Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Fragmentation Follows Structure: Top-Down Mass Spectrometry Elucidates the Topology of Engineered Cystine-Knot Miniproteins

Fragmentation Follows Structure: Top-Down Mass Spectrometry Elucidates the Topology of Engineered Cystine-Knot Miniproteins

  • Michael Reinwarth, 
  • Olga Avrutina, 
  • Sebastian Fabritz, 
  • Harald Kolmar


Over the last decades the field of pharmaceutically relevant peptides has enormously expanded. Among them, several peptide families exist that contain three or more disulfide bonds. In this context, elucidation of the disulfide patterns is extremely important as these motifs are often prerequisites for folding, stability, and activity. An example of this structure-determining pattern is a cystine knot which comprises three constrained disulfide bonds and represents a core element in a vast number of mechanically interlocked peptidic structures possessing different biological activities. Herein, we present our studies on disulfide pattern determination and structure elucidation of cystine-knot miniproteins derived from Momordica cochinchinensis peptide MCoTI-II, which act as potent inhibitors of human matriptase-1. A top-down mass spectrometric analysis of the oxidised and bioactive peptides is described. Following the detailed sequencing of the peptide backbone, interpretation of the MS3 spectra allowed for the verification of the knotted topology of the examined miniproteins. Moreover, we found that the fragmentation pattern depends on the knottin’s folding state, hence, tertiary structure, which to our knowledge has not been described for a top-down MS approach before.


Precise information concerning identity, structure, and pharmacokinetics of drug candidates is an important issue in the development of biopharmaceuticals, among them a vast number of bioactive peptides [1]. Therefore, reliable and reproducible analysis of their structure and topology needs well-elaborated high-throughput methods. Tandem mass spectrometry (MS/MS or MS2) has become a valuable tool for the identification and quantification of peptides and proteins. Recent experiments indicate that MS analysis can also be applied to their structural characterization since folded and unfolded molecules may give rise to different fragmentation patterns upon ionization [1][3]. Moreover, MS2-based methods can be applied to identify and characterize inter- and intramolecular disulfide bonds along with NMR and X-ray analysis [2]. Following the bottom-up methodology for the determination of disulfide-bond topologies in multi-disulfide proteins, the analytes are partially reduced and subjected to enzymatic digestion prior to the analysis via MS or MS/MS [4][6]. Although this approach provides decisive information on the primary structure, determination of disulfide connectivities via MS analysis of proteolytic fragments of non-reduced or partially reduced peptides with high disulfide content remains a complicated task.

In some instances, the available specific enzymatic cleavage sites within the peptide of interest do not necessarily lead to one-cystine-one-peptide fragment distributions after digestion of folded species, or no appropriate cleavage sites are present at all. Moreover, partial reduction and S-alkylation prior to proteolysis often results in a complex mixture of variants [4], [7], [8]. Furthermore, lack of information originating from incomplete sequence coverage and the fact that not every posttranslational modification (PTM) could be detected, are additional disadvantages [4], [5], [9][15].

In recent years, efforts have been made to apply a top-down methodology for the characterization of full-length native proteins and peptides via concurrent cleavage of disulfide bonds using MS/MS methods minimizing loss of information [8], [11][14], [16][18]. To this end, different MS fragmentation methods were applied, among them the widely employed collision-induced dissociation (CID) featuring low collision energies. Electron-transfer or capture dissociation (ETD or ECD, respectively) methods have also been used for the generation of MS/MS spectra [8], [17][21]. For CID the mobile proton theory, which states that positive charges are randomly distributed among all amino acids of an analyzed peptide, thus facilitating amide N-protonation and cleavage, is an important tool to interprete MS spectra [16], [22][27]. Regarding arginine-rich peptides this is not entirely true since the basic side chains are known to sequester charges [16], [22][24]. Thus, the ratio of charge to the number of arginines within the sequence must be over one to make the mobile proton theory applicable [16], [22][24]. Hence, the overall charge of the peptide plays an important role since protonated arginine side chains are known to promote disulfide cleavage [15], [19][24]. As a consequence, the collision energy required for the cleavage of disulfides depends on the ratio of cystine to arginine units within the sequence [16], [22][24]. Additional difficulties upon assigning peptidic fragments in the complex spectra progressively increase with the size of the peptide and the number of S-S bonds [8], [17], [18]. The side chains of several amino acids tend to neutral losses and the formation of uncommon product ions under ionization conditions due to specific proximity effects [2], [8], [17], [18], [28]. In particular, asymmetric cleavage of disulfide bonds results in the formation of both perthiocysteine (Ptc) and dehydroalanine (Dha), of which the latter is known to induce cleavage of the peptide backbone not at the amide bond, but promotes fragmentation between the amide nitrogen and the α-carbon, hence forming c-ions (Figure 1C) [2], [8], [16], [28], [29]. These effects lead to an increased number of signals in the spectrum resulting in a statistically high chance of false-positive assignments. Thus, regarding large disulfide-rich peptides and proteins, the gained spectra require further in-depth analysis due to their complexity [30].

Figure 1. Structure of synthetic open-chain MCoTI, structural overview and fragment ion formation.

(A) 3D structure of synthetic open-chain cystine knot oMCoTI (pdb: 2IT8).15 Active loop is shown in red. (B) A 6 Å close-up on the disulfide-tightened core of MCoTI. For both (A) and (B) disulfides are shown in yellow. (C) Overview of the generation of peptidic fragment ions upon CID. (D) Sequences of the parent MCoTI wild type (wt) and the miniproteins (1–3) used in this study. Positions with altered amino acid residues (regarding the wild type sequence) are marked red.

Following previous studies on the analysis of cysteine-rich proteins, Green and co-workers elaborated the idea of an additional fragmentation stage (MS3) and used it for a top-down approach to analyze chicken lysozyme [2], [22], [31], [32]. Therein, nozzle-skimmer ionization was used as the first dissociation source, while regular CID served as a second fragmentation stage [2]. Although this pseudo MS3 approach has delivered promising results, the MS3 utilizing two dissociation sources could facilitate the detailed evaluation of the disulfide pattern topology.

As model disulfide-rich macromolecules for investigation of structural features and disulfide bond connectivities by application of MS3 methodologies, engineered cystine-knot protease inhibitors were used in the present study. These bioactive peptides, also known as knottins, consist of about 30 amino acid residues and are characterized by a unique, ‘pseudo-knotted’ architecture [33][35].

Knottin’s structural core is defined by three β-strands which are interconnected by three disulfide bonds. The bonds between CysII and CysV as well as CysI and CysIV form a ring which is penetrated by a third cystine, connecting CysIII and CysVI (Figure 1B) [34][36]. Additional thermodynamic stability is provided via an extensive network of hydrogen bonds between the peptide backbone and the side chains of the amino acids located within the β–sheets [37], [38]. Therefore peptides containing this structural motif display an exceptional structural, thermal, and biological robustness [35], [39], [40].

Knottins are considered excellent scaffolds for the generation of tailor-made compounds for diagnostic and therapeutic applications since the surface-exposed loops consist of segments with high structural and numeral flexibility, while the conserved core only tolerates minor amino acid exchanges [32], [35], [41][43]. Engineered miniproteins derived from trypsin inhibitors of the bitter gourd Momordica cochinchinensis (MCoTI) have been successfully applied for the inhibition of proteinases of clinical relevance, among them human mast cell tryptase β, foot-and-mouth-disease virus 3C protease, and, most recently, cancer-related protease matriptase-1 (Figure 1D) [44][46]. Moreover, a cystine-knot peptide from cone snails, Ziconotide, has been approved for the treatment of severe and chronic pain [47]. Although optimized conditions for both chemical and recombinant synthesis of these miniproteins have been recently reported, oxidative folding towards the respective bioactive isomer still reveals no information on the disulfide topology [7], [48], [49]. Biological activity is an important indication for a correct fold, but for engineered variants with multiple amino acid exchanges this information may not necessarily prove the knotted cystine pattern [45].

To date, determination of disulfide bond connectivities for this class of peptides relies on NMR and X-Ray of native peptides as well as on MS analysis of chemically or enzymatically modified ones [17], [18]. Compared to the NMR and X-Ray technology, a major advantage of an analytical technique such as a mass spectrometry is the requirement of less analyte. Despite the fact that method development may be time consuming, the broad availability of MS instruments enable more laboratories to make structural determinations and, when using well-established methods, analysis can be performed in shorter time. Thus, Tam and coworkers recently described disulfide mapping of the cyclotide hedyotide B2 by MS2 analysis of the partially reduced and alkylated peptide [50]. Furthermore, Balaram et al. elucidated disulfide topology of four-cysteine native peptides as well as enzymatically treated conotoxines [17]. Camadro et al. described structural analysis of knotted Psalmopeotoxin I through combined tandem MS and 15NNMR [18]. In the present study we applied MS3 technology for the characterization of folded and oxidised open-chain MCoTI variants. For the confirmation of structural and topological characteristics we employed a top-down approach to deliver information on elements regarding the primary, secondary, and tertiary structure of the miniprotein.

Materials and Methods

Peptides 1, 2, and 3 have been synthesized as previously reported [44], [49].

Detailed information on mass spectrometric measurements is provided in the supporting information (Method S1). Briefly, the 4000 QTRAP LC/MS/MS and the 6500 QTRAP LC/MS/MS systems (AB Sciex Germany GmbH, Table S1) were used. For MS full scans experiments signal intensity was adjusted to achieve 1e6 counts per second (cps) maximal signal intensity at 10 µL per minute infusion from a syringe pump. Peptide stock solutions have been prepared by dissolving 0.5 mg of the respective peptide in 1 mL of a mixture of 20% methanol (Fluka Analytical, LC-MS CHROMASOLV) and 80% water (Merck KGaA, LC-MS LiChrosolv) with 0.2% formic acid (Fluka Analytical, analytic additive).

For MS2 experiments both declustering potential (DP) and collision energy (CE) have been optimized via ramping. Regarding MS3 measurements, resonance excitation energy (AF2) was additionally optimized (using ‘ramping’ functionality of Analyst software). All spectra were accumulated to ensure distinct differentiation between signals and noise.


Experimental design and general results

As model peptides we chose variants derived from the cystine-knot peptide MCoTI (Figure 1), a synthetic protease inhibitor combining the three-disulfide pattern with the open-chain amide backbone [44]. Since peptides of this family are known for their structural stability regarding biological and chemical decomposition, denaturation or disulfide scrambling should not take place under acidic conditions which were used for mass-spectrometric analysis [7]. The examined miniprotein variants comprised selected exchanges within the flexible loops [44], [49].

To investigate the influence of structural characteristics on the fragmentation pattern, tandem mass spectrometric analysis triple quadrupole systems with a linear ion trap were used (ESI, Table S1). The design of these systems allows for two dissociation steps upon selection of the respective precursor ions. Thus, we applied CID to both the oxidised three-disulfide peptide 1 (Figure 1) and its reduced precursor. In the resulting spectra, the unique peaks and those showing the typical shift of hydrogens or sulfhydryl groups (Figure 2 and Figures S1S8) were compared.

Figure 2. MS2 of MCoTI peptide 1.

(A) CID of the doubly charged ion of reduced peptide 1. (B) CID of the doubly charged ion of folded miniprotein. (C) CID of the five-fold charged ion of folded miniprotein.

Identification of the respective fragment was achieved via selection of the corresponding precursor for MS3 (Figure 2 and Figures S9S13). Ion path parameters and ionization energies, namely the declustering potential, the collision energy and the excitation energy for the MS3 were semi-automatically optimized (ramped) towards maximum intensity of the respective product ions. Similar fragmentation patterns were found for peptides 1, 2, and 3 that deviated from each other in the amino acid sequences of loop 1, 2, and 5 (Figure 1) [44].

While amino acid exchanges outside the inhibitory loop 1 that protrudes into the active site of the respective protease induced no altered fragmentation of the native peptide, as expected, relative intensities of the respective ions dropped upon arginine exchanges since protonated arginine side chains are known to promote disulfide cleavage [15], [19][21]. Indeed, the relative peak intensity of the C-terminal fragments (y14, Figure 2 and Figures S7 and S8) is higher for the arginine-containing variants (2 and 3) compared to 1 that lacks this moiety. In contrast, for the fragment c16y19 or c16z20, respectively, only a minor decrease of signal intensity was observed for peptide 3, since it still contained an additional arginine which was kept unchanged in each variant. Moreover, signal intensity of product ion series immediately drops upon cleavage of arginine, which is caused by the sequestered charge at the arginine side chains, thus decreasing the intensity of the following b-ions by leaving the majority of them neutral (Figures 3B and Figure S9).

Figure 3. MS3 of the CID-obtained major fragments of MCoTI 1.

(A) MS3 of c10 of the doubly charged ions of unfolded peptide. (B) MS3 of c16y19 of the doubly-charged ions of folded miniprotein. (C) MS3 of y14 of the doubly charged ions of unfolded peptide. (D) MS3 of y24 of the five-fold charged ions of folded miniprotein. Arrows above the spectra indicate intensity amplifications.

The MS2 spectra only showed minor differences between the unfolded and folded variants (Figure 2). Indeed, the fragmentation patterns were similar, apart from the expected unsymmetrical cystine dissociations resulting in Ptc or Dha moieties and different intensities for the fragments c16y19 (c16z20) and, particularly, c10y26.

CID spectra obtained from the lower charged precursors [M+2H]2+ or [M+3H]3+, respectively, were significantly enriched with three fragment species (c10, c16y19 and y14) covering the whole sequence (Figure 2). These major fragments were identified and sequenced using MS3 methodology. This pattern lacked Cys11, which was caused by the formation of the major c-ion fragment c10 (Figure 2). C-ion formation has been reported to occur upon cysteine fragmentation towards Dha [2].

For high-charged states, e.g. [M+5H]5+, MS2 spectra of the folded miniprotein displayed a different fragmentation pattern. Interestingly, we observed formation of several b-ions which could not be detected upon low-charge CID (Figure 2). While all b-ions comprising large linear fragments still possessed at least one disulfide, one fragment was identified lacking the central SD motif, thus resulting in the generation of a ‘bridged’ fragment (Figure 3D). However, assignment of the fragments was complicated by the multiple charge states within the spectra.

Elucidation of primary structure

For the sequential analysis, the major fragment ions c10, c16y19 and y14 were chosen for MS3 analysis. Assignment of b- and y-ions was complicated since several obviously predetermined breaking points dictated the shape of the spectra, while other fragments were hard to distinguish from noise (Figure 3). Additionally, ‘satellite’ peaks were observed for each sub-fragment ion, originated from neutral losses of amino acid side chains, e. g. lysine (−17 Da) [51].

In the case of the aminoterminal fragment c10, intensity of b- and c-ions decreased upon cleavage of Arg9 (which is equivalent to the formation of the b9/c9 ion), due to the sequestered charge at the guanidyl moiety (Figure 3A). Regarding the y-ions, the dominating fragment in the MS3 spectrum was the y6-ion which became the starting point for further b- ion fragmentation leading to the generation of b10y6, b9y6, b8y6, and b7y6 ions (Figure 3A). Moreover, the y-ion series was detected until formation of y3.

The c16y19, respectively, c16z20, fragment was analyzed by MS3 and sequencing revealed its water adduct. For this fragment we observed b4 as well as y4, y3, and y2-subfragments which provided detailed sequence information (Figure 3B). As expected, the spectrum was enriched by y-ions formed upon aspartate cleavages. Interestingly, the b4 ion fragment was not observed as water adduct anymore, thus indicating Arg12 being primarily responsible for the binding of water (Figure 3B).

The carboxyterminal fragment y14 delivered complete information regarding the primary structure. B-ion formation from b13 to b6, as well as y-ion formation of y13 to y6 (Figure 3C) was observed. However, this spectrum was also complicated by predetermined breaking points as well as neutral losses, due to several proximity effects or directed non-standard fragmentation like Dha-promoted c-ion formation.

Secondary structure

Determination of the cystine connections could not be achieved directly through detection of ‘bridged’ fragments for lower-charge states. This is due to the fact that the cysteine residues are close to each other within the sequence, e. g. having only one isoleucine in between CysIV and CysV. Moreover, analysis is complicated by the intrinsic knotted structure and the fact that disulfide dissociation occurs upon charge localization on basic side chains [8], [17], [18].

For this reason the four-cysteine fragment y14 was of particular interest as it contained an intact disulfide bridge. For the determination of the expected III–VI connection, we analyzed three related m/z fragments, namely, 1294, 1326 and 1358, via MS3. These ions correspond to a four-cysteine-three-sulfur (y14-32, Figure 4A), a four-cysteine-four-sulfur (y14, Figure 3C) and a four-cysteine-five-sulfur (y14+32, Figure 4B) fragment, respectively. MS3 analysis of those fragments of 1 suffered from the absence of an arginine compared to 2, thus these fragment series were displayed a higher intensity in the arginine-containing variant (Figures S11 and S12). Nevertheless, combinatorial interpretation of the resulting spectra clearly revealed the predicted connection of CysIII to CysVI (Figure 4). Involvement of CysIII was verified since it was found to be cut off exclusively as a Ptc moiety for all three parent fragments (Figure 4A), hence the remaining y ions were exclusively found as y-32 ions. After cleavage of the Ptc from the four-cysteine-three-sulfur fragment, a three-cysteine-one-sulfur one (y13-32, Figure 4A) was received, while the cleavage of Cys (y13) or Dha (y13+32) moieties did not occur (Figure 4A and Figures S11 and S12, red ‘peaks’). Indeed, the distribution of the sulfur atoms was exclusively found to be on CysIV and CysV. Hence, connections of IV–V, IV–VI, and V–VI could be excluded. As a result, the only possible intramolecular connection was the expected III–VI one. To provide further evidence of the connection CysIII to CysVI, analysis of the four-cysteine-five-sulfur fragment was performed. After generation of the b12 ion, a three-cysteine-three-sulfur fragment (b12+32, Figure 4B) was obtained, which is underpinned through the absence of the ions b12-32 and b12 (Figure 4B and Figures S11 and S12, red ‘peaks’). Thus, connections of III–IV and III–V could be excluded solely leaving III–VI one possible. Nonetheless, regarding the lower charged precursors, no exact conclusion could be made on the two remaining cystines, which still could be the native, knotted I–IV/II–V or the intuitively implausible, unknotted I–V/II–IV connection.

Figure 4. MS3 of y14-32 (A) and y14+32 (B) and the resulting combinatorial interpretation.

In red are inexistent peaks to provide evidence on the respective Ptc or Dha cleavage. Arrows above the spectra indicate intensity amplifications.

For verification of the knottin fold, the second disulfide connection was evaluated. However, the probability that a bridged fragment is formed upon CID is very low since at least three peptide bonds need to be cleaved while the respective cystine connection has to remain intact. For this purpose we focused on the higher charged states (5+) for the relationship between the protonated arginine residues and the fragmentation of disulfide bonds (see introduction section) [16], [23], [24].

However, we found the linear y24 fragment containing the connections CysIII–CysVI as well as CysII–CysIV or CysII–CysV, respectively (Figure 3D). Upon cleavage of the labile SD motif, this fragment comprised two linear fragments connected through one intermolecular cystine. For the proof of the II–V connection, this fragment was subjected to MS3 analysis. This procedure revealed an ion consisting of two linear peptide chains connected through the disulfide linkage between CysII and CysV. Direct MS3 analysis of this bridged fragment produced several ions belonging to the expected product connected via the linkage of CysII and CysV (Figure 5). Since peak intensities were relatively low due to the unlikeliness of the multiple, specific breakages necessary for the generation of this fragment, the complete fragment series could not be detected although further optimizations of the detector parameters could increase the signal to noise ratio. Nevertheless, the major product ions were identified.

Figure 5. Spectra and assigned structures of the bridged fragments.

(A) MS3 of the bridged fragment of 1. (B) MS3 of the bridged fragment of 2. Arrows above the spectra indicate intensity amplifications.

For peptide 2 this procedure of fragment analysis was interfered by another fragment ion with nearly identical mass (Figure 5), which might be avoided by tightening the isolation window of the precursor ions. Nevertheless, the expected product ion was verified, too. Additionally, fragmentation revealed a small fragment ion which was identified to originate from the II–V-connected peptide ion (Figure S13). This in-depth analysis was not performed for verification of peptide 3.

Tertiary structure

Comparison of the CID spectra of the positively charged folded and unfolded peptides revealed a distinct intensity increase of the peak corresponding to both the c16y19 as well as the c10y26 fragment for the folded miniprotein (Figure 2), which was confirmed by MS3 analysis. Generation of both fragments was compared for each miniprotein variant at different collision energies. Ratios depicted in Table 1 result from collision energies optimized towards a maximum TIC.

For miniproteins 1 and 3 ratios of over 6 regarding the formation of both fragments for the reduced peptide and the folded knottin were observed. For miniprotein 2 the observed ratio was approximately 20. Additionally, the dependency of these ratios as well as the maximum fragment formation on the applied collision energy in obedience to the folding state was elucidated. Interestingly, for 1 and 3 no influence of the folding state on the maximum fragment formation was observed. For miniprotein 2, instead, increasing collision energies lead to an adjustment in fragment formation. Thus, a semi-automatic optimization of the collision energy to determine the energy with a maximum formation of fragment c10y26 (Figure 6) was performed. Maxima at different collision energies were observed for the reduced peptide and the folded knottin. This accounted for high ratios at CE 50 as well as decreasing ratios for increasing collision energies.

Figure 6. Optimization of the collision energy towards a maximum formation of fragment c10y26 of 2.


The analyzed miniprotein variants displayed very similar fragmentation patterns, notwithstanding the fact that sequence alterations have been applied for all sequential and structural regions (Figure 1C). Interestingly, the resulting fragmentation pattern was not influenced by sequential changes. Therefore, the reproducibility of the results in this experimental setting should allow for the prediction of CID spectra for novel representatives of the MCoTI family.

Fragmentation was obviously guided by cysteines as well as aspartates. In non-mobile proton systems aspartate side chains are known to induce carboxyterminal cleavage [2], [16]. Hence, predetermined cleavage sites within the “DSD” motif of the respective peptide variants were verified. Regarding the cysteine residues, the fragmentation pattern was predetermined by CysII and CysIII of the sequence, although the folding state, hence secondary and tertiary structural elements, only had influence on signal intensity rather than fragment formation itself. This indicates a cysteine-induced cleavage of the peptide backbone resulting in the formation of c-ions for both reduced and oxidised cysteines [2], [16]. In contrast, the charge state played a key role in fragmentation. Higher charged precursors led to the formation of large b ion fragments. As a consequence, only a decreased amount of cysteine-guided fragmentation could be observed. Both findings strongly correlate with the fragmentation patterns reported by Green and co-workers [2], [16].

We achieved full sequence coverage for the reduced and oxidised peptides in this top-down approach. For all three peptides similar fragments were detectable and could be identified. Hence, for evidence of known sequences or those with minor exchanges, sequence analysis can be performed easily without sophisticated methods of sample preparation or spectra analysis.

Disulfide topology is the major determinant of secondary structure for cystine-knot miniproteins. Since no predetermined breaking points resulting in ‘bridged’ fragments were found in CID-based spectra, in-depth analysis towards determination of the cystine pattern was performed for relatively large, linear fragments containing intramolecular disulfides. MS3-based sequencing was performed and combinatorial interpretation of the resulting spectra revealed the expected disulfide connectivity (Figure 4). This analysis was based on the generation of Dha or Ptc moieties upon CID resulting from unsymmetrical breakage of cystines. This untypical manner of analysis has rarely been reported for the topology determination in a top-down approach since it is hardly applicable for large proteins [17], [18], [52]. The specific restrictions present for cystine-knot miniproteins, in particular, the short sequential distances between the respective cysteine residues and the overlap of cystine connections, almost exclude the generation of ‘bridged’ fragments which are not connected via the amide backbone and thus the exact determination of disulfide patterns [17], [18].

Interpretation of the spectra of the bridged fragments for the determination of a second cystine connection was complicated through their complex nature and low intensities, thus intensive manual analysis was required. Since multiple collisions are possible within the collision cell and the linear ion trap, the analyzed fragments indeed most likely were tertiary or quaternary fragment ions. Moreover, probabilities for the generation of those multiple fragmentation events are relatively low, peak intensities of the resulting ‘pseudo-MSn’ spectra are very low as well. This only allowed for detection of sequentially predetermined breaking points, thus no formation of fragment ion series was observed. Instead, several individual species with increased intensities were determined and the bridged fragments could be clearly identified for both 1 and 2, although the spectra of 2 were interfered by another fragment ion with a similar mass (Figure 5). Despite this complicated methodology for the ab initio determination of disulfide patterns, verification of knotted disulfide connectivity of oxidised peptides of the MCoTI family now can be routinely applied without time-consuming methods by fragment analysis of common CID spectra.

We observed increased intensities of the fragments corresponding to the surface-exposed loops in the CID spectra of the oxidatively folded miniproteins. Since some loops of knottins of the MCoTI family are subjected to tensed conformations upon oxidative folding, their increased energetic content may account for the increased peak intensities (Table 1). This is particularly reasonable for loop 1 (flanked by CysI and CysII) as its biological function, i.e. binding to the respective protease, strongly depends on the folding state of the peptide, which requires the correct disulfide pattern along with the surface-exposed and tensed inhibitory loop region [53]. Accordingly, these elements of the tertiary structure most likely account for the rise in peak intensity.

Further interest was spent on the investigation of the shifted maxima in collision energy regarding the formation of fragment c10y26 of reduced and folded 2. Interestingly, peptides 1 and 3 did not show this behavior. The additional basic amino acid Lys10 in the inhibitory loop 1 as well as a slightly different 3D structure may account for this interesting characteristic of 2. Applying this type of analysis, determination of the correct cystine-knot fold of miniproteins, at least with sequences similar to 1, can now be achieved through comparison of fragment formation in the folded and unfolded state. MS-based detection of the influence of chemically modified amino acids on the tertiary structure of proteins has often been reported, but to the best of our knowledge, an effect of the tertiary structure on the fragmentation pattern has not been described for a top-down approach before [1], [8], [54][56].


In this study a set of synthetic matriptase-1 inhibitors were analyzed through MS3 methodology. Full sequence coverage and determination of disulfide topology were achieved. Determination of primary structure is of particular interest for ‘one-bead-one-compound’ libraries which have already been applied for linear peptides and for medium throughput screening of bioactive molecules [57], [58]. It would be interesting to see whether this methodology can be applied to the on-bead sequence analysis of cystine-knot peptides although oxidative folding of the resin-bound peptides is still critical [59]. Additionally, an influence of secondary and tertiary structure on the fragmentation pattern was observed. As a consequence, the folding state and thus the knotted nature of the disulfides can easily be verified by detecting either formation of the respective fragments upon CID or via comparison of the ratios of the formation of loop 1 in the folded and unfolded state, which should be typically in the range of 10. Taking into consideration the reproducibility of this analysis, this methodology can be applied to the analysis of structure and disulfide bond connectivities of other miniproteins of the MCoTI family and may be useful for other peptides with related multidisulfide pattern and interesting pharmacological properties as e.g. cyclotides or conotoxines [17], [18]. The influence of tertiary structure on fragmentation patterns of MS/MS spectra may lead to information on the region of protein-protein interaction with ramifications for analysis of various biomolecular interactions as e.g. antibody-antigen binding or conformational changes of antigens upon antibody binding.

Supporting Information

Figure S1.

Initial spectra of 1. (A) MS1 of 1. (B) Zoom-In on M+2H]2+. (C) Zoom-In on M+5H]5+.


Figure S2.

Initial spectra of unfolded precursor of 1. (A) MS1 of unfolded precursor of 1. (B) Zoom-In on M+2H]2+.


Figure S3.

Initial spectra of 2. (A) MS1 of 2. (B) Zoom-In on M+2H]2+. (C) Zoom-In on M+5H]5+.


Figure S4.

Initial spectra of unfolded precursor of 2. (A) MS1 of unfolded precursor of 2. (B) Zoom-In on M+2H]2+.


Figure S5.

Initial spectra of 3. (A) MS1 of 3. (B) Zoom-In on M+2H]2+.


Figure S6.

Initial spectra of unfolded precursor of 3. (A) MS1 of unfolded precursor of 3. (B) Zoom-In on M+2H]2+.


Figure S7.

MS2 of MCoTI peptide 2. (A) CID of the triply charged ion of reduced precursor of peptide 2. (B) CID of the triply charged ion of folded miniprotein 2. (C) CID of the fivefold charged ion of folded miniprotein 2.


Figure S8.

MS2 of MCoTI peptide 3. (A) CID of the triply charged ion of reduced precursor of peptide 3. (B) CID of the triply charged ion of folded miniprotein 3.


Figure S9.

MS3 of major fragments of MCoTI peptide 2. (A) MS3 of c10 ion. (B) MS3 of b16z20 ion. (C) MS3 of y14 ion. Arrows above the spectra indicate intensity amplifications.


Figure S10.

MS3 of major fragments of MCoTI peptide 3. (A) MS3 of c10 ion. (B) MS3 of y14 ion. Arrows above the spectra indicate intensity amplifications.


Figure S11.

MS3 of y14-32 (A) and y14+32 (B) for the combinatorial interpretation of 2. In red are inexistent peaks to provide evidence on the respective Ptc or Dha cleavage. Arrows above the spectra indicate intensity amplifications.


Figure S12.

MS3 of y14-32 (A) and y14+32 (B) for the combinatorial interpretation of 3. In red are inexistent peaks to provide evidence on the respective Ptc or Dha cleavage. Asterisk indicates fragment from different parent ion with identical mass. Arrows above the spectra indicate intensity amplifications.


Figure S13.

MS3 of b14z20-32 fragment of MCoTI peptide 2.


Table S1.

Applied source and gas parameters.


Method S1.

Experimental procedure for mass spectrometric measurements.



We thank Iain Gibb and Dietrich Merkel (AB SCIEX) for their steady support of this project. AB SCIEX Germany GmbH, Darmstadt, generally provided equipment and expertise in MS measurements of the cystine-knot peptides.

Author Contributions

Conceived and designed the experiments: SF MR OA HK. Performed the experiments: MR SF. Analyzed the data: MR SF OA HK. Contributed reagents/materials/analysis tools: MR SF OA HK. Contributed to the writing of the manuscript: MR SF OA HK.


  1. 1. Berkowitz SA, Engen JR, Mazzeo JR, Jones GB (2012) Analytical tools for characterizing biopharmaceuticals and the implications for biosimilars. Nat Rev Drug Discov 11: 527–540.
  2. 2. Chen J, Shiyanov P, Schlager JJ, Green KB (2012) A pseudo MS3 approach for identification of disulfide-bonded proteins: uncommon product ions and database search. J Am Soc Mass Spectrom 23: 225–243.
  3. 3. Seidler J, Zinn N, Boehm ME, Lehmann WD (2010) De novo sequencing of peptides by MS/MS. Proteomics 10: 634–649.
  4. 4. Goransson U, Craik DJ (2003) Disulfide mapping of the cyclotide kalata B1. Chemical proof of the cystic cystine knot motif. J Biol Chem 278: 48188–48196.
  5. 5. Gundry RL, White MY, Murray CI, Kane LA, Fu Q, et al. (2009) Preparation of proteins and peptides for mass spectrometry analysis in a bottom-up proteomics workflow. Curr Protoc Mol Biol Chapter 10: Unit10 25.
  6. 6. Switzar L, Giera M, Niessen WM (2013) Protein digestion: an overview of the available techniques and recent developments. J Proteome Res 12: 1067–1077.
  7. 7. Reinwarth M, Nasu D, Kolmar H, Avrutina O (2012) Chemical synthesis, backbone cyclization and oxidative folding of cystine-knot peptides: promising scaffolds for applications in drug design. Molecules 17: 12533–12552.
  8. 8. Goyder MS, Rebeaud F, Pfeifer ME, Kalman F (2013) Strategies in mass spectrometry for the assignment of Cys-Cys disulfide connectivities in proteins. Expert Rev Proteomics 10: 489–501.
  9. 9. Castagnola M, Cabras T, Iavarone F, Vincenzoni F, Vitali A, et al. (2012) Top-down platform for deciphering the human salivary proteome. J Matern Fetal Neonatal Med 25: 27–43.
  10. 10. Kellie JF, Tran JC, Lee JE, Ahlf DR, Thomas HM, et al. (2010) The emerging process of Top Down mass spectrometry for protein analysis: biomarkers, protein-therapeutics, and achieving high throughput. Mol Biosyst 6: 1532–1539.
  11. 11. Lanucara F, Eyers CE (2013) Top-down mass spectrometry for the analysis of combinatorial post-translational modifications. Mass Spectrom Rev 32: 27–42.
  12. 12. Macht M (2009) Mass spectrometric top-down analysis of proteins. Bioanalysis 1: 1131–1148.
  13. 13. Peng Y, Yu D, Gregorich Z, Chen X, Beyer AM, et al. (2013) In-depth proteomic analysis of human tropomyosin by top-down mass spectrometry. J Muscle Res Cell Motil.
  14. 14. Chait BT (2006) Chemistry. Mass spectrometry: bottom-up or top-down? Science 314: 65–66.
  15. 15. Zhang H, Ge Y (2011) Comprehensive analysis of protein modifications by top-down mass spectrometry. Circ Cardiovasc Genet 4: 711.
  16. 16. Chen J, Shiyanov P, Zhang L, Schlager JJ, Green-Church KB (2010) Top-down characterization of a native highly intralinked protein: concurrent cleavages of disulfide and protein backbone bonds. Anal Chem 82: 6079–6089.
  17. 17. Gupta K, Kumar M, Balaram P (2010) Disulfide bond assignments by mass spectrometry of native natural peptides: cysteine pairing in disulfide bonded conotoxins. Anal Chem 82: 8313–8319.
  18. 18. Combes A, Choi SJ, Pimentel C, Darbon H, Waidelich D, et al. (2009) Determination with matrix-assisted laser desorption/ionization tandem time-of-flight mass spectrometry of the extensive disulfide bonding in tarantula venom peptide Psalmopeotoxin I. Eur J Mass Spectrom (Chichester, Eng). 15: 517–529.
  19. 19. Chrisman PA, Pitteri SJ, Hogan JM, McLuckey SA (2005) SO2-* electron transfer ion/ion reactions with disulfide linked polypeptide ions. J Am Soc Mass Spectrom 16: 1020–1030.
  20. 20. Guthals A, Bandeira N (2012) Peptide identification by tandem mass spectrometry with alternate fragmentation modes. Mol Cell Proteomics 11: 550–557.
  21. 21. Zubarev RA, Kruger NA, Fridriksson EK, Lewis MA, Horn DM, et al. (1999) Electron Capture Dissociation of Gaseous Multiply-Charged Proteins Is Favored at Disulfide Bonds and Other Sites of High Hydrogen Atom Affinity. J Am Chem Soc 121: 2857–2862.
  22. 22. Ginter JM, Zhou F, Johnston MV (2004) Generating protein sequence tags by combining cone and conventional collision induced dissociation in a quadrupole time-of-flight mass spectrometer. J Am Soc Mass Spectrom 15: 1478–1486.
  23. 23. Lioe H, O’Hair RA (2007) A novel salt bridge mechanism highlights the need for nonmobile proton conditions to promote disulfide bond cleavage in protonated peptides under low-energy collisional activation. J Am Soc Mass Spectrom 18: 1109–1123.
  24. 24. Wells JM, Stephenson Jr JL, McLuckey SA (2000) Charge dependence of protonated insulin decompositions. Int J Mass Spec 203: A1–A9.
  25. 25. Wysocki VH, Tsaprailis G, Smith LL, Breci LA (2000) Mobile and localized protons: a framework for understanding peptide dissociation. J Mass Spectrom 35: 1399–1406.
  26. 26. Boyd R, Somogyi A (2010) The mobile proton hypothesis in fragmentation of protonated peptides: a perspective. J Am Soc Mass Spectrom 21: 1275–1278.
  27. 27. Cheng C, Gross ML (2000) Applications and mechanisms of charge-remote fragmentation. Mass Spectrom Rev 19: 398–420.
  28. 28. Paizs B, Suhai S (2005) Fragmentation pathways of protonated peptides. Mass Spectrom Rev 24: 508–548.
  29. 29. Chrisman PA, McLuckey SA (2002) Dissociations of disulfide-linked gaseous polypeptide/protein anions: ion chemistry with implications for protein identification and characterization. J Proteome Res 1: 549–557.
  30. 30. Wu SL, Jiang H, Lu Q, Dai S, Hancock WS, et al. (2009) Mass spectrometric determination of disulfide linkages in recombinant therapeutic proteins using online LC-MS with electron-transfer dissociation. Anal Chem 81: 112–122.
  31. 31. Körner R, Wilm M, Morand K, Schubert M, Mann M (1996) Nano electrospray combined with a quadrupole ion trap for the analysis of peptides and protein digests. J Am Soc Mass Spectrom 7: 150–156.
  32. 32. Raska CS, Parker CE, Huang C, Han J, Glish GL, et al. (2002) Pseudo-MS3 in a MALDI orthogonal quadrupole-time of flight mass spectrometer. J Am Soc Mass Spectrom 13: 1034–1041.
  33. 33. Craik DJ, Daly NL, Waine C (2001) The cystine knot motif in toxins and implications for drug design. Toxicon 39: 43–60.
  34. 34. Gelly JC, Gracy J, Kaas Q, Le-Nguyen D, Heitz A, et al. (2004) The KNOTTIN website and database: a new information system dedicated to the knottin scaffold. Nucleic Acids Res 32: D156–159.
  35. 35. Gracy J, Le-Nguyen D, Gelly JC, Kaas Q, Heitz A, et al. (2008) KNOTTIN: the knottin or inhibitor cystine knot scaffold in 2007. Nucleic Acids Res 36: D314–319.
  36. 36. Chiche L, Heitz A, Gelly JC, Gracy J, Chau PT, et al. (2004) Squash inhibitors: from structural motifs to macrocyclic knottins. Curr Protein Pept Sci 5: 341–349.
  37. 37. Felizmenio-Quimio ME, Daly NL, Craik DJ (2001) Circular proteins in plants: solution structure of a novel macrocyclic trypsin inhibitor from Momordica cochinchinensis. J Biol Chem 276: 22875–22882.
  38. 38. Heitz A, Hernandez JF, Gagnon J, Hong TT, Pham TT, et al. (2001) Solution structure of the squash trypsin inhibitor MCoTI-II. A new family for cyclic knottins. Biochemistry 40: 7973–7983.
  39. 39. Craik DJ, Cemazar M, Wang CK, Daly NL (2006) The cyclotide family of circular miniproteins: nature’s combinatorial peptide template. Biopolymers 84: 250–266.
  40. 40. Wang CK, Hu SH, Martin JL, Sjogren T, Hajdu J, et al. (2009) Combined X-ray and NMR analysis of the stability of the cyclotide cystine knot fold that underpins its insecticidal activity and potential use as a drug scaffold. J Biol Chem 284: 10672–10683.
  41. 41. Kimura RH, Cheng Z, Gambhir SS, Cochran JR (2009) Engineered knottin peptides: a new class of agents for imaging integrin expression in living subjects. Cancer Res 69: 2435–2442.
  42. 42. Werle M, Kafedjiiski K, Kolmar H, Bernkop-Schnurch A (2007) Evaluation and improvement of the properties of the novel cystine-knot microprotein McoEeTI for oral administration. Int J Pharm 332: 72–79.
  43. 43. Werle M, Schmitz T, Huang HL, Wentzel A, Kolmar H, et al. (2006) The potential of cystine-knot microproteins as novel pharmacophoric scaffolds in oral peptide drug delivery. J Drug Target 14: 137–146.
  44. 44. Glotzbach B, Reinwarth M, Weber N, Fabritz S, Tomaszowski M, et al. (2013) Combinatorial Optimization of Cystine-Knot Peptides towards High-Affinity Inhibitors of Human Matriptase-1. PLoS One 8: e76956.
  45. 45. Sommerhoff CP, Avrutina O, Schmoldt HU, Gabrijelcic-Geiger D, Diederichsen U, et al. (2010) Engineered cystine knot miniproteins as potent inhibitors of human mast cell tryptase beta. J Mol Biol 395: 167–175.
  46. 46. Thongyoo P, Roque-Rosell N, Leatherbarrow RJ, Tate EW (2008) Chemical and biomimetic total syntheses of natural and engineered MCoTI cyclotides. Org Biomol Chem 6: 1462–1470.
  47. 47. Klotz U (2006) Ziconotide–a novel neuron-specific calcium channel blocker for the intrathecal treatment of severe chronic pain–a short review. Int J Clin Pharmacol Ther 44: 478–483.
  48. 48. Hofmeyer T, Bulani SI, Grzeschik J, Krah S, Glotzbach B, et al. (2013) Protein Production in Yarrowia lipolytica Via Fusion to the Secreted Lipase Lip2p. Mol Biotechnol.
  49. 49. Reinwarth M, Glotzbach B, Tomaszowski M, Fabritz S, Avrutina O, et al. (2013) Oxidative folding of peptides with cystine-knot architectures: kinetic studies and optimization of folding conditions. Chembiochem 14: 137–146.
  50. 50. Nguyen GK, Zhang S, Wang W, Wong CT, Nguyen NT, et al. (2011) Discovery of a linear cyclotide from the bracelet subfamily and its disulfide mapping by top-down mass spectrometry. J Biol Chem 286: 44833–44844.
  51. 51. Papayannopoulos IA (1995) The interpretation of collision-induced dissociation tandem mass spectra of peptides. Mass Spectrometry Reviews 14: 49–73.
  52. 52. Choi S, Jeong J, Na S, Lee HS, Kim HY, et al. (2010) New algorithm for the identification of intact disulfide linkages based on fragmentation characteristics in tandem mass spectra. J Proteome Res 9: 626–635.
  53. 53. Heitz A, Avrutina O, Le-Nguyen D, Diederichsen U, Hernandez JF, et al. (2008) Knottin cyclization: impact on structure and dynamics. BMC Struct Biol 8: 54.
  54. 54. Pan J, Borchers CH (2013) Top-down structural analysis of posttranslationally modified proteins by Fourier transform ion cyclotron resonance-MS with hydrogen/deuterium exchange and electron capture dissociation. Proteomics 13: 974–981.
  55. 55. Engen JR (2009) Analysis of protein conformation and dynamics by hydrogen/deuterium exchange MS. Anal Chem 81: 7870–7875.
  56. 56. Heck AJR (2008) Native mass spectrometry: a bridge between interactomics and structural biology. Nat Meth 5: 927–933.
  57. 57. Aina OH, Liu R, Sutcliffe JL, Marik J, Pan CX, et al. (2007) From combinatorial chemistry to cancer-targeting peptides. Mol Pharm 4: 631–651.
  58. 58. Lam KS, Liu R, Miyamoto S, Lehman AL, Tuscano JM (2003) Applications of one-bead one-compound combinatorial libraries and chemical microarrays in signal transduction research. Acc Chem Res 36: 370–377.
  59. 59. Atherton E, Clive DL, Sheppard RC (1975) Letter: Polyamide supports for polypeptide synthesis. J Am Chem Soc 97: 6584–6585.