Characterization of tryptolinophyllins from the treefrog Boana boans expands the structural and functional diversity of the trypytophyllin family of peptides

Abstract

Amphibian skin secretions constitute a rich source of bioactive peptides whose structural diversity is often amplified by post-translational modifications. In this study, we report the discovery and comprehensive characterization of a post-translational modification that has not been described previously in peptides from a vertebrate. Tryptophyllins are a heterogeneous group of short, tryptophan- and proline-containing peptides widely distributed among anuran species and are thought to contribute to skin protection against oxidative and UV-induced damage. Two peptides derived from tryptophyllins, termed tryptolinophyllins, were isolated from skin secretions of the giant gladiator treefrog Boana boans. They are distinguished by the presence of an N-terminal (S)-tryptoline-3-carboxylic acid (Tpi) residue, resulting from the conversion of tryptophan into a tetrahydro-β-carboline moiety via a Pictet–Spengler-type reaction. The primary structures of two tryptolinophyllins were elucidated by high-resolution tandem mass spectrometry and multidimensional NMR spectroscopy, and confirmed through total chemical synthesis of the corresponding diastereoisomers. To our knowledge, this represents the first confirmed tetrahydro-β-carboline post-translational modification in a vertebrate peptide, revealing an unexpected level of chemical innovation in amphibian peptide biosynthesis. Computational analyses of electronic structure identified the Tpi residue as the preferred locus for electron transfer, suggesting a potential role in redox chemistry. ABTS radical-cation scavenging assays demonstrated that the antioxidant activity of the tryptolinophyllins was reduced compared with unmodified tryptophyllins, highlighting a nuanced relationship between post-translational chemical diversification and functional output. Together, these findings expand the known repertoire of naturally occurring peptide modifications in vertebrates and underscore the exceptional chemical plasticity of amphibian skin secretions.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58272-w.

Introduction

The tryptophyllins are small peptides (≤ 14 amino acid residues) containing one or more tryptophan and one or more proline residues that were first identified in the skin of the frog Phyllomedusa rohdei (reclassified as Pithecopus rohdei)¹. Subsequently, multiple tryptophyllins have been identified in the skins of several species of frogs belonging to the sub-families Pelodryadinae (Australian treefrogs) and Phyllomedusinae (S. American treefrogs) within the extensive family Hylidae [for review,²]. Tryptophyllins have also been isolated from skin secretions of the most primitive extant anuran, the tailed frog Ascaphus truei (Ascaphidae)³ and from the Coruna frog Pelophylax perezi (Ranidae)⁴. In addition, a tryptophyllin (BmT-2) has been purified from the venom of the pit viper Bothrops moojeni⁵. Classification of the tryptophyllins together in a single peptide family is based solely upon common structural features which are quite different from those of the cationic amphipathic α-helical host-defense peptides present in skins of the majority of frog species and does not imply that its members have evolved from a common ancestral peptide.

The tryptophyllins identified to-date lack antimicrobial and cytotoxic activities but individual peptides have shown a range of biological properties including myotropic activities such as inhibition of bradykinin-induced vasodilatation of rat tail artery smooth muscle⁶ and restoration of vasorelaxation in oxidative stress-induced endothelial dysfunction⁷, inhibition of neuronal nitric oxide synthesis⁸, antiproliferation of prostate cancer cell lines^4,6 and activity at opioid receptors⁹. Amphibians occupy both terrestrial and aquatic environments, it has been speculated that the principal physiological role of the tryptophyllins lies in their antioxidant and radical-scavenging properties that serve to protect the skins of adult frogs against oxygen- and UV-induced damages following transition of tadpoles from water to land¹⁰.

One of our previous studies led to the isolation in high-yield of tryptophyllin BN in both carboxy terminal (WRPFPFL) and C-terminally amidated (WRPFPFL.NH₂) forms together with antimicrobial peptides related to figainin 2 and picturin 2 from norepinephrine-stimulated skin secretions of the Giant Gladiator treefrog Boana boans, also known as the Rusty treefrog, that belongs to the sub-family Hylinae within the family Hylidae¹¹. The analysis of the RP-HPLC elution fractions of the secretions revealed the presence of two peptides whose molecular mass, determined by MALDI-TOF mass spectrometry (MS), and partial amino acid sequence, determined by automated Edman degradation, suggest that they may represent post-translational modified forms of tryptophyllins BN. The present study provides an unambiguous determination of the primary structure of these components using high resolution tandem MS/MS and NMR spectroscopy. Primary structures were confirmed by chemical synthesis of the identified diastereoisomer. Finally, theoretical and experimental antioxidant studies were carried out on the C-terminally amidated tryptophyllin BN and on both modified tryptophyllin-related peptides.

Material and methods

Purification of tryptophyllins and tryptophyllin-related peptides

Collection of norepinephrine-stimulated skin secretions from two adult B. boans frogs, partial purification on Sep-Pak C₁₈ cartridges and purification to near homogeneity of tryptophyllins BN (C-terminally amidated and C-terminal carboxyl forms) and the putative post-translationally modified forms by reversed-phase (RP) HPLC have been described previously¹¹. For the convenience of the reader, full details are provided in the Supplementary material section. Relevant permits approving the collection and sampling of live animals were granted by the Wildlife Section, Forestry Division, Trinidad and by the University of the West Indies (UWI) Campus Ethics Committee (CEC234/07/16). All experiments were carried out by authorized investigators in accordance with relevant guidelines and regulations as previously reported¹¹.

Determination of monoisotopic mass of synthetic and skin-purified peptides by MALDI-TOF MS

MALDI-TOF MS was carried out using an UltrafleXtreme instrument (Bruker France SAS, Wissembourg, France). A full description of the procedure, including calibration of the instrument with peptides of known molecular mass in the 0.9–2 kDa range, has been provided previously¹². The accuracy of mass determinations was < 0.02%.

Determination of primary structure of skin-purified peptides by Edman sequencing

The amino acid sequences of the peptides were determined by automated Edman degradation. Briefly, peptides were solubilized in 30% (v/v) acetonitrile/H₂O (Sigma-Aldrich, Saint-Quentin-Fallavier, France) and deposited on a glass fiber disk previously conditioned with BioBrenPlus (Applied Biosystems, Courtaboeuf, France). The disk was then inserted into an automatic sequanator (Procise 494, Applied Biosystems). The analytical methodology consists of 3 steps: (i) the N-terminal amino acid of the peptide was coupled to phenylisothiocyanate and cleaved by anhydrous trifluoroacetic acid (TFA) from the peptide as anilinothiozolinone derivative, (ii) this 5-membered heterocyclic moiety was then converted into a more stable phenylthiohydantoin (PTH) by 25% (v/v) TFA/H₂O treatment, and (iii) finally, PTH-derivative was chromatographed on-line using analytical RP-HPLC (C₁₈, λ = 269 nm) and determined by their order and retention time in comparison to standard PTH-amino acids¹³.

De novo sequencing of skin-purified peptides by LC–MS/MS

Peptide samples were dissolved in 0.1% (v/v) formic acid (FA)/H₂O and 1 nmol was injected in a linear ion trap-Orbitrap mass spectrometer (LTQ Orbitrap Elite, Thermo Scientific, Les Ulis, France) equipped with a nano-ESI source coupled to a nanoliquid chromatography (Easy-nLC II, Thermo Scientific). The chromatographic separation was carried out on an analytical RP-HPLC C₁₈ column (NikkyoTechnos, Tokyo, Japan) using a linear gradient (2–50% over 20 min) of acetonitrile/FA (99.9:0.1, v/v) at 50 °C and a flow rate of 300 nL/min. The mass spectrometer operated in data dependent mode to automatically switch between Orbitrap-MS (resolution of 60,000 from m/z 100 to 1800) and LTQ-MS/MS acquisition (resolution of 15,000 from m/z 100 to 1500). The top-10 most intensive precursors were selected for fragmentation with a dynamic exclusion duration set at 20 s. The tandem MS/MS fragmentation was performed in higher energy collisional dissociation (HCD) mode with a collision energy of 30–35 V. The de novo sequencing from the MS raw data were performed with the PEAKS Studio 10.5 software (Bioinformatics Solutions Inc. Waterloo, ON, Canada).

Synthesis of tryptophyllins and tryptophyllin-related peptides

All Fmoc amino acid residues, O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU), preloaded 4-hydroxymethyl-phenoxymethyl-copolystyrene-1%-divinylbenzene resin (HMP) and Rink amide 4-methylbenzhydrylamine (MBHA) resin were purchased from Novabiochem (Fontenay sous Bois, France) or IRIS Biotech (Marktredwitz, Germany). N,N-Diisopropylethylamine (DIEA), piperidine, TFA, formaldehyde, triisopropylsilane (TIS), tert-butylmethylether (TBME), di-tert-butyl dicarbonate (Boc₂O) and (3R,3S)-2,3,4,9-tetrahydro-1H-β-carboline-3-carboxylic acid (THβC) were supplied from Sigma-Aldrich. N-methylpyrrolidone (NMP), dimethylformamide (DMF), dichloromethane (DCM) and acetonitrile were from Fisher Scientific (Illkirch, France).

Peptides were synthesized by Fmoc solid phase methodology on a Liberty microwave assisted automated peptide synthesizer (CEM, Saclay, France) using the standard manufacturer’s procedures at 0.1 mmol scale¹⁴. All Fmoc-amino acids (0.5 mmol, 5 eq.) were coupled on preloaded Fmoc-Leu-HMP resin or Rink amide resin, by in situ activation with HBTU (0.5 mmol, 5 eq.) and DIEA (1 mmol, 10 eq.) before Fmoc removal with 20% (v/v) piperidine in DMF. The non-natural amino acid THβC was added manually at the end of the synthesis after protection of its amino group by a Boc moiety. Briefly, Boc₂O (4 eq.) and DIEA (5 eq.) were added to a solution of THβC (2.3 eq., 5 mg/mL) in 50% (v/v) acetonitrile/H₂O. After 6 h of gentle agitation at 40 °C, the mixture is diluted in water and freeze dried. The crude product is purified by RP-HPLC (Gilson, Villiers le Bel, France) on a 22 × 250 mm Vydac 218TP1022 C₁₈ column (Avantor, Rosny-sous-Bois, France) using a linear gradient (10–90% over 45 min) of acetonitrile/TFA (99.9:0.1, v/v) at a flow rate of 10 mL/min. Boc-THβC is coupled to the peptidyl-resin by in situ activation with HATU (1 eq.), HOAt (1 eq.) and DIEA (2 eq.) in DMF. Alternatively, the N-terminally deprotected L- and D-Trp(Boc)-containing resin-bound peptides were subjected to on-resin Pictet–Spengler cyclization to afford the corresponding diastereoisomeric tryptophyllin-related peptides¹⁵. Practically, formaldehyde (37% w in H₂O, 4 eq.) was added to the peptidyl-resin in dichloromethane containing 1% TFA, and the mixture was agitated at room temperature for 4 h. Reaction progress was monitored by the Kaiser test¹⁶. To drive the reaction to completion, additional formaldehyde (6 eq.) was added, and agitation was continued for a further 2 h. Peptides were then deprotected and cleaved from the resin by adding 10 mL of an ice-cold mixture of TFA/TIS/H₂O (9.5:0.25:0.25, v/v/v) for 3 h at room temperature, followed by precipitation in TBME and centrifugation (4500 rpm, 15 min). THβC being a commercial racemic form (3R,3S), we obtained two diastereoisomers for each peptide that were easily purified by RP-HPLC on a 22 × 250 mm Vydac 218TP1022 C₁₈ column using a linear gradient (20–50% over 45 min) of acetonitrile/TFA (99.9:0.1, v/v) at a flow rate of 10 mL/min (Table 1). Other peptides were purified under identical chromatographic conditions. Peptides were then characterized by MALDI-TOF MS on a UltrafleXtreme (Bruker France SAS, Wissembourg, France) in the reflector mode using α-cyano-4-hydroxycinnamic acid as a matrix. Analytical RP-HPLC, performed on a 4.6 × 250 mm Vydac 218TP54 C₁₈ (Avantor) or Kinetex XB-C₁₈ (Phenomenex, Le Pecq, France) column, indicated that the purity of the peptides was > 99.7%.

Table 1.

Condensed structural formula and chemical data for synthetic diastereoisomer replicates of cmpds 1 and 2.

Code	X	Sequencea	HPLC		MS
			Rt (min)b	Purity (%)	Calcdc	Obsdd
Synth-cmpd 1a	NH2	[Tpi]RPFPFL.NH2	27.4	99.9	972.533	973.525
Synth-cmpd 1b			30.2	99.8	972.533	973.523
Synth-cmpd 2a	OH	[Tpi]RPFPFL	29.6	99.7	973.517	974.502
Synth-cmpd 2b			32.2	99.9	973.517	974.506

^aTpi, tryptoline-3-carboxylic acid, undetermined configuration. ^bRetention time determined by analytical RP-HPLC. ^cTheorical monoisotopic molecular weight. ^dm/z [MH]⁺ value assessed by MALDI-TOF MS. Diastereoisomers are numbered according to their elution order in RP-HPLC.

Characterization of synthetic and skin-purified peptides by LC–MS

Synthetic replicates of tryptophyllin-related peptides (Table 1) and skin-purified cmpd 2, in the micromolar concentration range, were prepared by 40-fold or 400-fold dilutions into a solution of acetonitrile/H₂O/TFA (2.5:97.4:0.1, v/v/v) and analyzed by ultra-high-performance liquid chromatography (UHPLC) coupled with high-resolution MS (HRMS). Analyses were performed on an UltiMate™3000 NanoRSLC System (Dionex, Courtaboeuf, France) connected to a 4-GHz maXis quadrupole TOF mass spectrometer (Bruker Daltonics, Bremen, Germany) equipped with an electrospray ion source. The LC loading pump was used for these experiments. The LC–MS setup was controlled by the Bruker HyStar software ver.3.2. Peptides (1 pmol) were chromatographed on a 2.1 × 100 mm (1.6 µm) CORTECS Shield C₁₈ column (Waters, Guyancourt, France) thermostated at 40 °C using a 2.5–37.5% linear gradient of acetonitrile/FA (99.9:0.1, v/v) over 10 min at 300 µL/min. Mass spectra were acquired in positive ion MS mode over a 150–3000 m/z range with a nebulizer gas pressure of 2.0 bars. The drying gas flow was 10 L/min, the temperature was 200 °C. The acquisition rate was 1 Hz corresponding to spectra summations of 5428. External calibration was performed with the ESI-L Low Concentration Tuning Mix (Agilent Technologies, Les Ulis, France). Mass spectra were processed using DataAnalysis 4.4 software (Bruker Daltonics).

Comparative analysis of synthetic and skin-purified peptides by MALDI-Q-Exactive MS

Experiments were performed on an AP-SMALDI-5AF (TransMIT GmbH, Giessen, Germany) ion source coupled to an Orbitrap Q-Exactive Plus mass spectrometer (Thermo Scientific). The S-lens RF level, capillary temperature and spray voltage parameters were 50, 400 °C and 4.3 kV, respectively. The instrument operated in positive ion MS mode. The mass range was set to 100–1500 m/z and the resolution to 140,000 for MS and MS/MS. One μL of skin-purified and synthetic (1–2 nmol) peptides (Table 1) was deposited on the target with 1 µL of α-cyano-4-hydroxycinnamic acid (20 mg/mL) as a matrix. External calibrations were carried out using a standard peptide mixture (Peptide Calibration Standard kit, Bruker Daltonique). Sections 35–40% laser energy was used and at least 10 laser shots were summed for each MALDI spectrum. The MS/MS fragmentation of precursor ion was performed in HCD mode with a collision energy of 20–35 V.

Characterization of synthetic and skin-purified peptides by NMR experiments

Synthetic replicates of tryptophyllin-related peptides (Table 1) and highly purified cmpd 2 were dissolved in 600 μL of CD₃OD 99.8% (EurisoTop, Saint-Aubin, France) at concentrations ranging from 1.24 to 1.54 mM. ¹H-¹H TOCSY, ¹H-¹H NOESY, ¹H-¹H ROESY and ¹H-¹³C-HSQC were performed on each sample at 298 K on an Avance III HD BRUKER 700 MHz spectrometer (Bruker France SAS) equipped with a cryoprobe. A ¹H-¹³C-HSQC spectrum was recorded exclusively for synth-cmpd 1a. ¹H and ¹³C chemical shifts were referenced to the residual methyl signal of CD₃OD (3.34 ppm and 49.86 ppm for ¹H and ¹³C, respectively). NMR data were processed using Bruker’s Topspin 3.2TM and analyzed with CCPNMR (version 2.2.2)¹⁷.

In silico antioxidant studies of tryptophyllins and tryptophyllin-related peptides

Both tryptophyllins BN (WRPFPFL, WRPFPFL.NH₂) and the four diastereoisomeric tryptophyllin-related peptides ([(S)-Tpi]RPFPFL, [(S)-Tpi]RPFPFL.NH₂, [(R)-Tpi]RPFPFL, [(R)-Tpi]RPFPFL.NH₂) were drawn with the molecular editor Avogadro¹⁸ and their geometry was first optimized with the Universal Force Field (UFF)¹⁹ implemented in the cross-platform tool. Subsequently, the quantum chemistry software ORCA^20,21 was used to generate conformers for each peptide employing the global optimizer algorithm (GOAT)²² with the semi-empirical method GFN2-xTB²³ combined with the implicit water solvent model ALPB²⁴. The lowest energy conformer was selected to further optimize each peptide. Density functional theory (DFT) calculations were run to optimize the peptides using B3LYP correlation functional, as previously described in a similar quantum chemistry-based antioxidant study on the peptide salamandrin-I²⁵. Given the limitations of plain B3LYP²⁶, we added Grimme’s D4 dispersion correction²⁷. This functional was combined with the triple-zeta basis set def2-TZVP and the conductor-like polarizable continuum model (CPCM) implicit water solvent. The vibrational frequencies were computed to ensure the absence of imaginary modes. We used a similar protocol to optimize the reactive oxygen species H₂O₂, HOO• and HO•, simply skipping the GOAT conformer generation step. Geometry of Trolox, an antioxidant reference standard, was directly optimized at the same final level of theory starting from the PubChem coordinates²⁸.

The electron acceptance index and the electron donation index, R_a and R_d respectively, were computed for each peptide. The detailed equations can be found in²⁹. Briefly, these quantities are derived from the vertical electron affinity and the vertical ionization energy. Computing these quantities required to run single point energies on the optimized geometry of each compound at the same level of theory (B3LYP-D4/def2-TZVP/CPCM(water)), adding one electron and removing one electron. The condensed-to-atom Fukui functions () and the local softness () were computed to compare the reactivity of the peptides and Trolox with the reactive oxygen species based on the Hirshfeld charges. All the values were consistently obtained in the B3LYP-D4/def2-TZVP/CPCM(water) framework. The detailed equations can be found in²⁵.

ABTS decolorization assay for synthetic replicates of C-terminally amidated tryptophyllin BN and both post-translationally modified tryptophyllin-related peptides

The antioxidant activity of synthetic replicates of C-terminally amidated tryptophyllin BN and both post-translationally modified tryptophyllin-related peptides was evaluated by ABTS (2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid, Sigma-Aldrich) assay as previously described³⁰. Briefly, stock solutions of Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid, 1.6 mM; Sigma-Aldrich) and peptides (10 mg/mL) were prepared in H₂O to achieve, in a 96 well preparation plate, a dilution range of 1600 to 93.64 µM and 10 to 1.25 mg/mL, respectively. The ABTS radical cation (ABTS•⁺) solution was generated by mixing 990 µL of an ABTS solution (7 mM in H₂O) with 10 µL of a K₂S₂O₈ solution (245 mM in H₂O, final concentration of 2.45 mM). The resulting solution turned dark blue and was incubated in the dark, at room temperature for 16 h. According to the Beer-Lambert law, with a molar extinction coefficient of 1.5 × 10⁴ L/mol/cm, the final concentration of ABTS•⁺ solution was then calculated to obtain an absorbance of 0.7 (± 0.02) at 734 nm, and diluted accordingly.

Antioxidant activity of C-terminally amidated tryptophyllin BN and both tryptophyllin-related peptides was then evaluated in a 96 well reading plate by mixing 10 µL of each peptide and 190 µL of the obtained ABTS•⁺ solution, and by measuring the corresponding absorbances on a scanning fluorometer Flexstation III (Molecular Devices, Sunnyvale, CA) at 734 nm and 30 °C, every minute for 1 h. Antioxidant capacity is expressed as the percentage inhibition of ABTS•⁺ absorbance or Trolox equivalent antioxidant capacity (TEAC) by comparison with the Trolox standard. TEAC values are calculated as the ratio of the slope of the percentage inhibition of absorbance versus concentration for the tested antioxidant to that obtained for Trolox, at a defined time point. The EC₅₀ value, defined as the concentration required to inhibit 50% of ABTS•⁺, was determined by regression analysis of the dose–response curve (% inhibition versus concentration). Data are expressed as mean ± SEM from three independent experiments performed in triplicate (n = 3).

Statistical analysis

Significant differences among multiple groups were analyzed using one-way analysis of variance (one-way ANOVA), followed by Dunnett’s multiple comparisons test. Statistical significance was defined as p < 0.05. All analyses were performed using GraphPad Prism version 10 (GraphPad Software, La Jolla, CA, USA).

Results

Purification of tryptophyllins and tryptophyllin-related peptides

The elution profile on a Vydac C₁₈ semi-preparative column of the pooled norepinephrine-stimulated skin secretions from B. boans is shown in Fig. 1A. Peak 1 contained the C-terminally amidated form of tryptophyllin BN ([M + H]⁺: 961.545; Fig. 1B). Peak 2 contained the C-terminal carboxylated form of the peptide ([M + H]⁺: 962.524) as previously reported¹¹ and a co-eluting unidentified compound (cmpd 1) with an experimental [M + H]⁺ of 973.546 (Fig. 1C). Peak 3 contained a second unidentified peptide (cmpd 2; [M + H]⁺: 974.530; Fig. 1D). The four peptides were purified to near homogeneity, as assessed by peak symmetry and MS, by further chromatographies on semipreparative Vydac C₄ and C₈ columns (chromatograms not shown). For the convenience of the reader, the designation of the HPLC peaks is consistently maintained throughout the article.

Fig. 1.

Purification of tryptophyllins and tryptophyllin-related peptides from skin secretions of the Giant Gladiator treefrog Boana boans. (A) RP-HPLC chromatogram with UV detection at 214 nm on a semipreparative Vydac C₁₈ column of skin secretions from B. boans after partial purification on Sep-Pak C₁₈ cartridges. The dashed line shows the concentration of acetonitrile in the eluting solvent. The peaks denoted 1–3 were subjected to further purifications on semipreparative Vydac C₄ and C₈ columns.

Adapted from Conlon et al.¹¹. (B–D) MALDI-TOF mass spectra of the respective contents of peaks 1 to 3 prior to further purifications.

Structural characterization of tryptophyllins and tryptophyllin-related peptides

The primary structure of the unmodified tryptophyllins BN have been unambiguously determined as WRPFPFL.NH₂ (peak1) and WRPFPFL (peak 2)¹¹. The amino acid sequence of cmpd 2 was determined directly by automated Edman degradation and corresponded to that of C-terminal carboxylated tryptophyllin BN, except for the N-terminal residue, whose PTH-derivative did not co-elute with the standard tryptophan (Supp. Fig. S1). Tandem MS/MS-based full-length peptide sequencing confirmed that cmpd 2 was C-terminally carboxylated, indicating that only its N-terminal moiety differed from that of C-terminal carboxylated tryptophyllin BN (Fig. 2). Cmpd 1 exhibited a fragmentation pattern similar to that of cmpd 2, but based on C-terminal amidated tryptophyllin BN, suggesting that the same N-terminal modification was present in both purified variants of tryptophyllin BN from the dermal secretions of B. boans (Supp. Fig. S2). The b1 fragment ion generated from cmpds 1 and 2 showed a deviation of + 12.00 (b1: 199.09) in comparison to both tryptophyllins BN (b1: 187.09), corresponding to the theoretical incorporation of an additional carbon atom (Fig. 2 and Supp. Fig. S2). Based on the strong sequence identity among these four peptides and the fact that no proteinogenic amino acid could produce such a MS/MS fragment ion, we hypothesized that the N-terminal residue of cmpds 1 and 2 corresponded to a modified tryptophan moiety. As a plausible candidate, we focused on a tetrahydro-β-carboline framework resulting from a ring closure via a methylene bridge between carbon 2 of the indole nucleus and the α-nitrogen of tryptophan (Fig. 3).

Fig. 2.

MS/MS fragmentation spectra of (A) carboxylated tryptophyllin BN ([M + 2H]²⁺: 481.7660) purified from peak 2 and (B) cmpd 2 ([M + 2H]²⁺: 487.7655) purified from peak 3.

Fig. 3.

Condensed structural formula of L-tryptophan (left panel) and of the homolog (3S)-1,2,3,4-tetrahydro-β-carboline-3-carboxylic acid (right panel) corresponding to the residue potentially incorporated at position 1 in cmpds 1 and 2.

To clarify further the nature of the modified peptides, we first synthesized authentic standards of C-terminal amidated and carboxylated tryptophyllin BN-related peptides using racemic (3R,3S)-1,2,3,4-tetrahydro-β-carboline-3-carboxylic acid (THβC), also named (3R,3S)-tryptoline-3-carboxylic acid (Tpi). Each modified peptide was obtained as a pair of separable diastereoisomers, denoted as synth-cmpds 1a and 2a (first eluting), and 1b and 2b (second eluting) (Table 1) and fully characterized by analytical RP-HPLC, MALDI-TOF MS and ¹H,¹³C NMR (Supp. Figs. S3–S18; Supp. Tables S1–S4).

Prior to purification to near homogeneity of the two tryptophyllins BN and cmpds 1 and 2, RP-HPLC analysis of the partially purified fraction centered on peak 2 showed contamination with a small amount of peak 3, as illustrated in trace A of Fig. 4. Under the same chromatographic conditions, cmpd 1 purified from peak 2 and synth-cmpd 1b displayed identical retention times, as did cmpd 2 purified from peak 3 and synth-cmpd 2b (Fig. 4).

Fig. 4.

RP-HPLC chromatograms with UV detection at 215 nm on an analytical Kinetex XB-C₁₈ column of (A) partially purified peak 2, (B, C) synth-cmpds 1a and 1b, respectively, and (D, E) synth-cmpds 2a and 2b, respectively. The arrows in panels C and E denote the retention times of peak 2 and peak 3, respectively, as referenced from panel A.

The fragmentation profiles of synth-cmpds 1b and 2b were then compared to that of cmpds 1 and 2 by MALDI-Q-Exactive MS (Supp Figs. S19 and S20). As expected, cmpd 1 and synth-cmpd 1b exhibited the same [M + H]⁺ signal at 973.54 (Supp Fig. S19, trace A1 and trace A2, respectively), and their fragmentation spectra were very similar (Supp Fig. S19, trace B1 and trace B2, respectively). More precisely, both amidated peptides exhibited 81 common fragmentation signals, including the b1, b2, b4, and b6 fragment ions (in blue), as well as y1, y2, y3, y4, y5, and y6 fragment ions (in red) (Supp. Fig. S19C). The b3 and b5 fragments were detected exclusively in the synthetic replicate. Likewise, cmpd 2 and synth-cmpd 2b shared the same monoisotopic mass, [M + H]⁺: 974.52, and a same fragmentation pattern (Supp. Fig. S20). Indeed, 83 common peaks were observed, including the b1, b2, and b4 fragment ions (in blue) and the y2, y3, y4, y5, and y6 fragment ions (in red) (Supp. Fig. S20C). Similarly to the amidated peptides, the b3 and b5 fragment ions appeared only in the synthetic peptide, while the b6 fragment ion was exclusive to cmpd 2. These correlations in fragmentation data between skin-purified and synthetic peptides support the presence of an N-terminal Tpi residue in the natural peptides.

The series of RP-HPLC purification steps yielded a quantity of pure cmpd 2, sufficient to proceed with its detailed characterization by LC–MS and ¹H-¹³C NMR. Under chromatographic conditions different from those used in Fig. 4, the diastereoisomers of each pair eluted in the same order as before, with retention times of 8.5 and 9.0 min for synth-cmpds 1a and 1b and 8.9 and 9.4 min for synth-cmpds 2a and 2b, respectively, as shown in the extracted ion chromatograms (XIC) corresponding to the m/z ions (z = 2) of each compound (Fig. 5A–D). Cmpd 2 exhibited a retention time of 9.4 min in the XIC for [M + 2H]²⁺: 487.7660, which matched precisely with the retention time of synth-cmpd 2b (Fig. 5E). No signal was detected in the XIC for [M + 2H]²⁺: 487.2740 corresponding to the other pair of diastereoisomers (Fig. 5F). Co-injection of cmpd 2 and synth-cmpd 2b yielded the same signal as when each was analyzed individually (Fig. 5G).

Fig. 5.

LC–MS profiles of synth-cmpds 1a, 1b and 2a, 2b in absence or presence of cmpd 2. (A, B) Extracted ion chromatogram (XIC) referred to [M + 2H]²⁺: 487.2740 for synth-cmpds 1a (Rt = 8.5 min) and 1b (Rt = 9.0 min). (C, D) XIC referred to [M + 2H]²⁺: 487.7660 for synth-cmpds 2a (Rt = 8.9 min) and 2b (Rt = 9.4 min). (E, F) XIC referred to [M + 2H]²⁺: 487.7660 and 487.2740 for cmpd 2, respectively. (G) XIC referred to [M + 2H]²⁺: 487.7660 for co-injection of cmpd 2 and synth-cmpd 2b.

The 1D and 2D NMR spectra obtained for synth-cmpds 1a, 1b and synth-cmpds 2a, 2b enabled the assignment of all observable chemical shifts (Supp. Figs. S7–S18; Supp. Tables S1–S4). Each diastereoisomer displayed a predominant form, comprising at least 70% of the total population, alongside one or more minor forms, indicated by peak splitting. The minor forms could not be confidently assigned due to their low signal intensity and overlapping peaks, which were likely the result of local structural variations caused by cis/trans equilibrium of the two prolines. The presence of a Tpi residue (chirality undetermined) at the N-terminal part of the four diastereomers was unambiguously observed at the atomic level on the spectra by the additional signals due to the methylene closing the third cycle (Fig. 6A, left) and the absence of signal from the indolic C2 proton (C2-H2) in comparison to a canonical tryptophan (Fig. 6B, left and right). Moreover, the overlap of the ¹H-¹³C HSQC and ¹H-¹³C HMBC spectra of synth-cmpd 1a clearly exemplified a correlation between the newly added methylene group (C1-H1) and C3-H3 of the Tpi moiety (Fig. 6C). Similarly, the section of the ¹H-¹³C HSQC spectrum for cmpd 2 (Supp. Figs. S21–S23), exhibited the characteristic signal of the additional CH₂ group and the absence of C2-H2 correlation (Fig. 6A, right; Fig. 6B, middle), confirming that cmpd 2 exhibited the identical and unique modification as the four synthetic replicates (Table 2).

Fig. 6.

(A) Section of the ¹H-¹³C HSQC spectra for all samples, highlighting the characteristic signal of the additional methylene (CH₂) group. Left panel: overlay for the four synthetic compounds synth-cmpds 1a, 1b, 2a and 2b in blue, cyan, red and green, respectively. Right panel: cmpd 2. (B) Left panel: aromatic regions of the ¹H-¹³C HSQC spectra for synth-cmpds 1a, 1b, 2a and 2b (same color code as in A). Middle panel: same region for cmpd 2. Right panel: same region for a standard tryptophan-containing tryptophyllin BN. (C) Overlay of a section of the ¹H-¹³C HSQC (blue) and ¹H-¹³C HMBC (red) spectra of synth-cmpd 1a, showing that the additional CH₂ signal is correlated with the C3/H3.

Table 2.

¹H and ¹³C chemical shifts (ppm) of cmpd 2.

residue	NH	Hα	Hβ	Cα	Cβ	Other H	Other C
Tpi1	–	3*: 4.27	4: 3.50/3.07	3: 56.31	4: 23.65	1: 4.44/4.44 9: 10.55, 6: 7.06 8: 7.36, 7: 7.15 5: 7.49	1: 42.87, 6: 121.02 8: 112.91, 7: 123.88 5: 119.33
Arg2	8.67	4.74	1.96/1.84	51.12	28.15	γ: 1.72/1.72 δ: 3.23/3.23 ε: 7.87	γ: 26.57 δ: 42.86
Pro3		4.38	2.10/1.83	60.51	28.37	γ: 1.93/1.93 δ: 3.75/3.66	γ: 26.52 δ: 49.34
Phe4	8.13	4.79	3.09/2.95	52.79	37.04	δ, ε, ζ: 7.20–7.35	δ, ε, ζ: 132–128
Pro5		4.34	2.05/1.83	60.76	28.71	γ: 1.91/1.83 δ: 3.72/3.41	γ: 26.45 δ: 49.19
Phe6	7.75	4.67	3.29/3.02	54.58	37.00	δ, ε, ζ: 7.20–7.35	δ, ε, ζ: 132–128
Leu7	8.03	4.44	1.68/1.68	52.17	41.57	γ: 1.72 δ: 0.97/0.96	γ: 26.52 δ: 24.21/22.92

Tpi, tryptoline-3-carboxylic acid, undetermined configuration. *The italicized numbers refer to the numbering of Tpi atoms as shown in Fig. 3. The chemical shifts in bold refer to the methylene group linking the indole moiety to the α-nitrogen within the Tpi residue.

To determine the α-carbon configuration of the Tpi residues located at the N-terminal extremity of cmpds 1 and 2, a solid-phase Pictet–Spengler reaction was performed starting from tryptophyllins BN bearing either an N-terminal L-Trp (S) or D-Trp (R) residue (Fig. 7). This approach proceeded quantitatively within 6 h, yielding four tryptophyllin-related peptides, designated PS-L 1 and PS-D 1 for the C-terminal amide peptides generated from L-Trp and D-Trp, respectively, and PS-L 2 and PS-D 2 for the corresponding C-terminal acid peptides (Table 3; Supp. Figs. S24–S27). Under the same chromatographic conditions used for the analysis of the four synthetic diastereoisomers obtained from racemic THβC (Fig. 4), the diastereoisomers PS-L 1 and PS-L 2 incorporating a putative (S)-Tpi residue eluted after PS-D 1 and PS-D 2, which possibly contained the (R)-Tpi moiety, and exhibited retention times identical to those of the synth-cmpds 1b and 2b (Tables 1 and 3).

Fig. 7.

Generalized synthetic Pictet–Spengler cyclization of L-tryptophan with formaldehyde to form (S)-Tpi-containing peptides, using polymer-bound L-tryptophan at ambient temperature under acidic conditions for 6 h.

Table 3.

Chemical data for synthetic diastereoisomer replicates of cmpds 1 and 2 synthesized by on-resin Pictet–Spengler reaction from L- and D-Trp-containing precursors and match with synth-cmpds 1 and 2 diastereoisomers.

Code	Sequencea	HPLC		MS		Compound reference match
		Rt (min)b	Purity (%)	Calcdc	Obsdd
PS-L 1	[(S)-Tpi]RPFPFL.NH2	30.2	99.9	972.533	973.546	synth-cmpd 1b
PS-D 1	[(R)-Tpi]RPFPFL.NH2	27.3	99.7	972.533	973.553	synth-cmpd 1a
PS-L 2	[(S)-Tpi]RPFPFL	32.2	99.9	973.517	974.556	synth-cmpd 2b
PS-D 2	[(R)-Tpi]RPFPFL	29.7	99.9	973.517	974.542	synth-cmpd 2a

^aTpi, tryptoline-3-carboxylic acid, determined configuration. ^bRetention time determined by analytical RP-HPLC. ^cTheorical monoisotopic molecular weight. ^dm/z [MH]⁺ value assessed by MALDI-TOF MS.

In silico antioxidant studies of tryptophyllins and tryptophyllin-related peptides

To investigate the antioxidant properties of both tryptophyllins BN and the four diastereoisomeric tryptophyllin-related peptides synth-cmpds 1a, 1b, 2a and 2b in comparison to Trolox, we computed a range of density functional theory (DFT)-derived descriptors. The rational is that the peptides are likely to show antioxidant properties if their electronic and reactivity properties are similar to that of Trolox. In particular, similar reactivity toward reactive oxygen species (ROS) such as H₂O₂, HOO• and HO•, implies similar antioxidant properties. Figure 8 displays the donor–acceptor map (DAM)—better acceptor for large R_a, better donor for small R_d—associated with the six peptides and Trolox, as well as the relative position of the frontier orbitals (HOMO: highest occupied molecular orbital; LUMO: lowest unoccupied molecular orbital). Interestingly, the nearest peptide to Trolox in the DAM was [(S)-Tpi]RPFPFL (synth-cmpd 2b, PS-L 2), the diastereoisomer with the N-terminal modification retaining the stereochemistry of the L-tryptophan, highlighting its importance for the antioxidant property (Fig. 8A). The orbital energy diagram gives consistent insights, regarding the reactivity of the peptides as electron acceptor or donor: [(S)-Tpi]RPFPFL (synth-cmpd 2b, PS-L 2) had the highest energy HOMO (i.e. best to act as a nucleophile) and WRPFPFL.NH₂ had the lowest energy LUMO (i.e. best to act as an electrophile) (Fig. 8B). The energy differences between all the peptides and Trolox HOMOs were smaller than the differences for LUMOs, indicating more similar donor than acceptor properties.

Fig. 8.

(A) Donor–acceptor map for Trolox, both tryptophyllins BN and the four diastereoisomeric tryptophyllin-related peptides. (B) Frontier energy level alignments between the six peptides and Trolox. Colour coding is consistent across both panels.

The analysis of the local softness of [(S)-Tpi]RPFPFL (synth-cmpd 2b, PS-L 2) revealed that the most reactive sites in the peptide were part of the (S)-Tpi residue, and that it could react with ROS according HSAB theory (Fig. 9). This strengthens the hypothesis that the tryptophan and its modified counterpart Tpi were mainly responsible for the antioxidant properties of the peptides. Notably, this observation holds true for all peptides modeled in our study (Supp. Figs. S28–S32) for all three types of reactivity investigated (electrophilicity, nucleophilicity and free-radical scavenging activity), with the exception of the electrophilic reactivity of WRPFPFL.NH₂ and [(R)-Tpi]RPFPFL (synth-cmpd 2a, PS-D 2). In these cases, the electrophilic reactivity is mainly localized on the phenylalanine residues (Supp. Figs. S29A, S31A).

Fig. 9.

Comparative study of the local chemical softness of [(S)-Tpi]RPFPFL (synth-cmpd 2b, PS-L 2) and Trolox in relation to HO•, HOO• and H₂O₂. Antioxidant activities as (A) electrophiles (s⁺/s⁻), (B) nucleophiles (s⁻/s⁺) and (C) free-radical scavengers (s⁰/s⁰).

ABTS decolorization assay for synthetic replicates of C-terminally amidated tryptophyllin BN and both post-translationally modified tryptophyllin-related peptides

Finally, we compared the antioxidant capacity of [(S)-Tpi]RPFPFL.NH₂ (synth-cmpd 1b, PS-L 1), [(S)-Tpi]RPFPFL (synth-cmpd 2b, PS-L 2) and WRPFPFL.NH₂, originally present in the norepinephrine-stimulated skin secretions from B. boans, with that of Trolox using the ABTS assay, which measures their ability to scavenge ABTS radical cations (ABTS•⁺). The assay is based on the spectrophotometric quantification of the residual ABTS•⁺ following incubation with the antioxidant³⁰. EC₅₀ values, representing the concentration required to reduce 50% of the radical cations, were estimated from the curves shown in Fig. 10, which depict linear dose–response relationships between the compounds and ABTS•⁺ scavenging activity after 30 min of incubation (Table 4). The individual regression lines highlighted clear differences in antioxidant effectiveness. The curve for the C-terminally amidated tryptophyllin BN, WRPFPFL.NH₂ lay above those of both tryptophyllin-related peptides (PS-L 1 and PS-L 2) across the same concentration range, indicating a higher antioxidant potency of the Trp-containing peptide compared with its Tpi-containing counterparts. Consistent with this observation, WRPFPFL.NH₂ exhibited the lowest EC₅₀ value among the peptides, whereas PS-L 2 showed the weakest activity. At a concentration of 10 mg/mL, the antioxidant effects of PS-L 1 and PS-L 2 were significantly lower than that of WRPFPFL.NH₂. In contrast, no statistically significant differences were observed among the TEAC values of the three peptides after 30 min, as determined by the ABTS•⁺ decolorization assay (Table 4). Noteworthy, the antioxidant capacities of all peptides were at least 50-fold lower than that of Trolox, which displayed the steepest slope and lowest EC₅₀.

Fig. 10.

The effects of concentration of Trolox, tryptophyllin BN and tryptophyllin-related peptides on the inhibition of the ABTS•⁺ after 30 min of incubation. (●) Trolox (r² = 0.996); (▼) C-terminally amidated tryptophyllin BN (WRPFPFL.NH₂, r² = 0.984); (▲) PS-L 1 (r² = 0.964); (■) PS-L 2 (r² = 0.977).

Table 4.

Antioxidant activity parameters of C-terminally amidated form of tryptophyllin BN and cmpds PS-L 1 and PS-L 2 relative to Trolox.

Code	Inhibition (%) at 10 mg/mLa	Slope	EC50 (mg/mL)	TEAC
Trolox	ND	211.10 ± 4.94	0.235 ± 0.007	1.0000 ± 0.0331
WRPFPFL.NH2	45.96 ± 0.40	2.09 ± 0.19	11.822 ± 1.206	0.0099 ± 0.0009
PS-L 1	26.48 ± 0.82****	1.94 ± 0.27	21.749 ± 3.087	0.0092 ± 0.0013
PS-L 2	22.86 ± 0.93****	2.16 ± 0.23	22.852 ± 2.541	0.0103 ± 0.0011

^aVs. WRPFPFL.NH₂, one-way ANOVA followed by Dunnett’s multiple comparisons test, **** p < 0.0001.

Discussion

Tryptophyllin peptides bearing an N-terminal Trp residue are rare². Tryptophyllin BN (WRPFPFL.NH₂), identified in our previous study¹¹, shows the greatest structural similarity to dentatidin 1.1 (WSPFWD.NH₂) and dentatidin 1.2 (WSPFWR.NH₂) from the Australian treefrog Litoria dentata (Pelodryadinae)⁹. Tryptophyllin BN was isolated in both C-terminal α-amidated and C-terminal carboxylic acid forms. Although most commonly α-amidated at the C-terminal extremity, some tryptophyllins, such as the tryptophyllin PsT-1 (KPPPWVPV) from Phyllomedusa sauvagei³¹ contain a C-terminal carboxyl group. As is the case with B. boans, tryptophyllins L 3.1 (FPWP.NH₂) and L 3.2 (FPWP), isolated from the skin secretion of Litoria rubella after mild electrical stimulation, are amidated and non-amidated variants of the same sequence³².

In the present study, we have characterized, from the same pooled skin extrudate of B. boans, two tryptophyllin BN-related peptides (cmpds 1 and 2) containing a non-canonical residue at the N-terminal position which differed by an experimental [M + H]⁺ mass deviation of + 12 compared with naturally occurring tryptophan. Such a shift has been previously reported in formaldehyde-induced modified peptides³³. Indeed, formaldehyde reacts with the primary amino group of a single or an N-terminal tryptophan residue to yield a Schiff base, which then cross-links with the indole side chain to form a heterotricyclic structure (Fig. 3)³⁴. Known as the Pictet–Spengler reaction [for review,³⁵], this condensation requires an aldehyde and acidic conditions, which were not applied in our process for collecting skin secretions. However, in the absence of an alternative explanation, we hypothesized that such conversion products were indeed present in the extrudate of B. boans. To investigate this, we have introduced the 1,2,3,4-tetrahydro-β-carboline scaffold at the N-terminal position of both tryptophyllin BN peptides, using the commercially available racemic (3R,3S)-tryptoline-3-carboxylic acid (Tpi), which was N-Boc pre-protected.

The four synthetic diastereoisomers were individually and comprehensively characterized by their retention times in RP-HPLC, their MS/MS fragmentation patterns and the NMR chemical shifts of their hydrogen and carbon atoms. Correspondences between the natural cmpds 1 and 2 and the later-eluting diastereoisomers (synth-cmpds 1b and 2b) of each pair were unambiguously confirmed by identical retention times in two different eluting systems, and identical fragmentation profiles. Finally, ¹H-¹³C-HSQC spectrum of cmpd 2, which was isolated in comparatively high abundance, exhibited the characteristic signal of the additional CH₂ group and the absence of C2-H2 correlation, confirming that the natural peptide exhibited the identical and unique modification as synth-cmpd 2b. The dispersion and splitting of NMR signals observed for all diastereoisomers (synth-cmpds 1a, 1b, 2a and 2b) indicate the coexistence of multiple conformational states in solution. This behavior is consistent with a slow cis/trans isomerization of proline residues on the NMR timescale, a well-known feature of short proline-rich peptides³⁶. Such conformational heterogeneity is frequently reported for amphibian skin peptides, which often display significant structural flexibility in solution rather than a single well-defined fold^37,38. The presence of minor conformers with low intensity and partially overlapping signals precluded their full assignment, a common limitation in NMR studies of dynamic amphibian peptides. Importantly, the conserved spectroscopic signatures of the Tpi residue across all compounds demonstrate that this unusual N-terminal modification is compatible with the intrinsic conformational plasticity characteristic of frog skin peptides^37,38. Nevertheless, despite the acquisition of NOESY spectra, the absolute configuration of the asymmetric center of the Tpi residue in cmpd 2, as well as the four diastereoisomers, could not be assigned, since all protons were magnetically equivalent and no stereochemically informative NOE correlations were observed.

To address this challenge, cmpds 1 and 2 were synthesized from precursors bearing either an N-terminal L- or D-tryptophan, enabling direct evaluation of the impact of its stereochemistry on the resulting peptides. The Pictet–Spengler reaction with an aldehyde typically forms a new stereocenter at C1 (Fig. 3) in a diastereoselective manner, controlled by the stereochemistry at C3^39,40; with formaldehyde, however, no new asymmetric center is generated, allowing the original tryptophan configuration to be fully preserved. PS-L 1, prepared from L-tryptophan, exhibited retention times identical to cmpd 1 purified from the skin secretion and to synth-cmpd 1b obtained with racemic THβC, whereas PS-L 2 showed the same correspondence with cmpd 2 and synth-cmpd 2b. Based on the retention of tryptophan configuration during the Pictet–Spengler reaction with formaldehyde, cmpds 1 and 2 isolated from pooled norepinephrine-stimulated skin secretions of B. boans can be assigned with confidence as containing Tpi residues with the (S) configuration, and are consequently derived from L-tryptophan. Moreover, based on the relative yields of the purified Tpi-containing peptides and their unmodified tryptophyllin BN counterparts isolated from the same skin secretion, the extent of conversion of the N-terminal Trp residue into Tpi is estimated to be approximately 10%.

To the best of our knowledge, the presence of a tryptoline-3(S)-carboxylic acid residue is unprecedented in the sequence of a peptide isolated from a eumetazoan. Nevertheless, frog skin secretions are known to contain peptides incorporating a variety of post-translationally modified and non-ribosomally synthesized amino acids. Modifications such as N-terminal pyroglutamyl residues, hydroxyproline, methionine sulfoxide, sulfated tyrosine, and D-amino acids, particularly D-alanine and D-alloisoleucine, generated through peptidyl aminoacyl L-to-D isomerization, have been documented in amphibian skin peptides⁴¹. The presence of a Tpi residue further expands the structural diversity generated by post-translational modifications in amphibian skin secretions. Interestingly, tetrahydro-β-carboline derivatives have previously been reported in amphibian skin secretions⁴². Although structurally distinct from the Tpi-containing peptides described here, this finding demonstrates that tetrahydro-β-carboline biosynthesis can occur in amphibian skin. In addition, the toxin aculeine B, isolated from the marine sponge Axinyssa aculeata (a parazoan), exhibits the same type of modification on its N-terminal tryptophan residue. Aculeines are a family of cytotoxic and neurotoxic peptides that are post-translationally modified by the attachment of a long-chain polyamine⁴³ to C1 of the N-terminal Tpi residue⁴⁴. Natural products containing a THβC template are biosynthesized through a Pictet–Spengler reaction between L-tryptophan and aldehydes^45,46. A detailed mechanism involving mediation by a Pictet–Spenglerase has been proposed³⁵. These enzymes have been shown to occur in higher plants that produce indole alkaloids but are absent in indole alkaloid unproductive plants⁴⁷. To date, only a few such enzymes have been characterized from microorganisms. McbB from the marine actinomycete Marinactinospora thermotolerans⁴⁸ and NscbB from the extremophilic actinomycete Nocardiopsis synnemataformans, isolated from a kidney transplant patient⁴⁹, are known to catalyze not only the Pictet–Spengler condensation but also the complete aromatization of the β-carboline scaffold⁵⁰. Finally, it has been documented that an enzyme-catalyzed Pictet–Spengler reaction leads to the formation of salsolinol from dopamine and acetaldehyde in the human brain⁵¹. In addition to the possible involvement of a Pictet–Spenglerase, the availability of suitable carbonyl substrates in frog skin secretions should also be considered. Volatile profiling of the skin secretions of the hylid frogs Hypsiboas pulchellus and Hypsiboas riojanus reveals the presence of several endogenous aliphatic aldehydes among a diverse array of volatile metabolites⁵². Since aldehydes constitute the essential reaction partners in Pictet–Spengler condensations, their occurrence in amphibian skin secretions demonstrates that the chemical prerequisites for the formation of THβC scaffold can naturally exist within the cutaneous secretory milieu. Although formaldehyde itself has not been reported from frog skin secretions, and no aldehydes have yet been identified in B. boans, the occurrence of endogenous aldehydes in anuran secretions raises the possibility that formaldehyde, or a related carbonyl compound, could participate in the biosynthesis of the Tpi-containing peptides described here. Whether the formation of these peptides proceeds through a spontaneous non-enzymatic Pictet–Spengler reaction involving endogenous formaldehyde or through the action of a dedicated Pictet–Spenglerase remains to be determined.

Tryptophyllins have been reported to exhibit both neuroprotective and cytoprotective properties and antioxidant activities that are likely related to the redox properties of the indole moiety⁴. Beyond tryptophyllins, antioxidant peptides are increasingly recognized as promising bioactive scaffolds owing to their safety and capacity to combine radical-scavenging activity with additional antimicrobial or anticancer properties⁵³. This general interest further supports the investigation of tryptophyllin-related sequences as potential multifunctional antioxidant systems. These observations motivated the combined use of in silico DFT calculations and ABTS•⁺ scavenging assays to assess the antioxidant potential of tryptophyllins and their Tpi-containing derivatives. DFT-based analyses suggested that [(S)-Tpi]RPFPFL (synth-cmpd 2b, PS-L 2) displayed electronic properties close to those of Trolox, including a high HOMO energy and favorable donor–acceptor characteristics, indicative of potential antioxidant activity⁵⁴. Local softness calculations, which reflect a molecule ability to easily accept or donate electrons and thus indicate high polarizability and reactivity, further identified the (S)-Tpi residue as the main reactive site toward ROS, supporting the central role of the tryptophan-derived moiety in radical scavenging. This strengthens the hypothesis that tryptophan (or its modified version Tpi) is mainly responsible for the antioxidant properties, which is consistent with a previous report²⁵. This is not surprising, since Trp (or Tpi) and Phe are the only aromatic residues in the peptides. However, it suggests that the stereochemistry of Tpi may substantially influence the peptide antioxidant activity. Such in silico approaches are particularly valuable for rapidly screening diastereomers and exploring structure–reactivity relationships without extensive material requirements, although they rely on simplified models that do not fully capture solvent effects, peptide flexibility, or aggregation phenomena that may modulate reactivity in solution. In contrast, ABTS•⁺ decolorization assays revealed that the naturally-occurring tryptophyllin WRPFPFL.NH₂ exhibited the highest radical-scavenging activity, while the Tpi-containing peptides (PS-L 1 and PS-L 2) were significantly less effective, and all peptides showed antioxidant capacities far below that of Trolox. This superior radical-scavenging activity may be related to its ability to act as an electron acceptor, as suggested by its lowest LUMO energy among the peptides investigated by DFT. Notably, the diastereomers PS-D 1 and PS-D 2 were even less active than their L-counterparts (data not shown), highlighting the influence of stereochemistry on radical-scavenging efficiency. The discrepancy between the two approaches illustrates a key limitation of in silico predictions, as electronic similarity to Trolox does not necessarily translate into comparable antioxidant behavior in solution, particularly when comparing small molecules with larger, structurally flexible peptides. Variations in molecular size, hydrophilicity, and steric accessibility of reactive sites strongly influence peptide interactions with the ABTS radical cation. Consequently, although Trolox is widely used as a reference antioxidant²⁵, its low molecular weight and high solubility make it fundamentally different from tryptophyllins and tryptophyllin-related peptides. Accordingly, TEAC values obtained for peptides should not be directly compared with those of small-molecule antioxidants such as quercetin (TEAC 6.24 ± 0.07;⁵⁵), as differences in molecular mass, diffusion, and reaction kinetics may overestimate the apparent efficiency of low-molecular-weight compounds^30,56.

Although the individual antioxidant activities of tryptophyllins and their Tpi-containing derivatives are modest when assessed in isolation, their co-occurrence in very high concentrations in frog skin secretions suggests that they may act collectively. Such a combined or cocktail-effect could contribute to the protection of adult frog skin against oxygen- and UV-induced damage during the transition from an aquatic to a terrestrial environment⁵⁷. In this context, additive or complementary mechanisms, rather than strong individual radical-scavenging activity, may underlie their biological relevance.

In conclusion, we have identified and comprehensively characterized two tryptophyllin BN-related peptides from B. boans skin secretions, containing a previously undescribed N-terminal (S)-tryptoline-3-carboxylic acid (Tpi) residue. We propose the term “Tryptolinophyllins” to designate these peptides, reflecting their origin as tryptophyllin-related sequences in which the N-terminal tryptophan is converted into a tetrahydro-β-carboline (tryptoline) moiety via a Pictet–Spengler post-translational modification. This represents the first example of such a modification in a vertebrate peptide and highlights the natural formation of tetrahydro-β-carboline-peptide hybrids. The unambiguous structural confirmation through synthetic replicates, MS/MS, and NMR demonstrates that the Tpi moiety is a true natural modification rather than an artifact, while the co-occurrence of amidated and carboxylated forms suggests that these peptides may arise either from distinct precursors or from differential post-translational processing of a common precursor¹¹. Computational analyses indicated that Tpi is the primary site for electron transfer, suggesting a role in the redox behavior of these peptides. Consistent with this prediction, ABTS•⁺ assays revealed modest radical-scavenging activity relative to the corresponding tryptophyllins, illustrating the nuanced relationship between electronic structure and functional output in complex peptides. Future studies will focus on elucidating the biosynthetic origin of the Tpi residue and determining whether its formation is enzyme-catalyzed or results from spontaneous Pictet–Spengler chemistry within the skin secretory milieu. Finally, the discovery of these tryptolinophyllins underscores the remarkable chemical versatility of amphibian skin secretions and suggests that, despite their modest individual activities, they may act collectively to protect frog skin against oxidative and UV-induced stress, providing insight into how post-translational modifications expand both structural diversity and functional adaptability in vertebrate peptide systems.

Author contributions

J.M.C., and J.L. conceived the study. L.C., P.T., K.L., and J.L. designed the methodology. B.L., L.C., A.D., H.M., T.R., N.N, G.G., C.D., and A.D.M. performed the experiments and collected the data. B.L., L.C., A.D., H.M., T.R., C.D., and K.L. curated the data and prepared the visualizations. L.C., P.T., K.L., J.M.C., and J.L. wrote the original draft of the manuscript. All authors contributed to the interpretation of the results and to the writing, review, and editing of the manuscript. J.L. supervised the project.

Funding

This study received no external funding.

Data availability

The NMR chemical shift assignments have been deposited in the Biological Magnetic Resonance Data Bank (BMRB) under accession number 53705. All other data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.