin mRNA in unfertilized oocytes and whole embryos starting at Nieuwkoop–Faber (NF) stage 23 (Fig. 3B).

Open in a separate window

Fig. 3.

Expression of leptin and LR mRNAs in the frog. (A) The tissue distribution in juvenile frogs of leptin and LR mRNAs was analyzed by quantitative RT-PCR (see Methods). Leptin and LR mRNA levels were normalized to the expression of the ribosomal protein L8 gene (rpL8). (B) The developmental expression of leptin mRNA in X. laevis was analyzed by semiquantitative RT-PCR (see Methods).

We also analyzed the tissue distribution of LR mRNA by RTqPCR. In mammals, alternative splicing results in at least six LR isoforms that differ in their C termini (6). Only one of these isoforms, the LRb, possesses the C-terminal region known to be necessary for intracellular signaling. We therefore used two primer sets, one to target the extracellular ligand binding region (expected to be present in all forms of the frog LR, exons 14/15) and the other to target the intracellular C-terminal region known to be necessary for signaling in mammals (exons 25/26). LR mRNA is widely distributed in the frog and is expressed in tissues where leptin is not (e.g., skin, pancreas, spleen, and skeletal muscle). Transcripts containing exons 25/26 were expressed in all tissues analyzed (except blood) and variation in their relative levels among tissues paralleled that of transcripts containing exons 14/15 (only data for RTqPCR analysis of exons 25/26 is shown in Fig. 3A; similar results were obtained with semiquantitative RT-PCR; data not shown). We found the highest LR mRNA expression in the pituitary followed by the brain. High levels of expression were also observed in skin, muscle, and testis; lower but detectable levels of expression were found in different visceral organs and the ovary (note the different scales of the graphs in Fig. 3A Left and Right).

Food Intake Assays.

Intracerebroventricular injection of rxLeptin was potently anorexigenic in juvenile frogs, with a dose of 0.02 ng/g of body weight causing maximal inhibition of food intake (n = 4; ANOVA, F = 11.44, P = 0.0002; Fig. 4Left). rxLeptin (i.c.v.) had no effect on foraging time in early prometamorphic tadpoles (n = 5–6 tadpoles per treatment) but produced a dose-dependent inhibition of feeding in midprometamorphic (Gosner stage 36–37) tadpoles (n = 5–6 tadpoles per treatment; ANOVA, F = 31.58, P < 0.0001; Fig. 4). Stage-dependent effects of i.c.v. rxLeptin injections on appetite are further supported by our finding that daily i.p. injections of rxLeptin for 1 week did not affect body weight (relative to saline-injected controls) in fed or fasted early prometamorphic tadpoles (data not shown). However, rxLeptin injections caused significant body weight reductions in midprometamorphic tadpoles [fed: saline 3.73 ± 0.48 g, rxLeptin 3.24 ± 0.35 g, P = 0.054 (t test); fasted: saline 2.90 ± 0.35 g, rxLeptin 2.41 ± 0.48 g, P = 0.043; n = 7–9 per group].

Open in a separate window

Fig. 4.

Effects of i.c.v. injection of rxLeptin on time spent foraging in early prometamorphic or midprometamorphic S. hammondii tadpoles and on meal size in X. laevis juveniles. Letters indicate Duncan’s pairwise differences among treatments (P < 0.05).

Structural Characterization of Frog ob and obr Genes.

The predicted leptin protein of X. laevis shares ≤35% sequence similarity with mammalian leptins, which is in striking contrast to the 95% sequence similarity reported between the mouse and chicken ob genes (greater similarity than that among several mammalian leptins) (21, 23). Despite the low sequence similarities between the ectotherm and mammalian leptins, molecular modeling shows a conserved four-helix structure that is not only characteristic of mammalian leptins but also of other long-chain helical cytokines (24). Thus, despite divergence in amino acid sequence among species, a conserved tertiary structure has been maintained by natural selection and is presumably constrained by the structure of the receptor binding pocket. Leptins of frog and human origin exhibited similar potencies on mouse and frog LRs when tested in transfection assays, and human and frog leptins had similar potencies in inhibiting food intake in juvenile X. laevis (E.J.C., unpublished data). Based on these structural and functional similarities with mammals, we have designated the gene that we isolated from frog leptin.

Vertebrate leptins, including frog and pufferfish, differ from other long-chain helical cytokines in that they possess two conserved cysteine residues predicted to form disulfide bonds known to be required for bioactivity of human leptin (24, 25). The frog and pufferfish sequences also have conserved Leu (163) and Glu (164) residues at the C terminus of helix D that contribute to the unique conformation of the leptin protein (24). Unlike the pufferfish, frog leptin has regions of higher conservation (50–70%) with mammalian leptins that correspond to receptor binding domains (e.g., midregions of helix A and C, AB loop, and the N terminus of helix D; ref. 26). However, there are two stretches of 16–20 aa in frog leptin that have almost no sequence similarity to mammalian leptins, including the C terminus of the helix A/AB loop region and the CD loop/helix E region (which does not interact with the LR but may play a role in appetite regulation; ref. 26). The conspicuous similarity of residues 61–66 (GLDFIP; Fig. 1B) in the AB loop among mammalian and frog leptins suggests that this sequence may be important for receptor binding.

The predicted frog LR protein shares only 37% amino acid sequence similarity with human LRb, but the frog gene structure is similar to mammalian obr genes, and phylogenetic analysis clearly placed the frog gene within a clade of mammalian and chicken obr genes. Stretches of highest sequence similarity correspond to regions that are known to be important for mammalian LR signaling. The sequences of the corresponding ligand binding and transmembrane domains show greater conservation, as do specific residues in the C-terminal domain of the mammal LRb shown to be necessary for intracellular signaling (6). Our in vitro transfection assays confirm that leptin binding to the frog LR activates STAT-3 signaling. The number of exons encompassing the 5′ UTR is greater in frog compared with human (8 vs. 2 exons). In the human, the first two exons also encode a protein (leptin receptor gene-related protein; ref. 27), which has been identified in Xenopus (GenBank accession no. AW766133; the first four exons of the frog LR gene overlap with the N-terminal sequence of the leptin receptor gene-related protein). Based on conserved structure and function, we conclude that the frog gene that we isolated is homologous to mammalian LR.

Tissue Distribution of Leptin and LR mRNAs.

The tissue distribution of leptin mRNA is broader in the juvenile frog compared with mammals. It is noteworthy that the expression levels in liver and fat, the two primary sites of leptin expression in mammals were lower than those in other frog organs that express leptin.

We found that LR transcripts containing exons that code for the C-terminal region of the molecule necessary for intracellular signaling are widely distributed, which suggests that leptin may have widespread and diverse physiological roles in amphibians. High expression levels were detected in the frog brain, which is a main site of LRb expression in mammals (6). We detected highest levels of expression of LR mRNA in the pituitary gland, which may be a target for leptin action in mammals (28). The relative levels of LR transcripts detected with both RTqPCR assays (targeting the ligand binding domain and the C terminus) were comparable among tissues. The LR was expressed in all of the same organs as leptin, suggesting that leptin may have a paracrine function in these tissues. By contrast, an exclusively hemocrine action of leptin is predicted in tissues where only the receptor is expressed.

Effect of rxLeptin on Food Intake Throughout Amphibian Development.

Leptin had potent inhibitory effects on appetite in midprometamorphic tadpoles and juvenile frogs. Murine leptin inhibited food intake in a fish (29), but our results in the frog show that a homologous leptin in an ectotherm has anorectic activity. Repeated rxLeptin injections also caused food-deprived midprometamorphic tadpoles to lose more weight than vehicle-injected controls, suggesting that leptin also increases energy expenditure as it does in mammals (2). Thus, the function of leptin as a regulator of food intake and energy balance evolved before the emergence of amniote vertebrates.

The anorectic action of leptin in the amphibian does not develop until midprometamorphosis, which corresponds to the formation of discrete fat bodies (E.J.C., unpublished data) and the maturation of the neuroendocrine system under the influence of thyroid hormone (22). Earlier, we showed that inhibitory inputs to hypothalamic feeding control centers involving corticotropin-releasing factor (CRF) neurons are formed at this stage of development (30). The inhibitory effect of leptin on food intake is at least partially mediated by CRF signaling in mammals (31); therefore, it is likely that the maturation of CRF neurons in brain regions controlling appetite is linked to the development of leptin’s effects on food intake in amphibians.

RTqPCR Analysis of Frog Leptin and LR mRNAs.

We used RTqPCR to determine the tissue distribution of leptin and LR mRNAs and to compare relative expression levels in juvenile frogs. We isolated RNAs from frog tissues using the TRIzol reagent (Invitrogen) and synthesized cDNA using Superscript II (Invitrogen) following the manufacturer’s instructions. We developed TaqMan assays and analyzed samples on an ABI 7500 fast real-time PCR machine using TaqMan Universal PCR Master Mix (Applied Biosystems). The primer/probe sets were designed to span exon/intron boundaries and are described in supporting information. Standard curves were generated by using cDNAs from tissues that exhibited the highest expression level for each gene. Leptin and LR mRNAs were normalized to the level of L8 mRNA.

For semiquantitative RT-PCR, we used HotStar Taq polymerase (Qiagen, Valencia, CA) following the manufacturer’s protocol. Oligonucleotide primers and the PCR conditions used are described in supporting information.

Production of Recombinant Frog Leptin.

We produced recombinant X. laevis leptin in E. coli using the pET 151/D-TOPO expression vector (Invitrogen) transformed into BL21 Star (DE3) cells (Invitrogen). We purified rxLeptin from inclusion bodies by fractionating bacterial lysate on a 12% SDS/PAGE gel, electroeluting into 20 mM Tris base/150 mM glycine/0.01% SDS (pH 7.0), and dialyzing overnight against 10 mM ammonium bicarbonate.

Cell Transfection.

We first determined whether rxLeptin could activate the mouse LRb in transfection assays. We then used transfection assay to test whether the frog LR that we isolated was functional. We cotransfected COS-7 cells (40,000 cells per well; 24-well culture plates) with either the pcDNA3.1 mouse or frog LR-receptor expression vector (200 ng) and a STAT-3 responsive luciferase reporter construct (GAS; 100 ng; ref. 37) using FuGENE 6 transfection reagent (Roche Biosciences). We also transfected cells with Renilla luciferase reporter construct (2 ng; Promega) for normalization of transfection efficiency. After transfection, cells were serum-deprived for 14–15 h before the addition of different concentrations of either recombinant human or rxLeptin. After 6 h, we harvested cells and measured both firefly and Renilla luciferase activities using a dual-luciferase reporter assay (Promega).

Food Intake Assays.

We administered i.c.v. injections of rxLeptin (0.2, 2, 20 ng/g body weight in PBS) into the third ventricle of juvenile X. laevis and monitored food intake as described in ref. 38. We also tested the effects of rxLeptin on food intake in amphibian larvae using tadpoles of the Western spadefoot toad Spea hammondii in which food intake (foraging) can be easily quantified. Using the protocol of Crespi and Denver (30), we tested animals at two stages of metamorphosis, early prometamorphosis when hind-limb buds are first visible (Gosner stage 30) and midprometamorphosis (Gosner stage 36–37). We measured the percent time foraging within a 2-min observation period and used ANOVA to detect treatment effects. We also injected 2 μg of rxLeptin/50 μl of PBS into the peritoneum (i.p.) every other day for 6–7 days in daily fed and food-deprived tadpoles to assess the effects of prolonged rxLeptin treatment on growth and development rate (n = 8 per treatment) during early and midprometamorphic stages. We analyzed the ratio of hind limb to body length with ANOVA and ln-transformed Gosner stage with ANCOVA (body weight as a covariate). We compared posttreatment body weight of saline- and rxLeptin-injected tadpoles with t tests. All animal experiments were conducted following institutional guidelines for animal care and use.

We thank Georgina Mang, Hynsuk Roh, Graham Boorse, Keith Williamson, Aaron Hoffman, and Miranda Hicks-Courant for technical assistance and discussions. The GAS-luciferase reporter and mouse LR expression vectors were generously provided by Dr. Martin Myers, and Dr. Craig Jaffe provided human recombinant leptin. This work was supported by National Science Foundation Grant IBN 0235401 (to R.J.D.). This work used the Molecular Core of the Michigan Diabetes Research Training Center, funded by National Institutes of Health Grant NIH5P60 DK20572.