The peptide secreted at the water to land transition in a model amphibian has antioxidant effects

Abstract

In addition to the morphophysiological changes experienced by amphibians during metamorphosis, they must also deal with a different set of environmental constraints when they shift from the water to the land. We found that Pithecopus azureus secretes a single peptide ([M + H]+ = 658.38 Da) at the developmental stage that precedes the onset of terrestrial behaviour. De novo peptide and cDNA sequencing revealed that the peptide, named PaT-2, is expressed in tandem and is a member of the tryptophyllins family. In silico studies allowed us to identify the position of reactive sites and infer possible antioxidant mechanisms of the compounds. Cell-based assays confirmed the predicted antioxidant activity in mammalian microglia and neuroblast cells. The potential neuroprotective effect of PaT-2 was further corroborated in FRET-based live cell imaging assays, where the peptide prevented lipopolysaccharide-induced ROS production and glutamate release in human microglia. In summary, PaT-2 is the first peptide expressed during the ontogeny of P. azureus, right before the metamorphosing froglet leaves the aquatic environment to occupy terrestrial habitats. The antioxidant activity of PaT-2, predicted by in silico analyses and confirmed by cell-based assays, might be relevant for the protection of the skin of P. azureus adults against increased O₂ levels and UV exposure on land compared with aquatic environments.

1. Introduction

Amphibians are characterized by a bi-phasic ontogeny marked by the metamorphosis from egg to tadpole and finally to the adult form. For most species, the transition to the adult life is also a transition from an aquatic to a terrestrial environment with many associated challenges. For example, the concentration of O₂ in air is much higher than that in water and terrestrial habitats lack an important ultraviolet (UV) radiation barrier, the water column [1]. Thus, in addition to the profound morphophysiological changes of the metamorphosis, amphibians must deal with a more oxidizing environment, since O₂ availability and UV radiation are closely related to the generation of reactive oxygen species (ROS) [2].

In addition to exogenous sources, aerobic organisms naturally produce ROS as products of aerobic metabolism. These reactive species act as signalling molecules in key physiological processes. Levels of ROS are managed by endogenous antioxidants in the so-called redox metabolism, which allows ROS to exert their signalling functions without detrimental effects. When this balance is disrupted, excessive ROS cause oxidative damage to biomolecules and ultimately oxidative stress. Disruptions of the redox metabolism and oxidative stress are associated with the pathophysiology of many human diseases, including neurological disorders [3].

The core functions of endogenous antioxidant systems are performed by antioxidant enzymes, such as glutathione peroxidase, peroxiredoxins, superoxide dismutase and catalase. The activity of enzymatic antioxidants are supported by non-enzymatic elements, such as glutathione, thioredoxin and NADPH. In the skin of amphibians, however, secreted ROS-scavenging peptides have been hypothesized to play key roles in managing ROS levels, representing another type of antioxidant defense [4,5]. These peptides are part of a complex mixture of compounds secreted by specialized glands in the skin of amphibians. The cutaneous secretion of amphibians plays a central role in their ecophysiology and have been extensively explored for biotechnological applications. One of the most studied biological roles of peptides derived from anuran skin secretion is their antimicrobial activity. While many compounds with other diverse pharmacological effects have been identified in amphibian's skin [6,7], there are still some with unknown activity.

Tryptophyllins are a large and structurally heterogeneous family of peptides in the skin secretion of Phyllomedusinae anurans [8–10]. These peptides are divided into three groups with common structural motifs but poorly understood biological roles [8]. In many cases, the activity of a novel amphibian skin peptide might be predicted by identifying counterpart peptides in the brain and gut of vertebrates. This interplay between peptides produced in three different tissues is attributed to an axis of communication between brain, gut and skin, named the brain–gut–skin triangle [11]. In the case of tryptophyllins, however, the lack of a mammalian peptide hormone counterpart hinders our understanding of the actual ecophysiological relevance of tryptophyllins. Therefore, tryptophyllins have been screened for diverse potential biological activities. The findings from these screens indicate that, while minor, some tryptophyllins have myotropic, vasorelaxant, vasoconstrictive, opioid-like, antiproliferative and antimicrobial activities [8,10,12]. Thus, a common prominent activity of peptides from this family remains to be discovered, which might shed some light on the biological significance of tryptophyllins for anurans and point to novel directions for their biotechnological application.

Similar to other organs and systems (e.g. maturation of lungs and gastrointestinal system), the integument of anurans is adjusted for the transition to land during metamorphosis to prevent water loss and protect against biotic and abiotic factors [13]. Considering that secreted peptides would be especially important during the transition from the aquatic to the terrestrial environment, we aimed to identify the first peptide or set of peptides secreted along the ontogeny of Pithecopus azureus. We found that P. azureus secretes a single peptide, called PaT-2, at the developmental stage that precedes the onset of terrestrial behaviour. Here, we describe the isolation of PaT-2 as a mature peptide, as well as its cDNA sequence, from the skin secretion of P. azureus. We found that PaT-2, a T-2 group tryptophyllin, was the first peptide detected in the skin of P. azureus along its ontogeny. Considering the lack of consensus on the activity of tryptophyllins and the content of Phe, Pro and Trp, which are common in antioxidant peptides [14], we tested whether PaT-2 presents antioxidant activity. PaT-2 and analogues were synthesized to evaluate the positional effect of relevant amino acids and assessed using in silico assays followed by cell-based assays.

2. Material and methods

(a) . Animals

The species studied in this work is P. azureus (Cope, 1862) [15], which was previously called Phyllomedusa azurea (Cope, 1862) [16]. Fertilized eggs (Brasília, Federal District, Brazil) of P. azurea and adult frogs (Mimoso, Goiás, Brazil) were collected in the Brazilian region and manipulated according to the rules of the Instituto Brasileiro do Meio Ambiente e dos Recursos Renováveis, IBAMA, under the license number 31066-1. After egg hatching, tadpoles were maintained in tanks with tap water treated with water conditioner (Prime Seachem) and fed with fish feed. The developmental stages of tadpoles were defined according to morphological changes suggested by Gosner [17]. Skin secretions were obtained from three individuals at each developmental stages (adult frogs and tadpoles at stages 37 and 42) by electrical stimulation and washing with ultrapure water. The secretions were then transferred to tubes, frozen and lyophilized. The dorsal skin tissues used for MALDI mass spectrometry imaging (MALDI-MSI) and cDNA sequencing were dissected after euthanasia with 2% liquid lidocaine hydrochloride solution added to the water surrounding tadpoles or injected directly into the brain of adult frogs. Frogs submitted to MALDI MSI and cDNA sequencing were not electrically stimulated.

(b) . MALDI mass spectrometry imaging

Dorsal skin tissues of P. azureus tadpoles at developmental stages 37 and 41 were immediately placed on a MALDI target plate after dissection. After air drying for 12 h, a solution of α-cyano-4-hydroxycinnamic acid (dissolved in 50% acetonitrile, 40% Milli-Q water and 10% trifluoroacetic acid, v/v/) was deposited on skin tissues and left to dry for 20 min. Mass spectrometry acquisitions were performed in a UltraFlex III mass spectrometer (Bruker Daltonics) operating in the positive reflective mode, with constant laser (solid state) intensity and 200 µm of resolution. The mass range analysed was m/z 600–5000 Da. The mass spectrometer was calibrated using a drop of calibration solution on a fragment of skin tissue corresponding to the same developmental stage under investigation. Equipment control, automatic scanning steps and analyses were performed using FlexImaging 4.0 software.

(c) . Ultra-fast liquid chromatography

The skin secretions of P. azurea were dissolved in ultrapure water and injected into an LC-20AD chromatography system (Shimadzu) and separated through a Shim-pack XR-ODS 2.0 mm i.d., 50 mm column (Shimadzu). The binary mobile phase system used ultrapure water (solvent A) and HPLC-grade acetonitrile (solvent B) both containing 0.1% (v/v) trifluoroacetic acid. A linear gradient of solvent B ranging from 5 to 75% for 15 min and 75 to 95% for 5 min was run over with a flow rate of 0.4 ml min⁻¹. Chromatographic fractions were monitored at 216 and 280 nm, collected in tubes, frozen and dried under vacuum. The experiment involving natural, synthetic and analogues (containing d-amino acids) of PaT-2 peptide was carryout as described above, but the gradient of solvent B was of 5–95% during 20 min.

(d) . MALDI-MS/MS analysis of chromatographic fractions

Dried chromatographic fractions were dissolved in ultrapure water, mixed with α-cyano-4-hydroxycinnamic acid (dissolved in 50% acetonitrile, 40% Milli-Q water and 10% trifluoroacetic acid, v,v) in a proportion of 1 : 3 (v/v), spotted on a MALDI-target plate and submitted to air drying. Acquisitions were performed in a MALDI UltraFlex III mass spectrometry (Bruker Daltonics) operating in the reflective positive mode, controlled by FlexControl 4.0 software. The mass range analysed was m/z 500–5500 Da. The precursor ions were selected and fragmented using the LIFT mode (MS/MS). Mass spectrometer was calibrated using calibration solution before analysis. Results were analysed using FlexAnalysis 3.5 software.

(e) . cDNA sequencing

The dorsal skin of eight adults of P. azurea had their total RNA extracted using the Trizol reagent (Invitrogen) according to manufacturer's instructions. Isolated RNA samples were quantified using Quant-iT RiboGreen RNA Reagent and Kit (Invitrogen). An aliquot of each sample containing 150 µg of total RNA was submitted to the transcript sequencing procedures by 454 Life Science/Roche Company (EUA) using the GS-FLX platform. The Geneious Prime software was used for blast search by entering the unformatted peptide sequence searching across entire sequenced reads.

(f) . Peptide synthesis

PaT-2 peptide and analogues were synthesized according to a solid phase synthesis technique adapted from [18] using Rink-amide MBHA (Peptides International, 200–400 Mesh and substitution grade of 0.37 meq g⁻¹) and l-amino acids derivatives or d-proline derivatives. After the final deprotection step, peptides were cleaved with a cocktail solution containing 9.4 ml trifluoroacetic acid, 500 µl water, 250 µl triisopropylsilane, 500 µl thioanisole, 500 mg phenol and 250 µl 1,2-ethanodiol. Thereafter, peptides were precipitated in cold diisopropyl ether aiming to get rid of synthesis by-products through filtration and then solubilized in acetonitrile : water (1 : 1, v : v) for lyophilization.

Another lot of peptides was similarly assembled by the SPPS methodology as a C-terminal (C_t) amide, according to the standard Fmoc/tBu orthogonal protection scheme [19] on a Symphony X peptide synthesis instrument (Gyros Protein Technologies, Inc. USA), following in-house methods [20]. Briefly, the peptide chain was grown in the C_t → N_t direction on a Fmoc-Rink-amide MBHA resin. Cleavage of the peptides from the resin was realized using a cleavage cocktail comprising TFA, TIS and deionized water (95 : 2.5 : 2.5 v/v/v) in the proportion of 1 ml cocktail per 100 mg of peptidyl-resin. Crude peptide was precipitated from the filtrate using cold tert-butylmethyl ether (MTBE) and pelleted by centrifugation at 3500 r.p.m. The supernatant was discarded, and the peptide pellet was resuspended in fresh MTBE and the centrifugation step repeated; this procedure was repeated three times, after which the peptide pellet was left in a vacuum desiccator overnight, before solubilization in 0.1 M aqueous acetic acid.

The crude peptides was purified by preparative reverse-phase high-performance liquid chromatography (RP-HPLC) on a Merck-Hitachi LaPrep Sigma system (VWR International, Portugal), using a 0 → 100% gradient of acetonitrile and 0.05% aqueous TFA. A final purity degree of 100% (mixture of cis and trans amide rotamers arising from presence of two Pro residues) was determined by analytical RP-HPLC on a Merck-Hitachi LaChrom Elite instrument (VWR International, Portugal). The structure of the target peptide was confirmed by observation of peaks associated with the quasi-molecular ions of the peptide and of the peptide dimer at m/z values of 658.80 and 1615.47 Da, respectively, by electrospray ionization-ion trap mass spectrometry (ESI-IT MS) analysis on a Thermo Finnigan LCQ DECA XP system (Thermo Fisher Scientific, USA). The pure peptide was quantitated by microvolume spectrophotometry, on a NanoDrop One. Finally, the peptide was freeze-dried on a VirTis BenchTop Pro 9 L instrument.

(g) . Effect of UV on the expression of FPPWL-NH₂

Nine specimens of P. azureus were submitted three times to 30 s electrical stimulation aiming to empty the granular gland. The crude secretions were individually collected in 50 ml of Milli-Q water, frozen in liquid N₂ and lyophilized. Amphibians were separated in three groups and submitted to different conditions: dark, day light and ultraviolet B radiation (UVB) at 311 nm using a fluorescent lamp (Philips, TL 20 W/01-RS UV-B Medical Narrowband—λ 305–315 nm with peak at λ 311 nm) installed in a homemade box. After 9 h, the skin secretions of each frog were obtained, frozen and lyophilized, as described above. All individual samples were dissolved at 0.1 mg ml⁻¹ (w/v), mixed with a reference peptide ([M + H]⁺ = 626.3120 Da) at final concentration of 5 µM and submitted to ultra-fast liquid chromatography-MS experiments. Chromatographies were performed by automatic injection of 2 µl of each sample in an ekspert^TM ultraLC 100-XL chromatography system (Eksigent, Dublin, CA, USA) coupled to a Kinetex 2.6 µm C₁₈ 100 Å (50 × 2.1 mm) LC Column connected to a TripleTOF 5600+ mass spectrometer (Sciex, Concord, ON, Canada) housing a DuoSpray Ion Source. Milli-Q H₂O containing formic acid 0.1% was used as solvent A and samples eluted across a linear gradient of solvent B (acetonitrile containing formic acid 0.1%) ranging from 5% to 95% with a flow rate of 0.4 ml min⁻¹ in 10 min. Ion source operated in the positive and the mass rage of acquisitions was between m/z 300–2000. The other parameters were: source temperature = 650.0°C; polarity = positive; number of cycles = 2043; period cycle time = 525 ms; pulser frequency = 13.569 kHz and accumulation time = 500.00 ms. The mass spectrometer was calibrated using APCI positive calibration solution before acquisitions. MultiQuant^tm 3.0.2 software (Sciex, Concord, ON, Canada) was used for quantification of peak areas corresponding PaT-2 peptide (658.3711 ± 0.005 Da) that were normalized to the areas corresponding to the reference peptide (626.3120 ± 0.005 Da) on each LC–MS acquisition.

(h) . In silico studies

The chemical structures of the amidated peptides PaT2 (FPPWL-NH₂), PaT2-a1 (FPLPW-NH₂) and PaT2-a2 (PWLFP-NH₂) were designed with the aid of Avogadro software [21], and pre-optimized via molecular mechanics, using the Amber force field [22]. Given the terminal amidation on the structures at the typical physiological pH range, positively charged structures were considered (with NH⁺₃ terminations). Small antioxidant peptides were also evaluated, as zwitterions, for comparison purposes: l-carnonise, glycyl-l-histidyl-l-lysine (GHK) and glutathione (GSH). The molecular structures were fully optimized in the framework of Kohn–Sham density functional theory (DFT), as implemented in the Gaussian 09 computational package [23]. The geometry optimizations were carried out using the B3LYP exchange-correlation functional [24–27] and 6–311G(d,p) basis set on all the atoms. Solvent (water) effects were simulated via the polarizable continuum model (PCM) [28].

Condensed-to-atoms Fukui indexes (CAFIs) [29] and local chemical softness [30] were evaluated to assess peptide reactivities and identify possible antioxidant sites. CAFIs describe how the local electronic populations change when the number of the electrons of the system is modified, allowing the identification of sites that are prone to interact with nucleophiles (f⁺), electrophiles (f⁻) or free-radicals (f⁰). These simple descriptors have been successfully employed for a variety of compounds with biological properties [31–33]. Hirshfeld's partition charge was employed to estimate the electronic populations to avoid negative indexes [34,35].

The antioxidant properties of the compounds were evaluated via the electron acceptance (R_a) and electron donation (R_d) indexes, as proposed by Martínez et al. [36,37]. Based on these parameters the antioxidant and anti-reductant activities of the compounds can be compared based on their position in the donor–acceptor maps (DAM) [37]. The CAFIs, local softness, R_a and R_d descriptors were obtained using the same level of theory (DFT/B3LYP/6–311G (d,p)) and computational package employed in the geometry optimizations. Details regarding these calculations are presented in the electronic supplementary material.

(i) . Cell viability assays

BV2 microglial cells (CRL-2468-ATCC EOC 13.31; microglia brain; mouse: Mus musculus). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Gibco, USA) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (Gibco, USA) and 1% (v/v) antibiotic solution (10 000 U ml⁻¹ penicillin and 10 mg ml⁻¹ streptomycin, Sigma-Aldrich, USA) at 37°C and 5% CO₂ in a humidified atmosphere. First, BV2 cells were seeded into 96-well culture plate at a density of 3 × 10³ cells well⁻¹ and maintained for 24 h. Then, cells were treated with peptides (PaT2, PaT2-a1 and PaT2-a2) diluted in DMEM medium at a concentration range from 0 to 1800 µM, in triplicates. The plates were incubated for 24 h. Cell viability was evaluated by MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide, Sigma-Aldrich, USA) dye reduction method. After exposure time, the medium was removed and 100 µl of MTT solution (0.75 mg ml⁻¹) diluted in DMEM was added to each well and incubated for 2 h at 37°C. DMSO (100 µl) was then added to dissolve the formazan salts produced by living cells. The absorbance was measured at 595 nm using a SpectraMax Plus 384 microplate reader (Molecular Devices, USA) [38].

(j) . Cellular production of reactive oxygen species and reactive nitrogen species

ROS and reactive nitrogen species (RNS) were quantified with intracellular fluorescent probes in BV2 microglial cells and SK-N-BE(2) (Homo sapiens, tissue brain; neuroblast, CRL-2271-ATCC) that were maintained under the same conditions described in the previous section. Cells were treated with phorbol 12-myristate 13-acetate (PMA, Sigma Aldrich, USA) to induce the overproduction of both ROS and RNS. Intracellular production of ROS was measured using DCFH-DA (Sigma Aldrich, USA) and RNS was measured using DF-FM (Sigma Aldrich, USA). Cells (BV2 and SK-N-BE(2)) were adhered to the plate for 2 h (1.0 × 10⁵ cells well⁻¹) and then subjected to the following treatments: medium (DMEM) only; 100 nM PMA; 100 nM PMA + 50 µM peptide; 100 nM PMA + 100 µM peptide; 50 µM peptide only; or 100 µM peptide only. The duration of the treatment was 30 min for BV2 cells and 60 min for SK-N-BE(2) cells. After the respective incubation times, probes (DAF-FM and DCFDA) were added according to the manufacturer's instructions. ROS and RNS production were evaluated by flow cytometry (FACSCalibur; BD Bioscience) with the excitation set at 488 nm. Emission was detected in the FL-1 (515–545 nm) channel. Ten thousand events were recorded for each sample, and data were analysed using FlowJo v. 10.7.

(k) . FRET-based live cell imaging of microglia

FRET-based live cell imaging of microglia was performed as previously described [39]. The human microglial cell line HMC3 (ATCC CRL­3304) was cultivated in DMEM GlutaMAX-I (Thermo Fisher Scientific) supplemented with 10% FBS (Thermo Fisher Scientific), 100 U ml⁻¹ penicillin and 100 µg ml⁻¹ streptomycin (Thermo Fisher Scientific), and maintained at 37°C and 5% CO₂ in a humidified incubator. For the FRET assay, microglia were plated on culture dishes (µ-Dish 35 mm, iBidi) and the FRET biosensors were transfected using jetPRIME (Polyplus Transfection) according to the manufacturer's protocol. Two gene-encoded biosensors were used. The first was the HyperRed biosensor, which was used to measure the generation of cytosolic hydrogen peroxide [40] using an Addgene plasmid (48249) as previously described [41]. The second biosensor was pDisplay FLIPE-600n^Surface biosensor [42], which was used to measure glutamate release from microglia using an Addgene plasmid (13545), as previously described [43,44]. When using this sensor, an increase in donor-to-FRET fluorescence ratio indicates the specific release of glutamate, without the interference of glutamate metabolism or glutamate uptake. These biosensors have been previously validated in microglial cells [39,41,45].

Transfected microglia were incubated with HBSS containing Ca²⁺ and Mg²⁺ and imaged before stimulation (5–10 min) to attain basal recording curve and also for fluorescent light adaptation). After that, 2 µg ml⁻¹ lipopolysaccharide (LPS) were added to culture dishes to reach a final concentration of 1 µg ml⁻¹ LPS and then recorded for additional 15 min. Image acquisition was performed using a Leica DMI6000B inverted microscope as previously described [43,46]. The excitation light source was a mercury metal halide bulb integrated with an EL6000 light attenuator. High-speed low vibration external excitation/emission filter wheels equipped with filter cubes for cyan fluorescent protein (CFP) (BP 427/10) and yellow fluorescent protein (YFP) (BP 504/12) working with specific dichroic (CG1 440/520) and a separate filter cube for monomeric red fluorescent protein (mRFP) (Ex. BP580/20; DM 595; Em. 630/55) mounted into a microscope filter carrousel (Leica fast filter wheels). Images were acquired with 2 × 2 binning using a digital CMOS camera (ORCA-Flash4.0 V2, Hamamatsu Photonics). Images were exported as 16-bit tiff files and processed in FIJI software. The background was dynamically subtracted from all slices from both channels. Segmentation was achieved on a pixel-by-pixel basis using a modification of the Huang algorithm. After background subtraction and thresholding, binary masks were generated for CFP and FRET images. Original CFP and FRET images were masked and ratio metric images were generated as 32-bit tiff images. Values for the mean grey values were generated using the multicalculation function in FIJI and exported as mentioned above.

(l) . Statistical analysis

One-way ANOVA followed by the Dunn's multiple comparation test evaluated statistical significance in live cell imaging experiments. All statistical analyses were carried out using the Graph Pad Prism 6.0 software. A 95% confidence interval was used and p < 0.05 was considered statistically significant difference in sampled groups.

3. Results and discussion

The skin secretion from adults of P. azureus was fractionated using reverse-phase liquid chromatography and each fraction was analysed using MALDI-TOF/MS. A molecule with [M + H]⁺ = 658.38 Da was found at two different retention times (figure 1a) and its primary structure was determined as FPPW(L/I)-NH₂ (figure 1b). The peptide was identified as a tryptophyllin, based on similarity searches in peptide databases and, thus, named PaT-2 [2]. The tryptophyllin peptide family comprises the first peptides identified in amphibian skin secretions [10]. Most tryptophyllins have a tryptophan residue at position 2 from the C-terminus and one or two prolines at positions 2 and 3 from N-terminus [11], such as PaT-2. To resolve the Leu/Ile ambiguity in PaT-2, we sequenced the cDNA from the skin of P. azureus and found a multimodular precursor encoding a signal peptide followed by an acidic piece and four copies of the FPPWL peptide in tandem containing sites of enzymatic cleavage and sites for C-terminal amidation (figure 1c). The occurrence of multiple copies of a peptide sequence in tandem in a transcript is uncommon in eukaryotes [47] and differs from transcripts for other tryptophyllins, in which there is a single peptide copy [8,12]. Multiple PaT-2 copies in a transcript could lead to fast accumulation of the peptide in response to stimuli, but the implications and significance of this transcript structure warrant further research. Considering that d-Pro is found in peptides from phyllomedusid frogs [7] and the unusual elution pattern of PaT-2 in the skin secretion of P. azureus, we synthesized three analogues containing only l-amino acids, d-Pro in the second (d-Pro2) or third (d-Pro3) residues and compared their elution profiles (figure 1d). The elution profile of the synthetic peptide containing two l-Pro matched that of the natural peptide, confirming that PaT-2 is constituted exclusively by l-amino acids. These two pieces of information, resolved L/I ambiguity and exclusive l-amino acids content, are crucial for further biological activity experiments.

Figure 1.

(a) Spatial distribution of the ion with [M + H]⁺ = 658.37 Da on the dorsal skin of Pithecopus azureus tadpoles at stages 37 and 41, as demonstrated by MALDI mass spectrometry imaging (MALDI-MSI), and the chromatographic profile of the skin secretions collected at the same developmental stages. The peptide FPPWL-NH₂ (theoretical mass [M + H]⁺ = 658.37 Da) eluted only in the secretion from tadpoles at stage 41. The vertical colour bar represents relative ion signal intensity. Scale bar, 2 mm. Spatial distribution of the ion with [M + H]⁺ = 658.37 Da the dorsal skin of P. azureus adults, as demonstrated by MALDI-MSI, and the chromatographic profile of the skin secretion from adult specimens. The absorbance was monitored at 216 (black) and 280 nm (red line) in arbitrary units (mAU). MALDI-TOF analysis of the fractions revealed that PaT-2 eluted at two distinct retention times on the chromatogram (marked with blue arrows). (b) De novo sequencing of FPPWL-NH₂. (c) Nucleotide and predicted amino acid sequences of the precursor encoding PaT-2. cDNA sequencing revealed that PaT-2 is expressed in tandem in the skin of P. azureus. Amino acids corresponding to the signal peptide, acidic piece and mature peptide are as green, orange and blue letters, respectively. The glycine residues used as signals for peptide amidation are shown as red letters. The asterisk marks the stop codon. (d) Ultra-fast liquid chromatography analysis of naturally occurring PaT-2, synthetic PaT-2 containing only l-amino acids (FPPWL-NH₂), synthetic PaT-2 containing a d-amino acid in the second position (FpPWL-NH₂) and synthetic PaT-2 containing a d-amino acid in the third position (FPpWL-NH₂). (e) FPPWL-NH₂ concentration in the skin secretion of adults of P. azureus before (baseline) and after exposure to UVB radiation, darkness or daylight for 9 h. The asterisk denotes a statistically significant difference between baseline and UV exposure groups. (f) Histological analyses of the skin of adults of P. azureus before (left) and after UVB-exposed animals using optical microscopy (H&E staining) and transmission electron microscopy (insets). (Online version in colour.)

We investigated how the expression of PaT-2 related to the ontogeny of P. azureus by screening dorsal skin samples for peptides using MALDI mass spectrometry imaging (MALDI-MSI). No ions were detected until stage 37 of development (figure 1a). In tadpoles at stage 41, PaT-2 was the only detectable peptide (figure 1a). These findings were confirmed by the chromatographic analysis of the skin secretions from tadpoles at stages 37 and 41 (figure 1a). Two major peaks, later identified as PaT-2, eluted in the secretion from tadpoles at stage 41, whereas none were found in that from stage 37 individuals (figure 1a). Dorsal skin glands are in an advanced stage of maturation at stage 41 [48]. The occurrence of PaT-2, the first detectable peptide in P. azureus, coincides with the maturity of granular glands [42], which precedes the transition to the terrestrial habitat. Tadpoles at stage 41 have completely developed anterior and posterior limbs, although the former remain inside the body until stage 42, when P. azureus actually shift from the water to the land [17]. The expression of a specific molecule in the skin secretion of tadpoles at stage 41 could represent a physiological preparation for the new challenges that froglets of P. azureus will face in the new habitat.

Compared with aquatic environments, the terrestrial habitat presents additional challenges to the skin, such as increased O₂ availability and the lack of a water column to filter UV radiation. Therefore, we tested whether the expression of PaT-2 in adults of P. azureus was responsive to UV-B radiation exposure. First, the skin glands were depleted by electrical stimulation and the animals were maintained in the dark, exposed to UV-B radiation or daylight for 9 h. Then, skin secretions were individually obtained and analysed by LC-MS/MS to quantify PaT-2. PaT-2 expression was not stimulated by any of the conditions, and, in the case of UVB, the exposure actually decreased the abundance of PaT-2 compared with that found in animals prior to gland depletion (figure 1e). Optical and transmission electron microscopy revealed that UVB-exposed animals did not suffer any histological damage (figure 1f). We observed that PaT-2 concentration in two specimens was higher than that in the other seven individuals at first extraction, suggesting that the synthesis of PaT-2 might be regulated by other variables not controlled for in this experiment, such as the exact age of animals and prior exposure to environmental factors. These findings indicate that the expression of PaT-2 is not a simple response to exposure to UV, at least in adults.

The coincidence between the transition from water to land with the expression of PaT-2 led us to hypothesize that PaT-2 could play a role in protecting the skin from oxidative damage. First, quantum chemistry calculations were made to assess the theoretical redox reactivity of PaT-2, as well as to locate reactive centres. Two analogues, PaT-2(a1) (FPLPW-NH₂) and PaT-2(a2) (PWLFP-NH₂), as well as other known antioxidants (l-carnonise; glycyl-l-histidyl-l-lysine, GHK and γ-l-glutamyl-l-cysteinylglycine, GSH) were assessed for comparative purposes (figure 2). High values of s⁺, s⁻ and s⁰ (and f⁺, f⁻ and f⁰) indicate which sites are prone to undergo chemical reactions by receiving, donating or without changes in the total number of electrons, respectively. Antioxidant properties of the sequences were dominated by W and F amino acids, being associated with electron donation (via W units) and electron release (via F for PaT-2 and PaT-2(a1); and W for PaT-2(a2)) from/to the peptide to/from the environment (figure 2b). Regarding free radicals (s⁰ and f⁰), the reactivity of PaT-2(a2) was centred on the W unit, while it is distributed over W and F units for PaT-2 and PaT-2(a1) (figure 2b). The existence of an effective π–π interaction between W and F amino acids could limit the activity of the peptides by steric hindrance. Then, antioxidant activities of the structures were also evaluated via electron accepting/donating properties of the sequences by means of ionization potentials (IP), electron affinities (EA), as well as the electron acceptance (R_a) and electron donation (R_d) indexes in relation to other common natural antioxidants (figure 2a,b). The results indicate a common mechanism for PaT-2 and PaT-2(a2), similar to that for the naturally occurring tripeptide GHK, whereas a different mode of action was found for PaT-2(a1) (figure 2a).

Figure 2.

(a) Computational analysis of redox potential and comparative study of PaT-2 and its analogues. (b) Chemical softness and CAFI representation (inset) for reactions towards nucleophiles (f⁺, s⁺), electrophiles (f⁻, s⁻) and free-radicals (f⁰, s⁰). Red and blue colours in the CAFI maps represent reactive and non-reactive sites, respectively (following a RGB scale). (c) Comparative analysis of the donor–acceptor properties, ionization potentials and electron affinities (left-bottom) of PaT-2, PaT-2(a1) and PaT-2(a2) sequences in relation to l-carnonise, glycyl-l-histidyl-l-lysine (GHK) and γ-l-glutamyl-l-cysteinylglycine (glutathione). (Online version in colour.)

The results from the in silico analyses motivated us to test the antioxidant effects of PaT-2 and their analogues in cell-based assays. First, we determined the concentration range of peptides that were not cytotoxic against BV2 microglia cells (figure 3a–c). Then, we tested the antioxidant effects of peptides on PMA-induced ROS and RNS production in microglia BV2 cells and neuroblastoma SK-N-BE(2) cells (figure 3d–g). PMA is an activator of NADPH oxidase [49], a membrane bound enzyme that produces ROS [50]. These cell lines were chosen based on the relevance of oxidative processes in the development and progression in neurodegenerative diseases. Indeed, abnormal ROS-mediated redox signalling and accumulation of oxidatively damaged biomolecules is a hallmark of several neurodegenerative disorders [51–54]. The production of reactive species in PMA-stimulated cells was attenuated by PaT-2 and its two analogues (figure 3d–g). Compared with PaT-2(a1) and PaT-2(a2), PaT-2 was more effective in reducing RNS and ROS generation than the analogues. Regarding the analogues, a general pattern from the experiments was that PaT-2(a2) was slightly more effective than PaT-2(a1) in mitigating PMA-induced reactive species production. This finding indicates that antioxidant effects are quantitatively affected, but not nullified, by scrambling the primary structure. In fact, at 100 µM, PaT-2 decreased the constitutive production of both RNS and ROS in unstimulated BV2 microglia cells (figure 3d,e). Considering the prominent effect of PaT-2, we further investigated its protective effects in human microglial cells expressing genetically encoded biosensor. We found that PaT-2 decreased LPS-induced ROS production and glutamate release in human microglia (figure 3h,i). Glutamate is an important excitatory amino acid neurotransmitter, whose signalling is crucial for normal functioning of the nervous system [35]. Excessive activation of glutamate receptors, however, might result in excitotoxicity and ultimately lead to neurodegeneration [55]. Thus, the novel tryptophyllin described in this study, PaT-2, significantly inhibited two key processes in neurodegenerative conditions, glutamate release and ROS production by human microglial cells.

Figure 3.

PaT-2 and its analogues (PaT-2(a1) and PaT-2(a2)) display antioxidant activity and low cytotoxicity in mammalian central nervous system cells. First, we assessed the cytotoxicity of PaT-2 (a), PaT-2(a1) (b) and PaT-2(a2) (c) on BV2 cells. Mouse microglia BV2 cells (d and e) and human neuroblastoma SK-N-BE(2) cells (f and g) were stimulated with phorbol 12-myristate 13-acetate (PMA), treated at the same time with PaT-2, PaT-2(a1) or PaT-2(a2) at 50 µM or 100 µM, and stained for the detection of reactive oxygen species (ROS, d and f) and reactive nitrogen species (RNS, e and g) prior to flow cytometry analysis. ^#Significantly different from control cells maintained in Dulbecco's modified Eagle's medium (DMEM) only; *significantly different from cells stimulated with PMA; ^§significantly different from control cells maintained in DMEM only. Human CHME3 microglia expressing a ROS biosensor (h; HyperRed) or a glutamate release FRET biosensor (i; FLIPE) were pre-incubated (30 min) with vehicle or PaT-2 (100 µM) and then recorded in saline before (Ctrl) and after 1 µg ml⁻¹ LPS treatment. Data are shown as mean ± s.e.m. of HyperRed fluorescence amplitude (N = 33–41 cells) or FLIPE donor/FRET ratio amplitude (N = 30–35 cells) from cells pooled across three different experiments). *Significantly different from control cells; ^§significantly different from LPS-treated cells. (Online version in colour.)

In conclusion, we characterized PaT-2, the first detectable peptide in the skin secretion of P. azureus along its ontogeny, the expression of which coincides with the maturation of skin glands and anticipates the transition from water to land. PaT-2 is expressed in tandem and presents antioxidant effects in mammalian central nervous system cells, and, thus, stands as a potential neuroprotective molecule [41]. Although the regulators of PaT-2 expression, as well as the actual significance of the peptide to the animal ecophysiology, remain elusive, our findings highlight the utility of an interdisciplinary approach for investigating peptides secreted by amphibians. Here, we combined knowledge of the natural history of the amphibian, ecophysiology, in silico tools and cell-based assays as a strategy to discover a novel peptide with potential biotechnological applications. This approach might be useful in bioprospecting molecules from natural sources with potential utility in preventing or treating human diseases.

Ethics

Fertilized eggs of P. azurea and adult frogs were collected in the Brazilian region and manipulated according to the rules of the Instituto Brasileiro do Meio Ambiente e dos Recursos Renováveis, IBAMA, under the license number 31066-1.

Data accessibility

All data are included in the manuscript. DNA sequence: Genbank accessions OK605281.

Authors' contributions

E.A.B.: conceptualization, data curation, investigation, methodology, writing—original draft; A.P.: conceptualization, formal analysis, funding acquisition, investigation, project administration, validation, writing—original draft; D.C.M.: conceptualization, data curation, formal analysis, investigation, methodology, validation, writing—original draft, writing—review and editing; L.A.: data curation, formal analysis, investigation, methodology; A.D.: conceptualization, data curation, investigation, methodology; A.É.S.-C: data curation, investigation, methodology, visualization, writing—original draft; W.F.C.: data curation, investigation, methodology; S.N.B.: conceptualization, formal analysis, funding acquisition, investigation, methodology, supervision, validation, writing—original draft, writing—review and editing; F.S.-A.: funding acquisition, methodology, supervision, validation; S.A.S.K.: funding acquisition, methodology, project administration, supervision; T.K.B.: conceptualization, formal analysis, funding acquisition, investigation, methodology, visualization, writing—original draft; C.C.P.: investigation, methodology, validation, writing—original draft; R.S.: data curation, formal analysis, investigation, methodology, validation, writing—original draft; C.T.: formal analysis, investigation, methodology, visualization, writing—original draft; writing—review and editing; F.C.D.A.L.: data curation, formal analysis, funding acquisition, investigation, methodology, visualization, writing—original draft, writing—review and editing; A.B.-N.: conceptualization, formal analysis, funding acquisition, investigation, methodology, supervision, validation, writing—original draft, writing—review and editing; A.S.: conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, writing—original draft, writing—review and editing; P.E.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, supervision, visualization, writing—original draft, writing—review and editing; P.G.: data curation, formal analysis, funding acquisition, investigation, methodology, project administration, supervision, visualization, writing—original draft, writing—review and editing; G.D.B.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, supervision, visualization, writing—original draft, writing—review and editing; J.B.R.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, visualization, writing—original draft, writing—review and editing; M.J.K.: funding acquisition, project administration, writing–revisions; J.R.S.A.L.: conceptualization, formal analysis, funding acquisition, investigation, project administration, supervision, writing—original draft. All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Competing interests

We declare we have no competing interests.

Funding

Computational resources were provided by GRID-Unesp, SICC/PRP-IFSP and CENAPAD/SP. A.B.N. and F.C.D.A.L. acknowledge funding from CNPq (420449/2018-3 and 428211/2018-6). This work was funded by Fundação para a Ciência e Tecnologia—Portugal (PTDC/BII-BIO/31158/2017 and CIRCNA/BRB/0281/2019), including a postdoctoral grant to A.P. (UIDB/50006/2020). This work also received financial support from Fundação para a Ciência e Tecnologia and Ministério da Ciência, Tecnologia e Ensino Superior—Portugal (UIDB/50006/2020), Funding Authority for Studies and Projects (FINEP, Brazil, 01.08.0457.00) and FAPESP (2018/07999—7).