Transient Crosslinking Mass Spectrometry: Taking Conformational Snapshots of Proteins

doi:10.21203/rs.3.rs-1976601/v1

Download PDF

Article

Transient Crosslinking Mass Spectrometry: Taking Conformational Snapshots of Proteins

https://doi.org/10.21203/rs.3.rs-1976601/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Protein structural analysis at the very moment of target binding or sensing incoming stimuli sheds light on how protein functions diversely with time or pathological conditions. To understand it, we need to intercept and see the intermediate conformation. Although conventional methods offer high resolution structural analysis, they do not address puzzling dynamic conformational changes. Herein, we developed a transient crosslinking mass spectrometry involving a novel photoreactive crosslinker that can capture intermediate conformers. The designed non-specific reactivity increased the crosslinking site diversity, thereby enhancing the resolution and broadening the scope of mass spectrometric-based structural analysis. A time-resolved crosslinking strategy was developed to take conformational snapshots for calmodulin, an important calcium sensor, and revealed the structural basis of its dynamic conformational response to calcium binding and target interaction. Therefore, the designed transient crosslinking makes short-lived conformers visible, which has the potential to tackle the question how variations in protein’s conformation change functions.

Chemical crosslinking

Conformational change

Intermediate conformer

Mass spectrometry

Variations in protein’s structure or interaction with other molecules regulate its biological functions. Besides the structural analysis in static states, there is in constant pursuit of more understanding of protein’s dynamic changes. The discovery of novel intermediate conformers is also an essential source for allosteric drug design. Methods including nuclear magnetic resonance (NMR) spectroscopy, X-ray crystallography and cryo-electron microscopy (cryo-EM)^1–3, have been used for protein structural analysis because of their high resolution. However, inherent limitations, such as requisite sample amounts and maximum observable protein sizes, restrict the scope of their applications^{4, 5}. Mass spectrometry (MS)-based techniques are attractive as alternative methods for for protein structural analysis because of their high sensitivity and specificity^{6, 7}. Native MS or affinity purification MS allow direct detection of protein complexes in their native states^{8, 9}. The combination with hydrogen-deuterium exchange (H/D exchange) or hydroxyl radical footprinting allows for protein structural mapping at the residue level^10–13.

Chemical crosslinking-MS has emerged as a complementary method for probing protein 3-dimentional (3D) structures and networks^14–19. A crosslinker of defined spacer arm is used to covalently link two spatially adjacent amino acid residues. The obtained crosslinks reveal variations in inter-residue distances at protein, key sites involved in interactions, and changes in solvent accessibility. Because most currently used crosslinkers target defined amino acids²⁰, the resolution is limited by the number of suitable sites within the crosslinking range. Typical crosslink reaction also takes time, which provide homogeneous protein conformations rather than transient ones. Hetero-bifunctional crosslinkers that combine a residue-specific reactive group (e.g., N-hydroxysuccinimide, NHS) with a photoreactive group (e.g., diazirine or aryl azide) expand the range of residues available for crosslinking and partially allow for transient crosslinking. However, different crosslinking moieties usually require separate initiation procedures²¹. Therefore, development of new crosslinkers for tracking dynamic conformational changes or transient protein-protein interactions (PPIs) remains a major challenge in the field.

In this study, we developed a transient crosslinking method involving a novel photoreactive crosslinker, which contains diazirine groups at both ends and can be fully activated by laser irradiation. There are therefore two non-specifically conjugate amino acids are expected within the spacer arm’s range. Additionally, spatial distances and solvent accessibility information were obtained by considering crosslinks and looplinks, which simplified the data interpretation process. MS/MS analysis of the crosslinked products yielded an identical crosslinker-specific reporter ion, which improved the identification confidence of crosslinks from complex sample. Calmodulin (CaM) is an important calcium (Ca²⁺) sensor that conformationally regulates various downstream target proteins in Ca²⁺-dependent manner^22–24. A time-resolved crosslinking strategy was developed to take conformational snapshots of CaM at different time points during Ca²⁺ binding and target peptide interaction process. The structural dynamics of CaM were modelled by molecular dynamics (MD) simulations with the guidance of analysing crosslinks from the captured intermediate conformers. Therefore, the designed crosslinker showed potential in protein conformational change analysis, PPI characterization and conformational conversion process monitoring.

Development of transient crosslinking and method validation

To overcome the constraints of reaction specificity and slow kinetics, a fully photoreactive diazirine crosslinker was developed, named as DCD (structure shown in Figure 1A). Theoretically, it can be activated by laser irradiation and its rapid reaction kinetic is expected to enable transient crosslinking with no specificity. The method feasibility was investigated by crosslinking of model peptide amyloid-related peptide SG (from human islet polypeptide) and model protein lysozyme. Upon one-second laser activation, a newly generated ion was observed at m/z 844.5 (Figure S2), which corresponds to the DCD crosslinked SG peptide (through the loss of two diazirine groups). For lysozyme, two sets of protein distributions were detected (~1 s activation, Figure S3A) and the deconvolution spectrum also suggested transient crosslinking was possible (Figure S3B). Liquid-nitrogen freezing enabled to extend the laser activation time while retain transient conformers and ensure in situ crosslinking. The efficiency was not affected by the sample volume within the tested range (Figure S3C) and 100 μL was selected for subsequent analysis. Total 60 s laser activation time was set by considering the balance between sample frozen and crosslinking efficiency. The effect of solution pH was also investigated (Figure S4A) and pH 5 was optimal, which demonstrated the potential biocompatibility of DCD. DMSO was necessary to solubilize DCD and its amount was evaluated to be controlled within 5% (Figure S4A) to avoid unexpected protein denaturation. A 1:20 (protein:crosslinker) molar ratio was selected based on the crosslinking of three model proteins, including bovine serum albumin (BSA), CaM and myoglobin (Figure S4B).

Importantly, the investigation of crosslinked sites verified the expected non-specific reactivity by crosslinking all amino acid residues, except cysteine (C) and tryptophan (W). The lack of cysteine residue reactivity (which was only observed in BSA) is due to the disulfide bond formation, while the lack of reactivity at tryptophan residues may be attributed to its low number in the sequences (three for CaM and two for myoglobin) combined with the hydrophobicity and aromatic structure of the sidechain. Interestingly, carboxylic acids, including aspartic acid (D) and glutamic acid (E), accounted for a considerable proportion of the crosslinked sites (Figure S4C), which was possibly due to the reactivity bias of generated diazo group towards the carboxyl group²⁵.

Crosslinking of protein and protein complex

CaM, which produces a significant conformational change upon Ca²⁺ binding, was selected as a model system and monitored by designed transient crosslinking (Figure 1B). In the Ca²⁺-free state, residues within the central linker of CaM (D79, D81, S82 and E85) formed abundant inter-crosslinks with the N- and C-terminal domains (NTDs and CTDs), indicating their spatial proximity and a “closed” conformation. In the Ca²⁺-bound state, the central linker formed few crosslinks with NTD and none crosslinks with CTD. The observation of increased looplinks indicated an increased distance between these parts of the protein and the formation of an “open” conformation (Figure 2A and Table S1). Notably, looplinks were formed by crosslinking amino acids on the same chain at one- or two-residue intervals. This reactivity pattern, not typically seen with amino acid-specific crosslinkers, is probably due to the rapid reaction kinetics and non-specific reactivity of DCD, which allows for facile crosslinking of nearby amino acids. It was realized that looplink modification sites overlapped with those of the monolinks (Figure S5), implying that such modification pattern revealed similar information of solvent accessibility. Moreover, upon MS/MS analysis, every crosslink product yielded an identical fragment ion at m/z 281.22 (highlighted in Figure 2B), corresponding to protonated DCD losing two diazirine groups, which was regarded as an indicator for crosslink product confirmation.

Next, DCD crosslinking was applied for monitoring interactions between Ca²⁺-CaM and target peptides, including CaMKII peptide and melittin. For the Ca²⁺-CaM–CaMKII peptide interaction, residues L1, K11, and G12 of the peptide formed numerous inter-molecular crosslinks with CaM covering the NTD, CTD and the central linker (Figure 2C, representative mass spectrum shown in Figure 2B, details in Table S2). Additionally, intra-molecular crosslink E12-D79 was formed within CaM, indicating that the NTD was located near the central linker. Resulted crosslink products suggested a compact conformation of CaM and its tightly interaction with CaMKII peptide, which was in line with the structure obtained from X-ray crystallography (Figure 2C) that CaM displayed a spherical conformation to wrap the CaMKII peptide. Remarkably, the maximum crosslinking range of the DCD was calculated as 31.8 Å (C_α–C_α distance, E was selected as a representative amino acid) and SASD between residues of most detected crosslinks were within this crosslinkable range based on the PDB structures of the apo-CaM, Ca²⁺-CaM and CaM-CaMKII peptide (Figure 2D), indicating the reliability of structural information provided by DCD crosslinking. In the case of CaM–melittin complex, melittin (from residues G1 to L6) were crosslinked with multiple sites of CaM, located within the NTD, CTD and central linker (Figure 2C and Table S2). MD analysis gave a possible structure that allowed SASD of most crosslinks fall within the limited crosslinking range (Table S3). Different with CaMKII peptide interaction, CaM presented a relative stretched conformation. Melittin fit the space formed by the CTD of CaM with its N-terminal pointed to the direction of the central linker and the NTD (Figure 2C). Notably, there still observed over-length crosslinks having theoretical SASDs outside the crosslinking range, which was probably attributed to the highly dynamic of CaM and the existence of different conformation in solution.

Providing Complementary Structural Information with Multiple Crosslinkers

Owing to crosslinking site diversity, combining DCD and conventional crosslinkers was expected to improve crosslinking coverage and provide complementary structural information. To explore this, we selected DSP (a homo-bifunctional NHS crosslinker) and sulfo-SDAD (a hetero-bifunctional crosslinker) for investigation (chemical structures are shown in Figure 3). When CaM was crosslinked by DSP, residue K95 emerged as a key position in most crosslinks (Table S4). However, limited structural information was revealed because DSP only formed crosslinks between lysine residues. The crosslinker sulfo-SDAD is expected to crosslink more sites by reacting with lysine and any other nearby amino acids. In addition to K95, the sulfo-SDAD crosslinking also centred on residues K14 and K22 (Table S4). A greater number of looplinks (bridging adjacent amino acids) were also observed, which may be attributed to the rapid reaction kinetics of carbene group. DCD crosslinking gave overlapped site or sequence information with that obtained from the sulfo-SDAD crosslinking, such as the crosslinks D79–Y100 and E12–D79 (Figure 3A, Table S1 and S4). Moreover, two diazirine groups of DCD completely liberated the reaction selectivity, thereby creating more sites that were unreachable with DSP and sulfo-SDAD. For crosslinking of CaM-melittin complex, no inter-molecular crosslinks were detected upon DSP crosslinking. In comparison with DCD results, sulfo-SDAD crosslinking yielded less inter-molecular crosslinks as one side of crosslinking was pinpointed at lysine residues and the structural information was insufficient, especially not fully covered the CTD of CaM (Figure 4B, Table S2 and S5). For CaM–CaMKII peptide complex, DSP crosslinking yielded only two inter-molecular crosslinks, while none was detected for sulfo-SDAD crosslinking. It was probably due to unsuitable inter-residue distances or a lack of reactive sites within the crosslinking range. In contrast, DCD crosslinking yielded numerous inter-molecular crosslinks, most of which were complementary to that of DSP and sulfo-SDAD crosslinking (Figure 4C, Table S2 and S5).

Time-resolved crosslinking strategy

Structure of CaM in Ca²⁺-free and Ca²⁺-saturated states has been well-resolved; however, the conformational change process between them remains untraceable. Therefore, a time-resolved crosslinking strategy was developed by initiating DCD crosslinking at different time points along Ca²⁺binding (Figure 4A). In the Ca²⁺-free state, crosslinks indicated a compact conformation with the NTD and CTD folded towards the central linker, such as E12–D79, F13–D79 and D79–Y100 (Figure S6A, and Table S1). The over-length crosslink D79–E128, with an SASD of 34.50 Å (based on PDB: 1CFD, Table S6), suggested that the CTD is dynamic in solution to produce a different inter-residue distance. MD simulations produced a compatible conformation with a measured SASD_MD of 20.81 Å, which indicates that the distance between the CTD and the central linker was closer than that obtained from X-ray crystallography (0 min in Figure 4B, and Table S6). By crosslinking at the moment of Ca²⁺ adding, the same crosslinks between the NTD and the central linker, such as E12–D79 and F13–D79, were detected, indicating that the NTD and the central linker remained in proximity (Figure S6B and Table S7). However, the D79–Y100 and D79–E128 crosslinks were not detected, which were expected and should not varied greatly according to SASD measurement based on the X-ray structures of Ca²⁺-free and Ca²⁺- bound CaM (Table S6 and S8). This suggests that Ca²⁺ binding induces a transient conformer, in which the CTD moves away from the central linker, creating inter-residue distances outside the DCD crosslinking range. MD simulations gave a transient conformer with measured SASD_MD of 37.45 Å and 41.00 Å for crosslinks D79–Y100 and D79–E128, respectively (transient in Figure 4B and Table S8). At the same time, the central linker was observed to converted to α-helix immediately. At the 1-min time point, additional over-length crosslinks (K14–D79 and K14–D81) were detected, highlighting the NTD dynamics (Figure S6C). MD simulations showed the changes occurred at α-helix from residues D2 to F19 positioned with measured SASD_MD of 24.35 Å and 26.43 Å for crosslinks K14–D79 and K14–D81, respectively (1 min in Figure 4B). Crosslinking products obtained after 2-min of Ca²⁺ binding was similar and SASD fit the structure of Ca²⁺-CaM, which implied that the conformation tended to be stable, varied slightly but no significant changes (Figure S6D-H and Table S7).

Time-resolved monitoring of protein complex formation

The conformational change of CaM during target peptide/protein interaction is of great importance for its biological functions. Time-resolved crosslinking was used to monitor the CaM–melittin complex formation. At the very moment of Ca²⁺addition, CaM adopted a similar transient conformation to that seen in Figure 4B at transient time point, wherein the CTD departed from the central linker and the NTD remained crosslinked to the central linker (crosslinks of E12-D79 and E12-D81). Meanwhile, both NTD and CTD of CaM formed inter-molecular crosslinks with several amino acid residues of melittin, including G1, G3, A4 and V5 (Figure S7A and Table S9). After 2 min, intra-molecular crosslinks were similar, while the spatial distance between melittin and CaM was supposed to decrease as indicated by increased inter-molecular crosslinks and more crosslinking sites within melittin and CaM (Figure S7B and Table S9). At the 5-min time point, a greater number of inter-molecular crosslinks were observed with similar crosslinking sites. Additional intra-crosslinks within CTD (Y100-E141) as well as between CTD and the central linker (Y99-E140) were detected (Figure S7C and Table S9), implying the conformational change occurred within these regions. Upon 10 min of Ca²⁺ mixing and after, melittin was supposed to interact with CaM tightly and stably as no significant change was observed for either crosslinking sites or number of inter-molecular crosslinks (Table S10 and S11). For CaM, intra-molecular crosslinks between CTD and NTD of CaM appeared at 10-min time point and were continuously detected afterwards (Figure S7D-G, Table S10 and S11), which is therefore proposed that the CTD continuously approached the NTD and tended to form a compact conformation. Based on MD simulation, it seemed impossible to find one single conformation at every time point that satisfied the SASD of all crosslink products (Table S12-S18). Thus, we proposed that the obtained crosslink products came from a mixture of different conformations. Figure 5 shows the most reasonable structures at each time point. It revealed that melittin interacted with CTD of CaM by passing through the cavity. From transient to 5-min time point, CaM adopted two major conformations that mainly differed by the α-helical content in the central linker. After 10 min, more intermediate conformations were proposed corresponding to each time point and most of CaM had an uncoiled central linker with less α-helical content. Such structural property caused the conformational variation of NTD and made it close to CTD. Additionally, the distribution of representative structures (insert in Figure 5) showed a consistency between those selected intermediate conformations with the timeline of MD simulation, which confirmed that the developed time-resolved crosslinking enabled to reveal the conformational change process. Notably that some of conformations appeared in unexpected time region, which was probably attributed to the highly dynamic of protein conformation in solution and more crosslinking information would be beneficial for modelling, if possible.

Structures of proteins and their interactions with other molecules determine their biological functions. Attempts are constantly being made to understand the dynamic changes of proteins. Chemical crosslinking mass spectrometry emerges as one of attractive methods for probing protein 3D structures. The generation of different types of crosslink products is informative for structural analysis by offering variations in spatial distance, solvent accessibility, and interaction interface of protein complex. Nevertheless, most currently used crosslinkers are designed to target specific amino acids residues, which has high reaction selectivity but sacrifices resolution of the method. The development of crosslinkers with different length of spacer arm or crosslinking active moieties can make up for the limitation. Moreover, the time-consuming reaction would miss possible intermediate conformations or homogenize the conformational difference generated by intermediate conformers. Therefore, the development of crosslinkers with rapid reaction kinetic and enlarged crosslinking range is still urgently required.

The crosslinker, DCD, designed in this study enabled to complete crosslinking at the very moment of laser activation. The application of liquid nitrogen assisted to preserve short-lived intermediate conformers while enhance the crosslinking efficiency. The linkage of crosslinker would reduce fragmentation efficiency and confuse sequencing by generating uncommon fragmentation patterns. For DCD crosslinking, resulted products yielded a crosslinker-specific reporter ion in MS/MS spectra through the cleavage of DCD, which greatly improved the annotation confidence and was beneficial for fragmentation. On this basis, crosslinking of model protein and protein complex confirmed the reliability of the method by showing that the SASD of most generated crosslink products (based on crystal structure) was within the crosslinking range of DCD. Additionally, over-length crosslinks also deserved attention as they were attributed to protein dynamics in solution and provided valuable clues for modelling unexpected conformational intermediates. DCD’s crosslinking site diversity revealed complementary structural information, and its combination with other crosslinkers with different spacer arms or reactive moieties significantly improved the analysis resolution.

CaM regulates various downstream target proteins through sensing a wide range of intracellular Ca²⁺ concentration and its conformational variation plays an important role in such precise regulation^{22, 26–28}. There are many well-resolved structures in different states, such as Ca²⁺-free, Ca²⁺-bound and complexed with target peptide/protein^{23, 29}. Nevertheless, it is still curious about the dynamic process and how the conformation is at the very moment when Ca²⁺ binding or target interactions. Current analysis greatly relies on the theoretical modelling based on MD simulations, while lacks of direct experimental evidence. Taking the advantages of transient crosslinking, we came up with a time-resolved crosslinking strategy, which enabled to take conformational snapshots at serious time points during conformational change process. The obtained crosslink products provided spatial distance restriction, which guided MD simulations to find most reasonable intermediate conformers. The connection of corresponding conformers revealed that the CTD had the most rapid kinetic and responded to Ca²⁺ binding within second-time scale. Meanwhile, the central linker converted to α-helix structure and followed up with the NTD conversion. The overall structure became relatively stable in minutes. In the case of interaction with melittin, besides the conformational change occurred at the CTD and NTD of CaM, the central linker acted as a flexible linker whose α-helical content varied greatly. Initially, intermediate conformers carrying an uncoiled and α-helical central linker was comparable, while the number of intermediates with an uncoiled structure gradually increased, indicating that melittin interaction induced the uncoil of central linker and such structure was more conductive to increase the structural flexibility and stabilize melittin interaction.

In summary, a fully photoreactive diazirine crosslinker, DCD, was developed. Its non-specific reactivity and crosslinking site diversity were beneficial for structural analysis, and its combination with other crosslinkers greatly improved the resolution of chemical crosslinking. A time-resolved crosslinking strategy was developed to take conformational snapshots and the resulted crosslinks guided the structural modelling to visualize the conformational change process. Therefore, the developed transient crosslinking MS enabled to see the intermediate conformers and has the potential to tackle the question how variations in protein’s conformation control its function.

Acknowledgement

This work was supported by the National Natural Science Foundation of China (81720108032, 81930109), the National Key Research and Development Programme of China (2021YFA1301300), the Project for Major New Drug Innovation and Development (2018ZX09711001-002-003), and the Project of State Key Laboratory of Natural Medicines at China Pharmaceutical University (SKLNMZZ202020).

1. Wlodawer, A., Minor, W., Dauter, Z. & Jaskolski, M. Protein crystallography for aspiring crystallographers or how to avoid pitfalls and traps in macromolecular structure determination. FEBS J. 280, 5705-5736 (2013).

2. Bai, X.-c., McMullan, G. & Scheres, S.H.W. How cryo-EM is revolutionizing structural biology. Trends Biochem. Sci. 40, 49-57 (2015).

3. Danev, R., Yanagisawa, H. & Kikkawa, M. Cryo-Electron Microscopy Methodology: Current Aspects and Future Directions. Trends Biochem. Sci. 44, 837-848 (2019).

4. Sekhar, A. & Kay, L.E. NMR paves the way for atomic level descriptions of sparsely populated, transiently formed biomolecular conformers. Proc. Natl. Acad. Sci. U. S. A. 110, 12867-12874 (2013).

5. Ishima, R. & Torchia, D.A. Protein dynamics from NMR. Nat. Struct. Biol. 7, 740-743 (2000).

6. Donnelly, D.P. et al. Best practices and benchmarks for intact protein analysis for top-down mass spectrometry. Nat. Methods 16, 587-594 (2019).

7. Zhou, M. et al. Higher-order structural characterisation of native proteins and complexes by top-down mass spectrometry. Chem. Sci. 11, 12918-12936 (2020).

8. Mitra, G. Application of native mass spectrometry in studying intrinsically disordered proteins: A special focus on neurodegenerative diseases. BBA-PROTEINS PROTEOM 1867, 140260 (2019).

9. Tamara, S., den Boer, M.A. & Heck, A.J.R. High-Resolution Native Mass Spectrometry. Chem. Rev. 122, 7269-7326 (2022).

10. Liu, X.R., Rempel, D.L. & Gross, M.L. Protein higher-order-structure determination by fast photochemical oxidation of proteins and mass spectrometry analysis. NAT PROTOC 15, 3942-3970 (2020).

11. Esben, T., E., N.Z. & D., R.K. Conformational Analysis of Complex Protein States by Hydrogen/Deuterium Exchange Mass Spectrometry (HDX-MS): Challenges and Emerging Solutions. TRAC-Trend Anal Chem 106, 125-138 (2018).

12. Zhang, M.M., Rempel, D.L. & Gross, M.L. A Fast Photochemical Oxidation of Proteins (FPOP) platform for free-radical reactions: the carbonate radical anion with peptides and proteins. Free Radocal Bio Med 131, 126-132 (2019).

13. Liu, X.R., Zhang, M.M. & Gross, M.L. Mass Spectrometry-Based Protein Footprinting for Higher-Order Structure Analysis: Fundamentals and Applications. Chem. Rev. 120, 4355-4454 (2020).

14. Iacobucci, C., Götze, M. & Sinz, A. Cross-linking/mass spectrometry to get a closer view on protein interaction networks. Curr Opin Biotech 63, 48-53 (2020).

15. Wang, J.-H. et al. Characterization of protein unfolding by fast cross-linking mass spectrometry using di-ortho-phthalaldehyde cross-linkers. Nat. Commun. 13, 1468 (2022).

16. Schneider, M., Belsom, A. & Rappsilber, J. Protein Tertiary Structure by Crosslinking/Mass Spectrometry. Trends Biochem. Sci. 43, 157-169 (2018).

17. Ding, Y.-H. et al. Modeling Protein Excited-state Structures from "Over-length" Chemical Cross-links. J. Biol. Chem. 292, 1187-1196 (2017).

18. Zheng, Q., Zhang, H., Tong, L., Wu, S. & Chen, H. Cross-Linking Electrochemical Mass Spectrometry for Probing Protein Three-Dimensional Structures. Anal. Chem. 86, 8983-8991 (2014).

19. Zheng, Q., Zhang, H., Wu, S. & Chen, H. Probing Protein 3D Structures and Conformational Changes Using Electrochemistry-Assisted Isotope Labeling Cross-Linking Mass Spectrometry. J. Am. Soc. Mass Spectrom. 27, 864-875 (2016).

20. Yu, C. & Huang, L. Cross-Linking Mass Spectrometry: An Emerging Technology for Interactomics and Structural Biology. Anal. Chem. 90, 144-165 (2018).

21. Gomes, A.F. & Gozzo, F.C. Chemical cross-linking with a diazirine photoactivatable cross-linker investigated by MALDI- and ESI-MS/MS. J. Mass Spectrom. 45, 892-899 (2010).

22. Sharma, R.K. & Parameswaran, S. Calmodulin-binding proteins: A journey of 40 years. Cell Calcium 75, 89-100 (2018).

23. Gong, D. et al. Modulation of cardiac ryanodine receptor 2 by calmodulin. Nature 572, 347-351 (2019).

24. Tadross, M.R., Dick, I.E. & Yue, D.T. Mechanism of Local and Global Ca2+ Sensing by Calmodulin in Complex with a Ca2+ Channel. Cell 133, 1228-1240 (2008).

25. Iacobucci, C. et al. Carboxyl-Photo-Reactive MS-Cleavable Cross-Linkers: Unveiling a Hidden Aspect of Diazirine-Based Reagents. Anal. Chem. 90, 2805-2809 (2018).

26. Chin, D. & Means, A.R. Calmodulin: a prototypical calcium sensor. Trends Cell Biol 10, 322-328 (2000).

27. Marcelo, K.L., Means, A.R. & York, B. The Ca2+/Calmodulin/CaMKK2 Axis: Nature's Metabolic CaMshaft. Trends Endocrin Met 27, 706-718 (2016).

28. Tao, L., Li, Y., Long, H., Ma, H. & Yang, X. The interactome and spatial redistribution feature of Ca 2+ receptor protein calmodulin reveals a novel role in invadopodia-mediated invasion. Cell Death Dis 9, 292 (2018).

29. Babu, Y.S. et al. Three-dimensional structure of calmodulin. Nature 315, 37-40 (1985).

30. Kuboniwa, H. et al. Solution structure of calcium-free calmodulin. Nat Struct Biol 2, 768-776 (1995).

31. Chattopadhyaya, R., Meador, W.E., Means, A.R. & Quiocho, F.A. Calmodulin structure refined at 1.7 Å resolution. J Mol Biol 228, 1177-1192 (1992).

32. Sosa, L.d.V. et al. The structure, molecular dynamics, and energetics of centrin–melittin complex. Proteins: Structure, Function, and Bioinformatics 79, 3132-3143 (2011).

33. Wang, J., Wang, W., Kollman, P.A. & Case, D.A. Automatic atom type and bond type perception in molecular mechanical calculations. J Mol Graph Model 25, 247-260 (2006).

34. Case, D.A. et al. The Amber biomolecular simulation programs. J Comput Chem 26, 1668-1688 (2005).

35. Maier, J.A. et al. ff14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from ff99SB. J Chem Theory Comput 11, 3696-3713 (2015).

36. Jorgensen, W.L., Chandrasekhar, J., Madura, J.D., Impey, R.W. & Klein, M.L. Comparison of simple potential functions for simulating liquid water. J Chem Phys 79, 926-935 (1983).

37. Darden, T., York, D. & Pedersen, L. Particle mesh Ewald: AnN⋅log(N) method for Ewald sums in large systems. J Chem Phys 98, 10089-10092 (1993).

38. York, D.M., Darden, T.A. & Pedersen, L.G. The effect of long‐range electrostatic interactions in simulations of macromolecular crystals: A comparison of the Ewald and truncated list methods. J Chem Phys 99, 8345-8348 (1993).

39. Ryckaert, J.P., Ciccotti, G. & Berendsen, H.J.C. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 23, 327-341 (1977).

40. Bucher, D., Pierce, L.C., McCammon, J.A. & Markwick, P.R. On the use of accelerated molecular dynamics to enhance configurational sampling in ab initio simulations. J. Chem. Theory Comput. 7, 890-897 (2011).

41. Bullock, J.M.A., Schwab, J., Thalassinos, K. & Topf, M. The Importance of Non-accessible Crosslinks and Solvent Accessible Surface Distance in Modeling Proteins with Restraints From Crosslinking Mass Spectrometry*. Mol. Cell Proteomics. 15, 2491-2500 (2016).

Chemicals and reagents

Urea, NH₄CO₃, formic acid (FA), trypsin, chymotrypsin, dimethyl sulfoxide (DMSO), dithiothreitol (DTT), iodoacetic acid (IAA), bis(2,5-dioxopyrrolidin-1-yl) octanedioate, crosslinkers dithiobis(succinimidylpropionate) (DSP) and sulfo-NHS-SS-diazirine (sulfo-SDAD) were all purchased from Thermo Fisher Scientific (Waltham, MA, USA). Ammonia was purchased from Nanjing Chemical Reagent Co. (Nanjing, China). 2-(3-methyl-3H-diazirin-3-yl) ethan-1-amine was purchased from Taizhou GreenChem Pharma-Tech Co., Ltd (Taizhou, China). NH₄Cl, NaHCO₃, Na₂SO₄ and tert-Butyl methyl ether (MTBE) were all purchased from Aladdin (Shanghai, China). Calmodulin (CaM), peptide (SSTNVG, denoted as SG) and peptide from CaMKII (sequence: LKKFNARRKLKGAILTTMLATRNFS), were purchased from Nanjing Peptide Industry (Nanjing, China). Melittin was purchased from Yien Chemical Technology (Shanghai, China). Acetonitrile (ACN), dichloromethane (DCM) and triethylamine were purchased from Merck (Darmstadt, Germany). Deionized water was prepared using a Milli-Q system of Millipore (Billerica, MA, USA). Bovine serum albumin, lysozyme and myoglobin were all purchased from Sigma Aldrich (St Louis, MO, USA).

Synthesis of DCD crosslinker

Bis(2,5-dioxopyrrolidin-1-yl) octanedioate (736 mg, 2.0 mmol) and 2-(3-methyl-3H-diazirin-3-yl) ethan-1-amine (416 mg,4.2 mmol) were mixed and dissolved in DCM (20 mL). Triethylamine (1011.9 mg, 10 mmol) was added and stirred for 18 h at room temperature. The mixture was washed with saturated NH₄Cl and subsequently saturated NaHCO₃, dried over anhydrous Na₂SO₄ and filtered. Solvent was removed under reduced pressure to afford the crude product, which was further purified by recrystallization using MTBE to obtain final product as a white solid. NMR of synthesized DCD was shown in Figure S1. DCD (423 mg, 63% yield). ¹H NMR of obtained DCD was shown in Figure S1A (300 MHz, CDCl₃) δ (ppm): 5.97 (s, 1H), 3.18-3.11 (q, 4H), 2.19-2.14 (t, 4H), 1.64-1.55 (m, 8H), 1.35-1.30 (m, 4H), 1.03 (s, 6H). ¹³C NMR was shown in Figure S1B (75 MHz, CDCl3) δ (ppm): 173.22, 36.46, 34.59, 34.13, 28.73, 25.41, 24.54, 19.78

Crosslinking reaction

Generally, samples, including model peptide and protein/protein complex, were mixed with DCD at a designed molar ratio (ranging from 1:10 to 1:20, sample:crosslinker) with indication. Liquid nitrogen was applied for frozen and crosslinking reaction was initiated by laser irradiation (wavelength: 355 nm, LWUVL355-120 uJ, Beijing Laserwave OptoElectronics Technology Co. Ltd.) for designed time period. Three biological repeats were performed for each crosslinking.

Model peptide/protein crosslinking. Peptide was prepared at 2 mM and DCD was added at a molar ratio of 1:10 (peptide:crosslinker). For crosslinking reaction efficiency monitoring, protein (lysozyme 75 μM) was prepared in serious volume and DCD was added at a molar ratio of 1:20 (protein:crosslinker). Liquid nitrogen was applied before laser activation.

Crosslinking condition optimization. For pH effect investigation, myoglobin was prepared in aqueous solution. A serious of pH was adjusted by FA or NH₃. For solvent composition effect monitoring, different solvents were applied, including 2.5 M NaCl, 10% methanol and 10% DMSO and DCD was added at a molar ratio of 1:20 (peptide:crosslinker). For optimization of crosslinker reaction, model proteins, including CaM, BSA and myoglobin, were prepared at concentration of 75 μM and mixed with DCD at a serious of ratios.

Crosslinking of protein complex. CaM was mixed with CaCl₂ at a molar ratio of 1:5 at room temperature for 30 min. Selected targets (melittin and binding peptide from CaMKII) was then mixed with Ca²⁺-CaM in the ratio of 1:2 and incubated at room temperature for 30 min to allow the formation of protein complex. DCD was added at a molar ratio of 1:20 (peptide:crosslinker), followed up with in situ crosslinking reaction.

Time-resolved crosslinking. CaM and CaCl₂ were mixed at a molar ratio of 1:5. Crosslinking reaction was initiated immediately after their incubation for a serious of time points.

Protein digestion.

Crosslinked samples were denatured by urea (8 M, dissolved in 25 mM NH₄HCO₃) and incubated at room temperature for 1h. Desalting was performed (3 kD ultrafiltration column) at 12,000 g for 20 min with 25 mM NH₄HCO₃ solution. The obtained samples were subjected to chymotrypsin (protein:enzyme ratio of 20:1) digestion for 4 hrs (37 ℃) and trypsin digestion overnight. Resulted peptide mixture underwent vacuum-dried and re-dissolved in H₂O containing 0.1% FA. C18 ZipTip desalting prior to LC-MS analysis.

LC-MS analysis

Liquid chromatographic separation was performed by an Easy-nLC 1200 (Thermo Scientific) equipped with a C18 column (Acclaim PepMap^TM RSLC, 75 μm × 25 cm). Mobile phase A was H₂O and ACN (80:20 by volume) containing 0.1% FA, and B was ACN. Gradient elution condition was set as follows: 3-8% B in 2 min; 8-28% B from 2 to 28 min; 28-38%B from 28 to 58 min; 38 -100%B from 58 to 61 min and maintained at100% B from 61 to 65 min. Column temperature was maintained at 56 ℃ and flow rate was 300 nL/min. An Orbitrap Eclipse Tribrid mass spectrometer (Thermo Scientific) was used for data acquisition. Parameters for MS were set as follows: m/z range 350-2000; positive ion voltage 2100 V; ion transfer tube temperature 300 ℃; activation type HCD; HCD collision energy 30%. A Q-TOF MS (Waters, SYNAPT G2-Si) was applied for model peptide and protein crosslinking investigation. MS parameters were as follows: m/z range 700–3000; capillary voltage +3 kV; sampling cone 40 V; source temperature 100 ℃; desolvation temperature 450℃; and scan time 0.5 s.

Data processing

The identification of crosslinking products was performed by pLink (version 2.3.10). The major parameter settings were as follows: missed cleavages 3; peptide mass range 400-6000 Da; peptide length 4-60; precursor tolerance 10 ppm; fragment tolerance 10 ppm. Results were filtered with a false discovery rate (FDR) of 0.05 at the spectrum level followed by an E-value cutoff of 10^-3. The spectra count was set as ≥ 2 to ensure the identification accuracy. MS/MS data was viewed by pLable (version 2.4). Crosslinking products shown as arcs on the sequence of proteins was generated by xiNET (https://crosslinkviewer.org/upload.php). Protein deconvolution was performed by MaxEnt from MassLynx (version 4.1, Waters).

Protein structural modeling

Structure Preparation. The Ca²⁺-free (PDB: 1CFD³⁰), Ca²⁺-bound (PDB: 1CLL³¹) and melittin-bound states of CaM were used for the MD simulation, in which the initial structure of melittin-bound CaM was modelled from a crystal structure of centrin-melittin complex (PDB: 3QRX³²). The antechamber and tleap modules in AMBER/20 ³³^,³⁴ were employed for the structure preparation. The protein (CaM) and peptide (melittin) were parameterized with the AMBER ff14SB force field ³⁵. Counterions of Na⁺ or Cl^- were added to the systems to balance the redundant charges. The protein was finally immersed in a cubic TIP3P water box extended 10 Å out of the solute ³⁶.

Molecular Mechanics (MM) Minimization. For MM minimization, the real-space cutoff distance was set to 10 Å (including the van der Waals and short-range electrostatic interactions), while the PME algorithm (particle mesh Ewald) ³⁷ was used to treat the long-range electrostatic interactions ³⁸. All the systems were optimized with four steps, namely (1) all the hydrogen atoms were relaxed with all the heavy atoms restrained at 5 kcal/mol∙Å² (5000 steps with 1000 cycles of steepest descent and 4000 cycles of conjugate gradient minimization); (2) The systems were minimized with the heavy atoms in the protein and ligand restrained at 5 kcal/mol∙Å² and with other atoms set free (5000 steps with 1000 cycles of steepest descent and 4000 cycles of conjugate gradient minimization); (3) 5000 steps of minimization were conducted with only the heavy atoms in backbone of the protein restrained at 5 kcal/mol∙Å² (1000 cycles of steepest descent and 4000 cycles of conjugate gradient minimization); (4) the whole system was released without any restraints for 10000 steps (5000 cycles of steepest descent and 5000 cycles of conjugate gradient minimization).

Conventional Molecular Dynamics (cMD) Simulation. In the phase of cMD simulation, the systems were performed with three steps of MD simulation, namely (1) the systems were heated from 0 k to 300 k during 50 ps in the NVT ensemble, in which heavy atoms in the protein backbone were restricted with 2 kcal/mol∙Å²; (2) 50 ps of density equilibration was carried out in the NTP ensemble (T=300 K and P=1 atm) with the same constraint as that of the heating process; (3) 1 μs cMD simulation was performed in the NPT ensemble without any restraint. In the cMD simulation, the time step was set to 2 fs. SHAKE algorithm ³⁹ was used to constrain the covalent bonds between the hydrogen atoms and the connected heavy atoms. The coordinates (the MD trajectory) were collected with an interval of 5 ps, thus a total of 20,000 frames were collected for each system.

Accelerated MD (aMD) Simulation. aMD simulation was further conducted for the CaM-melittin system because of the limited conformation-searching capability of cMD in this system. In the spirit of aMD simulation, the algorithm works by adding “boost energy” to the energy basins of the free energy landscape to help the system overcome the high energetic barriers located between different thermodynamic states ⁴⁰. Here, the “dual boost” protocol, working by adding boost potentials to both the “total” and “dihedral” potentials of the original system, was applied to the aMD simulation, in which the parameters were set according to the AMBER manual/2020 (https://ambermd.org/). 600 ns aMD simulation was performed for the system of CaM-melittin with the same other parameters as the cMD simulations (e.g. time step of 2 fs for the MD simulation, time interval of 5 ps for the coordinate collection, etc.). A total of 12,000 frames were collected in the aMD simulation. All the MM minimizations and c/aMD simulations were performed with the pmemd.cuda module in AMBER/20.

SASD Analysis. To match the conformation of CaM with the crosslink data, we calculated the solvent accessible surface distances (SASDs) of specific residue pairs using Jwalk (version 2.1) ⁴¹ or Model validation using crosslinking restraints (MNXL, http://mnxl.ismb.lon.ac.uk/mnxl/). Distance between the Cα atoms of the crosslink residues were measured with the voxel size of grid set to 1 Å.

There is NO Competing Interest.

Download PDF

Version 1

posted

You are reading this latest preprint version

Transient Crosslinking Mass Spectrometry: Taking Conformational Snapshots of Proteins

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Declarations

Acknowledgement

References

Methods

Additional Declarations

Supplementary Files

Status:

Version 1