Research Article Received: 3 September 2013
Revised: 13 October 2013
Accepted: 20 October 2013
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 178–184 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6772
Insulin, islet amyloid polypeptide and C-peptide interactions evaluated by mass spectrometric analysis Michael Landreh1, Gunvor Alvelius1, Jan Johansson2,3 and Hans Jörnvall1* 1
Department of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77 Stockholm, Sweden KI-Alzheimer’s Disease Research Center, NVS Department, Karolinska Institutet, S-141 86 Stockholm, Sweden 3 Department of Anatomy, Physiology and Biochemistry, Swedish University of Agricultural Sciences, S-751 23 Uppsala, Sweden 2
RATIONALE: Insulin, islet amyloid polypeptide (IAPP), and the C-peptide part of proinsulin are co-secreted from the
pancreatic beta cell granules. IAPP aggregation can be inhibited by insulin and insulin aggregation by C-peptide, but different binding and disaggregating interactions may apply for the peptide complexes. A more detailed knowledge of these interactions is necessary for the development strategies against diabetic complications that stem from peptide aggregations. METHODS: Mass spectrometry (MS) is utilized to investigate pH-dependencies, sequence determinants and association strengths of interactions between pairs of all three peptides. Electrospray ionization (ESI)-MS was used to monitor complex formation and interaction stoichiometries at different pH values. Collision-induced dissociation (CID) was employed to probe relative association strengths and complex dissociation pathways. RESULTS: IAPP, like C-peptide, removes insulin oligomers observable by ESI-MS. Both C-peptide and IAPP form stable 1:1 heterodimers with insulin. Complexes of the negatively charged C-peptide with the positively charged IAPP, on the other hand, are easily dissociated. Replacement of the conserved glutamic acid residues in C-peptide with alanine residues increases the stability, indicating that net charge alone does not predict association strength. Binding to insulin has been suggested to stabilize a helical fold in IAPP via charge and hydrophobic interactions, which is in agreement with the now observed high gas-phase stability and sensitivity to low pH. CONCLUSIONS: Combined, these results suggest that the C-peptide–insulin and IAPP–insulin interactions are mediated by a defined binding site, while such a feature is not apparent in the IAPP–C-peptide association. Hence, IAPP and Cpeptide are interacting in similar manners and with similar monomerizing effects on insulin, suggesting that both peptides can prevent insulin aggregation. Simultaneous interactions of all three peptides cannot be excluded but appear unlikely from the uneven pairwise binding strengths. Copyright © 2013 John Wiley & Sons, Ltd.
Insulin secretion is crucial for the maintenance of glucose homeostasis, and impairment of this process is connected with diabetic disorders.[1] In the secretory granules of the β-cells, the proinsulin precursor is cleaved into mature insulin and C-peptide.[2] After cleavage, insulin is stored in crystalline form with Zn2+ and as a fluid phase with other granule components.[3] Insulin and islet amyloid polypeptide (IAPP), which co-localize with C-peptide in the fluid phase of the granule, have a high propensity to form amyloid fibrils. Amyloid fibrils composed of IAPP have been observed in the pancreas of >90% of type 2 diabetic patients and are thought to play a role in β-cell death,[4] which has sparked the development of synthetic peptide inhibitors of IAPP fibrillation.[5] Insulin amyloid can deposit locally at the site of repeated insulin injections[6] and may be of further importance in diabetes of both types.[7]
178
* Correspondence to: H. Jörnvall, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77 Stockholm, Sweden. E-mail: [email protected]
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
IAPP is mostly unstructured in solution, but becomes α-helical upon membrane binding,[8] which accelerates fibril formation by promoting intermolecular contacts. IAPP dimerization accelerates aggregation in a similar manner and may occur through α-helical or β-strand intermediates.[9,10] Insulin, on the other hand, can inhibit IAPP aggregation.[11] Binding to the B-chain of insulin stabilizes the α-helical IAPP structure, but simultaneously blocks its self-association sites and thus stabilizes the monomeric form.[9,12] C-peptide, which does also not adopt a well-defined secondary structure in solution,[13] was reported to reduce IAPP aggregation, but only when present at a >10-fold excess.[14] This effect has been suggested to be mediated by interactions between the negatively charged C-peptide and the positively charged IAPP at the granular pH.[15] Sequences and suggested interaction points are summarized in Fig. 1. In addition, C-peptide can also form aggregates with time at high concentrations.[16,17] High and low concentrations both promote insulin aggregation,[18] which is driven by pH- or temperature-induced unfolding, or by sampling unfolded conformations in the absence of stabilizing interactions.[19] In both cases, local unfolding releases aggregation ’hot spots’ from the helical regions of the A- and B-chain and initiates fibril assembly in this manner.[20] Mass spectrometry has emerged as a tool for the
Copyright © 2013 John Wiley & Sons, Ltd.
Interactions of insulin, IAPP and C-peptide in ESI-MS
Figure 1. Sequences of the peptides used in this study. Top: Residue contacts involved in the interaction between IAPP and the insulin B-chain are highlighted in gray. + and – indicate charged residues as well as the charges on the N- and C-termini. The average molecular masses are indicated behind each peptide. investigation of protein aggregation. Protein interactions can be preserved in the gas phase and yield information about complex stoichiometries and relative binding stabilities.[21] Additionally, ion mobility and hydrogen/deuterium exchange studies can provide insights into aggregate structure.[22,23] Using electrospray ionization mass spectrometry (ESI-MS), it is possible to detect prefibrillar insulin oligomers and thus monitor the early steps of aggregation.[24] We have reported that molecular interactions with C-peptide can remove these insulin oligomers[25] and interfere with insulin fibrillation.[26] The disaggregating effect is mediated by conserved glutamic acid residues in the negatively charged C-peptide, and molecular interactions with positive charges on insulin lead to co-precipitation of the two peptides via mutual neutralization at certain ratios and pH conditions.[27] The monomerizing actions can help to counterbalance the low insulin solubility at granular pH and trap soluble insulin in the fluid phase.[7] Whether the interactions between insulin and IAPP prevent not only IAPP but also insulin from fibrillation has not been reported, likely due to the pronounced aggregation propensity of IAPP. In summary, heteromolecular interactions that interfere with peptide aggregation have been observed for all three peptide pairs, and can be relevant in the secretory granule. However, only the binding of insulin to IAPP has been characterized in detail by X-ray crystallography and solution nuclear magnetic resonance (NMR).[9,12] In the present study, we use gas-phase interactions in ESI-MS to probe pHdependency, binding strength and sequence determinants to extract information about the molecular basis for each of these pairwise interactions.
EXPERIMENTAL
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
RESULTS IAPP and C-peptide remove insulin oligomers observed by ESI-MS The ESI mass spectra of IAPP show no significant change in charge state distribution between pH 4 and 8 (Supplementary Fig. S1, see Supporting Information). The peptide appears mostly as monomer, but with minor dimer and trimer fractions, which is in agreement with solution studies.[28] C-peptide and insulin form oligomers close to their isoelectric points at pH 3.4 and 5.5, respectively.[27] The presence of equimolar amounts of IAPP visibly reduced the amount of insulin oligomers (Fig. 2), indicating that IAPP has a monomerizing effect on insulin. This finding is similar to the disaggregating effects of C-peptide on insulin oligomers[25] and supports the suggestion that IAPP binding may help to solubilize multimeric insulin.[9]
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
179
Recombinant insulin (human) was purchased from Sigma (St. Louis, MO, USA). C-peptide and the C-peptide mutants with Glu → Ala substitutions at position 11 only, 27 only, or 3, 11 and 27 together were purchased from GenScript (Piscataway, NJ, USA). C-peptide was dissolved in deionized water, desalted with SepPak C18 cartridges (Waters, Milford, MA, USA), and stored in lyophilized form at 20 °C. Insulin was dissolved in H2O adjusted to pH 5 with 10% acetic acid (Sigma) to a concentration of 1 mM and lyophilized directly. Human recombinant IAPP was purchased from Bachem (Bubendorf, Switzerland). Since human IAPP is known to
aggregate within 5–10 min in aqueous solution, the entire vial was dissolved in 100% dimethyl sulfoxide (DMSO) to a final concentration of 2 mM and stored at 20 °C. Aliquots from the peptide stock solutions were diluted in 0.5% HCl (pH 1) or 20 mM ammonium acetate (pH 4, 5, 6, 7, and 8) to a final concentration of 50 μM each and immediately subjected to MS analysis. Data was acquired using a QTOF Ultima API mass spectrometer (Waters) equipped with a Z-spray source operated in the positive-ion mode. Samples were introduced via metal-plated borosilicate glass capillary needles (Proxeon, Denmark). The source temperature was 80 °C, the capillary voltage 1.6 kV and the cone and RF lens 1 potentials were 100 and 38 eV, respectively. The mass spectrometer was operated in single-reflector mode to achieve sufficient resolution at reasonable sensitivity (10 000, full width half maximum definition). The mass scale was calibrated in linear mode using horse myoglobin (Sigma). Scans were acquired at a rate of 1 scan per 2 s between 500 and 3000 m/z. Collision gas was argon at 5.2 × 10–5 mbar. Mass spectra were analyzed using the Waters MassLynx 4.1 software. To characterize relative complex stabilities, the [M + 4H]4+ ion as the most abundant for each complex was chosen for tandem mass spectrometric (MS/MS) analysis. Collision voltages were automatically adjusted from 10 to 60 eV in increments of 5 eV. Collision-induced dissociation (CID) spectra were recorded for 3 min (1 scan per 0.5 s) for each collision voltage increment and subsequently averaged. The complexes dissociated into doubly and singly charged peptide ions. The signal ratio between complexes and complex plus free peptides was calculated from the [M + 4H]4+ complex ions and the sum of the [M + 2H]2+ and the [M + 3H]3+ peptide ions. Three separate sets of CID data were recorded for each complex and ratios are shown as mean ± standard deviation from these replicates. The intensity of the intact 1:1 complex divided by the sum of intensities of the [M + 1H]1+ and [M + 2H]2+ species of IAPP and the intact 1:1 complex was used for interactions of C-peptide, C-peptide variants, or insulin with IAPP. For the insulin–C-peptide interaction, the complex/peptide ratio was calculated using the [M + 1H]1+ and [M + 2H]2+ species of C-peptide and the intact 1:1 complex.
M. Landreh et al.
Figure 2. IAPP has a monomerizing effect on insulin. (A) IAPP is mostly monomeric, with minor di-, tri- and tetrameric fractions. (B) At pH 5, insulin oligomers composed of 2–6 molecules can be observed in the high m/z range. (C) Addition of an equimolar amount of IAPP removes insulin oligomers, instead leading to complexes with 1–2 IAPP and 1–3 insulin molecules (D, magnification). IAPP, C-peptide and insulin interactions are observable by ESI-MS
180
At pH 5, ESI-MS signals corresponding to the [M + 4H]4+ and [M + 5H]5+ charge states of 1:1 complexes between insulin and IAPP were observed. Additionally, peaks corresponding to insulin–IAPP heteromers with 1:2, 2:1 and 2:2 stoichiometries could be detected (Fig. 2). The complex formation and stoichiometries are similar to the interactions between Cpeptide and insulin, where insulin oligomer removal is accompanied by the formation of heteromers composed of one insulin and one or two C-peptide molecules (Fig. 3).[25,27] Heteromer formation between IAPP and insulin could be observed between pH 4 and 8, but was found to yield the strongest MS signals at pH 5 (Supplementary Fig. S2(A), see Supporting Information). ESI-MS spectra of solutions containing C-peptide and IAPP did not deviate noticeably from the oligomeric status or charge state distribution of either peptide between pH 4 and 8 (Supplementary Fig. S2(B), see Supporting Information), but showed strong signals for 1:1 heteromers between IAPP and C-peptide, as well as minor signals corresponding to complexes with 1:2, 2:1, and 2:2 stoichiometries (Fig. 3(A), Supplementary Fig. S2(B)). While the C-peptide signals were markedly reduced at pH 4, the C-peptide–IAPP heteromers persisted with only minor differences in intensity, indicating that the C-peptide– IAPP complex is not affected by the decreased solubility of
wileyonlinelibrary.com/journal/rcm
C-peptide close to its isoelectric point. When the pH was decreased to 1, peaks corresponding to C-peptide–insulin, insulin–IAPP and IAPP–C-peptide heteromers were almost completely abolished (bottom panels in Supplementary Figs. S2(A) and S2(B), see Supporting Information). In the case of insulin, this may be attributed to unfolding at low pH, which can affect the proposed IAPP binding site in the helical section of the B-chain.[19] While the interaction site(s) for insulin and C-peptide have not been identified, previous studies have shown that protonatable residues are involved in the association,[27] which may also be the case for the C-peptide–IAPP interactions.[15] Interestingly, the simultaneous presence of all three peptides did not change this association pattern significantly. C-peptide and IAPP form stable complexes with insulin, but not with each other Next, we compared the pairwise interactions between insulin, IAPP and C-peptide. We subjected the [M + 4H]4+ heterodimers observed for each peptide pair to CID at increasing collision voltages and monitored the fraction of intact complexes. Figure 4(A) shows the dissociation curves at increasing collision voltages; representative CID spectra illustrating the degree of dissociation for different precursor complexes at a collision voltage of 40 eV are shown in Fig. 4(B). This approach yields information about the relative strengths of the peptide
Copyright © 2013 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
Interactions of insulin, IAPP and C-peptide in ESI-MS
Figure 3. C-peptide and its Glu → Ala variants form complexes with IAPP in a similar manner. Spectra are shown for (A) IAPP with wt C-peptide; (B) IAPP with C-peptide with the glutamic acid residues at positions 3, 11, and 27 replaced with alanine; (C) IAPP with C-peptide with the glutamic acid residue at position 27 replaced with alanine; and (D) IAPP with C-peptide with the glutamic acid residue at position 11 replaced with alanine. Heteromers composed of two IAPP and one C-peptide, and one IAPP and two C-peptide molecules can be detected for every C-peptide variant.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
increase in gas-phase stability (Fig. 4(A)), which suggests that the high number of negative charges in C-peptide is not required for interactions with IAPP.
CID preferentially releases insulin–IAPP and insulin–C-peptide heterodimers from heterotrimeric insulin–IAPP or insulin–C-peptide complexes In addition to heterodimers, we also observed heterotrimers consisting of one IAPP with two insulin molecules or two IAPP with one insulin molecule. We used gas-phase dissociation of protein complexes to obtain information on their molecular architecture by performing CID analysis of the [M + 6H]6+ and the [M + 7H]7+ charge states of the 2:1 and 1:2 IAPP–insulin complexes, respectively, which were the strongest signals for each species (Figs. 5(A) and 5(B)). Both complexes were found to dissociate primarily into IAPP–insulin heterodimers. 2:1 IAPP–insulin complexes dissociate into 1:1 heterodimers and IAPP monomers and a minor fraction of IAPP dimers. Similarly, 1:2 IAPP–insulin complexes preferentially dissociated into 1:1 heterodimers and insulin monomers. In addition, a second dissociation pathway resulting in dimeric insulin and monomeric IAPP could also be detected.
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
181
interactions in the absence of solvent. In the gas phase, hydrophobic interactions are weakened, while electrostatic interactions are strengthened.[29] The complex between insulin and C-peptide as well as the complex between insulin and IAPP showed high stability in CID, with fragmentation occurring beginning at a collision voltage of 40 eV (Fig. 4(A)). This is consistent with reports that charge-based interactions are important for the formation of both complexes.[9,12,27] In contrast, the IAPP–C-peptide complex was almost completely dissociated at 40 eV and is therefore significantly less stable than the insulin–IAPP and insulin–C-peptide complexes. Since IAPP has a net charge at pH 5 of 3+ and C-peptide one of 4–, it is likely that their association occurs primarily via charge attraction. Therefore, the low gas-phase stability of the heterodimer is unexpected. To investigate the role of charges in complex formation, we tested whether the reduction of C-peptide charge by Glu → Ala mutations at position 11 only, position 27 only, or positions 3, 11, and 27 combined affects the gas-phase stability of the complexes with IAPP. All C-peptide mutants formed heteromers with IAPP. Charge-state distributions, signal intensities and stoichiometries were comparable to complexes formed with IAPP and wild-type (wt) C-peptide (Fig. 3). CID revealed that Glu → Ala substitutions in C-peptide cause a moderate
M. Landreh et al.
Figure 4. Gas-phase stabilities of 1:1 peptide complexes at pH 5. (A) Insulin–IAPP and insulin–C-peptide complexes have high stability when subjected to CID at increasing collision voltages, while IAPP–C-peptide complexes are easily dissociated. Replacement of one or three glutamic acid residues in C-peptide causes a moderate increase in stability. (B) CID spectra recorded at 40 eV for IAPP–C-peptide (top panel), C-peptide–insulin (middle panel) and IAPP–insulin complexes (bottom panel).
Figure 5. Collision-induced dissociation (CID) of heterotrimers: (A) CID of 2:1; (B) 1:2 complexes of IAPP and insulin; and (C) 1:2 complexes of C-peptide and insulin at a collision voltage of 30 eV. The 2:1 and 1:2 heterotrimers of insulin and IAPP (A and B) preferentially release the 1:1 heterodimer of both peptides upon CID. Similarly, the 1:2 C-peptide-insulin complex (C) dissociates into the 1:1 heterodimer and a minor population of dimeric insulin. The different dissociation products are indicated by black and grey arrows. The complex selected for CID is indicated by *.
182
We have previously reported that insulin and C-peptide can form a 2:1 complex composed of an insulin dimer with a bound C-peptide, which results in mutual charge neutralization and precipitation at pH 5.[25,27] At pH 6, however, the [M + 7H]7+ charge state of this complex is readily detectable by ESI-MS. When subjected to CID, these
wileyonlinelibrary.com/journal/rcm
heterotrimers dissociate into 1:1 insulin–C-peptide heterodimers and insulin monomers and a smaller population of insulin homodimers and monomeric C-peptide when subjected to CID (Fig. 5(C)). This bears close resemblance to the dissociation routes of 1:2 IAPP–insulin heterotrimers, but can be observed already at a lower collision voltage.
Copyright © 2013 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
Interactions of insulin, IAPP and C-peptide in ESI-MS
DISCUSSION
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
REFERENCES [1] D. J. Michael, H. Cai, W. Xiong, J. Ouyang, R. H. Chow. Mechanisms of peptide hormone secretion. Trends Endocrinol. Metab. 2006, 17, 408. [2] H. W. Davidson, C. J. Rhodes, J. C. Hutton. Intraorganellar calcium and pH control proinsulin cleavage in the pancreatic beta cell via two distinct site-specific endopeptidases. Nature 1988, 333, 93. [3] J. Suckale, M. Solimena. The insulin secretory granule as a signaling hub. Trends Endocrinol. Metab. 2010, 21, 599. [4] P. Westermark, A. Andersson, G. T. Westermark. Islet amyloid polypeptide, islet amyloid, and diabetes mellitus. Physiol. Rev. 2011, 91, 795. [5] L. M. Yan, M. Tatarek-Nossol, A. Velkova, A. Kazantzis, A. Kapurniotu. Design of a mimic of nonamyloidogenic and bioactive human islet amyloid polypeptide (IAPP) as nanomolar affinity inhibitor of IAPP cytotoxic fibrillogenesis. Proc. Natl. Acad. Sci. USA 2006, 103, 2046. [6] F. E. Dische, C. Wernstedt, G. T. Westermark, P. Westermark, M. B. Pepys, J. A. Rennie, S. G. Gilbey, P. J. Watkins. Insulin as an amyloid-fibril protein at sites of repeated insulin injections in a diabetic patient. Diabetologia 1988, 31, 158. [7] M. Landreh, J. Johansson, H. Jornvall. C-peptide: a molecule balancing insulin states in secretion and diabetes-associated depository conditions. Horm. Metab. Res. 2013, 45, 769. [8] R. P. Nanga, J. R. Brender, J. Xu, G. Veglia, A. Ramamoorthy. Structures of rat and human islet amyloid polypeptide IAPP (1–19) in micelles by NMR spectroscopy. Biochemistry 2008, 47, 12689. [9] J. J. Wiltzius, S. A. Sievers, M. R. Sawaya, D. Eisenberg. Atomic structures of IAPP (amylin) fusions suggest a mechanism for fibrillation and the role of insulin in the process. Protein Sci. 2009, 18, 1521. [10] N. F. Dupuis, C. Wu, J. E. Shea, M. T. Bowers. The amyloid formation mechanism in human IAPP: dimers have betastrand monomer-monomer interfaces. J. Am. Chem. Soc. 2011, 133, 7240.
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
183
In the present study, we have probed the peptide interactions involving insulin, IAPP and C-peptide using ESI-MS. All three peptides have previously been reported to interact in solution,[7] and, in all three cases, peptide charges have been suggested to be important for binding. However, reliable structure models are only available for the insulin–IAPP complex, which is held together by a mixture of charge and hydrophobic interactions.[9,12] While hydrophobic interactions are usually weakened in the gas phase, electrostatic interactions appear strengthened.[29] However, the different gas-phase stabilities of the three peptide pair interactions and the observation that a reduction in C-peptide charge stabilizes its interactions with IAPP show that net charge alone is a poor predictor of binding interactions in ESI-MS, as additional factors likely govern association strengths. Complexes with insulin, the only of the three peptides to adopt a well-defined three-dimensional (3D) structure, have a markedly higher stability in ESI-MS than the interactions between the two flexible peptides, C-peptide and IAPP. Instead, the latter peptides sample a variety of structures in the gas phase.[30] NMR and X-ray crystallographic investigations have shown that insulin binds IAPP in an α-helical conformation through a salt bridge between E13 in the insulin B-chain and R11 in IAPP, as well as hydrophobic stacking interactions of aromatic side chains.[9,12] Complex formation is therefore mediated by a 3D complementary binding interface. Flexible protein structures rapidly collapse into a tight gas-phase conformation following desolvation, while structures supported by hydrogen bonding and salt bridges can be preserved.[31] The high gas-phase stability of the insulin–IAPP heteromer is therefore indicative of a folded complex. Low pH, on the other hand, can abolish this interaction via e.g. pH-dependent unfolding of the B-chain and/or protonation of the glutamic acid residue at the interface. The abundance and stability of C-peptide–insulin complexes are comparable to that of the insulin–IAPP interaction. This can be indicative of a similar binding mechanism, since IAPP and C-peptide can transiently adopt defined secondary structures[12,16] which could be stabilized through interactions with insulin and in this manner form a defined binding site. Our results also show that binding of C-peptide to IAPP is not critically dependent on charged C-peptide residues and has low gas-phase stability. This is compatible with the notion that interactions between two flexible peptides do likely not involve a stable association via a defined binding site, but rather form transient contacts in solution at a high rate. This process may in part be mediated by charge attraction, as indicated by the observation that low pH interferes with their interaction, but cannot be pin-pointed to specific glutamic acid residues or C-peptide net charge, as revealed by the chargereduced C-peptide analogues studied here. A high association rate in solution could be favored if no specific binding site is required and can account for the observation of large heterodimer signals in ESI in spite of a low stability. In line with this, the low specificity and stability of the C-peptide–IAPP interactions can also serve to explain why high amounts of C-peptide are required to interfere with IAPP fibrillation.[14] It should also be noted that we were able to observe heterotrimeric complexes composed of two IAPP molecules and one insulin molecule, which dissociated into dimeric IAPP and insulin monomers in CID. This suggests that
dimeric IAPP can still interact with insulin and therefore provides evidence that IAPP and insulin may not only compete for the same binding site, but may also interact via different sites on the IAPP monomer. In fact, it has been suggested that IAPP can dimerize via its aggregation-prone C-terminal region,[30] leaving the helical N-terminus free to interact with insulin.[12,32] In conclusion, we demonstrate that ESI-MS can be used to distinguish peptide interactions. While the behavior of C-peptide–insulin complexes bears some resemblance to that of IAPP–insulin complexes formed via a defined interaction site, the interactions between IAPP and C-peptide appear to be non-specific. Our findings can be related to the situation in the secretory granule, where a high excess of C-peptide co-localizes with IAPP and insulin around the crystalline granule core. Strong interactions between IAPP and insulin may help to prevent IAPP aggregation,[11] while IAPP and C-peptide in turn have a monomerizing effect on insulin.[9,27] For C-peptide and IAPP, on the other hand, a similarly stable association would not serve a purpose, since IAPP aggregation can be effectively controlled by insulin, likely leaving their pairing redundant.
M. Landreh et al. [11] J. L. Larson, A. D. Miranker. The mechanism of insulin action on islet amyloid polypeptide fiber formation. J. Mol. Biol. 2004, 335, 221. [12] L. Wei, P. Jiang, Y. H. Yau, H. Summer, S. G. Shochat, Y. Mu, K. Pervushin. Residual structure in islet amyloid polypeptide mediates its interactions with soluble insulin. Biochemistry 2009, 48, 2368. [13] C. E. Munte, L. Vilela, H. R. Kalbitzer, R. C. Garratt. Solution structure of human proinsulin C-peptide. FEBS J. 2005, 272, 4284. [14] S. Janciauskiene, S. Eriksson, E. Carlemalm, B. Ahren. B cell granule peptides affect human islet amyloid polypeptide (IAPP) fibril formation in vitro. Biochem. Biophys. Res. Commun. 1997, 236, 580. [15] P. Westermark, Z. C. Li, G. T. Westermark, A. Leckstrom, D. F. Steiner. Effects of beta cell granule components on human islet amyloid polypeptide fibril formation. FEBS Lett. 1996, 379, 203. [16] J. Lind, E. Lindahl, A. Peralvarez-Marin, A. Holmlund, H. Jornvall, L. Maler. Structural features of proinsulin C-peptide oligomeric and amyloid states. FEBS J. 2010, 277, 3759. [17] H. Jornvall, E. Lindahl, J. Astorga-Wells, J. Lind, A. Holmlund, E. Melles, G. Alvelius, C. Nerelius, L. Maler, J. Johansson. Oligomerization and insulin interactions of proinsulin C-peptide: Threefold relationships to properties of insulin. Biochem. Biophys. Res. Commun. 2010, 391, 1561. [18] L. Nielsen, R. Khurana, A. Coats, S. Frokjaer, J. Brange, S. Vyas, V. N. Uversky, A. L. Fink. Effect of environmental factors on the kinetics of insulin fibril formation: elucidation of the molecular mechanism. Biochemistry 2001, 40, 6036. [19] Q. X. Hua, M. A. Weiss. Mechanism of insulin fibrillation: the structure of insulin under amyloidogenic conditions resembles a protein-folding intermediate. J. Biol. Chem. 2004, 279, 21449. [20] M. I. Ivanova, S. A. Sievers, M. R. Sawaya, J. S. Wall, D. Eisenberg. Molecular basis for insulin fibril assembly. Proc. Natl. Acad. Sci. USA 2009, 106, 18990. [21] J. Marcoux, C. V. Robinson. Twenty years of gas phase structural biology. Structure 2013, 21, 1541. [22] D. M. Williams, T. L. Pukala. Novel insights into protein misfolding diseases revealed by ion mobility-mass spectrometry. Mass Spectrom. Rev. 2013, 32, 169. [23] M. Landreh, J. Astorga-Wells, J. Johansson, T. Bergman, H. Jornvall. New developments in protein structure-function
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
analysis by MS and use of hydrogen-deuterium exchange microfluidics. FEBS J. 2011, 278, 3815. E. J. Nettleton, P. Tito, M. Sunde, M. Bouchard, C. M. Dobson, C. V. Robinson. Characterization of the oligomeric states of insulin in self-assembly and amyloid fibril formation by mass spectrometry. Biophys. J. 2000, 79, 1053. J. Shafqat, E. Melles, K. Sigmundsson, B. L. Johansson, K. Ekberg, G. Alvelius, M. Henriksson, J. Johansson, J. Wahren, H. Jornvall. Proinsulin C-peptide elicits disaggregation of insulin resulting in enhanced physiological insulin effects. Cell. Mol. Life Sci. 2006, 63, 1805. M. Landreh, J. B. Stukenborg, H. Willander, O. Soder, J. Johansson, H. Jornvall. Proinsulin C-peptide interferes with insulin fibril formation. Biochem. Biophys. Res. Commun. 2012, 418, 489. M. Landreh, G. Alvelius, H. Willander, J. B. Stukenborg, O. Soder, J. Johansson, H. Jornvall. Insulin solubility transitions by pH-dependent interactions with proinsulin C-peptide. FEBS J. 2012, 279, 4589. J. L. Larson, E. Ko, A. D. Miranker. Direct measurement of islet amyloid polypeptide fibrillogenesis by mass spectrometry. Protein Sci. 2000, 9, 427. J. M. Daniel, S. D. Friess, S. Rajagopalan, S. Wendt, R. Zenobi. Quantitative determination of noncovalent binding interactions using soft ionization mass spectrometry. Int. J. Mass Spectrom. 2002, 216, 1. N. F. Dupuis, C. Wu, J. E. Shea, M. T. Bowers. Human islet amyloid polypeptide monomers form ordered betahairpins: a possible direct amyloidogenic precursor. J. Am. Chem. Soc. 2009, 131, 18283. K. Breuker, F. W. McLafferty. Stepwise evolution of protein native structure with electrospray into the gas phase, 10 ( 12) to 10(2) s. Proc. Natl. Acad. Sci. USA 2008, 105, 18145. S. Gilead, H. Wolfenson, E. Gazit. Molecular mapping of the recognition interface between the islet amyloid polypeptide and insulin. Angew. Chem. Int. Ed. Engl. 2006, 45, 6476.
SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article at the publisher’ website.
184 wileyonlinelibrary.com/journal/rcm
Copyright © 2013 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
Revised: 13 October 2013
Accepted: 20 October 2013
Published online in Wiley Online Library
Rapid Commun. Mass Spectrom. 2014, 28, 178–184 (wileyonlinelibrary.com) DOI: 10.1002/rcm.6772
Insulin, islet amyloid polypeptide and C-peptide interactions evaluated by mass spectrometric analysis Michael Landreh1, Gunvor Alvelius1, Jan Johansson2,3 and Hans Jörnvall1* 1
Department of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77 Stockholm, Sweden KI-Alzheimer’s Disease Research Center, NVS Department, Karolinska Institutet, S-141 86 Stockholm, Sweden 3 Department of Anatomy, Physiology and Biochemistry, Swedish University of Agricultural Sciences, S-751 23 Uppsala, Sweden 2
RATIONALE: Insulin, islet amyloid polypeptide (IAPP), and the C-peptide part of proinsulin are co-secreted from the
pancreatic beta cell granules. IAPP aggregation can be inhibited by insulin and insulin aggregation by C-peptide, but different binding and disaggregating interactions may apply for the peptide complexes. A more detailed knowledge of these interactions is necessary for the development strategies against diabetic complications that stem from peptide aggregations. METHODS: Mass spectrometry (MS) is utilized to investigate pH-dependencies, sequence determinants and association strengths of interactions between pairs of all three peptides. Electrospray ionization (ESI)-MS was used to monitor complex formation and interaction stoichiometries at different pH values. Collision-induced dissociation (CID) was employed to probe relative association strengths and complex dissociation pathways. RESULTS: IAPP, like C-peptide, removes insulin oligomers observable by ESI-MS. Both C-peptide and IAPP form stable 1:1 heterodimers with insulin. Complexes of the negatively charged C-peptide with the positively charged IAPP, on the other hand, are easily dissociated. Replacement of the conserved glutamic acid residues in C-peptide with alanine residues increases the stability, indicating that net charge alone does not predict association strength. Binding to insulin has been suggested to stabilize a helical fold in IAPP via charge and hydrophobic interactions, which is in agreement with the now observed high gas-phase stability and sensitivity to low pH. CONCLUSIONS: Combined, these results suggest that the C-peptide–insulin and IAPP–insulin interactions are mediated by a defined binding site, while such a feature is not apparent in the IAPP–C-peptide association. Hence, IAPP and Cpeptide are interacting in similar manners and with similar monomerizing effects on insulin, suggesting that both peptides can prevent insulin aggregation. Simultaneous interactions of all three peptides cannot be excluded but appear unlikely from the uneven pairwise binding strengths. Copyright © 2013 John Wiley & Sons, Ltd.
Insulin secretion is crucial for the maintenance of glucose homeostasis, and impairment of this process is connected with diabetic disorders.[1] In the secretory granules of the β-cells, the proinsulin precursor is cleaved into mature insulin and C-peptide.[2] After cleavage, insulin is stored in crystalline form with Zn2+ and as a fluid phase with other granule components.[3] Insulin and islet amyloid polypeptide (IAPP), which co-localize with C-peptide in the fluid phase of the granule, have a high propensity to form amyloid fibrils. Amyloid fibrils composed of IAPP have been observed in the pancreas of >90% of type 2 diabetic patients and are thought to play a role in β-cell death,[4] which has sparked the development of synthetic peptide inhibitors of IAPP fibrillation.[5] Insulin amyloid can deposit locally at the site of repeated insulin injections[6] and may be of further importance in diabetes of both types.[7]
178
* Correspondence to: H. Jörnvall, Department of Medical Biochemistry and Biophysics, Karolinska Institutet, S-171 77 Stockholm, Sweden. E-mail: [email protected]
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
IAPP is mostly unstructured in solution, but becomes α-helical upon membrane binding,[8] which accelerates fibril formation by promoting intermolecular contacts. IAPP dimerization accelerates aggregation in a similar manner and may occur through α-helical or β-strand intermediates.[9,10] Insulin, on the other hand, can inhibit IAPP aggregation.[11] Binding to the B-chain of insulin stabilizes the α-helical IAPP structure, but simultaneously blocks its self-association sites and thus stabilizes the monomeric form.[9,12] C-peptide, which does also not adopt a well-defined secondary structure in solution,[13] was reported to reduce IAPP aggregation, but only when present at a >10-fold excess.[14] This effect has been suggested to be mediated by interactions between the negatively charged C-peptide and the positively charged IAPP at the granular pH.[15] Sequences and suggested interaction points are summarized in Fig. 1. In addition, C-peptide can also form aggregates with time at high concentrations.[16,17] High and low concentrations both promote insulin aggregation,[18] which is driven by pH- or temperature-induced unfolding, or by sampling unfolded conformations in the absence of stabilizing interactions.[19] In both cases, local unfolding releases aggregation ’hot spots’ from the helical regions of the A- and B-chain and initiates fibril assembly in this manner.[20] Mass spectrometry has emerged as a tool for the
Copyright © 2013 John Wiley & Sons, Ltd.
Interactions of insulin, IAPP and C-peptide in ESI-MS
Figure 1. Sequences of the peptides used in this study. Top: Residue contacts involved in the interaction between IAPP and the insulin B-chain are highlighted in gray. + and – indicate charged residues as well as the charges on the N- and C-termini. The average molecular masses are indicated behind each peptide. investigation of protein aggregation. Protein interactions can be preserved in the gas phase and yield information about complex stoichiometries and relative binding stabilities.[21] Additionally, ion mobility and hydrogen/deuterium exchange studies can provide insights into aggregate structure.[22,23] Using electrospray ionization mass spectrometry (ESI-MS), it is possible to detect prefibrillar insulin oligomers and thus monitor the early steps of aggregation.[24] We have reported that molecular interactions with C-peptide can remove these insulin oligomers[25] and interfere with insulin fibrillation.[26] The disaggregating effect is mediated by conserved glutamic acid residues in the negatively charged C-peptide, and molecular interactions with positive charges on insulin lead to co-precipitation of the two peptides via mutual neutralization at certain ratios and pH conditions.[27] The monomerizing actions can help to counterbalance the low insulin solubility at granular pH and trap soluble insulin in the fluid phase.[7] Whether the interactions between insulin and IAPP prevent not only IAPP but also insulin from fibrillation has not been reported, likely due to the pronounced aggregation propensity of IAPP. In summary, heteromolecular interactions that interfere with peptide aggregation have been observed for all three peptide pairs, and can be relevant in the secretory granule. However, only the binding of insulin to IAPP has been characterized in detail by X-ray crystallography and solution nuclear magnetic resonance (NMR).[9,12] In the present study, we use gas-phase interactions in ESI-MS to probe pHdependency, binding strength and sequence determinants to extract information about the molecular basis for each of these pairwise interactions.
EXPERIMENTAL
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
RESULTS IAPP and C-peptide remove insulin oligomers observed by ESI-MS The ESI mass spectra of IAPP show no significant change in charge state distribution between pH 4 and 8 (Supplementary Fig. S1, see Supporting Information). The peptide appears mostly as monomer, but with minor dimer and trimer fractions, which is in agreement with solution studies.[28] C-peptide and insulin form oligomers close to their isoelectric points at pH 3.4 and 5.5, respectively.[27] The presence of equimolar amounts of IAPP visibly reduced the amount of insulin oligomers (Fig. 2), indicating that IAPP has a monomerizing effect on insulin. This finding is similar to the disaggregating effects of C-peptide on insulin oligomers[25] and supports the suggestion that IAPP binding may help to solubilize multimeric insulin.[9]
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
179
Recombinant insulin (human) was purchased from Sigma (St. Louis, MO, USA). C-peptide and the C-peptide mutants with Glu → Ala substitutions at position 11 only, 27 only, or 3, 11 and 27 together were purchased from GenScript (Piscataway, NJ, USA). C-peptide was dissolved in deionized water, desalted with SepPak C18 cartridges (Waters, Milford, MA, USA), and stored in lyophilized form at 20 °C. Insulin was dissolved in H2O adjusted to pH 5 with 10% acetic acid (Sigma) to a concentration of 1 mM and lyophilized directly. Human recombinant IAPP was purchased from Bachem (Bubendorf, Switzerland). Since human IAPP is known to
aggregate within 5–10 min in aqueous solution, the entire vial was dissolved in 100% dimethyl sulfoxide (DMSO) to a final concentration of 2 mM and stored at 20 °C. Aliquots from the peptide stock solutions were diluted in 0.5% HCl (pH 1) or 20 mM ammonium acetate (pH 4, 5, 6, 7, and 8) to a final concentration of 50 μM each and immediately subjected to MS analysis. Data was acquired using a QTOF Ultima API mass spectrometer (Waters) equipped with a Z-spray source operated in the positive-ion mode. Samples were introduced via metal-plated borosilicate glass capillary needles (Proxeon, Denmark). The source temperature was 80 °C, the capillary voltage 1.6 kV and the cone and RF lens 1 potentials were 100 and 38 eV, respectively. The mass spectrometer was operated in single-reflector mode to achieve sufficient resolution at reasonable sensitivity (10 000, full width half maximum definition). The mass scale was calibrated in linear mode using horse myoglobin (Sigma). Scans were acquired at a rate of 1 scan per 2 s between 500 and 3000 m/z. Collision gas was argon at 5.2 × 10–5 mbar. Mass spectra were analyzed using the Waters MassLynx 4.1 software. To characterize relative complex stabilities, the [M + 4H]4+ ion as the most abundant for each complex was chosen for tandem mass spectrometric (MS/MS) analysis. Collision voltages were automatically adjusted from 10 to 60 eV in increments of 5 eV. Collision-induced dissociation (CID) spectra were recorded for 3 min (1 scan per 0.5 s) for each collision voltage increment and subsequently averaged. The complexes dissociated into doubly and singly charged peptide ions. The signal ratio between complexes and complex plus free peptides was calculated from the [M + 4H]4+ complex ions and the sum of the [M + 2H]2+ and the [M + 3H]3+ peptide ions. Three separate sets of CID data were recorded for each complex and ratios are shown as mean ± standard deviation from these replicates. The intensity of the intact 1:1 complex divided by the sum of intensities of the [M + 1H]1+ and [M + 2H]2+ species of IAPP and the intact 1:1 complex was used for interactions of C-peptide, C-peptide variants, or insulin with IAPP. For the insulin–C-peptide interaction, the complex/peptide ratio was calculated using the [M + 1H]1+ and [M + 2H]2+ species of C-peptide and the intact 1:1 complex.
M. Landreh et al.
Figure 2. IAPP has a monomerizing effect on insulin. (A) IAPP is mostly monomeric, with minor di-, tri- and tetrameric fractions. (B) At pH 5, insulin oligomers composed of 2–6 molecules can be observed in the high m/z range. (C) Addition of an equimolar amount of IAPP removes insulin oligomers, instead leading to complexes with 1–2 IAPP and 1–3 insulin molecules (D, magnification). IAPP, C-peptide and insulin interactions are observable by ESI-MS
180
At pH 5, ESI-MS signals corresponding to the [M + 4H]4+ and [M + 5H]5+ charge states of 1:1 complexes between insulin and IAPP were observed. Additionally, peaks corresponding to insulin–IAPP heteromers with 1:2, 2:1 and 2:2 stoichiometries could be detected (Fig. 2). The complex formation and stoichiometries are similar to the interactions between Cpeptide and insulin, where insulin oligomer removal is accompanied by the formation of heteromers composed of one insulin and one or two C-peptide molecules (Fig. 3).[25,27] Heteromer formation between IAPP and insulin could be observed between pH 4 and 8, but was found to yield the strongest MS signals at pH 5 (Supplementary Fig. S2(A), see Supporting Information). ESI-MS spectra of solutions containing C-peptide and IAPP did not deviate noticeably from the oligomeric status or charge state distribution of either peptide between pH 4 and 8 (Supplementary Fig. S2(B), see Supporting Information), but showed strong signals for 1:1 heteromers between IAPP and C-peptide, as well as minor signals corresponding to complexes with 1:2, 2:1, and 2:2 stoichiometries (Fig. 3(A), Supplementary Fig. S2(B)). While the C-peptide signals were markedly reduced at pH 4, the C-peptide–IAPP heteromers persisted with only minor differences in intensity, indicating that the C-peptide– IAPP complex is not affected by the decreased solubility of
wileyonlinelibrary.com/journal/rcm
C-peptide close to its isoelectric point. When the pH was decreased to 1, peaks corresponding to C-peptide–insulin, insulin–IAPP and IAPP–C-peptide heteromers were almost completely abolished (bottom panels in Supplementary Figs. S2(A) and S2(B), see Supporting Information). In the case of insulin, this may be attributed to unfolding at low pH, which can affect the proposed IAPP binding site in the helical section of the B-chain.[19] While the interaction site(s) for insulin and C-peptide have not been identified, previous studies have shown that protonatable residues are involved in the association,[27] which may also be the case for the C-peptide–IAPP interactions.[15] Interestingly, the simultaneous presence of all three peptides did not change this association pattern significantly. C-peptide and IAPP form stable complexes with insulin, but not with each other Next, we compared the pairwise interactions between insulin, IAPP and C-peptide. We subjected the [M + 4H]4+ heterodimers observed for each peptide pair to CID at increasing collision voltages and monitored the fraction of intact complexes. Figure 4(A) shows the dissociation curves at increasing collision voltages; representative CID spectra illustrating the degree of dissociation for different precursor complexes at a collision voltage of 40 eV are shown in Fig. 4(B). This approach yields information about the relative strengths of the peptide
Copyright © 2013 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
Interactions of insulin, IAPP and C-peptide in ESI-MS
Figure 3. C-peptide and its Glu → Ala variants form complexes with IAPP in a similar manner. Spectra are shown for (A) IAPP with wt C-peptide; (B) IAPP with C-peptide with the glutamic acid residues at positions 3, 11, and 27 replaced with alanine; (C) IAPP with C-peptide with the glutamic acid residue at position 27 replaced with alanine; and (D) IAPP with C-peptide with the glutamic acid residue at position 11 replaced with alanine. Heteromers composed of two IAPP and one C-peptide, and one IAPP and two C-peptide molecules can be detected for every C-peptide variant.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
increase in gas-phase stability (Fig. 4(A)), which suggests that the high number of negative charges in C-peptide is not required for interactions with IAPP.
CID preferentially releases insulin–IAPP and insulin–C-peptide heterodimers from heterotrimeric insulin–IAPP or insulin–C-peptide complexes In addition to heterodimers, we also observed heterotrimers consisting of one IAPP with two insulin molecules or two IAPP with one insulin molecule. We used gas-phase dissociation of protein complexes to obtain information on their molecular architecture by performing CID analysis of the [M + 6H]6+ and the [M + 7H]7+ charge states of the 2:1 and 1:2 IAPP–insulin complexes, respectively, which were the strongest signals for each species (Figs. 5(A) and 5(B)). Both complexes were found to dissociate primarily into IAPP–insulin heterodimers. 2:1 IAPP–insulin complexes dissociate into 1:1 heterodimers and IAPP monomers and a minor fraction of IAPP dimers. Similarly, 1:2 IAPP–insulin complexes preferentially dissociated into 1:1 heterodimers and insulin monomers. In addition, a second dissociation pathway resulting in dimeric insulin and monomeric IAPP could also be detected.
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
181
interactions in the absence of solvent. In the gas phase, hydrophobic interactions are weakened, while electrostatic interactions are strengthened.[29] The complex between insulin and C-peptide as well as the complex between insulin and IAPP showed high stability in CID, with fragmentation occurring beginning at a collision voltage of 40 eV (Fig. 4(A)). This is consistent with reports that charge-based interactions are important for the formation of both complexes.[9,12,27] In contrast, the IAPP–C-peptide complex was almost completely dissociated at 40 eV and is therefore significantly less stable than the insulin–IAPP and insulin–C-peptide complexes. Since IAPP has a net charge at pH 5 of 3+ and C-peptide one of 4–, it is likely that their association occurs primarily via charge attraction. Therefore, the low gas-phase stability of the heterodimer is unexpected. To investigate the role of charges in complex formation, we tested whether the reduction of C-peptide charge by Glu → Ala mutations at position 11 only, position 27 only, or positions 3, 11, and 27 combined affects the gas-phase stability of the complexes with IAPP. All C-peptide mutants formed heteromers with IAPP. Charge-state distributions, signal intensities and stoichiometries were comparable to complexes formed with IAPP and wild-type (wt) C-peptide (Fig. 3). CID revealed that Glu → Ala substitutions in C-peptide cause a moderate
M. Landreh et al.
Figure 4. Gas-phase stabilities of 1:1 peptide complexes at pH 5. (A) Insulin–IAPP and insulin–C-peptide complexes have high stability when subjected to CID at increasing collision voltages, while IAPP–C-peptide complexes are easily dissociated. Replacement of one or three glutamic acid residues in C-peptide causes a moderate increase in stability. (B) CID spectra recorded at 40 eV for IAPP–C-peptide (top panel), C-peptide–insulin (middle panel) and IAPP–insulin complexes (bottom panel).
Figure 5. Collision-induced dissociation (CID) of heterotrimers: (A) CID of 2:1; (B) 1:2 complexes of IAPP and insulin; and (C) 1:2 complexes of C-peptide and insulin at a collision voltage of 30 eV. The 2:1 and 1:2 heterotrimers of insulin and IAPP (A and B) preferentially release the 1:1 heterodimer of both peptides upon CID. Similarly, the 1:2 C-peptide-insulin complex (C) dissociates into the 1:1 heterodimer and a minor population of dimeric insulin. The different dissociation products are indicated by black and grey arrows. The complex selected for CID is indicated by *.
182
We have previously reported that insulin and C-peptide can form a 2:1 complex composed of an insulin dimer with a bound C-peptide, which results in mutual charge neutralization and precipitation at pH 5.[25,27] At pH 6, however, the [M + 7H]7+ charge state of this complex is readily detectable by ESI-MS. When subjected to CID, these
wileyonlinelibrary.com/journal/rcm
heterotrimers dissociate into 1:1 insulin–C-peptide heterodimers and insulin monomers and a smaller population of insulin homodimers and monomeric C-peptide when subjected to CID (Fig. 5(C)). This bears close resemblance to the dissociation routes of 1:2 IAPP–insulin heterotrimers, but can be observed already at a lower collision voltage.
Copyright © 2013 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
Interactions of insulin, IAPP and C-peptide in ESI-MS
DISCUSSION
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
REFERENCES [1] D. J. Michael, H. Cai, W. Xiong, J. Ouyang, R. H. Chow. Mechanisms of peptide hormone secretion. Trends Endocrinol. Metab. 2006, 17, 408. [2] H. W. Davidson, C. J. Rhodes, J. C. Hutton. Intraorganellar calcium and pH control proinsulin cleavage in the pancreatic beta cell via two distinct site-specific endopeptidases. Nature 1988, 333, 93. [3] J. Suckale, M. Solimena. The insulin secretory granule as a signaling hub. Trends Endocrinol. Metab. 2010, 21, 599. [4] P. Westermark, A. Andersson, G. T. Westermark. Islet amyloid polypeptide, islet amyloid, and diabetes mellitus. Physiol. Rev. 2011, 91, 795. [5] L. M. Yan, M. Tatarek-Nossol, A. Velkova, A. Kazantzis, A. Kapurniotu. Design of a mimic of nonamyloidogenic and bioactive human islet amyloid polypeptide (IAPP) as nanomolar affinity inhibitor of IAPP cytotoxic fibrillogenesis. Proc. Natl. Acad. Sci. USA 2006, 103, 2046. [6] F. E. Dische, C. Wernstedt, G. T. Westermark, P. Westermark, M. B. Pepys, J. A. Rennie, S. G. Gilbey, P. J. Watkins. Insulin as an amyloid-fibril protein at sites of repeated insulin injections in a diabetic patient. Diabetologia 1988, 31, 158. [7] M. Landreh, J. Johansson, H. Jornvall. C-peptide: a molecule balancing insulin states in secretion and diabetes-associated depository conditions. Horm. Metab. Res. 2013, 45, 769. [8] R. P. Nanga, J. R. Brender, J. Xu, G. Veglia, A. Ramamoorthy. Structures of rat and human islet amyloid polypeptide IAPP (1–19) in micelles by NMR spectroscopy. Biochemistry 2008, 47, 12689. [9] J. J. Wiltzius, S. A. Sievers, M. R. Sawaya, D. Eisenberg. Atomic structures of IAPP (amylin) fusions suggest a mechanism for fibrillation and the role of insulin in the process. Protein Sci. 2009, 18, 1521. [10] N. F. Dupuis, C. Wu, J. E. Shea, M. T. Bowers. The amyloid formation mechanism in human IAPP: dimers have betastrand monomer-monomer interfaces. J. Am. Chem. Soc. 2011, 133, 7240.
Copyright © 2013 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/rcm
183
In the present study, we have probed the peptide interactions involving insulin, IAPP and C-peptide using ESI-MS. All three peptides have previously been reported to interact in solution,[7] and, in all three cases, peptide charges have been suggested to be important for binding. However, reliable structure models are only available for the insulin–IAPP complex, which is held together by a mixture of charge and hydrophobic interactions.[9,12] While hydrophobic interactions are usually weakened in the gas phase, electrostatic interactions appear strengthened.[29] However, the different gas-phase stabilities of the three peptide pair interactions and the observation that a reduction in C-peptide charge stabilizes its interactions with IAPP show that net charge alone is a poor predictor of binding interactions in ESI-MS, as additional factors likely govern association strengths. Complexes with insulin, the only of the three peptides to adopt a well-defined three-dimensional (3D) structure, have a markedly higher stability in ESI-MS than the interactions between the two flexible peptides, C-peptide and IAPP. Instead, the latter peptides sample a variety of structures in the gas phase.[30] NMR and X-ray crystallographic investigations have shown that insulin binds IAPP in an α-helical conformation through a salt bridge between E13 in the insulin B-chain and R11 in IAPP, as well as hydrophobic stacking interactions of aromatic side chains.[9,12] Complex formation is therefore mediated by a 3D complementary binding interface. Flexible protein structures rapidly collapse into a tight gas-phase conformation following desolvation, while structures supported by hydrogen bonding and salt bridges can be preserved.[31] The high gas-phase stability of the insulin–IAPP heteromer is therefore indicative of a folded complex. Low pH, on the other hand, can abolish this interaction via e.g. pH-dependent unfolding of the B-chain and/or protonation of the glutamic acid residue at the interface. The abundance and stability of C-peptide–insulin complexes are comparable to that of the insulin–IAPP interaction. This can be indicative of a similar binding mechanism, since IAPP and C-peptide can transiently adopt defined secondary structures[12,16] which could be stabilized through interactions with insulin and in this manner form a defined binding site. Our results also show that binding of C-peptide to IAPP is not critically dependent on charged C-peptide residues and has low gas-phase stability. This is compatible with the notion that interactions between two flexible peptides do likely not involve a stable association via a defined binding site, but rather form transient contacts in solution at a high rate. This process may in part be mediated by charge attraction, as indicated by the observation that low pH interferes with their interaction, but cannot be pin-pointed to specific glutamic acid residues or C-peptide net charge, as revealed by the chargereduced C-peptide analogues studied here. A high association rate in solution could be favored if no specific binding site is required and can account for the observation of large heterodimer signals in ESI in spite of a low stability. In line with this, the low specificity and stability of the C-peptide–IAPP interactions can also serve to explain why high amounts of C-peptide are required to interfere with IAPP fibrillation.[14] It should also be noted that we were able to observe heterotrimeric complexes composed of two IAPP molecules and one insulin molecule, which dissociated into dimeric IAPP and insulin monomers in CID. This suggests that
dimeric IAPP can still interact with insulin and therefore provides evidence that IAPP and insulin may not only compete for the same binding site, but may also interact via different sites on the IAPP monomer. In fact, it has been suggested that IAPP can dimerize via its aggregation-prone C-terminal region,[30] leaving the helical N-terminus free to interact with insulin.[12,32] In conclusion, we demonstrate that ESI-MS can be used to distinguish peptide interactions. While the behavior of C-peptide–insulin complexes bears some resemblance to that of IAPP–insulin complexes formed via a defined interaction site, the interactions between IAPP and C-peptide appear to be non-specific. Our findings can be related to the situation in the secretory granule, where a high excess of C-peptide co-localizes with IAPP and insulin around the crystalline granule core. Strong interactions between IAPP and insulin may help to prevent IAPP aggregation,[11] while IAPP and C-peptide in turn have a monomerizing effect on insulin.[9,27] For C-peptide and IAPP, on the other hand, a similarly stable association would not serve a purpose, since IAPP aggregation can be effectively controlled by insulin, likely leaving their pairing redundant.
M. Landreh et al. [11] J. L. Larson, A. D. Miranker. The mechanism of insulin action on islet amyloid polypeptide fiber formation. J. Mol. Biol. 2004, 335, 221. [12] L. Wei, P. Jiang, Y. H. Yau, H. Summer, S. G. Shochat, Y. Mu, K. Pervushin. Residual structure in islet amyloid polypeptide mediates its interactions with soluble insulin. Biochemistry 2009, 48, 2368. [13] C. E. Munte, L. Vilela, H. R. Kalbitzer, R. C. Garratt. Solution structure of human proinsulin C-peptide. FEBS J. 2005, 272, 4284. [14] S. Janciauskiene, S. Eriksson, E. Carlemalm, B. Ahren. B cell granule peptides affect human islet amyloid polypeptide (IAPP) fibril formation in vitro. Biochem. Biophys. Res. Commun. 1997, 236, 580. [15] P. Westermark, Z. C. Li, G. T. Westermark, A. Leckstrom, D. F. Steiner. Effects of beta cell granule components on human islet amyloid polypeptide fibril formation. FEBS Lett. 1996, 379, 203. [16] J. Lind, E. Lindahl, A. Peralvarez-Marin, A. Holmlund, H. Jornvall, L. Maler. Structural features of proinsulin C-peptide oligomeric and amyloid states. FEBS J. 2010, 277, 3759. [17] H. Jornvall, E. Lindahl, J. Astorga-Wells, J. Lind, A. Holmlund, E. Melles, G. Alvelius, C. Nerelius, L. Maler, J. Johansson. Oligomerization and insulin interactions of proinsulin C-peptide: Threefold relationships to properties of insulin. Biochem. Biophys. Res. Commun. 2010, 391, 1561. [18] L. Nielsen, R. Khurana, A. Coats, S. Frokjaer, J. Brange, S. Vyas, V. N. Uversky, A. L. Fink. Effect of environmental factors on the kinetics of insulin fibril formation: elucidation of the molecular mechanism. Biochemistry 2001, 40, 6036. [19] Q. X. Hua, M. A. Weiss. Mechanism of insulin fibrillation: the structure of insulin under amyloidogenic conditions resembles a protein-folding intermediate. J. Biol. Chem. 2004, 279, 21449. [20] M. I. Ivanova, S. A. Sievers, M. R. Sawaya, J. S. Wall, D. Eisenberg. Molecular basis for insulin fibril assembly. Proc. Natl. Acad. Sci. USA 2009, 106, 18990. [21] J. Marcoux, C. V. Robinson. Twenty years of gas phase structural biology. Structure 2013, 21, 1541. [22] D. M. Williams, T. L. Pukala. Novel insights into protein misfolding diseases revealed by ion mobility-mass spectrometry. Mass Spectrom. Rev. 2013, 32, 169. [23] M. Landreh, J. Astorga-Wells, J. Johansson, T. Bergman, H. Jornvall. New developments in protein structure-function
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
analysis by MS and use of hydrogen-deuterium exchange microfluidics. FEBS J. 2011, 278, 3815. E. J. Nettleton, P. Tito, M. Sunde, M. Bouchard, C. M. Dobson, C. V. Robinson. Characterization of the oligomeric states of insulin in self-assembly and amyloid fibril formation by mass spectrometry. Biophys. J. 2000, 79, 1053. J. Shafqat, E. Melles, K. Sigmundsson, B. L. Johansson, K. Ekberg, G. Alvelius, M. Henriksson, J. Johansson, J. Wahren, H. Jornvall. Proinsulin C-peptide elicits disaggregation of insulin resulting in enhanced physiological insulin effects. Cell. Mol. Life Sci. 2006, 63, 1805. M. Landreh, J. B. Stukenborg, H. Willander, O. Soder, J. Johansson, H. Jornvall. Proinsulin C-peptide interferes with insulin fibril formation. Biochem. Biophys. Res. Commun. 2012, 418, 489. M. Landreh, G. Alvelius, H. Willander, J. B. Stukenborg, O. Soder, J. Johansson, H. Jornvall. Insulin solubility transitions by pH-dependent interactions with proinsulin C-peptide. FEBS J. 2012, 279, 4589. J. L. Larson, E. Ko, A. D. Miranker. Direct measurement of islet amyloid polypeptide fibrillogenesis by mass spectrometry. Protein Sci. 2000, 9, 427. J. M. Daniel, S. D. Friess, S. Rajagopalan, S. Wendt, R. Zenobi. Quantitative determination of noncovalent binding interactions using soft ionization mass spectrometry. Int. J. Mass Spectrom. 2002, 216, 1. N. F. Dupuis, C. Wu, J. E. Shea, M. T. Bowers. Human islet amyloid polypeptide monomers form ordered betahairpins: a possible direct amyloidogenic precursor. J. Am. Chem. Soc. 2009, 131, 18283. K. Breuker, F. W. McLafferty. Stepwise evolution of protein native structure with electrospray into the gas phase, 10 ( 12) to 10(2) s. Proc. Natl. Acad. Sci. USA 2008, 105, 18145. S. Gilead, H. Wolfenson, E. Gazit. Molecular mapping of the recognition interface between the islet amyloid polypeptide and insulin. Angew. Chem. Int. Ed. Engl. 2006, 45, 6476.
SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article at the publisher’ website.
184 wileyonlinelibrary.com/journal/rcm
Copyright © 2013 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2014, 28, 178–184
