FOCUS: MASS SPECTROMETRY-BASED STRATEGIES
FOR NEUROPROTEOMICS AND PEPTIDOMICS: RESEARCH ARTICLE
Screening Method for the Discovery of Potential Bioactive
Cysteine-Containing Peptides Using 3D Mass Mapping
Luuk N. van Oosten,
1Mervin Pieterse,
1Martijn W. H. Pinkse,
1Peter D. E. M. Verhaert
1,2,31
Department of Biotechnology, Delft University of Technology, 2628 BC, Delft, The Netherlands 2
Department of Biomedical Sciences, Antwerp University, 2610, Antwerp, Belgium 3
CEBMMS (Center of Excellence in Biological and Medical Mass Spectrometry), Department of Clinical Sciences, Lund University, 221 85, Lund, Sweden
Abstract.Animal venoms and toxins are a valuable source of bioactive peptides with
pharmacologic relevance as potential drug leads. A large subset of biologically active peptides discovered up till now contain disulfide bridges that enhance stability and activity. To discover new members of this class of peptides, we developed a workflow screening specifically for those peptides that contain inter- and intra-molecular disul-fide bonds by means of three-dimensional (3D) mass mapping. Two intrinsic proper-ties of the sulfur atom, (1) its relatively large negative mass defect, and (2) its isotopic composition, allow for differentiation between cysteine-containing peptides and pep-tides lacking sulfur. High sulfur content in a peptide decreases the normalized nominal mass defect (NMD) and increases the normalized isotopic shift (NIS). Hence in a 3D plot of mass, NIS, and NMD, peptides with sulfur appear in this plot with a distinct spatial localization compared with peptides that lack sulfur. In this study we investigated the skin secretion of two frog species; Odorrana schmackeri and Bombina variegata. Peptides from the crude skin secretions were separated by nanoflow LC, and of all eluting peptides high resolution zoom scans were acquired in order to accurately determine both monoisotopic mass and average mass. Both the NMD and the NIS were calculated from the experimental data using an in-house developed MATLAB script. Candidate peptides exhibiting a low NMD and high NIS values were selected for targeted de novo sequencing, and this resulted in the identification of several novel inter- and intra-molecular disulfide bond containing peptides.
Keywords:3D mass mapping, Orbitrap analysis, Disulfide bridge containing peptides, Animal toxin, MATLAB
script
Received: 7 April 2015/Revised: 18 September 2015/Accepted: 24 September 2015
Introduction
T
he venoms and toxins of animals are a major source ofbiologically active peptides and currently over 5000 of these peptides are recorded in the UniProt protein database
[1, 2]. The nature of activity ranges from antimicrobial to
neurotoxic, and this specific collection of peptides has proven to be an excellent source of potential new drug leads [3–5]. The classical route for the discovery of a bioactive peptide conven-tionally starts with determination of biological activity of the
whole crude venom. Identification of individual peptides that are responsible for a certain activity is only achieved after extensive purification. Subsequent primary structure character-ization either involves (de novo) sequencing via Edman
deg-radation or tandem mass spectrometry (MSn). Cloning of
mes-senger RNA encoding the bioactive peptides is another pow-erful approach [6], but only possible if some information of the primary sequence is known upfront.
Currently, the most widely used tool to study mixtures of non-tryptic endogenous peptides involves de novo sequencing
from high resolution MSn experiments using the ‘shotgun’
approach [7]. Characterization of animal toxin peptides via this approach is generally complicated because of the high number and high variety of post-translational modifications (PTMs), which makes de novo sequencing a difficult and
time-consuming task [8,9]. Fragmentation spectra can be‘manually’
Electronic supplementary material The online version of this article (doi:10. 1007/s13361-015-1282-z) contains supplementary material, which is available to authorized users.
interpreted or analyzed using specialized algorithms for auto-mated de novo sequencing such as PEAKS (Bioinformatics
Solutions Inc., Waterloo, ON, Canada), PepNovo [10], and
pNovo+ [11]. Full structure elucidation is only possible when
complete fragmentation occurs and the spectrum is of sufficient quality with respect to signal-to-noise ratio. Instrumental pa-rameters often need to be optimized for every peptide individ-ually to yield spectra that meet these standards. Owing to the fact that a toxin sample can contain dozens of different peptides with accompanying truncated or partially degraded products, manually acquiring optimized tandem MS spectra for each individual peptide species and interpreting all their fragmenta-tion spectra is a tremendously labor-intensive task. It is desirable to have several selection criteria to specifically target those peptides that meet these criteria upfront, before engaging in the time consuming task of de novo sequencing.
Remarkably, many of the biologically active peptides contain one or multiple disulfide bridges, which are believed to contribute to the peptides’ stability and activity. This is illustrated by the high prevalence of cysteines (6.91% of all residues) in the reviewed database from the animal toxin annotation program at UniProt compared with the whole UniProtKB/Swiss-Prot (release 2015_1) database (1.37% of all residues). The presence of disulfide bonds could thus be used as a selection criterion to target for unknown peptides in animal toxins and venoms with bioactivity potential.
A conventional way of determining the presence of disulfide bonds in a peptide is by simply counting the mass shift before and
after reduction [12] or after reduction and alkylation [13].
Disadvantages of this approach are that it requires comparison of two separate analyses and that it is difficult to detect dimeric peptides in complex mixtures. Prior to sequence analysis of pep-tides, disulfide bonds are often reduced to improve the quality of MS2spectra [14]. Additionally, free thiol groups of the cysteines can be derivatized using specific reagents such as iodoacetamide or selenamide [15,16]. The main advantages of this type of sample preparation are that disulfide bridges cannot randomly reform and fragmentation of the modified peptide could yield higher sequence coverage [15,16]. An additional advantage is that the derivatizing group could be engineered in such a way that cysteine-containing peptides can be specifically detected using tandem mass tags
(TMT) [17,18] in MS or even purified using isotope-coded
affinity tags (iCAT) [19]. A disadvantage is that these labels are often relatively large and will negatively influence fragmentation, making de novo sequencing even more difficult. Specific isolation, such as in the case of iCAT, requires substantial amounts of starting material, which is not easy to obtain in the case of animal venoms. Another disadvantage with these labeling strategies is that it requires reduction of the disulfide bonds and, thereby, secondary structure information of dimeric peptides is lost.
In order to quickly assign the presence of cysteines in a peptide of unknown nature without the need of reduction of disulfide bond, we propose here to use a high resolution mass filtering approach. Two intrinsic characteristics of the sulfur atom can be used to select peptides with cysteine residues. Sulfur has a relatively large negative mass defect (the difference between the isotopic and nominal or integer mass) and it has a positive
isotopic shift (difference between average and monoisotopic
mass) [20]. Normalization of these two shifts results in two
non-additive and independent peptide properties, which were previously introduced as normalized nominal mass defect (NMD) and normalized isotopic shift (NIS) by Artemenko et al. [21]. Certain types of peptides are localized in distinct regions on a 2D plot of NMD versus NIS, due to different chemical com-position. For example, NMD tends to increase due to very basic (arginine, lysine) and aliphatic (leucine, isoleucine, valine) amino acids, but decreases in the presence of acidic (aspartic acid and glutamic acid) or sulfur-containing (cysteine and methionine) residues. Furthermore, it was shown that different families of
peptides can be distinguished from the 2D plot [21]. It should
be noted here that although this is an elegant approach, it has so only been used as a data representation method after de novo sequencing [21]. With the advent of modern high resolution (HR) mass spectrometry devices, it now becomes possible to use mass defect filtering of unknown peptides in complex samples.
In this study, we expand the mass mapping approach to select for cysteine-containing peptides prior to de novo sequencing, based on NMD, NIS, and overall mass. Although it is common practice to determine the monoisotopic mass of a peptide with high precision using modern high resolution mass spectrometers, the average mass is not a directly measurable value. The average mass can be calculated from the ion abundance of all detected isotopomers. We developed a data-dependent analysis workflow to experimentally determine the monoisotopic and average masses of unknown peptides, in order to calculate NIS and NMD values. By making a three-dimensional (3D) plot of NMD, NIS, and monoisotopic mass, peptides containing sulfur are differentiated from non-sulfur containing peptides, and those
can then be further targeted for extensive MSnbased analysis
and full primary structure elucidation. We demonstrate this workflow via the analysis of peptides in the secretions produced by the granular skin glands of two anuran amphibian species, Odorrana schmackeri and Bombina variegata.
Experimental
Granular Gland Skin Secretions
Skin secretion of Odorrana schmackeri and Bombina variegata
were a kind gift from Professor Chris Shaw, Queen’s University,
Belfast, UK. The crude lyophilized skin secretions (~100μg)
were resuspended in 100μL of 50 mM ammonium bicarbonate
containing 1 M urea and 25% acetonitrile. Insoluble material was removed by centrifugation at 17,800 relative centrifugal force (RCF) for 15 min. The supernatant was split in two parts; one part was directly measured by LC-MS and the other part was treated with 20 mM tris(2-carboxyethyl)phosphine (TCEP; Sigma-Aldrich, Munich, Germany) to reduce disulfide bonds.
LC-MS
Samples were analyzed by on-line nanoflow LC-MS using an Agilent 1200 HPLC system and a LTQ-Orbitrap Velos
(ThermoFisher Scientific, Bremen, Germany) mass spectrom-eter. Samples were injected onto an in-house made trap column
with dimensions 20 mm L × 100μm i.d., filled with particles of
5 μm diameter (Reprosil Pur C4; Dr. Maisch GmbH,
Ammerbuch-Entringen, Germany). After trapping, the peptides were separated on an in-house made analytical column (C4,
dimensions 140 mm L × 75μm i.d., 5 μm particle size). The
vented column setup was adjusted to an analytical column flow rate of 150–200 nL/min, solvent A was 0.6% acetic acid in Milli-Q, solvent B was 0.6% acetic acid in 80% acetonitrile in Milli-Q (v/v). The effluent of the column was directly sprayed into the ion source of the LTQ-Orbitrap Velos mass spectrom-eter. For determination of the monoisotopic and average mass, the most abundant multiply charged ions were automatically selected for HR zoom scans. Full Fourier transform MS scan width was set to the range 400–1500 m/z in profile mode with a resolution of 30,000 at 400 m/z. Multiply charged ions with a
minimal intensity of 2 × 105were selected for a zoom scan with
a resolution of 30,000 at 400 m/z, acquired in centroid mode
with an isolation width set from–1.50 to +2.25 Dalton (Da)
with respect to the selected precursor m/z value. For each zoom
scan four microscans (AGC target value set to 1 × 105,
maxi-mal injection time set to 500 ms) were combined. For the acquisition of HCD peptide fragmentation spectra, a second data-dependent analysis was performed. Fragmentation spectra were acquired at a resolution of 30,000 at m/z 400 starting with a fixed first mass at m/z 100. For each precursor, three separate HCD spectra were acquired at normalized HCD energy of 26, 28, and 30. The isolation window was set to 2.5 m/z, signal
threshold for selection was 1 × 105, AGC target was 5 × 105,
and maximum injection time was set to 200 ms.
3D Mass Mapping
The Xcalibur (ThermoFischer) *.raw files were converted to *.mzXML files using the ReAdW software (ver. 2.0, available athttp://sourceforge.net/projects/sashimi/files/) and *.mzXML files were subsequently imported into the in-house developed MATLAB script (MATLAB and Bioinformatics Toolbox, re-lease 2014a, The MathWorks, Inc., Natick, MA, USA). This script determines from each zoom scan spectrum the charge of the ion and calculates its monoisotopic mass. The average mass of each peptide is calculated from the peak height (ion intensi-ty) and corresponding m/z values of all observed isotopomers above 3% in a zoom scan. This MATLAB script is available as
Supplementary Material (Script-I).
For the determination of theoretical NMD and NIS values of all reviewed peptide/protein sequences from the animal venom and toxin annotation program [1,2], all sequences smaller than 35 amino acids were downloaded in FASTA format, which allowed for direct import into MATLAB. From the primary amino acid sequence, the elemental composition was deter-mined. Using this theoretical monoisotopic mass, nominal mass and average mass were calculated. The experimental
nominal mass was calculated using Equation1, and NIS and
NMD values were calculated using Equations 2 and 3,
respectively. For a reference see Artemenko et al. [21]. This
MATLAB script is available as Supplementary Material (Script-II).
Nominal mass¼ integer 0:9995 monoisotopicmassð Þ ð1Þ
NIS¼ 1000 averagemass – monoisotopicmassð Þ=monoisotopicmass
ð2Þ
NMD¼ 1000 monoisotopicmass – nominalmassð Þ=monoisotopicmass
ð3Þ
Database Searching
Database searching of HCD fragmentation spectra acquired from both O. schmackeri and B. variegata was done using MASCOT (MASCOT server 2.2, Matrixscience Inc, Boston, MA, USA) against the UniProt entries anura (Tax ID 8342): protein and peptides sequences with 35 amino acids or less (downloaded in January 2015). The search parameters were set to the following; no enzyme, 5 ppm parent tolerance, 0.02 Da fragment tolerance, no enzyme specificity. C-terminal amidation was included as variable modification. De novo sequencing of selected peptides was done by combining information obtained by manual inter-pretation of the fragmentation spectra from different charge states of the peptide.
Results and Discussion
Disulfide bridges are a common PTM encountered in animal toxin and venom peptides. To illustrate this, over 85% of all reviewed peptides in the UniProt Toxin and Venom database [1,2] contain at least one disulfide bond. For all these peptide sequences up to 35 amino acids (a total of 1050 sequences), the mass, NIS and NMD were calculated (using Supplementary Information, Script-II) and plotted (Figure1) using and a color code to differentiate the number of cysteines. A general trend is observed in which peptides with a molecular weight between 1000 and 2000 Da and no cysteines have low NIS (0.55–0.65)
and high NMD values (0.50–0.75) (Figure1aandd). Peptides
with one or more cysteines shift to higher NIS and lower NMD
values (Figure1b), which is accompanied by increasing
mo-lecular weight. Based on the number of cysteines, distinct
clusters can be discerned in the 3D plot in Figure1. Using such
a 3D mass map, threshold values for both NMD and NIS were inferred that can be used to select peptides for de novo sequencing.
Figure1b–dillustrate the general trend that peptides with a
NMD below 0.55, and a NIS higher than 0.65 are likely to contain cysteine residues. For some peptides with a high
number of cysteines, NIS values can exceed 0.8 and NMD can even decrease to below 0.35. It should be noted here that besides cysteines, methionine residues contain a sulfur atom as well. Upon investigation of the amount of methionine resi-dues present in this collection of sequences, it was established that peptides containing solely methionine as sulfur source also have a distinct localization in the plot, which differs from the cysteine-containing peptides (Supplementary Information,
Figure I). This could due to the nature of the peptide (i.e.,
belonging to specific peptide families) or due to the fact that most methionine-containing peptides have only a single methi-onine-residue, whereas cysteine-containing peptides typically have multiple cysteine residues. As demonstrated for this col-lection of venom and toxin peptides, 3D mass mapping allows for filtering of cysteine-containing peptides from non-cysteine-containing peptides, even in cases where the latter contain a methionine residue. Our experimental determination of NMD and NIS values of unknown peptides is based on high resolu-tion MS and zoom-scans. Determinaresolu-tion of NIS and NMD values directly from the survey full scan MS spectrum would require a further advanced algorithm combining isotopic peak pattern and peak shape recognition. The use of a
data-dependent MS2based strategy based on zoom-scans works
with much less complicated data processing.
The skin secretion of Odorrana schmackeri is relatively well characterized and so far the primary structures of 27 peptides are listed in the UniProt protein database. The majority (25) of them have a C-terminal disulfide bond,
often referred to as ‘Rana-box’ [22]. Skin secretion of
O. schmackeri was reduced with TCEP and separated by nano-HPLC during a 120 min analysis. The base peak
intensity chromatogram is depicted in Figure 2a. The MS
system was programmed to acquire full scan spectra from m/z 400 to 1500. For each eluting multiply charged peptide exceeding the preset ion intensity threshold, a high resolu-tion zoom-scan was acquired. A zoom-scan is the acquisi-tion of the ion in a small m/z window that is written as a separate scan event in the raw data file. To avoid spectral contamination by other ions, an optimized zoom-scan
win-dow of–1.50 to +2.25 Da was used. Within the mass range of
all peptides analyzed in this study and the multiple charging nature of electrospray ionization, all isotopomers above 1%–2% of a single peptide are visible within this window. The *.raw data-file was converted to an *.mzXML data-file which was
0.3 0.4 0.5 0.6 0.7 0.6 0.7 0.8 0.9 0 1000 2000 3000 4000 NMD NIS mass (Da) 0 cys 1 cys 2 cys 3 cys 4 cys >4 cys 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 NMD NIS 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0 500 1000 1500 2000 2500 3000 3500 4000 4500 mass (Da) NMD 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 0 500 1000 1500 2000 2500 3000 3500 4000 4500 mass (Da) NIS
(c)
(a)
(d)
(b)
Figure 1. (a) Three-dimensional plot of NMD versus NIS versus monoisotopic mass (Da) of peptides smaller than 35 amino acids in UniProt animal venom and toxin database. Inset in (a) shows color scale used to label amount of cysteine residues in a sequence. (b) Two-dimensional plot (side view) of peptide mass versus NMD. (c) Side view of NIS versus NMD. (d) Side view of peptide mass versus NIS
imported into MATLAB, where all zoom scans were converted into a MATLAB structure. A script was programmed to process all zoom scans for calculation of both average and monoisotopic
mass (Supplementary Information, Script-I).
The processing consists of several steps. First, all zoom scans of poor quality indicated by a low total ion count are discarded. For all remaining zoom scans, noise peaks are re-moved by discarding all intensity peaks below 3% of the most abundant peak. Using the remaining peaks, the charge state is determined, followed by calculation of average and monoiso-topic mass. As an example, the most abundant peptide in the O. schmackeri skin secretion eluting at 68 min was selected. This peptide was identified as Odorranain-C7HSa (UniProt accession: B4ERK5_ODOSH) with amino acid sequence SLLGTVKDLLIGAGKSAAQSVLKGLSCKLSKDC, and
Figure2bshows the data-dependent zoom scan of the [M +
4H]4+. From the abundance of all isotopomers, the calculated
NMD and NIS values are 0.56 and 0.62, respectively. These values are close to the theoretical values of 0.55 and 0.63, as calculated based on the chemical composition derived from the amino acid sequence.
To determine the accuracy of our workflow, a separate measurement was performed in which high resolution HCD fragmentation spectra were acquired. A MASCOT database
search was performed against anura protein entries listed in the UniProt database (search results are provided in
Supplementary Information, Table1). A total of 35 protein
identifications were made using 182 peptide spectral matches. Using the identified peptides, theoretical NIS and NMD values were calculated and compared with the exper-imental determined values. For a total of 104 peptides, we compared experimental monoisotopic mass and average mass with the theoretically calculated values (results are
provided in Supplementary Information, Table2). From this
comparison, the error between experimental and theoretical
monoisotopic mass and average mass were –0.73 and
+26.2 ppm, respectively (see Figure 2c), indicating that
our computation of average mass and subsequent calcula-tion of NIS and NMD values are relatively accurate.
Figure 2d and e show, respectively, the 3D plot of mass,
NMD, and NIS and the 2D plot of NMD and NIS values for O. schmackeri. The majority of detected peptides have NMD values in the range of 0.45 to 0.50 and have NIS values in the range of 0.60 to 0.70, indicating a high pro-portion of disulfide bond containing peptides in this spe-cies’ skin secretion. This is also in agreement with the high number of disulfide bonds containing peptides identified in previous studies. 0 0 10 20 30 40 50 60 70 80 90 100 110 Time (min) 100 Relative Abundance 120 50 0.3 0.4 0.5 0.6 0.7 0.6 0.7 0.8 0.9 0 1000 2000 3000 4000 5000 6000 7000 NMD NIS mass (Da) 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 NMD NIS -200 -150 -100 -50 0 50 100 150 200 -6 -4 -2 0 2 4 6 ppm ppm monoisotopic mass average mass 826.0 826.5 827.0 827.5 828.0 828.5 829.0 0 50 100 Relative Abundance 826.966 100.0% 827.21598.9% 827.465 69.9% 826.716 57.0% 827.715 36.2% 827.966 18.3% 828.216 7.4% m/z NMD: 0.56 (0.55) NIS: 0.62 (0.63)
(a)
(c)
(b)
(d)
(e)
Figure 2. (a) Base peak intensity chromatogram of nanoflow HPLC-MS analysis of skin secretion of Odorrana schmackeri. (b)
Zoom scan of [M + 4H]4+ of peptide Odorranain-C7HSa (UniProt accession: B4ERK5_ODOSH) with amino acid sequence
SLLGTVKDLLIGAGKSAAQSVLKGLSCKLSKDC. Experimentally determined and theoretical values (in parenthesis) for NIS and NMD are given. (c) Box plots showing error distribution in the determination of monoisotopic and average mass (n = 104). (d) 3D plot of peptide mass, NIS, and NMD of unreduced O. schmackeri. (e) 2D plot of NIS and NMD
To further validate our approach, a peptide with a NMD of 0.49 and a NIS of 0.63, which had not been identified by the database search, was chosen for further primary structure elu-cidation. This was achieved through extended study of the fragmentation spectra of different charge states of the peptide’s
precursor ion (see Figure3 for the deconvoluted HCD
frag-mentation spectra of the peptide’s [M + 3H]3+
and [M + 4H]4+
ion). Interestingly, the two fragmentation spectra show relative-ly large differences in the pattern and degree of fragmentation. In the high mass-over-charge range the tandem MS spectrum of the 3+ ion shows a predominant series of b ions, whereas the fragmentation spectrum of the 4+ precursor reveals more prev-alent y-ions in this region. In addition, the 4+ spectrum (Figure3a) shows a relatively large degree of internal fragments around m/z 500–800, which are not seen for the 3+ spectrum
(Figure3b). Based on a combined analysis of both spectra a
complete amino acid sequence of TSRCYVGYRHK[I/L]VCS is proposed, which was not possible from studying the individ-ual fragmentation spectra. This underlines the difficulty of de novo sequencing non-tryptic peptides. A BLAST search against this sequence reveals a high sequence similarity of this peptide with the antimicrobial peptide Odorranain-T2-HN1 with amino acid sequence TSRCYVGYRRKIVCS (UniProt accession number: E7EKE4). The only difference between the newly identified O. schmackeri peptide and the earlier reported one from O. hainanensis is a histidine residue at position 10, instead of an arginine. It is therefore not unlikely that the newly discovered peptide has antimicrobial activity as
well. This illustrates that even in a well-studied species, our 3D mass map screening approach helps to identify novel structures of potentially bioactive peptides.
To explore the potential of our cysteine-containing peptide discovery strategy even further, the same workflow was applied to the skin secretion of Bombina variegata. From the skin of this particular frog, several peptides have been characterized previ-ously, including the large family of antimicrobial bombinins and
bombinin-like peptides [23,24]. However, only a few frog skin
peptides with one or more internal disulfide bonds have been described up till now (i.e., the insulin releasing peptides kinino-gen 1 and 2 [25], prokineticin [26], and the trypsin and thrombin inhibitor BSTI [27]. In Figure4, the 3D and the 2D map of mass versus NMD are shown for both the reduced and the unreduced B. variegata sample. From this plot, several candidate peptides with low NIS and high NMD values were selected for a targeted
MS2analysis. Table1lists the results from this analysis, and
many of the sequenced peptides have high similarity with kinogen 1 and 2 associated peptides. Also peptides with simi-larity to prokineticin and BSTI were found, although complete sequencing was not possible from the acquired spectra. Besides the peptides listed above, several novel structures were found and two of these will be discussed in more detail below.
Example I
One of the most peculiar examples is the peptide that has an apparent mass of 7163.273 Da (unreduced sample). This mass
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 m/z 0 100 0 100 Relative Abundance Relative Abundance [M+H]+ 1771.874 [M+H]+ 1771.870 y''14 1670.823 y''13 1583.791 y''12 1427.690 y''9 1062.550 b8 930.414 y''10 1161.617 b10 1223.519 b12 1464.753 b11 1351.669 b13 1563.822 b14 1666.831 b9 1086.515 b6 710.330 b10 1223.519 y''9 1062.550 a6 682.333 y''3 308.128 y''1 106.050 y''2 209.059 y''2 209.059 y''4 421.212 y''5 549.306 b11 1351.669 y''6 686.366 a13 1535.827 a12 1436.754 a5 583.265 GYRHKIV 854.498 GYRHKIVC 957.506 GYRHKI 755.430 b5 611.259 iY 136.076 b4 448.196 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 m/z
(b)
(a)
T S R C Y V G Y R H K L V C S
y5 y6 y9 y4y3 y2 y1 b6 b8 b9 b10b11b12b13b14T S R C Y V G Y R H K L V C S
y14y13y12 y10 y9 y2 b4 b5 b10b11b12b13b14 b12 1464.748 b14 1666.831 b13 1563.822Figure 3. HCD fragmentation spectra of (a) [M + 4H]4+at m/z 443.724 and (b) of the [M + 3H]3+at m/z 591.297 of peptide
does not appear on the mass maps after reducing all the disulfide bonds in the sample. Consistently, two novel peptides with masses of 4983.389 and 2187.956 Da appeared in the reduced sample, both of them not present in the unreduced sample. The mass difference between single unreduced peptide and the two reduced (complementary) peptides is 8.071 Da, which corre-sponds to the mass increase upon reduction of four disulfide bonds. The HCD fragmentation spectrum of the large
unreduced peptide (Figure 5a) shows a relatively low degree
of fragmentation. From the fragmentation spectrum of the re-duced peptide with molecular weight of 4983.389 Da
(Figure 5b), a near complete sequence assignment could be
made (except for distinction between isoleucine and leucine). From the fragmentation spectrum of the reduced peptide with
molecular weight of 2187.956 Da (Figure5c), a near complete
sequence assignment could be made with only the 4–6N-termi-nal residues unknown. More importantly, several of the se-quence ions observed for this reduced peptide are identical to fragment ions seen in the fragmentation spectrum of the large
unreduced peptide (inset of Figure 5a), confirming the
unreduced peptide is a dimeric peptide of the two reduced peptides. Also, the number of identified cysteine residues (eight in total) confirms the expected presence of four disulfide bonds. The carboxy-terminal end of the small peptide has strong similarity with the consensus sequence CCXXXXCN of
3-finger toxins, found in snakeα-neurotoxins and frog
cyto-toxins [31,32]. BLAST analysis of the two sequences against
the NCBI nonredundant protein database shows highest se-quence similarity to the secreted Ly-6/uPAR-related protein-1-like from Xenopus tropicalis (NCBI accession number: XP_002939220.1). For the larger chain, fragmentation at the
last four C-terminal residues was not complete (Figure 5b).
Based on the high similarity with the X. tropicalis sequence, a sequence assignment for the last four residues can be proposed.
Figure5dshows the sequences of the two Bombina variegata
peptides and the Ly-6/uPar-related peptide and the xenoxins, cytotoxins identified from the dorsal gland secretions of Xenopus laevis. Besides identical positions for all the cysteines, the sequences of B variegata and X. tropicalis show several other conserved residues. The sequence of X. tropicalis
0.3 0.4 0.5 0.6 0.7 0.6 0.7 0.8 0.9 0 1000 2000 3000 4000 5000 6000 7000 8000 NMD NIS mass (Da) 0.3 0.4 0.5 0.6 0.7 0.6 0.7 0.8 0.9 0 1000 2000 3000 4000 5000 6000 7000 8000 NMD NIS mass (Da) 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0 1000 2000 3000 4000 5000 6000 7000 8000 NMD mass (Da) 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0 1000 2000 3000 4000 5000 6000 7000 8000 NMD mass (Da)
(a)
(b)
(d)
(c)
Figure 4. 3D mass mapping results of unreduced and reduced skin secretion of Bombina variegata. Shown are 3D plot of mass, NMD, and NIS of (a) unreduced and (b) reduced peptides, and 2D plot of mass versus NMD of (c) unreduced and (d) reduced peptides
originates from DNA sequencing and evidence at protein level or function is not known yet. The similarity with the xenoxins previously shown to be present in the skin secretion of X. laevis [33,34] is much lower (Figure5d). Interestingly, the prototype and other described members of this family of 3-finger toxin proteins are all known to be composed of a single polypeptide chain. Remarkably the B. variegata variant discovered here is composed of two chains, forming an intermolecular S–S-linked heterodimer. Most likely this heterodimer is synthesized by processing of a single polypeptide chain. Reanalysis of the
MS and MS2data did not show evidence for the presence of
this full length single polypeptide and it also did not yield in the detection of other heterodimeric forms of this peptide, suggest-ing that the observed heterodimer is specifically formed and not the result of specific degradation.
Example II
We discovered two more novel heterodimeric B. variegata peptides (with masses of 3469.924 and 3541.942 Da). The sequences of these two homologous dimeric peptides are, re-spectively, (LQSLHKLRWPGKPLLLCENENGKLRPLamide-CLS and
LQSLHKLRWPGKPLLLCENENEKLRPLamide-CLS) (Table1). The sequences are given with all leucines, but
distinction between isoleucine and leucine could not be made.
Figure6shows the fragmentation spectra of the reduced and
unreduced form of the peptide with mass 3469.924 Da. The spectra for the peptide with mass 3541.942 Da are given in
Supplementary Figure 2. Both dimeric peptides consist of a
chain with 27 amino acids, which is amidated at the C-terminal residue. They only differ from each other with either a glutamic acid or glycine at position 22. By analysis of the HCD fragmentation spectra of the charge states 4+ to 6+, a complete sequence assignment could be made (except the I/L
distinguishing) and Figure6a shows the HCD fragmentation
spectrum with a, b, and y ion annotation of the 5+ charge state. The other chain of both heterodimers is identical for both and consists of only three amino acids. This chain was not recovered in the analysis of the reduced sample, but the fragmentation spectra of both unreduced dimers evidently showed that in the small chain the Cys is the N-terminal amino acid flanked by an
Ile/Leu and Ser at the C-terminus (Figure6band Supplementary
Figure2B). This was deduced from the loss of three single
amino acids from the precursor ion, which included the C-terminal amidated Ile/Leu of the large chain, and the losses of a C-terminal Ser. In the low m/z region, in the reduced peptide, internal fragments corresponding to CysGlu and (ILe/Leu)Cys were observed, whch were not observed for the unreduced peptide. Simultaneously, for the unreduced dimeric peptide, a y2 ion at m/z 219.134 was observed, corresponding to the two C-terminal amino acids, (Ile/Leu)Ser of the small chain. The two latter observations prove that both peptides are indeed a disulfide bond-linked dimeric peptide. BLAST analysis against anura sequences in the NCBInr database showed no significant similarity with any other known frog skin peptide, indicating this is a novel structure.
Ta bl e 1. Cysteine-Containing P eptides Identified from the Skin Secretion o f B. va rieg ata Mon o is otop ic ma ss (D a) a NMD NIS Number of SS-l inks Amin o acid seq ue nc e b, c Identi ty or similarit y Ref. 155 7.732 0. 47 0. 66 1 D LCTFT S PG KVKCY-am ide N ovel 211 1.971 0. 46 0. 68 1 A THNMHWHR KPCNGPLLC Similar to K inin oge n -2-a ssociate d p ep tide [ 28 ] 227 2.856 0. 38 0. 65 1 F NQEDCFHEYLQNDAYC H N ovel 229 9.091 0. 47 0. 71 2 L YNALWPCKHCNK CKPGLLC Similar to K inin oge n -1-as sociate d p ep tide [ 25 , 28 ] 232 9.882 0. 38 0. 67 1 G FNQEDCFHEYI QNDAYCH N ovel 249 3.259 0. 50 0. 65 1 F KAPYNIH WHCKPGLLCKNIN S im ilar to Kinin oge n -2-a ssociate d p ep tide [ 28 ] 303 4.483 0. 48 0. 66 1 L KYLYWPCKPGMPC E N V D (898. 477) Kininogen-2-ass o ci ated peptide [ 28 ] 380 3.893 0. 50 0. 65 1 L GSE (385. 245)LKYLYWPC KPGM (1 554.7 35) Kininogen-2-ass o ci ated peptide [ 28 ] 321 2.546 0. 48 0. 65 1 D MYELKGYKSAHGR PRVCPPGEQCPLWV-a m ide S imilar B radykinin inhibitor peptide DV-28 [ 29 ] 346 9.924 0.55 0.63 1 C hain 1 : LQSL HKLRWPGKPLLLCENENGKLRPL-amide Chain 2 : C LS Novel 354 1.942 0.54 0.62 1 C hai n 1 : LQSLHKLRWPGKPLLLCENENEKLRPL-ami de Chain 2 : C LS Novel 641 3.914 0. 46 0. 69 10 NFVCPPQGS(5494. 6008) Similar to B STI [ 27 , 30 ] 652 4.986 0. 46 0. 68 10 NFVCPPQGS(5605. 6698) Similar to B STI [ 27 , 30 ] 716 3.294 0.46 0.68 8 C hai n 1 : LKCNTLEGRGVQ AT L C PPGKE T C MTHSVLL N GNT N L M KGCATFSRCS Chain 2 : S VGSNR L S ESLYCC NLNL C N Novel 815 9.709 0. 46 0. 66 10 AVLTGT (374. 138)D VQCGSGTCCA T S VW S(5668. 750) Similar to P ro kineticin B v8 [ 26 ] aMe as ur ed m ono iso topi c m ass es o f u nre du ce d pe pt ide bSe que nc es ar e g ive n with al l le uci ne s; exac t d istin ct ion b et wee n le uc ine an d iso le uci ne cou ld n ot be ma de cIn ca se s w he re on ly pa rtia l se que nc e coul d b e d eriv ed , re ma ini ng m ass es ar e g iv en in p ar en the se s
0 10 20 30 40 50 60 70 80 90 100 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 m/z Relative Abundance 1698.776 5 6 5 . 2 0 1 1 661.281 1585.694 1265.628 0 10 20 30 40 50 60 70 80 90 100 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 m/z Relative Abundance 0 10 20 30 40 50 60 70 80 90 100 Relative Abundance 1024.902 z=7 1006.180 z=7 159.113 z=1 465.725 z=2 343.690 z=2 187.108 z=1 1071.521 z=4 1208.355 z=5 m/z 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1164.703 z=6 509.241 z=2 990.164 z=7 714.368 z=1 1248.774z=5 1510.436z=4 656.792 z=1 150 200 250 300 350 400 450 500 550 600 650 700 m/z 465.725 z=2 343.690 z=2 401.203 z=2 509.241 z=2 500.235 z=2 456.719 z=2 556.778z=2 I/L 113.084 S 87.032 E 129.042 357.687 z=2 401.203 z=2 714.368 z=1 656.792 z=1 159.113 z=1 187.108 z=1 129.102 z=1 133.061 z=1 490.278 S 87.032
(a)
(c)
(b)
Bombina variegata peptides LKCNTLEGRGVQATLCPPGKETCMTHSVLLNGNTNLMKGCATFSRCS xxxxxSHSESLYCCNLNLCN
XP_002939220.1 LKCNTLTEGRKEVTECPPGQTLCMTHSLTNNNKTDLTKGCTTFGRCARRDVFLYESEKLYCCNLDLCN
Xenoxin-1 (XEN1_XENLA) LKCVNLQANGIKMTQECAKEDTKCLTLRSLKKTLKFCASGRTCTTMKIMSLPGEQITCCEGNMCNA Xenoxin-2 (XEN2_XENLA) LKCVNLQANGIKMTQECAKEDNKCLTLRSLKKTLKFCASDRICKTMKIMSLPGEKITCCEGNMCNA Xenoxin-3 (XEN3_XENLA) LKCVNLQANGVKMTQECAKEDTKCLTLRSLKKTLKFCASDRICKTMKIASLPGEQITCCEGNMCNA
(d)
x(4-6) [M+H]+ 2188.961 bn-1 2056.908 bn-2 1953.890 bn-3 1840.814 bn-4 1726.771 bn-5 1613.687 bn-6 1499.643 bn-7 1396.634 bn-8 1293.624 bn-9 1130.560 bn-10 1017.477 bn-11 930.444 bn-13 714.369 bn-14 577.310 bn-15 490.278X X X X S H S E S L Y C C N L N L C N
y4 y3 y1 bn-9 bn-7bn-6bn-5bn-4bn-3bn-2bn-1 bn-12 801.402 bn-8 bn-10 bn-11 bn-12 bn-13 bn-14 bn-15 y1 133.061 NL LN 228.135 CNL LNC 331.144 y3 349.155 y4 463.198 [M+H]+ 4984.386 y45 4743.206 y46 4871.303 y44 4640.193 y41 4312.024 y40 4182.979 y31 3297.530 y30 3200.472 b27 2856.363 y21 2216.061 y22 2353.121 y19 2029.958 y32 3400.535 y4 452.192 y8 874.355 y9 931.376 y11 1190.511 b16 1687.862 y10 1059.470 y17 1803.792 y18 1916.878 y20 2129.030 b15 1584.853 b2 242.186 y43 4526.154 y42 4425.103 b35 3681.795 b36 3794.781 b34 3567.753 b3 354.195 y5 599.260 y6 700.308 y7 771.346 y12 1303.593 y13 1417.637 y14 1518.687 y23 2454.169 y24 2585.206 y25 2688.220 y27 2918.306 b29 3068.514 b28 2955.434 y26 2789.267 b 37 3925.941L K C N T L E G R G V Q A T L C P P G K E T C M T H S V L L N G N T N L M K G C A T F S R C S
y45 b2 y46 y44 b3 y43y42y41 y32 b15 y31 b16 y30 y27y26y25y24 y23y22y21y20 b27 y19 b28 y18y17 y14y13 b34 y12 b35 y11 b36 y10 b37 y9 y8 y7 y6 y5 y4 b45 b45 4776.329Figure 5. (a) HCD fragmentation spectra of [M + 7H]7+at m/z 1024.332 of unreduced B. variegata peptide of monoisotopic mass
7163.273 Da. Inset shows short series of sequence ions. (b) HCD fragmentation spectrum of [M + 6H]6+at m/z 831.572 of reduced
B. variegata peptide of monoisotopic mass 4983.389 Da. (c) HCD fragmentation spectrum of [M + 3H]3+at m/z 730.326 of reduced
B. variegata peptide with monoisotopic mass of 2187.956. (d) Sequences of the B. variegata peptides, secreted Ly-6/uPAR-related protein 1-like from X. tropicalis (NCBI refseq: XP_002939220.1), and 3 xenoxins from skin secretion of X. laevis, with obvious similarities
Conclusions
The normalized nominal mass defect and normalized isotope shift of peptides containing cysteine residues renders them a specific spatial distribution in 2D and 3D mass maps. We developed a MATLAB script to compute both values from
experimentally obtained high resolution (orbitrap) mass spec-trometry data. Together with 3D and 2D mass mapping, the script considerably enhances the efficiency with which disulfide bridge containing peptides can be detected in complex biological samples, such as amphibian skin defensive secretions, and can be
200 m/z 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Relative Abundance y12 1384.741 a15 1740.085 3021.713 a14 1627.000 y13 1497.825 y25 2910.669 y4 497.358 y9 1039.603 y1 131.118 y7 796.517 3151.812 a16 1853.168 y8 925.561 y22 2573.489 y23 2710.555 b18 2113.210 3151.809 1b26 3340.820 1b12 1444.855 1a15 1740.081 1b2 242.150 1y4 497.356 1590.778 3096.749 2910.661 1b15 1768.076 1y6 682.472 1y3 384.272 2010.013 2227.262 2764.522 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 y6 682.474 y3 384.273 b17 1984.173 y21 2445.393 1y7 796.516 1y9 1039.601 y10 1168.646
L Q S L H K L R W P G K P L L L C E N E N G K L R P L
-amide y26 b2 y25y24y23y22y21y20 y18y17y16 b18 y9 b17 y10 a16 b16 y11 a15 b15 y12 y13 y14 y15 b19 y8 y7 y6 y5 y4y3 y2 y1 b15 1768.078 y26 3038.724 b2 242.151 y11 1271.656 b19 2227.270 b20 3300 3350 3400 3450 m/z 1b26 3340.820 2b2 3365.883 3470.921 -S-OH -105.038 Relative Abundance(b)
(a)
1y10 1168.644 [M+H]1+ 1b11 1316.759 [M+H]1+ 3470.921 [M+H]1+ -L -113.08 1y26 3357.840 1b13 1541.908L Q S L H K L R W P G K P L L L C E N E N G K L R P L
-amideC L S
-Lamide -130.101 1y 26 1b 2 1y25 1y101y91y81y71y61y51y41y31y21y1 2b 2 2y 2 1b 121b131b141b151b16 a16 1853.168 210 215 220 225 230 235 240 m/z 1b2 242.150 1b2-NH3 225.124 PL 211.144 2y2 219.134 1y2 228.171 1b2 242.151 y2 228.171 PL 211.145 1b2-NH3 225.124 LC 217.101 CE 233.060 210 215 220 225 230 235 240 m/z 200 m/z 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400Figure 6. HCD fragmentation spectra of heterodimeric disulfide bond linked peptide from B. variegata with molecular mass of
3469.924 Da. (a) Deconvoluted HCD spectrum of the large reduced chain, ([M + 6H]6+at m/z 526.141 (monoisotopic peak) was
selected for fragmentation). (b) Deconvoluted HCD spectrum of unreduced dimeric peptide, ([M + 6H]6+at m/z 579.329
considered as a tool to discover peptides that require this structural element for biological activity and stability. After illustrating the concept by analyzing a subset of the UniProt database representing a collection of animal toxin sequences, we validated our approach experimentally on the skin secretion of two frog species. One of these species (O. schmackeri) is known to contain many such S–S linked peptides; the other (B. variegata) is much less studied in this respect. The workflow we elaborated indeed selects previously reported peptides with intramolecular S–S bonds. In addition, from B. variegata skin secretion novel intra-and inter-molecular disulfide bridge-forming peptides were dis-covered, confirming the usefulness of this novel approach.
Acknowledgments
The authors thank Joska van Oosten (http://www.joskaasworst.
com/) for the design of the graphical abstract.
Open Access
This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unre-stricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
References
1. Jungo, F., Bairoch, A.: Tox-Prot, the toxin protein annotation program of the Swiss-Prot protein knowledgebase. Toxicon45(3), 293–301 (2005) 2. Jungo, F., Bougueleret, L., Xenarios, I., Poux, S.: The
UniProtKB/Swiss-Prot Tox-UniProtKB/Swiss-Prot program: a central hub of integrated venom protein data. Toxicon60(4), 551–557 (2012)
3. Escoubas, P., King, G.F.L.: Venomics as a drug discovery platform. Expert. Rev. Proteome6(3), 221–224 (2009)
4. King, G.F.: Venoms as a platform for human drugs: translating toxins into therapeutics. Expert. Opin. Biol. Ther.11(11), 1469–1484 (2011) 5. Uhlig, T., Kyprianou, T., Martinelli, F.G., Oppici, C.A., Heiligers, D.,
Hills, D., Calvo, X.R., Verhaert, P.D.E.M.: The emergence of peptides in the pharmaceutical business: From exploration to exploitation. EuPA Open Proteome4, 58–69 (2014)
6. Chen, T., Farragher, S., Bjourson, A.J., Orr, D.F., Rao, P., Shaw, C.: Granular gland transcriptomes in stimulated amphibian skin secretions. Biochem. J.371(Pt 1), 125–130 (2003)
7. Swanson, S.K., Washburn, M.P.: The continuing evolution of shotgun proteomics. Drug Discov. Today10(10), 719–725 (2005)
8. Escoubas, P., Quinton, L., Nicholson, G.M.: Venomics: unravelling the complexity of animal venoms with mass spectrometry. J. Mass Spectrom. 43(3), 279–295 (2008)
9. Ueberheide, B.M., Fenyö, D., Alewood, P.F., Chait, B.T.: Rapid sensitive analysis of cysteine rich peptide venom components. Proc. Natl. Acad. Sci. U. S. A.106(17), 6910–6915 (2009)
10. Frank, A., Pevzner, P.: PepNovo: de novo peptide sequencing via proba-bilistic network modeling. Anal. Chem.77(4), 964–973 (2005)
11. Chi, H., Chen, H., He, K., Wu, L., Yang, B., Sun, R.X., Liu, J., Zeng, W.F., Song, C.Q., He, S.M., Dong, M.Q.: pNovo+: de novo peptide sequencing using complementary HCD and ETD tandem mass spectra. J. Proteome Res.12(2), 615–625 (2013)
12. Quinton, L., Demeure, K., Dobson, R., Gilles, N., Gabelica, V., De Pauw, E.: New method for characterizing highly disulfide-bridged peptides in complex mixtures: application to toxin identification from crude venoms. J. Proteome Res.6(8), 3216–3223 (2007)
13. Neitz, S., Jürgens, M., Kellmann, M., Schulz-Knappe, P., Schrader, M.: Screening for disulfide-rich peptides in biological sources by carboxyamidomethylation in combination with differential matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Commun. Mass Spectrom.15(17), 1586–1592 (2001)
14. Zhang, Y., Cui, W., Zhang, H., Dewald, H.D., Chen, H.: Electrochemistry-assisted top-down characterization of disulfide-containing proteins. Anal. Chem.84(8), 3838–3842 (2012)
15. Samgina, T.Y., Vorontsov, E.A., Gorshkov, V.A., Artemenko, K.A., Nifant'ev, I.E., Kanawati, B., Schmitt-Kopplin, P., Zubarev, R.A., Lebedev, A.T.: Novel cysteine tags for the sequencing of non-tryptic disulfide peptides of anurans: ESI-MS study of fragmentation efficiency. J. Am. Soc. Mass Spectrom.22(12), 2246–2255 (2011)
16. Wang, Z., Zhang, Y., Zhang, H., Harrington, P.B., Chen, H.: Fast and selective modification of thiol proteins/peptides by N-(phenylseleno)phthalimide. J. Am. Soc. Mass Spectrom.23(3), 520–529 (2012)
17. Thompson, A., Schäfer, J., Kuhn, K., Kienle, S., Schwarz, J., Schmidt, G., Neumann, T., Johnstone, R., Mohammed, A.K., Hamon, C.: Tandem mass tags: a novel quantification strategy for comparative analysis of complex protein mixtures by MS/MS. Anal. Chem.75(8), 1895–1904 (2003) 18. Murray, C.I., Uhrigshardt, H., O'Meally, R.N., Cole, R.N., Van Eyk, J.E.:
Identification and quantification of S-nitrosylation by cysteine reactive tan-dem mass tag switch assay. Mol. Cell. Proteome11(2), M111.013441 (2012) 19. Gygi, S.P., Rist, B., Gerber, S.A., Turecek, F., Gelb, M.H., Aebersold, R.: Quantitative analysis of complex protein mixtures using isotope-coded affinity tags. Nat. Biotechnol.17(10), 994–999 (1999)
20. Sleno, L.: The use of mass defect in modern mass spectrometry. J. Mass Spectrom.47(2), 226–236 (2012)
21. Artemenko, K.A., Zubarev, A.R., Samgina, T.Y., Lebedev, A.T., Savitski, M.M., Zubarev, R.A.: Two dimensional mass mapping as a general method of data representation in comprehensive analysis of complex molecular mixtures. Anal. Chem.81(10), 3738–3745 (2009)
22. Park, J.M., Jung, J.E., Lee, B.J.: Antimicrobial peptides from the skin of a Korean frog, Rana rugosa. Biochem. Biophys. Res. Commun.205(1), 948– 954 (1994)
23. Simmaco, M., Barra, D., Chiarini, F., Noviello, L., Melchiorri, P., Kreil, G., Richter, K.: A family of bombinin-related peptides from the skin of Bombina variegata. Eur. J. Biochem.199(1), 217–222 (1991)
24. Simmaco, M., Kreil, G., Barra, D.: Bombinins, antimicrobial pep-tides from Bombina species. Biochim. Biophys. Acta1788(8), 1551– 1555 (2009)
25. Marenah, L., Flatt, P.R., Orr, D.F., McClean, S., Shaw, C., Abdel-Wahab, Y.H.: Skin secretion of the toad Bombina variegata contains multiple insulin-releasing peptides including bombesin and entirely novel insulinotropic structures. Biol. Chem.385(3/4), 315–321 (2004) 26. Mollay, C., Wechselberger, C., Mignogna, G., Negri, L., Melchiorri, P.,
Barra, D., Kreil, G.: Bv8, a small protein from frog skin and its homologue from snake venom induce hyperalgesia in rats. Eur. J. Pharmacol.374(2), 189–196 (1999)
27. Mignogna, G., Pascarella, S., Wechselberger, C., Hinterleitner, C., Mollay, C., Amiconi, G., Barra, D., Kreil, G.: BSTI, a trypsin inhibitor from skin secretions of Bombina related to protease inhibitors of nematodes. Protein Sci.5(2), 357–362 (1996)
28. Chen, T., Orr, D.F., Bjourson, A.J., McClean, S., O'Rourke, M., Hirst, D.G., Rao, P., Shaw, C.: Novel bradykinins and their precursor cDNAs from European yellow-bellied toad (Bombina variegata) skin. Eur. J. Biochem.269(18), 4693–4700 (2002)
29. Wang, L., Chen, Y., Yang, M., Zhou, M., Chen, T., Sui, D.Y., Shaw, C.: Peptide DV-28 amide: an inhibitor of bradykinin-induced arterial smooth muscle relaxation encoded by Bombina orientalis skin kininogen-2. Peptides31(5), 979–982 (2010)
30. Chen, T., Shaw, C.: Identification and molecular cloning of novel trypsin inhibitor analogs from the dermal venom of the Oriental fire-bellied toad (Bombina orientalis) and the European yellow-bellied toad (Bombina variegata). Peptides24(6), 873–880 (2003)
31. Fleming, T.J., O'hUigin, C., Malek, T.R.: Characterization of two novel Ly-6 genes. Protein sequence and potential structural similarity to alpha-bungarotoxin and other neurotoxins. J. Immunol. 150(12), 5379–5390 (1993)
32. Ploug, M., Ellis, V.: Structure-function relationships in the receptor for urokinase-type plasminogen activator. Comparison to other members of the Ly-6 family and snake venom alpha-neurotoxins. FEBS Lett.349(2), 163– 168 (1994)
33. Kolbe, H.V., Huber, A., Cordier, P., Rasmussen, U.B., Bouchon, B., Jaquinod, M., Vlasak, R., Délot, E.C., Kreil, G.: Xenoxins, a family of peptides from dorsal gland secretion of Xenopus laevis related to snake venom cytotoxins and neurotoxins. J. Biol. Chem.268(22), 16458–16464 (1993)
34. Macleod, R.J., Lembessis, P., James, S., Bennett, H.P.: Isolation of a member of the neurotoxin/cytotoxin peptide family from Xenopus laevis skin which activates dihydropyridine-sensitive Ca2+ channels in mammalian epithelial cells. J. Biol. Chem. 273(32), 20046– 20051 (1998)
