JBC Papers in Press. Published on August 14, 2018 as Manuscript RA118.004322 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.RA118.004322 Native cross-links in elastin Elastin is heterogeneously cross-linked Christoph U. Schräder1§, Andrea Heinz1,2, Petra Majovsky3, Berin Karaman1$, Jürgen Brinckmann4,5, Wolfgang Sippl1, and Christian E. H. Schmelzer1,6* From the 1Institute of Pharmacy, Martin Luther University Halle-Wittenberg, Halle (Saale), Germany; the 2Department of Pharmacy, University of Copenhagen, Copenhagen, Denmark; the 3Proteome Analytics Research Group, Leibniz Institute for Plant Biochemistry, Halle (Saale), Germany; the 4Institute of Virology and Cell Biology, University of Lübeck, Lübeck, Germany; the 5Department of Dermatology, University of Lübeck, Lübeck, Germany; the 6Fraunhofer Institute for Microstructure of Materials and Systems IMWS, Halle (Saale), Germany Running title: Native cross-links in elastin Present addresses: §Department of Biochemistry and Molecular Biology, University of Calgary, Calgary AB, Canada; $Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Biruni University, Istanbul, Turkey D o w n *To whom correspondence should be addressed: Dr. Christian E. H. Schmelzer, Fraunhofer Institute for lo a Microstructure of Materials and Systems IMWS, Walter-Hülse-Str. 1, 06120 Halle (Saale), Germany; de d [email protected]; Tel. +49-345-5589116; FAX. +49-345-5589101. fro m h Keywords: extracellular matrix protein, desmosine, elastin, elastic fibers, lysyl oxidase, ttp ://w lysinonorleucine, allysine aldol, mass spectrometry, protein cross-linking, protein structure w w .jb c .o rg ABSTRACT were found to be cross-linked via b/ y Elastin is an essential vertebrate protein lysinonorleucine, allysine aldol, and desmosine. g u e responsible for the elasticity of force-bearing Unexpectedly, we identified a high number of st o tissues such as those of the lungs, blood vessels, intramolecular cross-links between lysine n J a and skin. One of the key features required for residues in close proximity. In summary, we nu a the exceptional properties of this durable show on the molecular level that elastin ry 1 1 biopolymer is the extensive covalent cross- formation involves random cross-linking of , 2 0 linking between domains of its monomer tropoelastin monomers resulting in an 1 9 molecule tropoelastin. To date, elastin’s exact unordered network, an unexpected finding molecular assembly and mechanical properties compared with previous assumptions of an are poorly understood. Here, using bovine overall beaded structure. elastin, we investigated the different types of cross-links in mature elastin to gain insight into its structure. We purified and proteolytically Elasticity and resilience are fundamentally cleaved elastin from a single tissue sample into important characteristics of dynamic tissues. soluble cross-linked and non-cross-linked Within vertebrates – with the exception of peptides that we studied by high-resolution MS. cyclostomata (1) – elastic fibers enable organs such This analysis enabled the elucidation of cross- as lungs, arteries or skin to stretch and bend with links and other elastin modifications. We found recoil and thus to maintain their physiological that the lysine residues within the tropoelastin function. The fibers consist of an outer mantle of sequence were simultaneously unmodified and fibrillin-rich microfibrils and a dense elastin core involved in various types of cross-links with that comprises over 90% of the volume. Elastin is different other domains. The Lys-Pro domains an extremely long-lived protein with remarkable were almost exclusively linked via properties including a very low elastic modulus. It lysinonorleucine, whereas Lys-Ala domains is hydrophobic and insoluble, but its hydration is a 1 Native cross-links in elastin requirement for the elastic properties (2). The in collagens, but DES and IDES are unique to structural basis for understanding its reversible elastin (20). While the types of cross-links present elasticity has been elusive. However, a fundamental in elastin are well studied, very little is known about feature is the generation of covalent cross-links the exact pattern of cross-linking and hence the between the tropoelastin (TE) monomers. The overall molecular organization of elastin. A better resultant network distributes the stress and strain comprehension of its overall structure is required to forces throughout the biopolymer during understand the mechanical properties of elastic deformation. TE consists of alternating fibers as well as their mechanisms of formation and hydrophobic and more hydrophilic domains. In the breakdown, especially as the latter contributes to hydrophobic domains, the small nonpolar amino pathologies such as pulmonary emphysema or acids Gly, Leu, Val and Pro dominate, whereas the cardiovascular diseases. However, the analysis of hydrophilic domains contain Lys-Ala (KA) and elastin’s cross-links is very challenging due to their Lys-Pro (KP) motifs, which are involved in cross- diversity, the tremendous number of possible linking. After secretion, TE undergoes self- combinations and the repetitive nature of the association by interactions between hydrophobic precursor’s primary structure. Complicated to deal domains of the monomers (3,4), leading to the with are further elastin’s insolubility and splice formation of distinct globular aggregates on the cell variants, the resistance to specific proteases and the surface (5). Lysyl oxidase (LOX) and LOX-like presence of other post-translational modifications D o w enzymes then catalyze the oxidative deamination of (PTMs). n lo the ε-amino group of Lys residues to the highly The only exact cross-linking sites were ad e d reactive α-aminoadipic acid-δ-semialdehyde, also determined more than two decades ago by Robert fro termed allysine (further denoted as Lya) (6). In KA Mecham’s group (21). In this study, Brown- m h domains, Lys residues occur as pairs or triplets Augsburger et al. created an incompletely cross- ttp separated by two or three Ala residues (KAAK, linked elastin by impeding the activity of ://w w KxAAK), whereas in KP domains the Lys residues LOX/LOXLs in pigs. This trick made it possible to w are separated by at least one Pro residue (KxPK, specifically digest elastin and to investigate some .jbc .o KPxK, KPK) (7,8). After oxidation, covalent cross- cross-link-containing peptides by amino acid rg links are formed spontaneously by non-enzymatic analysis and Edman sequencing. Thus, the authors by/ g condensation of either two Lya residues via aldol demonstrated the involvement of domains 10, 19 ue s condensation forming allysine aldol (AA) or by the and 25 in three cross-links. These results are also t o n reaction of a Lya residue with the ε-amino group of the basis of the proposed head-to-tail model (22). Ja n u another Lys residue via Schiff base reaction However, these cross-links represent only a small a ry resulting in the formation of part of the matter as elastin features a total of 10 KA 1 1 dehydrolysinonorleucine (ΔLNL) (9,10). Such and 6 KP domains, which can be covalently linked , 20 1 reducible cross-links then further condense with with each other by a multitude of different cross- 9 each other – partly with participation of unmodified links. To supplement the current knowledge and to Lys residues or with other intermediates – to form attain a better understanding of the elastogenesis, the trifunctional cross-links we examined in this study the cross-links in dehydromerodesmosine and cyclopentenosine (11- unaltered and mature elastin. The application of 13) as well as the tetrafunctional cross-link state of the art high-resolution mass spectrometry desmosine (DES) and its isomer isodesmosine (HRMS) as well as in-house developed software (IDES) (14-16). The majority of TE’s Lys residues (16,23,24) enabled us to gain new structural insight are modified in the course of maturation (17). The into the transition from TE to elastin. different types of cross-links coexist in mature elastin and over time, dehydrolysinonorleucine and Results dehydromerodesmosine further get reduced to Both KP and KA domains are incompletely cross- lysinonorleucine (LNL) and merodesmosine, linked in mature elastin respectively (see (18) for review). The most abundant cross-linking amino acids in elastin are The identity of linear (non-cross-linked) the bifunctional and tetrafunctional amino acids peptides can indirectly provide important (19) shown in Fig. 1. LNL and AA are also found information about putative cross-linking sites. After cleaving mature elastin by pancreatic elastase (PE), 2 Native cross-links in elastin we identified 310 linear elastin peptides covering The majority of Lys residues are modified 78% of TE’s sequence. The exact coverage and Amino acid analysis of isolated elastin revealed cleavage sites of PE are shown in Fig. 2. The a high correlation with the theoretical composition majority of linear peptides were released from of bovine TE with the exception of Pro and Lys hydrophobic domains, which are not involved in residues (see Fig. S1). Pro was reduced by 10% and cross-linking. We identified partial hydroxylation coincided with the detection of the same amount of of 16 Pro residues located in GxPG motifs, where x hydroxyproline. Lys, however, was decreased by was Val, Ile, Leu or Phe, and in one case within the 92%, indicating a high cross-linking degree. motif LPA. We found no hydroxyprolines within any of the cross-linking motifs. We observed that linear peptides containing Bifunctional cross-links are formed inter- and motifs capable of being involved in cross-linking intramolecularly in KA and KP domains occurred at lower frequency than those released The identification of cross-linked peptides can from hydrophobic domains. Nevertheless, we provide insights into the spatial proximity of certain found that Lys residues belonging to KP domains 4, amino acids and thus into the overall protein 8, 13 and 35 as well as to KA domains 6, 15, 19, 21, structure. Among the different types of cross-links 23, 27, 29 and 31 were partly unmodified. This in elastin are the bifunctional cross-links LNL and demonstrates that most KA and KP domains were D AA. We identified 41 peptides, each containing one ow not completely involved in cross-linking. However, n of the two cross-links. We detected no peptides with lo the abundance of those peptides was rather low ad ΔLNL, indicating the intermediate was fully e d indicating that the degree of cross-linking was reduced to LNL. The sequences revealed the fro generally high. In most cases, the C-terminal Lys of m remarkable fact that cross-links were not only h tlhinee apra iprwepistied eosc caunrrdi ncgo Lnysesq rueesnidtluye si tw wasa sf onuenidth ienr formed intermolecularly but also intramolecularly. ttp://w The latter was proven for the Lys pairs in KP w deaminated nor cross-linked. The other, N-terminal w Lys residue, which typically belongs to the domains 4, 8, 12 and 35 (KxxK motif) as well as in .jbc KA domains 21 and 27 (AKAAKx motif; x being a .o AAKAA motif, was partly unmodified only in KA rg domains 6, 19 and 23. Since elastin was isolated large hydrophobic residue here), which were all by/ linked by an intramolecular LNL. A representative g from adult animals, we conclude that some Lys ue product ion spectrum of one of the intramolecularly s residues are not oxidatively deaminated in the t o cross-linked peptides is shown in Fig. 3 (A). n course of the elastic fiber maturation and remain Ja Interestingly, for nearly all KP and some KA motifs, n u unmodified throughout life. We found no peptides a we identified peptides that show that both adjacent ry containing free Lya residues, even though we 1 considered this modification in the sequencing Lpayrst icripesaitdioune s ofc ano thceor nddeonmsea intso (LFNigL. 2w).i thTohuet 1, 201 approach. Furthermore, we found no linear peptides 9 condensation to ΔLNL requires one Lys to be covering cross-linking motifs of KP domains 10 oxidatively deaminated, whereas the other must and 12, KA domains 17 and 25 and no peptides have remained unmodified. As it can be seen in Fig. originating from the C-terminal RKRK-motif, 1, LNL is a symmetric molecule and hence there is which is highly conserved among species. no way to determine, which of two original Lys It is worth mentioning that the MS analysis of residues was previously oxidized by LOX. Besides linear peptides revealed nine amino acid LNL we found AA to cross-link the adjacent Lys substitutions with respect to the canonical amino residues in the KA domains 6 and 15. The aldol acid sequence of bovine TE (IF 1, UniProt condensation requires two Lya residues, which #P04985), displayed in Tab. 1. While six of the means that K-105, K-109, K-271 and K-275 residue replacements were already described earlier undergo enzymatic deamination by LOX. Another in the UniProt database (25,26), we here report interesting finding is that while two residues three new substitutions in bovine elastin. All nine separate the Lys residues in the identified LNL- mutations were present in hydrophobic regions of containing domains, three residues separate the two elastin and hence not affecting any sites of cross- Lys residues in the AA-containing domains. linking. Strikingly, we identified numerous interpeptidal cross-links, which connect KP and KA motifs. Tab. 3 Native cross-links in elastin 2 summarizes all identified intermolecularly and conducive site specificity (29-33). Thus, in this bifunctionally cross-linked peptides. Such peptides study we used pancreatic elastase, which rapidly consist of two continuous peptide chains that are cleaves mature elastin. covalently connected only by the cross-link. For all For the discovery of DES/IDES-peptides we interpeptidal cross-links, it is impossible to employed an LC-MS/MS-based workflow, which determine whether the two chains are linking two we recently introduced. This approach is based on domains of the same or two different monomers in the detection of emerging marker fragments when vivo. An example is the cross-linked peptide pyridinium-containing molecules such as AGKAGYPT_AKLGAGGA, the fragment DES/IDES are fragmented (24). We enriched the spectrum is shown in Fig. 3 (B). It represents two detected peptides by LC and fraction collection and AA-cross-linked strands of domains 14 and 15, then thoroughly investigated them by tandem MS. respectively. Even though the peptide consists of We interpreted the complex spectra manually two non-continuous chains, it is also possible that and/or in combination with the in-house software they, due to their spatial proximity, originate from PolyLinX, which we developed earlier for the the same monomer and were simply cleaved by PE. identification of polyfunctionally cross-linked Other peptides prove the formation of cross-links peptides (16). between domains 4 and 12 (both KP), 6 and 14 This approach enabled the exact sequence (both KA) as well as 12 (KP) and 27 (KA). determination of three DES/IDES-cross-linked D o w Furthermore, we found many peptides that elastin peptides. The isomers DES and IDES could n lo comprised one chain with a C-terminal Lys residue not be distinguished, since they yield the same ad e d of a cross-linking domain and another very short product ions in collision-induced dissociation (34), fro peptide chain. The latter were only two to five which is the fragmentation method of choice for m h amino acid residues in length, providing often not peptide identification. An annotated product ion ttp sufficient information for an unambiguous spectrum of a DES/IDES-containing peptide along ://w w assignment. For instance, Lys-61 of domain 4 (KP) with its reporter ions released at higher collision is w was found to be intramolecularly linked via LNL to shown in Fig. 4. This peptide as well as the two .jbc .o Lys-64 of the same domain (indicated in Fig. 2), others (Fig. S2A and Fig. S2B) were found to be rg intra- or intermolecularly linked via LNL to Lys composed of three chains. The amino acid by/ g residues located in the chains AKAA and KF (both sequences of these chains are reported in Tab. 3. ue s KA domains) and intra- or intermolecularly linked The identified DES/IDES-peptides have in t o n via AA to AKAA (KA domain). However, the common that they contain one peptide sequence Ja n u motifs AKAA and KF occur in different KA with a large hydrophobic residue C-terminal to one a ry domains of bovine TE twelve and four times, Lys residue followed by some other small residue 1 1 respectively. such as Ala, Gly or Pro. They cover KA domains , 20 1 15, 27 and 17 or 31. Hence, these domains are 9 KA domains are involved in the formation of involved in the formation of DES/IDES. We could desmosine or isodesmosine not elucidate the other domains involved, as the respective sequences were again too short for an The discovery and exact sequence assignment. Nevertheless, the short sequences rich determination of DES/IDES-containing peptides is in Ala residues reveal that they originate from KA considerably more difficult than the identification domains. We found no DES/IDES-peptides with of AA- and LNL-containing peptides. Both DES chains assignable to any of the KP domains. A and IDES can theoretically link up to four peptide schematic overview based on a domain map chains, even though it was proposed that DES/IDES summarizing all identified cross-links is depicted in may only link two peptide chains together in vivo Fig. 5. (27,28). The sequence determination is further complicated on the bioinformatics level by the Pancreatic elastase cleaves in regions of cross- complex fragmentation behavior of such peptides in linking tandem MS and the unpredictable proteolytic cleavage of elastin. In recent years, we studied The finding of DES/IDES-containing peptides, numerous elastases with regard to their cleavage connecting three discontinuous peptide chains, characteristics, but none of them showed any raises the question of whether this indicates the 4 Native cross-links in elastin formation of the cross-link between three domains. numerous cross-linked peptides derived from Alternatively, proteolytic cleavage must have almost all KA and KP domains of naturally cross- occurred between the two adjacent Lys of one linked elastin. The peptides revealed yet unknown cross-link motif. To prove this possibility, we intra- and intermolecular cross-links present in both conducted molecular docking and molecular KA and KP domains and shed new light on elastin’s dynamics (MD) simulations (see Methods section molecular assembly and structure. for details). As shown in Fig. 6, we found that Gerber and Anwar suggested that the docked DES/IDES-peptides, joining two peptide condensation of two bifunctional cross-links leads chains, are able to come into close proximity of the to the formation of DES and IDES and that these catalytic domain of PE. This would allow for cross-links are not necessarily the ultimate result. enzymatic cleavage of the Ala-Ala bonds and The authors rather stated that about half of the explain the presence of the observed peptides. previously formed bifunctional cross-links do not further condense (27). Our experimental data Discussion support these suppositions by revealing the existence of a high number of LNLs and AAs in Elastogenesis is the complex and still poorly elastin of full-grown animals. With the detection of understood sequence of events that eventually leads intramolecular bifunctional cross-links, we to the deposition of elastic fibers in our connective obtained clear evidence that these amino acids are D tissues. One of the many steps is the enzyme- ow formed as intermediates in the course of the n induced cross-linking, which is fundamental for the lo maturation. We conclude that some of them do not ad conversion of TE to the functional polymer elastin. e d While recent studies on TE and in vitro cross- meet due to steric hindrance and in turn remain in fro the elastic fiber. However, damage of the fibers m linked TE provided valuable information on h structural properties of elastin’s precursor in cparoutseeodl ytibcy c leeaxvcaegses idvuer ingm eacghinagn icoarl dissetraesses maoyr ttp://w solution (35-39), very little is known with respect w cause subsequent condensations. w to the organization of mature elastin in the elastic The observation of intramolecular LNL in KP .jbc fiber core. Earlier reports that dealt with the .o domains raises the question whether these domains rg identification of cross-linked elastin peptides are involved in DES/IDES formation. It was earlier by/ focused on the identification of a single or very few g proposed that their contribution is unlikely because ue peptides because of technical limitations (40,41). s of steric constraints caused by the Pro residues (21). t o Such peptides were enriched in labor-intensive Several studies have shown that Pro impedes the n Ja approaches and subsequently analyzed by Edman nu formation of α-helical conformations, which are a degradation (27). The most recent of these studies ry thought to be required to bring the Lys side chains 1 was carried out in 1995 by Brown-Augsburger et al. in close proximity (42,43). However, Baig et al. 1, 20 on porcine elastin (21). The authors deduced three 1 identified a peptide, belonging either to domain 4 9 cross-links connecting domains 10 (KP), 19 and 25 or 8 (both KP) in bovine elastin, that was involved (both KA). However, it remained unknown to what in the formation of DES/IDES (44). Furthermore, extent the respective Lys residues are incorporated quantitative analyses of the DES/IDES content in in these very cross-links, whether the lowering of amphibian elastin, whose cross-linking sites are the LOX activity influenced the cross-linking found almost exclusively KP motifs (8,45), showed qualitatively and which role the other 13 cross- that its level is comparable to those of teleosts (46), linking domains play in the maturation. On the in which KA motifs prevail. These facts together basis of this single cross-link in pig elastin along with our results suggest that KP domains may with low-resolution structural data of recombinant directly contribute to DES/IDES formation, even human TE without PTMs, a head-to-tail model for though the extent could be less when compared to elastin’s assembly was proposed by Baldock et al., KA domains. suggesting an ordered, beaded assembly (22). Yet Another intriguing question concerns the there is no experimental evidence for the interaction homogeneity of the cross-linking, i.e. whether the of the monomers in the suggested manner. same domains are always cross-linked with each In our study, we addressed the above-mentioned other. The opposite is the case: the peptide questions using an untargeted sequencing approach sequences reveal a high diversity in the formation via LC-MS/MS. We report the identification of 5 Native cross-links in elastin of cross-links in elastin of the same tissue for connected almost exclusively via LNL as also several unambiguously assignable Lys residues. described above for intramolecular cross-links. This behavior appears to be independent of the However, we identified one peptide consisting of involved domain type. We discovered for instance the chains AKFGA and KF linked via an LNL. that Lys-61 of domain 4 (KP) was cross-linked to Thus, Lys residues adjacent to Phe can undergo the sequence chain AKAA by both AA and LNL, oxidation induced by LOX. This exception to the respectively. This means in turn that for the rule demonstrates how spontaneous and formation of AA, both Lys residues underwent unregulated the cross-linking takes place. Another deamidation by LOX, whereas in the case of LNL prime example for the high complexity of the cross- only one Lys residue was modified to Lya. It was linking represents Lys-275 of domain 15 (KA). We previously shown that neighbor residues play a found this Lys residue to be a) unmodified; b) cross- critical role for the susceptibility of Lys residues by linked via LNL to a chain KF; c) via LNL to another LOX (47). While adjacent Ala residues favor the chain KI/KL; d) via AA with K-252 in domain 14; oxidation of Lys residues, the susceptibility e) via AA with a chain AKA; f) intramolecularly decreases with the neighboring residues in the cross-linked with the other Lys (Lys-271) of the following order Ala > Val > Leu > Lys > Phe > Tyr. same domain and g) involved in a DES/IDES cross- This suggests for the observed peptides that the Lys link. All these findings are insofar remarkable as residue of the first chain (Lys-61) was the one being they show that a single Lys residue can be involved D o w partially oxidized. We further verified this by the in many different cross-links or even remains n lo identification of linear peptides containing Lys-61. unmodified. This in turn proves that not only a large ad e d A close proximity of the respective residue with the variety of combinations exists, but also that they in fro aldehyde group of AKAA then resulted in the vivo actually coexist in the same tissue, which is m h formation of both LNL and AA. surprising as it is in contrast with prior assumptions ttp Based on these findings for the specificity of of a rather ordered cross-linking in mature elastin. ://w w LOX, it was previously proposed that Lys residues, The Lys residue K-252 in domain 14 of TE is an w which are followed by a bulky amino acid such as exception insofar as it is located within the .jbc .o Leu, Ile, Phe or Tyr, do not undergo oxidation in sequence motif AGKAG, which is unlike typical rg elastin (48). This assumption is used for explaining KA and KP domains. Furthermore, K-252 is not by/ g the formation of DES and IDES in vivo by the accompanied by another Lys residue. A previous ue s condensation of one Lys and one Lya residue on one study by Baig et al. revealed the involvement of K- t o n and two Lya residues on another TE molecule (21). 252 in the formation of DES/IDES suggesting that Ja n u Our peptide data are in line with this assumption. even Lys residues which do not occur pairwise can a ry For example, we found intramolecularly cross- form tetrafunctional cross-links (44). Although we 1 1 linked Lys residues, localized within the same did not identify DES/IDES peptides proving that, , 20 1 domain, to be cross-linked either by AA or by LNL. we confirm the involvement of K-252 in various 9 We observed a correlation between the type of AA cross-links together with Lys residues of KA cross-linking amino acid formed and the sequence domains 6 and 15. The proximity of domains 14 and motif involved. We found that Lys residues of the 15 suggest this cross-link could be formed also motif KxxKy, where y was often a bulky amino intramolecularly and that PE cleaved between the acid, were exclusively cross-linked via LNL. In domains. This is in line with a previous work, which contrast, AA was only formed between the two Lys showed that domains 14 and 15 of in vitro cross- residues of KxxxK motifs, when the Lys residues linked human TE were intramolecularly cross- were not C-terminally flanked by aromatic linked to a high degree (35). residues. Our results further show that KP domains The study of mature elastin by mass also undergo intrachain formation of LNL. We also spectrometry requires its prior hydrolysis. We used observed that Lys residues with adjacent Gly or Ala PE as a tool to generate soluble peptides with residues were linked to other Lys residues also lengths suitable for LC-MS based investigations. followed by Gly or Ala via AA, showing that both PE preferentially cleaves on the carboxyl side of Lys residues were oxidized prior to the Ala, Val, Ile, but also other residues (49). Hence, we condensation. On the other hand, we found that Lys observed multiple cleavage sites near or within KA residues followed by bulky residues were motifs, suggesting that PE predominantly cleaves 6 Native cross-links in elastin the oligo-Ala sequences in elastin. The released novel amino acid substitutions in elastin, arising cross-linked peptides in turn contain only a few Ala from point mutations, are of minor importance for residues and thus the respective chains are often not the overall properties of elastin and reflect the clearly assignable. We found this for many genetic pool of the investigated bovine intermolecularly cross-linked peptides in this study. subpopulation. The three DES/IDES-cross-linked peptides were The observed lack of the C-terminal domain found to be derived exclusively from KA domains. with its polybasic RKRK motif is consistent with The repetitive nature of the domains and the above- other studies dealing with mass spectrometric mentioned unfavorably located cleavage sites did investigation of peptides released from elastin by not allow for a distinct domain annotation for all proteolytic degradation (29,37). This motif is joining peptide chains. The finding that three supposed to mediate cell-adhesion of TE via peptide chains joining the cross-links was binding to integrin α β (55,56). It has been v 3 unexpected. Although we have recently shown that proposed that it is either cleaved afterwards or the formation of DES from three Lya-containing undergoes posttranslational modifications (57). peptides is actually possible (16), it is commonly In summary, we demonstrate with this study the accepted that DES and IDES only link two chains high heterogeneity of the cross-linking pattern in in mature elastin. The reason for the peptide unaltered, mature elastin. Members of the LOX composition most likely lies in the protease family merely catalyze the oxidative deamination D o w treatment. Several studies have shown the potential of Lys residues. This process is followed by n lo of different elastases to cleave within DES/IDES unregulated condensation reactions not involving ad e d peptides (40) or even to release free DES/IDES any catalysts, which mainly depend on the spatial fro from intact elastin (50). Our docking simulations proximity of the Lys and Lya residues. We provide m h between a DES-cross-linked peptide and PE extensive data on the contribution of 8 KA, 5 KP ttp revealed the ability of the protease to cleave and one further domain to intra- and intermolecular ://w w between residues within the cross-link. This can cross-linking, showing the occurrence of multiple w lead to the formation of these three cross-linked condensation reactions at the same residues of .jbc .o peptides. However, our results do not allow different monomers. Hence, we conclude that the rg concluding whether the oligo-Ala peptides attached assembly of elastin takes place in a more by/ g to DES/IDES arise from a single domain or not. unpredictable manner than previously assumed. It ue s The second abundant PTM in elastin is the may be possible that a first ordered alignment and t o n hydroxylation of Pro residues (51,52) and cross-linking of TE molecules as proposed by the Ja n u represents another source of variability. We head-to-tail model defines the overall longitudinal a ry observed partial modification for approx. 20% of all structure of elastic fibers and is followed by an 1 1 Pro residues and an overall hydroxylation degree of unordered and spontaneous formation of further , 20 1 10%. The enzyme responsible for the intracellular cross-links, which determine the lateral structure of 9 modification is prolyl 4-hydroxylase. It modifies the fibers. With regards to the functional Pro residues in GxPA or GxPG motifs and is an significance of an unordered network of covalently important element in collagenesis. We have bound tropoelastin molecules, it has been shown recently shown that prolyl hydroxylation is a that the extension of disordered, cross-linked feature present in elastin for various vertebrates, but peptide chains during mechanical stress leads to a varies with species and tissue (52). While the exact global decrease of the conformational energy in a role of the PTM in elastin is unknown, several ‘rubber-like’ polymer (58). Furthermore, a recent studies on elastin-mimetic peptides suggest that it study using MD simulations of hydrophobic may lead to structural changes and thus influence sequence motifs of elastin supports our model of a the assembly of elastic fibers and their stability disordered aggregate (59). (53,54). Overall, our results reveal surprising new facets Applying the SPIDER algorithm to our data further of the cross-linking in elastin that confute previous allowed us to gain insight into amino acid assumptions on a regularly ordered cross-linking substitutions within the primary TE sequence, pattern in elastin. which may also influence the physicochemical properties of elastin. We assume that all observed 7 Native cross-links in elastin Experimental Procedures gradients 10% to 40% B in 40 min, to 90% B in the next 5 min and by maintenance at 90% B for Materials 10 min. The flow rate was 200 µL min-1 and Porcine pancreatic elastase (PE) was obtained fractions were collected in 30 s intervals. from Elastin Products Company (Owensville, MO, USA). Ammonium bicarbonate, cyanogen bromide, Liquid chromatography coupled to mass 2-mercaptoethanol, urea, dithiothreitol (DTT), spectrometry sodium azide, sodium chloride and trypsin from Collected fractions and purified digests were porcine pancreas were obtained from Sigma separated on an Ultimate 3000 RSLCnano system Aldrich (Steinheim, Germany). Iodoacetamide (Thermo Fisher Scientific). Peptide mixtures were (IAA) was purchased from AppliChem GmbH loaded on the trap column (Acclaim PepMap RP- (Darmstadt, Germany), hydrochloric acid was C18, 300 µm x 5 mm, 5 µm, 100 Å) and washed purchased from Grüssing GmbH (Filsum, with water containing 0.1% TFA for 15 min (30 µl Germany) and analytical grade Tris, trifluoroacetic min-1), before the peptides were separated on the acid (TFA) and formic acid (FA) were obtained separation column (Acclaim PepMap RP-C18, 75 from Merck (Darmstadt, Germany). HPLC-grade μm x 250 mm, 2 µm, 100 Å) using gradients from acetonitrile (VWR Prolabo, Leuven, Belgium) and 1% to 35% B (90 min), 35% to 85% B (5 min) doubly distilled water were used. All other D followed by 85% B (5 min), with solvent A: 0.1% ow chemicals were of analytical grade or better. n FA in H O and solvent B: 0.08% FA in ACN. The lo 2 ad nano-HPLC system was coupled to an Orbitrap e Isolation of bovine aortic elastin Fusion Tribrid mass spectrometer (Thermo Fisher d fro m Bovine aorta was obtained from a single, healthy Scientific) equipped with a Nanospray Flex Ion h full-grown (3-year-old) Fleckvieh cow at a local Source. Data were acquired using data-dependent ttp://w slaughterhouse. Prior to isolation, the aortic sample MS/MS mode: Each high-resolution full-scan w w was cut into small pieces with a lateral length of (automatic gain control (AGC) target value 4 x 105, .jb c between 5 to 10 mm and each piece. Isolation of maximum injection time 50 ms) in the Orbitrap (m/z .o rg intact aortic elastin was carried out as described 300 to 1500, R = 120,000) was followed by high- b/ y previously (60). Purified elastin was dried at room resolution product ion scans in the Orbitrap g u e temperature and stored at -26 °C until analysis. (collision-induced dissociation, CID), 35% s t o normalized collision energy, R = 15,000, AGC n J a In-solution digestion of elastin and enrichment of target value 5 x 104, maximum injection time 200 nu a cross-linked peptides ms) within 5 s, starting with the most intense signal ry 1 Elastin was dispersed in 50 mM Tris-HCl, pH in the full-scan mass spectrum (quadrupole 1, 2 0 isolation window 2 Th). Dynamic exclusion for 1 8.5, at a concentration of 5 mg mL-1 followed by 9 60 s (mass window ± 2 ppm) was enabled to allow reduction with 10 mM DTT at 56 °C and alkylation analysis of less abundant species. Data acquisition with 30 mM IAA for 1 h at room temperature. PE was controlled with Xcalibur 3.0.63. was added to a final enzyme-to-substrate ratio of 1:50 (w/w) and the samples were subsequently Manual tandem MS experiments incubated at 37 °C for 24 h. Proteolysis was stopped by adding TFA to a final concentration of 0.5% DES-/IDES-containing peptides were found to (V/V). Samples were stored at -26 °C prior to be more hydrophilic than other peptides based on further analysis. Fractionation of elastin digests was the identification of specific reporter ions. This performed using an Agilent 1100 system enabled the selective enrichment of those (Waldbronn, Germany) coupled to a Fraction tetrafunctionally cross-linked peptides by Collector II (Waters, Manchester, UK). For collecting early eluting fractions using C18 chromatographic separation, the sample was loaded chromatography. Selected fractions containing onto a ReproSil-Pur 120 column (C18, 3 µm, 2 mm putative tetrafunctionally cross-linked peptides i.d. x 150 mm; Dr. Maisch GmbH, Ammerbuch- were further investigated by manual MS/MS Entringen, Germany) and eluted using a solvent experiments using static nanoelectrospray system of solvent A (0.1% FA in H O) and B (0.1% ionization. Optimization of the higher-energy 2 FA in ACN/H O 80:20 (V/V)) by applying the collisional dissociation (HCD) conditions and 2 8 Native cross-links in elastin extended acquisition times allowed improving the isoform 1 (UniProt #P04985-1) was used as basis spectra quality. Experiments were carried out on an for the peptide identification in StavroX and Orbitrap Velos Pro (Thermo Fisher Scientific) mass PolyLinX. spectrometer. The capillary and spray voltages were set to 50 V and 3.5 kV, respectively. Tandem mass Amino acid analysis spectra starting at m/z 50 were acquired with a An insoluble elastin pellet was hydrolyzed in 6 N resolution of 100,000. A method recently HCl at 110 °C for 24 h. Dried hydrolyzates were introduced by our group was used to detect DES- redissolved in sodium citrate loading buffer (pH /IDES-containing peptides based on specific 2.2), and amino acid analysis was performed by ion product ions generated when elevated collision exchange chromatography with postcolumn energies up to 161 eV are applied (24). Classical derivatization with ninhydrin (Biochrom 30; tandem MS experiments for identification purposes Biochrom, United Kingdom). were carried out by incrementally raising the collision energy up to 50 eV until optimum product Molecular docking and dynamics ion spectra were obtained. Conformations of the peptide Peptide sequencing (AkAAkFGAA_AAkAAAkAAA) were sampled by using the Monte Carlo search method and the D Non-cross-linked and intramolecularly cross- ow AMBER EHT12 force field implemented in MOE n linked peptides were identified by automated de lo (MOE 2012.10; Molecular Operating ad novo sequencing followed by matching to the e Swiss-Prot database using the software PEAKS Environment; http://www.chemcomp.com). In d fro total, 156 conformations of the DES peptide using m (version 7; Bioinformatics Solutions, Waterloo, h Canada). The enzyme was set to ‘none’ and the aann de nuesregdy fwoirn dfuorwth eorf 2d0o ckkcianlg m sotul-d1 iwese.r eG gOeLneDr a5te.2d ttp://w search was taxonomically restricted to Bos taurus. w (62) was used to dock the peptide into the binding w Hydroxylation of Pro (mass shift of +15.9949 Da), pocket of porcine pancreatic elastase (PDB ID .jbc and oxidative deamination of Lys (mass shift .o 1H9L - PORCINE PANCREATIC ELASTASE rg of -1.0316 Da) were chosen as variable COMPLEXED WITH ACETYL-VAL-GLU-PRO- by/ modifications. Mass error tolerances for precursor g ILE-COOH, 1.67 Å resolution). The active site ue and fragment ions were set to 8 ppm and 0.015 Da, s Ser214 and all residues within a 15 Å radius were t o respectively. The peptide score threshold was used to define the binding pocket. Goldscore was n Ja decreased until a false discovery rate of ≤ 1% on the nu used as scoring function. 10 poses for each of the a peptide level was reached. Amino acid substitutions ry 156 conformations were calculated. All docking 1 were identified using the SPIDER algorithm (61). poses were subjected to a cluster analysis using the 1, 2 0 A minimum of two identified peptides having a 1 peptide heavy atoms for calculating RMSD values. 9 mutation was used as constraint for unbiased Three favorable clusters were retrieved and identification. Intramolecular cross-links were subjected to MD simulations using the AMBER14 detected in PEAKS by using a modified sequence package (63). Parameters for non-standard amino of bovine elastin IF 1. More precisely, cross-linking acids, Lya and Lys residues connecting to form motifs were reduced to a single Lys residue and the DES ring, were first prepared. The modified amino mass difference from this residue to the acids were built and energy minimized using HF/6- corresponding intramolecularly cross-linked motif 31G* method. Then, parameter libraries of these was implemented as a variable PTM. Processed residues were generated. Before performing the spectra which did not result in satisfying sequence MD simulations, force field and charges for the information or only in sequence tags were exported DES-containing peptides were assigned as MGF files and further analyzed using the (AMBER03 force field for all standard amino acid programs StavroX (23) and PolyLinX to identify residues and the extra-generated parameters for the intermolecular cross-links (16). LNL and AA were non-standard amino acid residues). Water (TIP3P both included as cross-links with mass decreases of model) and counter ions were then added and the 17.0265 Da and 20.0738 Da, respectively. Lys protein-peptide complexes were energy minimized residues were considered as potential sites of cross- (63). MD simulations were performed at 310 K, linking, and the amino acid sequence of bovine TE thus the temperature of the system was gradually 9 Native cross-links in elastin increased from 0 to 310 K during the first few ps. measurements. The work was supported by the Then the temperature was kept at 310 K by German Research Foundation (DFG) grant HE applying Langevin dynamics with a collision 6190/1-2 (A.H.), by the LEO Foundation Center for frequency of 1 ps-1. In the final step, free MD Cutaneous Drug Delivery (2016-11-01) (A.H.), by simulation was carried out from 100 ps to 20 ns the European Regional Development Fund of the using NPT ensemble, which means that a constant European Commission (C.U.S.) and by the pressure was set at 1 bar and the temperature was Fraunhofer Internal Programs under Grant No. kept constant at 310 K. A time step of 2 fs with Attract 069-608203 (C.E.H.S.). SHAKE algorithm was also applied. The Particle- Mesh-Ewald method was used for computing Conflict of interest—The authors declare that electrostatic interactions, while non-bonded they have no conflicts of interest with the contents interaction was calculated using a cut-off radius at of this article. 10 Å. Acknowledgements—We thank Dr. Christian Ihling (MLU Halle-Wittenberg) for assistance with MS D o w n lo References a d e d 1. Sage, H. (1982) Structure-function relationship in the evolution of elastin. J Invest Dermatol 79 Suppl fro m 1, 146s-153s h 2. Li, B., Alonso, D. O., Bennion, B. J., and Daggett, V. (2001) Hydrophobic hydration is an important ttp source of elasticity in elastin-based biopolymers. J Am Chem Soc 123, 11991-11998 ://w w 3. Rauscher, S., Baud, S., Miao, M., Keeley, F. W., and Pomes, R. (2006) Proline and glycine control w.jb protein self-organization into elastomeric or amyloid fibrils. Structure 14, 1667-1676 c.o 4. Tamburro, A. M., Bochicchio, B., and Pepe, A. (2003) Dissection of human tropoelastin: exon-by-exon brg/ chemical synthesis and related conformational studies. Biochemistry 42, 13347-13362 y g u 5. Clarke, A. W., Arnspang, E. C., Mithieux, S. M., Korkmaz, E., Braet, F., and Weiss, A. S. (2006) es Tropoelastin massively associates during coacervation to form quantized protein spheres. Biochemistry t on J 45, 9989-9996 an u a 6. Moon, H. J., Finney, J., Ronnebaum, T., and Mure, M. (2014) Human lysyl oxidase-like 2. Bioorg ry 1 Chem 57, 231-241 1 , 2 7. Chung, M. I., Miao, M., Stahl, R. J., Chan, E., Parkinson, J., and Keeley, F. W. (2006) Sequences and 0 1 9 domain structures of mammalian, avian, amphibian and teleost tropoelastins: Clues to the evolutionary history of elastins. Matrix Biol 25, 492-504 8. He, D., Chung, M., Chan, E., Alleyne, T., Ha, K. C., Miao, M., Stahl, R. J., Keeley, F. W., and Parkinson, J. (2007) Comparative genomics of elastin: Sequence analysis of a highly repetitive protein. Matrix Biol 26, 524-540 9. Franzblau, C., Faris, B., and Papaioannou, R. (1969) Lysinonorleucine. A new amino acid from hydrolysates of elastin. Biochemistry 8, 2833-2837 10. Lent, R. W., Smith, B., Salcedo, L. L., Faris, B., and Franzblau, C. (1969) Studies on the reduction of elastin. II. Evidence for the presence of alpha-aminoadipic acid delta-semialdehyde and its aldol condensation product. Biochemistry 8, 2837-2845 11. Francis, G., John, R., and Thomas, J. (1973) Biosynthetic pathway of desmosines in elastin. Biochem J 136, 45-55 12. Akagawa, M., Yamazaki, K., and Suyama, K. (1999) Cyclopentenosine, major trifunctional crosslinking amino acid isolated from acid hydrolysate of elastin. Arch Biochem Biophys 372, 112-120 13. Nakamura, F., Yamazaki, K., and Suyama, K. (1992) Isolation and structural characterization of a new crosslinking amino acid, cyclopentenosine, from the acid hydrolysate of elastin. Biochem. Biophys. Res. Commun. 186, 1533-1538 10