Mass spectrometry approaches for the discovery and full primary structure analysis of peptides secreted by the amphibian skin

Abstract

Most, if not all, frog species are known to secrete a mixture of more or less toxic compounds from skin glands as defense mechanism. This skin secretion of amphibian constitutes a rich source of bioactive peptides with potential pharmaceutical applications. Research in this area includes peptidome (primary structure), transcriptome and bioassays analyses (Chapter 1). Although the complete genome of a frog species has been mapped, the analysis of frog peptides remains difficult. This is mainly because during evolution each frog species has gathered its own set of bioactive peptides encoded in its genes. In this thesis we focus on MS-based methods that help to characterize the peptidome of frog venoms. In Chapter 2 an overview is given of mass spectrometry methods and approaches to study frog skin secretions. Many of the frog skin peptides are post-translationally modified after their synthesis and mass spectrometry is an excellent tool to characterize them. This is illustrated in Chapter 2 for several frequently occurring post-translational modifications previously found on frog peptides. The skin secretion of the Chinese odorous frog Odorrana schmackeri had been studied before at the transcriptome level, leading to the discovery of full length open reading frames. The amino acid sequence of the mature peptide can be predicted from these open reading frames obtained by molecular cloning. After obtaining transcriptome results, mass spectrometry is the next logical step to proof the existence of the predicted peptides. Miniaturized nano liquid chromatography in combination with mass spectrometry is the preferred tool to do this as it requires minimal amounts of sample. In Chapter 3 a combination of transcriptome and peptidome analyses lead to the discovery and characterization of several members of the nigrocin peptide family found in the skin secretion of Odorrana schmackeri. The results from the MS measurements showed that all nigrocin peptides have a disulfide-bridge and from their peptide fragmentation spectra the complete amino acid sequence for each of the three nigrocins peptides could be confirmed. During these measurements of the nigrocin peptides, three other peptides caught our attention because of their a-typical peptide fragmentation behavior. Using both CID and ETD fragmentation analyses we could fully de novo sequence them and establish that they are O-linked glycosylated peptides (Chapter 4). All three peptides had the same primary structure and O-glycosylation at a Threonine residue. They differed in the monosaccharides composition (2, 3 and 4 units). Interestingly, the unmodified peptide (i.e. with no glycosylation) was not detected in the venom. Interestingly the very same primary structure has been found by molecular cloning in other Ranid frogs, however, obviously the glycosylation could not be (and was not) predicted from the cDNA/mRNA analysis. To our knowledge, this is the first example of a glycosylation as a PTM in amphibian skin secretion. The bioactivity of the peptide and the role of the glycosylations have yet to be determined. From the same analysis of the skin secretion of Odorrana schmackeri, yet two other peptides caught our interest. The majority of the peptides from Ranidae frogs contain a so called Rana box motif, which is a disulfide bridge in the C-terminal domain of the peptide. However, we discovered two relatively large peptides with a disulfide bridge at the N-terminal side, which is very uncommon for frog skin peptides. These peptides, with 34 residues, could be de novo sequenced using chemical treatments with DTT, iodoacetamide and bromoethylamine; high resolution CID and ETD fragmentation on full length peptides as well as on selected tryptic fragments; and labeling purified tryptic fragments with heavy and light dimethyl. The entire primary sequence, including Leu and Ile discriminations, was confirmed by Edman degradation. These two peptides share conservative motifs with calcitonin, adrenomedullin and calcitonin gene related peptides, but yet cannot be categorized in either of these existing peptide families. These unusual calcitonin-like peptide sequences (OsCLP1 and 2) present in the skin secretion of an anuran species may represent a novel peptide family within the calcitonin gene related peptide superfamily. Another type of venom which was studied during this research was that of the genus Phyllomedusa, which are known for their highly poisonous skin secretions. Specimens of the walking leaf frog, Phyllomedusa burmeisteri and Phyllomedusa rohdei were captured during hunting trips in Brazil. Not much was known yet about the peptides present in their toxic secretions. Their skin secretions were collected by a non-invasive method and afterwards all specimen were immediately released in their natural surroundings. The skin secretions of both frogs were used to optimize our MS-based methods and approaches and this has led to the discovery of several novel peptides. In Chapter 6 a generic method is presented to detect peptides with intra- or an inter-molecular disulfide bonds prior to MS based sequencing of their primairy structure. The method is based upon the comparative analysis by LC-MS of the untreated crude venom and the crude venom in which all S-S bonds are broken with a reducing agent, such as dithiotreitol. The LC-MS data are converted into a so called peptide display, in which the retention time is plotted against the mass-over-charge ratio. By comparing the peptide display of both analyses, peptides with a disulfide bond(s) are easily recognized, because their mass increases by 2 Da per broken disulfide bond, and their retention times usually also slightly shift. Beside the two Phyllomedusa species, 3 other types of frog venoms were subjected to this method and it was observed that each venom shows different profiles. O. schmackeri, for example, is highly rich in intramolecular disulfide bonds. The skin secretion of P. burmeisteri and P. rohdei contain several large peptides containing 3 S-S bonds, Kassina senegalensis contains overall a less complex peptide content and only a few peptides with a S-S bridge, whereas Bombina variegata venom yielded several peptides with single or double S-S bonds. Inter-molecular disulfide bonds were also observed by this approach in the venom of P. burmeisteri, i.e., the distinctin homo- and heterodimers. In Chapter 7 we describe studies involving the shotgun cloning of the distinctin chains from the skin secretion. cDNA cloning surprisingly revealed that the chains A and B are encoded by separate mRNAs. Oxidation experiments with the synthetic chains did not show any preferential formation over the dimerization, suggesting either that there is an alternative mechanism for dimer formation in vivo or that synthetic chains have different behavior than native peptides. The P. burmeisteri and P. rohdei peptides described in Chapter 6 that contain 3 disulfide bonds as determined by the 2-D differential peptide display method are further characterized in Chapter 8. Several MS-based approach were used to de novo sequence these large 6 kDa peptides, including complementary CID and ETD peptide fragmentation and high resolution orbitrap detection. These peptides showed several conservative motifs of the Kazal peptides, a group of protease inhibitors. An interesting observation made during these analyses is that peptides modified with bromoethylamine show an much better ETD fragmentation pattern compared to the same peptides modified with iodoacetamide, which is the most commonly used reagent to block free cysteines in proteomics. Another important issue to mention is that for this, and all the other MS based methods described in this thesis, only little amount of material was used, which is crucial for frog skin researches on limited samples which are difficult to acquire. Moreover, the methodologies showed in this thesis can be applied to other peptidomic samples. Finally in Chapter 9 all important findings done during this research are discussed, several concluding remarks are given and a future outlook is presented.