In vitro cross-linking of elastin peptides and molecular characterization of the resultant biomaterials

Biochimica et Biophysica Acta 1830 (2013) 2994–3004

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbagen

In vitro cross-linking of elastin peptides and molecular characterization of the resultant biomaterials Andrea Heinz a,⁎, Christoph K.H. Ruttkies a, Günther Jahreis b, Christoph U. Schräder a, Kanin Wichapong a, Wolfgang Sippl a, Fred W. Keeley c, Reinhard H.H. Neubert a, Christian E.H. Schmelzer a a b c

Institute of Pharmacy, Martin Luther University Halle-Wittenberg, Faculty of Natural Sciences I, Halle (Saale), Germany Max Planck Research Unit for Enzymology of Protein Folding, Halle (Saale), Germany Hospital for Sick Children, Molecular Structure and Function Program, Toronto, Canada

a r t i c l e

i n f o

Article history: Received 29 October 2012 Received in revised form 22 December 2012 Accepted 16 January 2013 Available online 30 January 2013 Keywords: Desmosine Native cross-links Dehydrolysinonorleucine Allysine aldol Dehydromerodesmosine Mass spectrometry

a b s t r a c t Background: Elastin is a vital protein and the major component of elastic fibers which provides resilience to many vertebrate tissues. Elastin's structure and function are influenced by extensive cross-linking, however, the cross-linking pattern is still unknown. Methods: Small peptides containing reactive allysine residues based on sequences of cross-linking domains of human elastin were incubated in vitro to form cross-links characteristic of mature elastin. The resultant insoluble polymeric biomaterials were studied by scanning electron microscopy. Both, the supernatants of the samples and the insoluble polymers, after digestion with pancreatic elastase or trypsin, were furthermore comprehensively characterized on the molecular level using MALDI-TOF/TOF mass spectrometry. Results: MS2 data was used to develop the software PolyLinX, which is able to sequence not only linear and bifunctionally cross-linked peptides, but for the first time also tri- and tetrafunctionally cross-linked species. Thus, it was possible to identify intra- and intermolecular cross-links including allysine aldols, dehydrolysinonorleucines and dehydromerodesmosines. The formation of the tetrafunctional cross-link desmosine or isodesmosine was unexpected, however, could be confirmed by tandem mass spectrometry and molecular dynamics simulations. Conclusions: The study demonstrated that it is possible to produce biopolymers containing polyfunctional cross-links characteristic of mature elastin from small elastin peptides. MALDI-TOF/TOF mass spectrometry and the newly developed software PolyLinX proved suitable for sequencing of native cross-links in proteolytic digests of elastin-like biomaterials. General significance: The study provides important insight into the formation of native elastin cross-links and represents a considerable step towards the characterization of the complex cross-linking pattern of mature elastin. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Elastin is one of the most important proteins of the extracellular matrix of vertebrates. As the core protein of elastic fibers, elastin Abbreviations: AA, allysine aldol; ACN, acetonitrile; CHCA, α-cyano-4-hydroxycinnamic acid; CID, collision-induced dissociation; DES, desmosine; Δ-LNL, dehydrolysinonorleucine; Δ-MD, dehydromerodesmosine; EP, elastin-like polypeptide; FA, formic acid; HPLC, high performance liquid chromatography; IDES, isodesmosine; k, allysine residue; LID, laser-induced dissociation; LOX, lysyl oxidase; MALDI, matrix-assisted laser desorption/ionization; MD, molecular dynamics; MS, mass spectrometry; PDF, probability density function; PQQ, pyrroloquinoline quinone; PyBOP, benzotriazol-1-yl-oxytripyrrolidinophosphonium hexafluorophosphate; RMSD, root mean square deviation; SEM, scanning electron microscopy; SVM, support vector machine; TFA, trifluoroacetic acid; Tris, 2-amino-2(hydroxymethyl)1,3-propanediol ⁎ Corresponding author at: Institute of Pharmacy, Martin Luther University Halle Wittenberg, Wolfgang-Langenbeck-Str. 4, 06120 Halle (Saale), Germany. Tel.: +49 345 5525220; fax: +49 345 5527292. E-mail address: [email protected] (A. Heinz). 0304-4165/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.bbagen.2013.01.014

provides elasticity and resilience to tissues including aorta, lung, skin, ligaments, tendon and cartilage and is, thus, critical for their long-term function. It is secreted in the form of its monomeric precursor tropoelastin that consists of alternating highly hydrophobic and more hydrophilic K-containing domains, of which the more hydrophobic regions are responsible for self-aggregation and tensile properties of elastin and the latter are involved in cross-linking to form an insoluble and durable polymer highly resistant to proteolytic degradation [1–3]. While it is recognized that elastic fiber networks in different tissues are organized differently, e.g. concentric fenestrated lamellae in the medial layer of the aorta or honeycomb-like structures in elastic cartilage, and that their function is strongly influenced by their composition, organization and architecture, virtually nothing is known about the cross-linking pattern in different human tissues and the exact domains of tropoelastin molecules that are involved in cross-linking. The only report on the exact location of a cross-link in elastin has

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

been made by Brown-Augsburger et al. who used Edman degradation to identify a cross-linked structure in porcine elastin formed through the association of three tropoelastin chains of domains 10, 19 and 25 [4]. With respect to the formation of elastic fibers, it is known that microfibrils act as a scaffold for the deposition of tropoelastin after its secretion into the extracellular space where coacervation and structural alignment occur prior to cross-linking. Initially, allysine (α-aminoadipic acid-δ-semialdehyde) is produced through oxidative deamination of the ε-amino group of a K residue by the enzyme lysyl oxidase (LOX) or LOX-like proteins. Intra- and intermolecular cross-links are subsequently formed either by non-enzymatic condensation of two allysine residues via aldol condensation which produces allysine aldol (AA) or by reaction of an allysine residue with the ε-amino group of another lysine residue via Schiff base reaction which creates dehydrolysinonorleucine (Δ-LNL). Such reducible cross-links then further condense to form the stable and non-reducible trifunctional cross-links merodesmosine and cyclopentenosine, the tetrafunctional cross-links desmosine (DES) and isodesmosine (IDES) as well as pentafunctional cross-links such as allodesmosine and pentasine [1–3,5–12]. The structures and formation pathways of prominent elastin cross-links are shown in Fig. 1. Owing to its unique structure, elastin does not undergo significant turnover in healthy tissues [13]. Pathological conditions and diseases, including emphysema, chronic obstructive pulmonary disease, atherosclerosis and actinic elastosis (photoaging), however, have been identified to be associated with changes in the structure, distribution and abundance of elastin which often involves the destruction of elastic fibers [2,3,14]. Thus, elucidating the molecular-level structure of human elastin including the cross-linking pattern in different tissues and comparing the structures of elastin from healthy and diseased tissues would help to better understand the biomechanical properties of elastin as well as elastic-tissue diseases and their functional consequences. In recent years, mass spectrometry (MS) has become important for the determination of low resolution 3D structures of proteins using a technique called MS3D (MS in three dimensions) [15]. In MS3D experiments, proteins are chemically cross-linked, digested by site-specific proteases and the resultant cross-linked peptides are identified using

2995

bioinformatics methods and used to derive inter-atomic distance constraints. Available software solutions such as xQuest [16], VIRTUALMSLAB [17], CLPM [18] or pLink [19], however, mainly deal with chemically introduced and only bifunctional cross-links. Since elastin shows unique tri-, tetra- and pentafunctional cross-links in addition to bifunctional cross-links, there is a need for the development of new bioinformatics tools that are able to deal with MS data of such peptides. However, it is very challenging analytically as well as computationally to identify cross-linked peptides from elastin since the protein is very hydrophobic and has a highly repetitive sequence, with 78% of human tropoelastin (isoform 2) composed just of the four amino acids G, A, V and P. Hence, peptides released during proteolysis often have similar or even identical masses and show similar product ion spectra when subjected to tandem MS (MS/MS). Moreover, alternative splicing, the hydroxylation of P and especially the many types of cross-links together with the high number of the highly similar cross-linking domains result in a huge number of possible combinations for cross-links. This is reflected in a tremendous number of theoretical peptides in an enzymatic digest of elastin which is further multiplied since proteases with broad specificities have to be used to solubilize elastin. Further challenges arise from the fact that the cross-linking of two tropoelastin domains can take place between chains that are oriented either in a parallel or antiparallel manner. Moreover, the fragmentation behavior of such cross-linked species is different from the typical backbone fragmentation of linear peptides and has not yet been investigated. For instance, in dissociation experiments some fragments can remain cross-linked whereas others do not. Overall, based on product ion spectra of cross-linked peptides derived from mature elastin it is hard and in some cases probably impossible to assign the peptides to unique positions in the involved monomeric sequences. Therefore, in this study small deca- and undecapeptides based on cross-linking domains of human elastin were used as model substrates to produce defined cross-linked species characteristic of mature elastin. Based on these cross-linking domain peptides, mass spectrometric and bioinformatics methods designed for the identification of cross-linked species were developed successfully. This work represents an important

Fig. 1. Formation of bi-, tri- and tetrafunctional cross-links in elastin (adapted from [3]).

2996

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

step towards the elucidation of the not yet characterized cross-linking pattern of human elastin, which will be helpful to better understand elastin-degrading diseases as well as the structure-function relationship of elastin from different tissues.

buffer at pH 7.5 and digested with pancreatic elastase for 24 h at 37 °C. The enzyme-to-substrate ratio was 1:100 (w/w).

2. Materials and methods

Insoluble residues of the cross-linked samples and cross-linked EP20-24-24 were totally hydrolyzed at a concentration of 1 mg mL−1 in 6 M HCl at 110 °C for 24 h, respectively, to investigate whether the tetrafunctional cross-link DES was present in the samples. DES was purchased from Elastin Products Company and dissolved in deionized water at a concentration of 10 μg mL − 1 as a reference. The totally hydrolyzed samples were evaporated to dryness using a SpeedVac and derivatized by propionylation. For propionylation, 100 μL of 0.1 N NH4HCO3 and 100 μL of a solution of 2-propanol containing 23% (V/V) propionic anhydride (Sigma-Aldrich) and 9% (V/V) NH4OH were added to the samples. After reaction for 1 h, the samples were dried using a SpeedVac and redissolved in ACN/H2O (1:1, V/V). MS experiments for DES detection were carried out as described under 2.6.

2.1. Synthesis of elastin cross-linking domain peptides The elastin cross-linking domain peptides Ac-AAAkAAAKAA, Ac-SAAkVAAKAQL and Ac-GAGVkPGKVPG (where k denotes an allysine residue) were synthesized in a SYRO II peptide synthesizer (MultiSynTech, Witten, Germany) on preloaded (0.15 mmol) A-, L- or G- 2-Cl-Trt-resins (Novabiochem, Merck, Darmstadt, Germany) by Fmoc chemistry using a standard protocol. In each cycle, Fmocprotected amino acids as building blocks were pre-activated with PyBOP and N,N-diisopropylethylamine (in four-fold excess) in dimethylformamide and coupled for 2 h at room temperature. The side chains of the amino acids were protected with Fmoc-allysine ethylene acetal, Fmoc-Lys(Boc)-OH, Fmoc-Ser(tBu)-OH and Fmoc-Gln(Trt)-OH (Merck). Detachment of the peptides from the resin and deprotection were performed with 35% (V/V) trifluoroacetic acid (TFA) in dichloromethane for 30 min at room temperature. Purification was carried out by preparative RP-HPLC on an Abimed system (Langenfeld, Germany) with an Interchrom Modulo-Cart Strategy 5, C18-2, 250-10 (Interchim, Montluçon, France) using a water–acetonitrile (ACN) gradient containing 0.1% TFA and a detection wavelength of 220 nm. The purity of the peptides was verified as described previously [20]. After lyophilization, the crystalline, aldehyde-containing peptides were stored at −20 °C in a desiccator over P2O5. 2.2. In vitro cross-linking of elastin cross-linking domain peptides and elastin-like polypeptide 20-24-24 The elastin cross-linking domain peptides were dissolved in 50 mM borate buffer, pH 9.4, at a concentration of 10 mg mL−1 and were incubated at 40 °C for 48 h. After cross-linking, which took place spontaneously, samples were centrifuged at 14,100 g for 5 min, and the supernatant of each sample was aliquoted and stored at −26 °C prior to MS analysis. The insoluble cross-linked product at the bottom of the reaction tube was washed with a solution of water and ethanol (1:1, V/V), dried at room temperature and stored at -26 °C prior to further analysis. Elastin-like polypeptide EP20-24-24 [21], which includes both hydrophobic and cross-linking domains of tropoelastin, was dissolved in 50 mM Tris buffer (pH 7.5) containing 1.5 M NaCl at a concentration of 5 mg mL−1 and was then heated to 30 °C to allow for coacervation. Cross-linking was performed over night at 40 °C by adding pyrroloquinoline quinone (PQQ; Sigma-Aldrich, Steinheim, Germany) and CuSO4 to final concentrations of 4 mM and 2 mM, respectively. The insoluble polymer that had formed was washed twice with water and ethanol, respectively, and then dried at room temperature prior to enzymatic digestion. 2.3. Proteolysis of tropoelastin and cross-linked materials Insoluble residues of the cross-linked peptide samples and cross-linked EP20-24-24 were dispersed in 50 mM Tris buffer (pH 7.5) at a concentration of 1 mg mL−1, respectively. The crosslinked peptide samples were digested with either pancreatic elastase (Elastin Products Company, Owensville, MO, USA) or proteomics grade trypsin (Sigma-Aldrich). Cross-linked EP20-24-24 was digested with proteomics grade trypsin (Sigma-Aldrich). Digestions were performed for 24 h at 37 °C using an enzyme-to-substrate ratio of 1:100 (w/w), respectively. Recombinant human tropoelastin isoform 2 [22] was dissolved at a concentration of 0.5 mg mL−1 in 50 mM Tris

2.4. Total hydrolysis of cross-linked samples and desmosine derivatization

2.5. Scanning electron microscopy The insoluble residues of the cross-linked peptide samples and cross-linked EP20-24-24 were analyzed by scanning electron microscopy (SEM). The samples were mounted on aluminum stubs with doublesided carbon tape and coated with a thin layer of platinum-palladium in a Cressington 208HR High Resolution Sputter Coater (Cressington Scientific Instruments, Watford, UK) prior to analysis. SEM was performed using the environmental scanning electron microscope ESEM XL 30 FEG (Philips, Amsterdam, Netherlands) with a beam acceleration voltage between 5 kV and 20 kV, a resolution of ≥2 nm and at a pressure of about 10−4 Pa. 2.6. MALDI-TOF/TOF MS analysis Analysis of the supernatants and the totally hydrolyzed insoluble residues of the cross-linked samples was carried out by MALDI-TOF/TOF MS and MS/MS using a 4800 MALDI-TOF/TOF Analyzer (AB Sciex, Darmstadt, Germany) equipped with a Nd:YAG laser with a repetition rate of 200 Hz. One microliter of sample was mixed with 9 μL of a solution of 10 mg mL−1 α-cyano-4-hydroxycinnamic acid (CHCA) in a mixture of ACN and 0.1% TFA (1:1, V/V). 0.5 μL of the resulting mixture was manually spotted onto the sample plate and allowed to dry at room temperature. Mass spectra in the m/z range 600 to 5000 were acquired in the positive ionization and reflectron mode by accumulating data from 1200 laser shots per spot. Detected species were automatically subjected to MS/MS experiments. Data acquisition was performed using the 4000 Series Explorer (AB Sciex), and the MS/MS data was processed from raw data into de-isotoped peak lists using Mascot Distiller (Matrix Science, London, UK). 2.7. NanoHPLC/MALDI-TOF/TOF MS analysis Analysis of the digested residues of the cross-linked samples and EP20-24-24 was carried out using an UltiMate 3000 RSLCnano system (Dionex/Thermo Fisher, Idstein, Germany) and a 4800 MALDI-TOF/TOF Analyzer. Chromatographic separation of the peptides was performed by loading 1 μL of the sample onto a trapping column (Acclaim PepMap 100 C18, 5 μm, 100 Å, 300 μm I.D. × 5 mm; Dionex) and washing for 9 min with a solvent containing 98% of 0.1% FA and 2% ACN at a flow rate of 7 μL min−1. The trapped peptides were then eluted onto the separation column (Acclaim PepMap RSLC C18, 2 μm, 100 Å, 75 μm I.D. × 150 mm; Dionex), which had been equilibrated with 88% solvent A (0.1% FA) and 12% solvent B (80% ACN and 20% of 0.1% FA). The peptides were separated using a solvent system of solvent A and solvent B: linear gradient of 12% to 40% solvent B in 51 min, linear gradient of 40%

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

to 90% solvent B in 10 min, maintenance at 90% solvent B for 5 min, linear gradient of 90% to 12% solvent B in 15 min and equilibration for 5 min prior to the next run. The column was kept at 40 °C, and the flow rate was 300 nL min−1. Fractions were mixed in a 1:3.7 ratio with matrix solution (5 mg mL−1 CHCA in 50% ACN and 50% of 0.1% TFA) and spotted in 30-s intervals onto a 384-spot MALDI target using a Probot microfraction collector (Dionex). Mass spectra in the m/z range 600 to 5000 were acquired in positive ionization mode and reflectron mode by accumulating data from 1000 to 2000 laser shots per spot. Peptides with signal-to-noise ratio above 30 at the MS mode were automatically subjected to MS/MS analysis by laser-induced dissociation (LID) and high-energy collision-induced dissociation (CID) using 1 keV as collision energy and air as collision gas. Data acquisition was performed using the 4000 Series Explorer (AB Sciex) and the MS/MS data was pre-processed using Mascot Distiller (Matrix Science). 2.8. Peptide sequencing The program PolyLinX, which was newly developed in the scope of this work, was used to sequence not only linear and bifunctionally cross-linked peptides, but also tri- and tetrafunctionally cross-linked species in the enzymatic digests of the cross-linked elastin peptide samples and EP20-24-24, respectively. 2.9. Molecular dynamics simulations The conformational behavior of DES-containing peptides composed of three molecules of Ac-AAAkAAAKAA, Ac-SAAkVAAKAQL or Ac-GAGVkPGKVPG, respectively, was sampled by using the Monte Carlo search method implemented in MOE2011.10 [23]. In total, nine DES peptide structures, i.e. three different starting conformations of each peptide, were subjected to molecular dynamics (MD) simulations using AMBER 11 [24]. Before performing MD simulation, force fields and charges for the peptides were assigned (Amber03 force field for all standard amino acid residues and generated parameters for the non-standard amino acid residues k and K, forming the DES ring). Water (TIP3P model) and counter ions were added in the radius within 10 Å from the molecular surface of the DES peptides. To relax the system, energy minimization was carried out. After that, the position-restrained phase of MD simulation was performed by restraining the complex with the weak force constraint (10 kcal mol−1) during the first 100 ps. The temperature of the system was gradually increased from 0 K to 310 K during the first few ps and was then kept at 310 K. In the final step, free MD simulation was carried out from 100 ps to 500 ns by maintaining a pressure of 1 bar and keeping the temperature at 310 K. A time step of 2 fs with SHAKE algorithm [25] was applied, and the cut-off of non-bonded interaction was set at 10 Å. 3. Results and discussion The sequences for the elastin peptides used in this study were chosen on the basis of the known cross-linking motifs present in

2997

tropoelastin, which contain K residues separated either by two or three A residues (KA domains) or by a P residue (KP domains), respectively [3]. Three peptides corresponding to these domains were synthesized containing reactive k residues allowing for spontaneous cross-linking of the samples (Table 1). The advantage of short peptides with only two K residues over longer elastin domains or even tropoelastin, which contain many cross-linking domains, is the formation of a limited number of well defined cross-links. This makes it possible to first discover cross-linked species based on their mass and subsequently get an insight into their fragmentation behavior during MS/MS analysis. The structures of possible defined cross-linked products occurring upon incubation of the peptide solutions are shown in Fig. 2. After 48 h of incubation at 40 °C the samples showed a clear supernatant and a white, gel-like, insoluble residue. Both were further analyzed. Recombinantly produced EP20-24-24 [21] is 200 amino acid residues long and was chosen as substrate to produce a more complex cross-linked model system for the identification of cross-linked structures characteristic of elastin (Table 1). EP20-24-24 does not contain k residues and therefore needs to be oxidized prior to cross-linking. Coacervation of EP20-24-24, oxidation by PQQ and cross-linking has been described previously, and the material is well characterized [26].

3.1. Development of software to detect and sequence cross-linked species The supernatants of the cross-linked peptide samples were analyzed by MALDI-TOF/TOF MS and MS/MS. The tandem mass spectra were used to develop PolyLinX, software specifically designed for the analysis of cross-linked elastin peptides. The software can be obtained by sending an e-mail to the corresponding author. For calculating peptide candidates, all possible peptides with lengths between 2 and 50 residues are generated based on the FASTA sequence of one of the thirteen known isoforms of tropoelastin. The software allows taking into account posttranslational modifications such as hydroxylation of P or oxidation of K to k. All linear peptide sequences, apart from redundant species, are saved in a database together with their masses and locations within the protein. These sequences are then cross-linked in silico by combining two or three linear peptides and incorporating a maximum of six K or k residues in up to three bifunctional cross-links, two trifunctional cross-links or one tetrafunctional cross-link and their combinations, respectively. All cross-linked peptides are added to the database. Experimentally determined precursor masses within a user-defined mass tolerance window (in ppm) are subsequently used to filter the peptide database. For each candidate peptide, a fragment ion list containing b and y ions of linear and cross-linked species is then generated and matched to the respective pre-processed experimental spectrum. Based on the agreement of the nominal masses between theoretical and experimentally determined masses for the fragments, scores are calculated to differentiate between correct and incorrect matches. The scoring relies on the calculation of a preliminary score [27] and an ions score [28] which are both used to compute the final probabilistic score that can adopt values between 0 and 1. A library for support vector machines,

Table 1 Substrates used for cross-linking reactions. Allysine-containing peptides derived from human elastin and their occurrence in elastin are shown based on tropoelastin isoform 2. Small letter k denotes allysine residues. The N-termini of the peptides are protected by acetyl groups. Peptide

Peptide sequence

P1

Ac-AAAkAAAKAA

P2 P3 EP20-24-24

Ac-SAAkVAAKAQL Ac-GAGVkPGKVPG FPGFGVGVGGIPGVAGVPGVGGVPGVGGVPGVGISPEAQAAAAAKAAKYGVGTPAAAAAKAAAKAA QFGLVPGVGVAPGVGVAPGVGVAPGVGLAPGVGVAPGVGVAPGVGVAPAIGPEAQAAAAAKAAKYGV GTPAAAAAKAAAKAAQFGLVPGVGVAPGVGVAPGVGVAPGVGLAPGVGVAPGVGVAPGVGVAPAIGP

Mr 883.48

1097.61 1006.54 16981.39

Residues

Domain

360–369 446–455 608–617 514–524 129–139 N.A.

19 29 23 25 8–9 20–24 + 21–24

2998

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

Fig. 2. Different types of cross-linked structures that may be formed upon incubation of allysine-containing elastin peptides at elevated temperature. Reactive allysine residues are indicated by white circles, whereas the black circles indicate unmodified lysine residues. (A) shows a linear peptide, (B) an intramolecular, bifunctional cross-link, (C) an intermolecular, bifunctional cross-link (dehydrolysinonorleucine), (D) an intermolecular, bifunctional cross-link (allysine aldol), (E) two intermolecular, bifunctional cross-links (dehydrolysinonorleucines), (F) an intermolecular, trifunctional cross-link (dehydromerodesmosine), (G) an intermolecular, tetrafunctional cross-link (desmosine) and (H) one possible cross-linked product of higher mass generated through the formation of multiple allysine aldols and one dehydromerodesmosine. For clarity, peptide chains are mostly only shown in parallel orientation, however, can be oriented both in a parallel or antiparallel manner in reality.

LIBSVM [29], was trained with scores obtained for a set of foreground data containing correct sequences and background data containing randomly created decoy sequences for the experimental MS/MS data. Training and model evaluation was performed by cross-validation. Analysis of the peptides interestingly revealed a correlation between the types of peptides, i.e. linear or differently cross-linked species, and the distribution of their fragment ions when fragmented by LID and CID in MALDI-TOF/TOF MS/MS experiments. This finding enabled the development of an SVM classifier that allows differentiating between linear and differently cross-linked species. Linear peptides, for instance, show abundant fragments all over the m/z range of the spectrum (Fig. 3A), whereas bi- and trifunctionally cross-linked species mainly show fragments in the low and high m/z range of the spectra (Fig. 3B) and tetrafunctionally cross-linked peptides show mainly higher mass fragments (Fig. 3C). One of the reasons for the characteristic fragmentation behavior of peptides containing the tetrafunctional DES or IDES is the intrinsic positive charge of the molecule (Fig. 1). Since many fragmentation reactions are charge driven processes, it is likely that the bond of the CH2 group next to the quaternary nitrogen is ruptured on laser-induced or collisional activation which produces a primary carbocation that can donate a proton to a basic site, e.g. the amide nitrogen of a peptide bond. This frequent pathway for amide bond cleavage would result in the formation of a double bond at the end of the hydrocarbon chain of the former cross-linked structure and a protonated oxazolone at the peptide chain [30]. Depending on whether the proton affinity of the peptide is greater than that of the oxazolone, a y or b ion would be formed, respectively. It is likely that the free proton donated by the carbocation is mainly transferred to the higher mass fragments which would explain the occurrence of predominantly high mass fragments in the product ion spectrum of the DES or IDES cross-linked peptides.

For classification, the product ion spectra are evenly divided into three segments (Fig. 3), and the contribution of the peak intensities in each segment to the overall intensity of the spectra is determined. Probability density functions are used to avoid strict separation of the segments which has the advantage that peaks occurring at the border between two segments have a weighted contribution to each segment. For the outer segments of the fragment spectra gamma distributions are used, and for the inner segment a normal distribution is used. A three dimensional feature vector is finally calculated which is shown for linear peptides and bi- and tetrafunctionally cross-linked peptides in Fig. 3. The LIBSVM was trained based on 25 tandem mass spectra of tetrafunctionally cross-linked peptides, 31 tandem mass spectra of bifunctionally cross-linked species and 50 tandem mass spectra of linear peptides from the tropoelastin digest with pancreatic elastase described under 2.3. Due to the relatively low amount of data, validation was performed using 10-fold cross-validation. To avoid unreliable classification due to insufficient quality of some MS/MS data, a filter was incorporated into the classification module. Using this filter, tandem mass spectra showing less than five peaks with intensities of more than 5% of the intensity of the highest peak in the spectrum (apart from the precursor peak) are not classified using the LIBSVM but rejected. 3.2. Characterization of in vitro cross-linked samples of elastin cross-linking domain peptides With the help of PolyLinX, it was possible to identify and sequence a variety of bi-, tri- and interestingly even tetrafunctionally cross-linked species in the supernatants of the samples (Table 2). Comparing the samples which either contained P1, P2 or P3 with the sample containing both P1 and P2, a greater variety of cross-linked structures was identified in the latter sample due to the presence of cross-linked species

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

2999

Fig. 3. MALDI-TOF/TOF tandem mass spectra showing the characteristic fragmentation pattern of (A) linear peptides, (B) bi- and trifunctionally cross-linked peptides and (C) tetrafunctionally cross-linked peptides and the three-dimensional feature vectors calculated from their intensity distributions (shown in red for linear peptides, green for bi- and trifunctionally cross-linked peptides and blue for tetrafunctionally cross-linked peptides).

formed through the reaction of both peptides. Identified bifunctionally cross-linked species include intramolecular Δ-LNLs (Fig. 2B), intermolecular Δ-LNLs (Fig. 2C) and intermolecular AAs (Fig. 2D) which are characterized by a mass decrease of 18.01 Da (due to the loss of water) compared to the masses of either one or two non-cross-linked peptide molecules. Peptides containing two bifunctional Δ-LNLs (Fig. 2E) and trifunctional dehydromerodesmosines (Δ-MDs; Fig. 2F) were also detected and show a mass loss of 36.02 Da as compared to the mass of two linear peptides. It was surprising to detect species containing tetrafunctional DES or IDES (mass decrease of 54.04 Da due to a loss of three water molecules) generated by reaction of three peptide molecules in the samples (Fig. 2G). While only one DES/IDES species was formed through the reaction of three molecules of the same peptide in the samples which contained either P1, P2 or P3, four DES/IDES species were likely to occur in the sample of higher complexity containing P1 and P2 (Table 2). Three of these four DES/IDES-containing, cross-linked peptides were identified. Overall, the presence of DES/IDES was verified by investigating the peptides using tandem MS. According to the proposed fragmentation behavior of tetrafunctional cross-links (discussed in Section 3.1), which is different to that of linear or bifunctionally cross-linked peptides, the majority of experimentally derived fragment ions could be assigned (data not shown). However, due to a lack of characteristic fragment ions, MS/MS did not allow for determination of whether the cross-link is DES or IDES. Further MS experiments showed that the supernatants of the cross-linked samples contained a series of cross-linked species up to 15 kDa whose masses result from the repeated addition of always one peptide molecule to the respective lower mass species under loss of a single water molecule (Fig. 2H). Cross-links between tropoelastin molecules during elastogenesis are typically formed after coacervation and self-alignment of the

molecules through hydrophobic interactions of specific tropoelastin domains, which allow for close proximity of juxtaposed K residues [31]. DES/IDES residues typically link two tropoelastin chains, i.e. two KA domains are involved in the reaction of three k residues and

Table 2 Cross-linking experiments carried out using elastin cross-linking domain peptides. Cross-linked products were identified from the supernatants of the samples by MALDI-TOF/TOF MS and MS/MS. Sample

Cross-linked products identified

Mr of cross-linked product

P1

Intramolecular Δ-LNL Intermolecular Δ-LNL or AA 2 × intermolecular Δ-LNL or Δ-MD DES/IDES Intramolecular Δ-LNL Intermolecular Δ-LNL or AA Intermolecular Δ-LNL DES/IDES Intramolecular Δ-LNL Intermolecular Δ-LNL or AA 2 × intermolecular Δ-LNL or Δ-MD DES/IDES Intermolecular Δ-LNL or AA (P1) 2 × intermolecular Δ-LNL or Δ-MD (P1) Intermolecular Δ-LNL (P2) Intermolecular Δ-LNL or AA(P1, P2) 2 × intermolecular Δ-LNL or Δ-MD (P1, P2) DES/IDES 3 × P1 2 × P1, 1 × P2 1 × P1, 2 × P2

865.47 1748.94 1730.93 2595.39 1079.60 2177.21 2159.20 3237.78 988.53 1995.08 1977.07 2964.59 1748.94 1730.93 1079.60 1963.07 1945.06

P2

P3

P1 and P2

2595.39 2809.52 3023.65

3000

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

one K residue [3,4,32]. Since the peptides used in this study only contain one k and one K residue, respectively, three molecules (three cross-linking domains) are required to form DES/IDES. Moreover, based on experimental findings [32–34] it has been hypothesized that DES/IDES cross-links are less likely if not unlikely to form in KP domains of the tropoelastin molecule [2,4,35]. This has been attributed to the lack of α-helical structure in these domains, which has been discussed to be required for arrangement of K residues on the same side of the helix prior to cross-linking [3,35]. However, previous studies have shown that KP domains, e.g. the one derived from exon 8, are able to form polyproline II left-handed helices (PPII) in aqueous solution, which also allow the K residues to come into proximity prior to cross-linking [36,37]. In contrast, this study showed that the formation of DES/IDES is possible, at least in vitro in short cross-linking domain peptides containing a KPGK motif. In this context, it would be interesting to investigate cross-links present in amphibians or teleosts more in detail since the elastins of these species are rich in KP domains and have been found to show a lower DES content than for instance mammalian elastins [35,38]. As the presence of DES/IDES in the cross-linked samples was unexpected, their stability within cross-linked structures composed of three peptide molecules was further investigated by MD simulations for all three model peptides, respectively (Fig. 4). Analysis revealed that the pyridinium ring was stable as indicated by small fluctuation of RMSD values during the simulation (2 Å–4 Å), while the peptide residues were more flexible (RMSD ~8 Å–12 Å), adopting different conformations. The conformations of the DES-containing peptides composed of three molecules of either Ac-SAAkVAAKAQL, Ac-AAAkAAAKAA or Ac-GAGVkPGKVPG, respectively, derived from the last snapshot (500 ns) of the MD simulation revealed that hydrogen bonds can be formed between backbone carbonyl and amide groups and stabilize the DES-containing peptides (Fig. 4A–C). Moreover, the conformations of the peptides are stabilized by the interaction between the carbonyl group of the backbone peptide residues and the pyridinium ring system. In the case of Ac-AAAkAAAKAA hydrogen bonds are found for instance between residues of different peptide molecules (Fig. 4B), while for Ac-GAGVkPGKVPG, the conformation of the pyridinium ring is stabilized by the hydrogen bond network between k and close G and V residues as shown in Fig. 4C. As described above, the formation of DES/IDES in peptide species containing three molecules Ac-GAGVkPGKVPG may be facilitated by the presence of polyproline II left-handed helices in the peptide molecules which bring K residues closer together and allow them to react with each other [36,37]. Overall, the formation of DES/IDES is most likely generally enabled by the short length of the peptides which reduces steric hindrance and facilitates the reaction of k and K residues of different molecules. Interestingly, it was found that the elastin cross-linking domain peptides are able to form an insoluble cross-linked product whose elastic properties and potential use for biomedical applications is currently under further investigation. The formation of this material was unexpected since such biomaterials have so far only been prepared from longer, recombinantly produced elastin peptides or even entire tropoelastin [26,31,39–41]. The presence of hydrophobic domains adjacent to the cross-linking domains has been described to be important for cross-linking, since these domains allow for coacervation, self-assembly and alignment of the larger elastin domains or tropoelastin molecules prior to cross-linking [21]. The present study, however, revealed that it is possible to produce an insoluble biopolymer from peptides lacking hydrophobic domains. This may be due to the very reactive already oxidized K residues in the peptides and the lack of steric hindrance in short peptides which facilitates the interaction of k and K residues and, hence, the formation of cross-links. The insoluble materials of the samples were analyzed by SEM and the images reveal the presence of a porous biomaterial with a rough, characteristically structured surface. Higher magnification showed

Fig. 4. Conformation of the identified desmosine peptides formed by reaction of three molecules of (A) Ac-SAAkVAAKAQL, (B) Ac-AAAkAAAKAA or (C) Ac-GAGVkPGKVPG, respectively, after 500 ns of MD simulation. The central pyridinium ring is shown in cyan, the backbone of the peptide is displayed as magenta ribbon and hydrogen bonds are shown as green dashed lines. The residues of the three peptide molecules are labeled with the respective one letter code and either (A) for the first peptide molecule, (B) for the second peptide molecule or (C) for the third peptide molecule, respectively.

the lateral assembly of fine, twisted fibers with a diameter of around 500 nm sticking out of the surface of the larger macroassembly (Fig. 5A–C). Total hydrolysis of the material using 6 M HCl and subsequent propionylation of the sample allowed for the unambiguous identification of DES in the cross-linked sample. Tetrapropionylated DES was identified based on its mass (750.39 Da) and MALDI-TOF/TOF

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

3001

Fig. 5. Scanning electron micrographs of cross-linked insoluble biomaterials that formed after incubation of the elastin cross-linking domain peptide Ac-AAAkAAAKAA at elevated temperature (A–C) and after incubation of EP20-24-24 with PQQ (D–F). The white bar represents 20 μm (A, B, D, E) or 2 μm (C, F).

fragment spectrum in comparison to a propionylated DES standard (Fig. 6). Digestions of the insoluble, cross-linked material were carried out using either pancreatic elastase as an enzyme with broad specificity or trypsin, which only cleaves C-terminal to R residues and unmodified and non-cross-linked K residues. Subsequent analysis using MALDITOF/TOF MS(/MS) revealed the presence of different cross-linked species including AAs, Δ-LNLs and DES/IDES (data not shown). The results show that in the higher molecular cross-linked material all types of cross-linked species are present. The structure of this material is most likely a combination of the structures shown in Fig. 2G and H which includes the presence of all types of cross-links, however, with many more peptide molecules involved. 3.3. Characterization of more complex cross-linked samples for software validation Spiking experiments were performed to test whether PolyLinX is able to identify cross-linked species in more complex samples, which contain a variety of unknown linear peptides in addition to cross-linked species. This resembles an enzymatic digest of elastin

better than the samples with only the short cross-linked peptides. The tropoelastin digest described under 2.3 containing linear peptides was spiked with supernatant of one of the cross-linked samples to give a final concentration of 1 mg mL − 1, and the samples were analyzed using MALDI-TOF/TOF MS/MS. In addition to the identification of linear peptides, PolyLinX successfully identified all cross-linked species that were also detected in the much less complex samples described under 3.2. Cross-linked EP20-24-24 was used as more complex elastin-like model system and was analyzed by MALDI-TOF/TOF MS(/MS) after enzymatic digestion with trypsin and total hydrolysis using HCl, respectively. After total hydrolysis and propionylation, DES was identified in the sample as tripropionylated species (694.37 Da). From the tryptic digest, it was possible to identify a variety of linear peptides of which many were found to have free k residues. In total, at least four of the eight K residues that are present in EP20-24-24 were found to be oxidized to k. It is likely that two more residues are oxidized to k which can, however, not be confirmed since it is not possible to unambiguously assign two residues to positions in the sequence of EP20-24-24 due to exact repetitions in its sequence [26]. Although it was shown that DES is present in cross-linked EP20-24-24, this result suggests that the cross-linking

3002

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

Fig. 6. MALDI TOF/TOF tandem mass spectrum of tetrapropionylated desmosine (M+ 750.39) detected in a sample containing cross-linked Ac-AAAkAAAKAA (top). The fragment pattern shows clear conformity with a tetrapropionylated desmosine standard (bottom) and verifies the presence of desmosine in the sample.

degree is not very high and, consequently, the abundance of crosslinked peptides in the digests low in comparison to linear ones. However, using PolyLinX it was possible to successfully identify the peptide AAkYGVGTPAAAAAkAAAK (residues 449–467 based on tropoelastin isoform 2) with an intramolecular AA. Interestingly, the same peptide, however with a bis(sulfosuccinimidyl)suberate cross-link between the two K residues 451 and 463, has been previously identified from a Lys-C digest of an elastin polypeptide corresponding to domains 17–27 of human tropoelastin [42]. It has been hypothesized that the K residues 451 and 463 are well accessible and possibly show a greater flexibility as compared to other K residues which allows them to get into proximity prior to cross-linking [42]. The presence of a double β-turn between the residues GVGTPA may also bring K451 and K463 closer together [43]. Fig. 7 shows the respective MALDI-TOF/TOF product ion spectrum with annotated b and y ions. Intermolecularly cross-linked species, however, were not detected which could be explained by the low abundance of such species in the digest. Overall, the samples with the cross-linked short peptides, which contained predominantly cross-linked species in high concentration, were easier to analyze with respect to cross-linked species. To overcome this problem, current efforts are put into the enrichment of cross-linked species which increases their concentration and makes it easier to detect them by MS. Moreover, higher concentrations of cross-linked species will significantly improve the quality of the MS/MS data and facilitate sequencing in combination with PolyLinX. 4. Conclusion In this study, short elastin cross-linking domain peptides containing reactive allysine residues but lacking hydrophobic flanking domains

were successfully cross-linked in vitro under formation of numerous soluble cross-linked peptides including several desmosine-containing species. The novel approach furthermore demonstrated the feasibility of producing insoluble elastin-like biopolymers containing polyfunctional cross-links characteristic of mature elastin. The cross-linked samples were comprehensively analyzed on the molecular level using MALDITOF/TOF MS(/MS), and the data was used to develop a bioinformatics tool for identification and sequencing of polyfunctional cross-links. With the help of the newly developed software PolyLinX, it was possible for the first time to not only discover and successfully sequence bifunctionally but also tri- and tetrafunctionally cross-linked peptides provided the cross-linked species occur in sufficient abundance. The results of this study form an important basis for the elucidation of the cross-linking pattern of mature elastin. Moreover, PolyLinX is easily adaptable and could be used to analyze similarly cross-linked matrix proteins such as collagens.

Acknowledgements The work was supported by the Deutsche Forschungsgemeinschaft (DFG) by grant HE 6190/1-1 (A.H.). The authors thank Dr. Frank Heyroth (Interdisciplinary Center for Materials Science, Martin Luther University Halle-Wittenberg, Germany) for assistance with scanning electron microscopy, Dr. Alex G. Harrison (University of Toronto, Toronto, Canada) and Dr. Steffen Neumann (Leibniz Institute for Plant Biochemistry, Halle, Germany) for helpful discussions and Dr. Anthony S. Weiss (University of Sydney, Australia) for providing tropoelastin.

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

3003

Fig. 7. Annotated MALDI-TOF/TOF fragment spectrum of the peptide AAkYGVGTPAAAAAkAAAK containing an intramolecular AA identified in a trypsin digest of EP20-24-24.

References [1] M. Akagawa, K. Suyama, Mechanism of formation of elastin crosslinks, Connect. Tissue Res. 41 (2000) 131–141. [2] S.M. Mithieux, A.S. Weiss, Elastin, Adv. Protein Chem. 70 (2005) 437–461. [3] B. Vrhovski, A.S. Weiss, Biochemistry of tropoelastin, Eur. J. Biochem. 258 (1998) 1–18. [4] P. Brown-Augsburger, C. Tisdale, T. Broekelmann, C. Sloan, R.P. Mecham, Identification of an elastin cross-linking domain that joins three peptide chains. Possible role in nucleated assembly, J. Biol. Chem. 270 (1995) 17778–17783. [5] D. Bedell-Hogan, P. Trackman, W. Abrams, J. Rosenbloom, H. Kagan, Oxidation, cross-linking, and insolubilization of recombinant tropoelastin by purified lysyl oxidase, J. Biol. Chem. 268 (1993) 10345–10350. [6] K. Reiser, R.J. McCormick, R.B. Rucker, Enzymatic and nonenzymatic cross-linking of collagen and elastin, FASEB J. 6 (1992) 2439–2449. [7] N.R. Davis, R.A. Anwar, On the mechanism of formation of desmosine and isodesmosine cross-links of elastin, J. Am. Chem. Soc. 92 (1970) 3778–3782. [8] D.R. Eyre, M.A. Paz, P.M. Gallop, Cross-linking in collagen and elastin, Annu. Rev. Biochem. 53 (1984) 717–748. [9] J. Rosenbloom, W.R. Abrams, R. Mecham, Extracellular matrix 4: the elastic fiber, FASEB J. 7 (1993) 1208–1218. [10] R.B. Rucker, J. Murray, Cross-linking amino acids in collagen and elastin, Am. J. Clin. Nutr. 31 (1978) 1221–1236. [11] F. Sato, H. Wachi, M. Ishida, R. Nonaka, S. Onoue, Z. Urban, B.C. Starcher, Y. Seyama, Distinct steps of cross-linking, self-association, and maturation of tropoelastin are necessary for elastic fiber formation, J. Mol. Biol. 369 (2007) 841–851. [12] F. Nakamura, K. Yamazaki, K. Suyama, Isolation and structural characterization of a new cross-linking amino-acid, cyclopentenosine, from the acid hydrolysate of elastin, Biochem. Biophys. Res. Commun. 186 (1992) 1533–1538. [13] Z. Urbán, C.D. Boyd, Elastic-fiber pathologies: primary defects in assembly-and secondary disorders in transport and delivery, Am. J. Hum. Genet. 67 (2000) 4–7. [14] U. Saarialho-Kere, E. Kerkelä, L. Jeskanen, T. Hasan, R. Pierce, B. Starcher, R. Raudasoja, A. Ranki, A. Oikarinen, M. Vaalamo, Accumulation of matrilysin (MMP-7) and macrophage metalloelastase (MMP-12) in actinic damage, J. Invest. Dermatol. 113 (1999) 664–672. [15] S.L.N. Mayne, H.G. Patterton, Bioinformatics tools for the structural elucidation of multi-subunit protein complexes by mass spectrometric analysis of protein-protein cross-links, Brief. Bioinform. 12 (2011) 660–671.

[16] O. Rinner, J. Seebacher, T. Walzthoeni, L.N. Mueller, M. Beck, A. Schmidt, M. Mueller, R. Aebersold, Identification of cross-linked peptides from large sequence databases, Nat. Methods 5 (2008) 315–318. [17] L.J. de Koning, P.T. Kasper, J.W. Back, M.A. Nessen, F. Vanrobaeys, J. Van Beeumen, E. Gherardi, C.G. de Koster, L. de Jong, Computer-assisted mass spectrometric analysis of naturally occurring and artificially introduced cross-links in proteins and protein complexes, FEBS J. 273 (2006) 281–291. [18] Y. Tang, Y.F. Chen, C.F. Lichti, R.A. Hall, K.D. Raney, S.F. Jennings, CLPM: a cross-linked peptide mapping algorithm for mass spectrometric analysis, BMC Bioinforma. 6 (2005). [19] B. Yang, Y.-J. Wu, M. Zhu, S.-B. Fan, J. Lin, K. Zhang, S. Li, H. Chi, Y.-X. Li, H.-F. Chen, S.-K. Luo, Y.-H. Ding, L.-H. Wang, Z. Hao, L.-Y. Xiu, S. Chen, K. Ye, S.-M. He, M.-Q. Dong, Identification of cross-linked peptides from complex samples, Nat. Methods 9 (2012) 904–906. [20] A. Heinz, M.C. Jung, L. Duca, W. Sippl, S. Taddese, C. Ihling, A. Rusciani, G. Jahreis, A.S. Weiss, R.H.H. Neubert, C.E.H. Schmelzer, Degradation of tropoelastin by matrix metalloproteinases - cleavage site specificities and release of matrikines, FEBS J. 277 (2010) 1939–1956. [21] C.M. Bellingham, K.A. Woodhouse, P. Robson, S.J. Rothstein, F.W. Keeley, Self-aggregation characteristics of recombinantly expressed human elastin polypeptides, Biochim. Biophys. Acta 1550 (2001) 6–19. [22] S.L. Martin, B. Vrhovski, A.S. Weiss, Total synthesis and expression in Escherichia coli of a gene encoding human tropoelastin, Gene 154 (1995) 159–166. [23] Molecular Operating Environment (MOE), 2011.10, Chemical Computing Group Inc., 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2011. [24] D.A. Case, T.A. Darden, T.E. Cheatham III, C.L. Simmerling, J. Wang, R.E. Duke, R. Luo, R.C. Walker, W. Zhang, K.M. Merz, B. Roberts, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossváry, K.F. Wong, F. Paesani, J. Vanicek, J. Liu, X. Wu, S.R. Brozell, T. Steinbrecher, H. Gohlke, Q. Cai, X. Ye, J. Wang, M.-J. Hsieh, G. Cui, D.R. Roe, D.H. Mathews, M.G. Seetin, C. Sagui, V. Babin, T. Luchko, S. Gusarov, A. Kovalenko, P.A. Kollman, AMBER 2011, University of California, San Francisco, 2010. [25] J.-P. Ryckaert, G. Ciccotti, H.J.C. Berendsen, Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes, J. Comput. Phys. 23 (1977) 327–341. [26] C.M. Bellingham, M.A. Lillie, J.M. Gosline, G.M. Wright, B.C. Starcher, A.J. Bailey, K.A. Woodhouse, F.W. Keeley, Recombinant human elastin polypeptides self-assemble into biomaterials with elastin-like properties, Biopolymers 70 (2003) 445–455. [27] J.K. Eng, A.L. McCormack, J.R. Yates III, An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database, J. Am. Soc. Mass Spectrom. 5 (1994) 976–989.

3004

A. Heinz et al. / Biochimica et Biophysica Acta 1830 (2013) 2994–3004

[28] D.N. Perkins, D.J. Pappin, D.M. Creasy, J.S. Cottrell, Probability-based protein identification by searching sequence databases using mass spectrometry data, Electrophoresis 20 (1999) 3551–3567. [29] C.-C. Chang, C.-J. Lin, LIBSVM: a library for support vector machines, ACM Trans. Intell. Syst. Technol. 2 (2011) 1–27. [30] A.G. Harrison, To b or not to b: the ongoing saga of peptide b ions, Mass Spectrom. Rev. 28 (2009) 640–654. [31] F.W. Keeley, C.M. Bellingham, K.A. Woodhouse, Elastin as a self-organizing biomaterial: use of recombinantly expressed human elastin polypeptides as a model for investigations of structure and self-assembly of elastin, Philos. Trans. R. Soc. Lond. B Biol. Sci. 357 (2002) 185–189. [32] J.A. Foster, L. Rubin, H.M. Kagan, C. Franzblau, E. Bruenger, L.B. Sandberg, Isolation and characterization of crosslinked peptides from elastin, J. Biol. Chem. 249 (1974) 6191–6196. [33] H. Sage, W.R. Gray, Studies on the evolution of elastin .3. The ancestral protein, Comp. Biochem. Physiol. B 68 (1981) 473–480. [34] G.E. Gerber, R.A. Anwar, Structural studies on cross-linked regions of elastin, J. Biol. Chem. 249 (1974) 5200–5207. [35] M.I. Chung, M. Miao, R.J. Stahl, E. Chan, J. Parkinson, F.W. Keeley, Sequences and domain structures of mammalian, avian, amphibian and teleost tropoelastins: clues to the evolutionary history of elastins, Matrix Biol. 25 (2006) 492–504.

[36] B. Bochicchio, A. Pepe, Role of polyproline II conformation in human tropoelastin structure, Chirality 23 (2011) 694–702. [37] A.M. Tamburro, B. Bochicchio, A. Pepe, Dissection of human tropoelastin: exon-byexon chemical synthesis and related conformational studies, Biochemistry 42 (2003) 13347–13362. [38] H. Sage, W.R. Gray, Studies on the evolution of elastin .1. Phylogenetic distribution, Comp. Biochem. Physiol. B 64 (1979) 313–327. [39] S.M. Mithieux, J.E.J. Rasko, A.S. Weiss, Synthetic elastin hydrogels derived from massive elastic assemblies of self-organized human protein monomers, Biomaterials 25 (2004) 4921–4927. [40] S.M. Mithieux, S.G. Wise, M.J. Raftery, B. Starcher, A.S. Weiss, A model two-component system for studying the architecture of elastin assembly in vitro, J. Struct. Biol. 149 (2005) 282–289. [41] S.G. Wise, S.M. Mithieux, M.J. Raftery, A.S. Weiss, Specificity in the coacervation of tropoelastin: solvent exposed lysines, J. Struct. Biol. 149 (2005) 273–281. [42] L.B. Dyksterhuis, C. Baldock, D. Lammie, T.J. Wess, A.S. Weiss, Domains 17-27 of tropoelastin contain key regions of contact for coacervation and contain an unusual turn-containing crosslinking domain, Matrix Biol. 26 (2007) 125–135. [43] A.M. Tamburro, A. Pepe, B. Bochicchio, Localizing alpha-helices in human tropoelastin: assembly of the elastin “puzzle”, Biochemistry 45 (2006) 9518–9530.