English

• Rapidly tracking subtle differences and dynamic transitions among various conformational states adopted by substrate proteins represents primary challenges facing structural and chemical biology. To this end, forced protein unfolding in the absence of solvent has increasingly become powerful and impactful in the realm of functional interrogation of target proteins with variable and dynamic domains [1-6]. In principle, snapshotting time-resolved unfolding intermediates which are aligned with the on-pathway folding progress is vital for extending the versatility of gas-phase unfolding technique. Collisional activation in the gas phase is one such analytical regime that permits the induction of forced protein stepwise unfolding in a controllable manner.

  Native ion mobility-mass spectrometry (IM-MS) [7] has been well suited to the gas-phase collision induced-unfolding (CIU) regime [8], allowing for the precise and quantitative comparison of unfolding patterns and stabilities as carried by diverse proteoforms at both the intact protein and complex levels [9]. Since the early integration of CIU with native IM-MS, the past decade has seen its rapidly expanding roles in characterizing therapeutic and many other protein systems. The great needs from academic and industrial applications of CIU-IM-MS have catalyzed the development of novel CIU setups in multiple commercial instruments, and the birth of powerful data visualization algorithms and associated software tools, highlighting the CIUSuite [10]. Recent collaborative efforts from our group seek to extend the utility, speed and throughput of CIU-IM-MS with an expanding range of small to large protein examples, applicable to both isolated protein standards and crude conjugate antibodies [11]. Thanks to pioneers' efforts [12], increasing factors contributing to subtle conformational differences the target proteins bear can be distinguished [13], including external stress [14], disulfide isoform [2], glycosylation [15], batch-to-batch variation [16] and biological heterogeneity [17] arising from conjugate vaccine treatment.

  Despite extensive applications, relatively few examples exist where the molecular evidence for CIU process [18] is adequately provided to track how protein secondary structures dynamically change in response to slow heating during collisional activation [19, 20]. Notably, the Ruotolo group previously demonstrated the positive correlation of CIU transition numbers to domain numbers of a series of monomeric proteins [3], albeit with some exceptions. The same group continuously devotes to the mechanistic research on protein unfolding pathway in the absence of solvent, including revealing the unfolding pathway of serum albumin [21], a protein primarily with α-helix but lack of the element of beta-sheet. While it is apparent that such efforts will enhance our fundamental understanding of CIU process, utilizing IM-MS alone is not sufficient to accomplish the task of mechanistic tracking due to limited structural information attainable from collisional cross section (CCS)-based measurements [18, 19, 22, 23].

  Molecular dynamics (MD) simulations have the capability to replicate the behavior of atoms and molecules under varying conditions, encompassing aspects such as conformational changes, kinetic properties, and energy variations [18, 24-28]. As a result, they can unveil the intrinsic nature and behavior of molecular systems and provide valuable experimental data that is otherwise challenging to obtain. Nonetheless, simulating large molecules, such as proteins, presents its own set of challenges, including dealing with large time scales, limited computational resources, and complex data processing. To tackle these challenges, researchers commonly integrate experimental data and theoretical models, employ parallel computing and high-performance computing technology to enhance simulation efficiency, and continually refine simulation methods and algorithms to confront these obstacles.

  Herein, after selecting a model system of secondary structure-separated multidomain protein, CRM₁₉₇ [29], we systematically examine its solvent-free unfolding pathway with traditional CIU-IM-MS analytical regime. CRM₁₉₇, a non-toxic mutant (G52E) of diphtheria toxin, has been widely used as a carrier protein for conjugate vaccine development [30], such as in approved pediatric vaccines against Haemophilus influenzae type b and pneumococcal disease [31], as well as in novel bioconjugate vaccines directed against abusable small molecule drugs like opioids [32] and dissociative-hypnotics [11]. Its static 3D structure has been well elucidated with traditional biophysical tools, confirming the presence of three distinctive domains with diverse secondary structure elements. Namely, catalytic domain region (D1), transmembrane domain region (D2) and receptor domain region (D3) bear α-helix/β-sheet-comparable, α-helix-rich and β-sheet-rich secondary structures, respectively (Fig. 1A, Fig. S1 and Table S1 in Supporting information). The physicochemical characteristics of CRM₁₉₇, including higher order structure, have been found to be highly similar across different production methods and expression systems [33]. Previous studies on molecular events of the G52E mutation-associated toxicity difference have revealed the importance of an active-site loop that contributes to the increased flexibility and instability of CRM₁₉₇ [30, 34].

  Figure 1

  

  Figure 1.  CRM₁₉₇ unfolding as a function of cofactor NCA. (A) CRM₁₉₇ (PDB ID: 4ae0), (B) CRM₁₉₇—NCA complex (PDB ID: 4ae1), and (C) crystal structure for CRM₁₉₇—NCA complex showing the evidence of noncovalent specific interactions of NCA with catalytic domain D1, forming a large π-π conjugated hydrophobic pocket with surrounding amino acids such as H21, Y65 and F140 (the key amino acids surrounding NCA within a distance of < 7.0 Å). (D) Representative native MS spectrum for CRM₁₉₇ and CRM₁₉₇—NCA complex (CRM: NCA = 1: 50). (E, F) CIU fingerprints for CRM₁₉₇ (15+) and CRM₁₉₇—NCA complex (15+). Theoretical CCS values were derived from both the trajectory method (TJM) and projection approximation (PA) method. (G) CIU₅₀-based conformational stability comparison between CRM₁₉₇ and CRM₁₉₇—NCA complex. CIU₅₀ data are derived from averaging CIU fingerprints of three independent measurements (Student's t-test, P < 0.05, **P < 0.01).

  DownLoad: Full-Size Img PowerPoint

  Notably, nicotinamide (NCA) is a small molecule that can specifically bind to the hydrophobic pocket region in D1, forming a large π-π conjugated system with the surrounding key amino acids (Y20, H21, G22, Y54, T56, Y65, F140, and E148) (Figs. 1B and C). To delineate the dynamic structures of CRM₁₉₇ and the effect of NCA complexation, we firstly examine its solvent-free unfolding pathway with CIU-IM-MS analytical regime. Native MS experiments, illustrated in Fig. 1D and Fig. S2A (Supporting information), confirm the binding of NCA to CRM₁₉₇; however, this interaction is relatively weak, as a significant mass shift is observed only with a high NCA to CRM₁₉₇ ratio. Tandem MS results further substantiate the specificity of the NCA-CRM₁₉₇ interaction, as distinctive NCA peaks are absent at low trap voltages, as shown in Fig. S2B (Supporting information). Upon NCA incubation, the overall charge state distributions for both CRM₁₉₇ monomer (14+ to 17+) and dimer (22+ to 24+) remain unchanged as well as the relative level of dimer species, indicating very minor shift on the natively folded structure by NCA complexation just as previously reported [30].

  While protein unfolding fingerprints have been reported to be highly sensitive to domain numbers [3], we herein sought to uncover the contribution of secondary structure to the overall fingerprints and its unfolding pathways. A series of CIU experiments (Figs. 1E and F, Fig. S3 in Supporting information) were conducted with the CRM₁₉₇ protein sample and the comparative NCA incubation sample. CIU fingerprints from both 15+ and 16+ ion species generate four conformational features (can be defined as five with less critical parameters, namely F1-F5) for both CRM₁₉₇ and CRM₁₉₇—NCA complex, correlating well with the well-defined 3D structure of three major domains. The subtle conformer F4 may be a result of dynamic transitions between two adjacent domains, which unfold at the last stage but yet remain to be determined for the exact domain number. To delineate the structural information, CCS values were calculated across all conformational features. As expected, the experimental CCSs of conformer F1 were 41.83 nm² and 42.05 nm² for CRM₁₉₇ 15+ and CRM₁₉₇—NCA 15+, respectively, both of which lay in the range of TJM and PA-based theoretical CCS values. The slight difference of < 1% in CCSs of conformer F1 align well with previous reports using other tools [30], further confirming the maintenance of the native structure of CRM₁₉₇ with NCA complexation. Surprisingly, the CCS differences induced by NCA complexation show significant dependency on unfolding energy. Data in Fig. S4 (Supporting information) clearly indicate that late-stage unfolding conformers produce much larger CCS differences than conformer F1, suggesting a potentially regulatory role of NCA during CRM₁₉₇ unfolding in the absence of solvent.

  The CIU₅₀ value [20, 35] by definition represents the unfolding energy at which 50% of protein conformations undergo transition to adjacent conformers and it frequently serves as a critical parameter to characterize the conformational stability. Intriguingly, we observed a significant reduction of conformational stability (Fig. 1G and Fig. S3C) upon NCA binding, by a reducing percentage of CIU₅₀ values from 22.8% to 4.1% for four conformational transitions of 15+ ion species. MD results (Fig. S3D, details shown below) indicate that NCA disrupts the interactions between Y55 and Y66, thereby affecting the D1 stability on the side of Y55 and overall conformational stability of CRM₁₉₇. Contrary to the trends observed in CCS comparisons, the alteration of the CIU₅₀ value upon NCA complexation exhibits an inverse relationship with the unfolding energy, with the most pronounced decrease in stability observed during the initial transition. Similar trends were observed for 16+ ion species. Besides, the calculation of NCA and CRM₁₉₇—NCA complex intensity over collisional energy (Fig. S5 in Supporting information) suggests that NCA gradually separates from the protein complex system with a trap voltage over ~40 V, corresponding to the appearance of conformer F2. On top of these observations, overall comparison of CIU unfolding heatmap (Fig. S6 in Supporting information) yields striking differences with root mean square deviation (RMSD) values of 31.40% and 29.66% for 15+ and 16+, respectively. Both RMSD values represent NCA-induced disparities in overall CRM₁₉₇ unfolding pathway, as they are fairly higher than baseline RMSDs being ~10% (Fig. S6). Taking these CIU-IM-MS data together, we suspect that NCA initially binds to the compact hydrophobic pocket with least perturbation on CRM₁₉₇'s native structure. But during gas-phase activation, it may undergo dynamic translocation and releasing processes induced by the significant conformational expansions, most of which can be distinguished via collisional activation.

  In order to deduce the changes in the secondary structure during the unfolding process, we performed classical MD simulations to predict the CIU pathway. The gas-phase MD simulations of CRM₁₉₇ and CRM₁₉₇—NCA at charge states of 15+ and 16+ were firstly carried out over a linear temperature gradient from 300 K to 750 K for 1000 ns, mimicking the collisional heating that happens during CIU. In contrast to previous methods of modeling collisional dissociation and unfolding in the gas phase [36], maintaining a constant temperature for the simulation throughout additional 500 ns enables sampling of multiple protein states at a specific temperature. During the temperature ramping process (Fig. S7 in Supporting information), we determined that 750 K provided the best correlation with the experimental CCS measurements. Consequently, this temperature was chosen for the constant heating phase of our simulations. To ensure comprehensive sampling and robustness of our results, we performed three independent simulations for each system in our study. This trifold simulation approach generated a combined total of 30, 000 structural models across different thermodynamic conditions. It is through this extensive dataset that we were able to confidently reproduce the experimental CCS distribution and the unfolding process observed in CIU experiments.

  The initial phase of our study examined the effect of protonation states on the unfolding behavior of CRM₁₉₇ by analyzing three distinct low-energy protonation states. The data presented in Fig. 2 and Fig. S8 (Supporting information) demonstrate that variations in these protonation states do not significantly influence the unfolding pathways or the structural integrity of CRM₁₉₇. Representative MD results on the unfolding of CRM₁₉₇ and CRM₁₉₇—NCA at 15+ are shown in Figs. 2A and B, mimicking the conformational changes occurring during the gas-phase heating process, as guided by experimental findings. Throughout the gas-phase heating process, a total of 5 conformational ensembles can be accordingly identified and captured in line with IM-MS datasets. The MD simulation results show good agreement with the experimental measurements, demonstrated by the acceptable CCS deviations largely being ~3% for both CRM₁₉₇ and CRM₁₉₇—NCA complex (Fig. S9 in Supporting information). Notably, CRM₁₉₇—NCA unfolding was observed to occur at lower temperatures than CRM₁₉₇ alone, aligning with the experimentally observed unfolding at lower collision voltages. Furthermore, domain-specific MD results (Fig. S10 in Supporting information) reveal that the incorporation of NCA leads to an acceleration in the time required by the protein to reach a steady state during heating unfolding processes (~500 ns for CRM₁₉₇ and ~450 ns for CRM₁₉₇—NCA), which align closely with our experimental findings on CIU₅₀ values.

  Figure 2

  

  Figure 2.  Capturing sheet-to-helix dynamic transitions during CRM₁₉₇ and CRM₁₉₇—NCA unfolding in the absence of solvent. CCS changing curves for CRM₁₉₇ (A) and CRM₁₉₇—NCA (B) derived from temperature-jumping MD simulation from 300 K to 750 K. (C) Representative models for CRM₁₉₇ unfolding intermediates. (D-F) Quantitative comparison of secondary structures of five representative unfolding features. (G-J) Overall occupancies of α-helix and β-sheet throughout the unfolding process for each domain. For CRM₁₉₇, the duration of gas-phase heating for F1, F2, F3/F4 and F5 was 0–180 ns, 190–420 ns, 420–504 ns and 504–1000 ns, respectively. For CRM₁₉₇—NCA, the duration of gas-phase heating for F1, F2, F3/F4 and F5 was 0–380 ns, 400–470 ns, 470–550 ns and 550–1000 ns, respectively. All data were obtained from three parallel simulations.

  DownLoad: Full-Size Img PowerPoint

  Moreover, the MD results indicate that the D3 domain exhibits the highest degree of unfolding (Fig. S10 in Supporting information). This is presumably due to the abundant β-sheet content within D3, which is consistent with the general understanding that β-sheets are less stable than α-helix under thermal stress. In contrast, the final unfolding extents of the D1 and D2 domains, both rich in α-helices, do not show a significant difference. These findings suggest that the presence of NCA affects the stability and unfolding kinetics of CRM₁₉₇, with a pronounced effect on the β-sheet-rich D3 domain, thereby providing a molecular-level explanation for the differential stability of the domains during unfolding. Taken together, these observations validate the confidence and reliability of MD simulations, assuring the following further analysis at the secondary structure level.

  The protein conformational diversity increases with incremental temperature. Although each conformational feature contains abundant structures with similar CCS values, there exist differences among them. Therefore, cluster analysis was performed to identify and group sets of similar structures for each conformational feature, utilizing the algorithm of density-based spatial clustering of applications with noise (DBSCAN). Simultaneously, to ensure that the representative conformations obtained from clustering were highly representative (over 70% of total frames, 10, 000 structural models, three replicates totaling 30, 000 structural models), we conducted tests to determine the optimal cutoffs based on the RMSD data of the protein backbone. Among them, cluster 1 is the optimal conformational set. Fig. 2C and Fig. S9C illustrate the changes in the representative structural models of the protein during the gas-phase heating process. We can visually observe a transformation in the global structure from compact to more extended, with only α-helix and loop regions remaining, and complete conversion of β-sheets. The similar phenomenon was also observed in the remaining clustered datasets (Fig. S11 in Supporting information). We quantified the changes in secondary structure that occur during the unfolding process. Surprisingly, even at relatively low temperatures, β-sheets have already completed their transformation, and it seems that the addition of NCA aids in this process. In contrast to CRM₁₉₇ which achieves complete transformation at conformer F3, CRM₁₉₇—NCA achieves full transformation at conformer F2. Meanwhile, the transition between α-helix and loop structures exists in a dynamic state of competitive equilibrium. It is observed that the secondary structure tends to be stable starting at the conformer F3. With increasing temperature, the regular secondary structure of α-helix undergoes a transition to a loop structure through thermal collisions, ultimately reaching equilibrium (Figs. 2E and F).

  Furthermore, the β-sheet undergoes a continuous transformation into α-helix during the whole unfolding process, as demonstrated in Figs. 2G and I. Surprisingly, the content of the α-helix in the D3 gradually increases from 0% to ~20%. Eventually, the α-helical occupancy in all three domains remains at a stable level of around 20%, while the β-sheet structure undergoes a complete transformation (Figs. 2G–J). The conversion from β-sheet to α-helix further elucidates that during the unfolding process in the gas phase, the stability of the α-helix is significantly greater than that of the β-sheet. Similar trends were observed for the 16+ ion (Figs. S12–S15 in Supporting information). In addition, MD data (Figs. S15–S17 in Supporting information) also supports that the NCA complexation promotes the unfolding process following sheet-to-helix transition pathway. Furthermore, we also verified this through a series of experiments (Figs. S2 and S18 in Supporting information) that when the voltage reached approximately 40 V, the m/z between CRM₁₉₇ and CRM₁₉₇—NCA stabilized, and the NCA was observed at low-mass end of MS/MS spectra, indicating that NCA has completely dissociated from the protein and coexists with the protein in the system.

  Notably, our additional experiments (Figs. S19 and S20 in Supporting information) on a single domain protein, ubiquitin, with both secondary structure elements provide further support for the greater stability of α-helix over β-sheet during the unfolding process, confirming the occurrence of structural transitions from β-sheet to α-helix in diverse protein systems. On top of previous methods that experienced challenges in predicting the subtle changes in the substructure of protein complexes during the CIU process [28, 37], we herein have tentatively drafted an unfolding atlas for a multidomain protein and its noncovalent complex at the secondary structure- and domain-resolved molecular level. Consequently, the newly discovered unfolding map drives us to capture the new pathway of preferential unfolding of sheet-rich domains, and to confirm that α-helix exhibits higher stability compared to β-sheet for resisting solvent-free forced unfolding.

  In summary, utilizing a multi-domain protein model system with domain-separated secondary structures, we systematically studied its solvent-free unfolding pathway using native CIU-IM-MS and MD simulations. Consequently, a nearly complete unfolding atlas for CRM₁₉₇ was established, which drives the discovery of a first secondary structure-level unfolding pathway for a multidomain protein complex. Collective data reveal a gradual transformation of β-sheet structures into α-helix structures throughout the unfolding process. Additionally, there is substantial evidence supporting the greater stability of α-helix compared to β-sheet structures in resisting heating activation in the absence of solvents. Therefore, we envision a general unfolding pipeline for single-/multi-domain proteins where the preferential unfolding of β-sheet-rich domains precedes unfolding of the α-helix-rich domains. We expect that this sheet-to-helix dynamic transition is central to gas-phase protein unfolding pathway, potentially serving as a novel blueprint to govern the fundamental understanding of multidomain protein unfolding in the absence of solvent. The study has identified an unprecedented paradigm of dynamic conformational changes in the multidomain protein CRM₁₉₇, unveiling the dynamic alterations in secondary structure during the unfolding process, and paving an avenue for further exploration into the dynamic regulation of additional multidomain protein structures and functions.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper

  CRediT authorship contribution statement

  Xia Xu: Writing – review & editing, Writing – original draft, Data curation. Guiqian Yang: Data curation. Zhen Zheng: Writing – original draft, Data curation, Conceptualization. Cody J. Wenthur: Writing – original draft, Supervision. Jinyu Li: Writing – review & editing, Supervision, Conceptualization. Gongyu Li: Writing – review & editing, Writing – original draft, Supervision, Conceptualization.

  Acknowledgments

  We thank the support by the National Key R&D Program of China (No. 2022YFA1305200, to GL), National Natural Science Foundation of China (No. 22104064 to GL, No. 22173020 to JL), and the US National Institute of Mental Health (No. R01MH122742, to CJW) for financial and instrumental support.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110401.

  
  
• 1. [1]

     S.J. Hyung, C.V. Robinson, B.T. Ruotolo, Chem. Biol. 16 (2009) 382–390.

  2. [2]

     Y. Tian, L. Han, A.C. Buckner, B.T. Ruotolo, Anal. Chem. 87 (2015) 11509–11515. doi: 10.1021/acs.analchem.5b03291

  3. [3]

     Y. Zhong, L. Han, B.T. Ruotolo, Angew. Chem. Int. Ed. Engl. 53 (2014) 9209–9212. doi: 10.1002/anie.201403784

  4. [4]

     M. Kohtani, T.C. Jones, R. Sudha, M.F. Jarrold, J. Am. Chem. Soc. 128 (2006) 7193–7197. doi: 10.1021/ja056745s

  5. [5]

     J. Luo, Q. Wan, S. Chen, Chin. Chem. Lett. (2024) 109836. doi: 10.1016/j.cclet.2024.109836

  6. [6]

     M. Chen, P. Zhu, Q. Wan, et al., Anal. Chem. 95 (2023) 7495–7502. doi: 10.1021/acs.analchem.2c05414

  7. [7]

     E. Desligniere, S. Ollivier, A. Beck, et al., Anal. Chem. 95 (2023) 4162–4171. doi: 10.1021/acs.analchem.2c05265

  8. [8]

     S. Hong, M.F. Bush, J. Am. Soc. Mass. Spectrom. 30 (2019) 2430–2437. doi: 10.1007/s13361-019-02326-z

  9. [9]

     D.S. Chan, M.E. Kavanagh, K.J. McLean, et al., Anal. Chem. 89 (2017) 9976–9983. doi: 10.1021/acs.analchem.7b02329

  10. [10]

      J.D. Eschweiler, J.N. Rabuck-Gibbons, Y. Tian, B.T. Ruotolo, Anal. Chem. 87 (2015) 11516–11522. doi: 10.1021/acs.analchem.5b03292

  11. [11]

      Z. Zheng, M. Ma, Y. Jia, et al., Anal. Chem. 95 (2023) 10895–10902. doi: 10.1021/acs.analchem.3c00223

  12. [12]

      V. Gabelica, A.A. Shvartsburg, C. Afonso, et al., Mass Spectrom. Rev. 38 (2019) 291–320. doi: 10.1002/mas.21585

  13. [13]

      S.M. Dixit, D.A. Polasky, B.T. Ruotolo, Curr. Opin. Chem. Biol. 42 (2018) 93–100.

  14. [14]

      Y. Yao, X. Wang, J. Guan, et al., Nat. Commun. 14 (2023) 2339–2350.

  15. [15]

      K.T. Schjoldager, Y. Narimatsu, H.J. Joshi, H. Clausen, Nat. Rev. Mol. Cell Biol. 21 (2020) 729–749. doi: 10.1038/s41580-020-00294-x

  16. [16]

      D.J.B. Hunter, A. Bhumkar, N. Giles, E. Sierecki, Y. Gambin, Biotechnol. Bioeng. 115 (2018) 1904–1914. doi: 10.1002/bit.26604

  17. [17]

      A. D'Urzo, A. Konijnenberg, G. Rossetti, et al., J. Am. Soc. Mass. Spectrom. 26 (2015) 472–481. doi: 10.1007/s13361-014-1048-z

  18. [18]

      S. Nash, R.W. Vachet, J. Am. Chem. Soc. 144 (2022) 22128–22139. doi: 10.1021/jacs.2c09573

  19. [19]

      J. Seo, W. Hoffmann, S. Warnke, et al., Angew. Chem. Int. Ed. Engl. 55 (2016) 14173–14176. doi: 10.1002/anie.201606029

  20. [20]

      A. Phetsanthad, G. Li, C.K. Jeon, B.T. Ruotolo, L. Li, J. Am. Soc. Mass. Spectrom. 33 (2022) 944–951. doi: 10.1021/jasms.2c00004

  21. [21]

      J.D. Eschweiler, R.M. Martini, B.T. Ruotolo, J. Am. Chem. Soc. 139 (2017) 534–540. doi: 10.1021/jacs.6b11678

  22. [22]

      M.L. Feuerstein, M. Hernandez-Mesa, A. Kiehne, et al., J. Am. Soc. Mass. Spectrom. 33 (2022) 1951–1959. doi: 10.1021/jasms.2c00196

  23. [23]

      L. Huang, P.K. So, Y.W. Chen, Y.C. Leung, Z.P. Yao, J. Am. Chem. Soc. 142 (2020) 13756–13767. doi: 10.1021/jacs.0c04088

  24. [24]

      J. Li, W. Lyu, G. Rossetti, et al., J. Phys. Chem. Lett. 8 (2017) 1105–1112. doi: 10.1021/acs.jpclett.7b00127

  25. [25]

      E. Christofi, P. Barran, Chem. Rev. 123 (2023) 2902–2949. doi: 10.1021/acs.chemrev.2c00600

  26. [26]

      Z. Hall, A. Politis, M.F. Bush, L.J. Smith, C.V. Robinson, J. Am. Chem. Soc. 134 (2012) 3429–3438. doi: 10.1021/ja2096859

  27. [27]

      C.E. Bartman, H. Metwally, L. Konermann, Anal. Chem. 88 (2016) 6905–6913. doi: 10.1021/acs.analchem.6b01627

  28. [28]

      L. Konermann, E. Aliyari, J.H. Lee, J. Phys. Chem. B, 125 (2021) 3803–3814. doi: 10.1021/acs.jpcb.1c00944

  29. [29]

      M.S. Kent, H. Yim, J.K. Murton, et al., Biophys. J. 94 (2008) 2115–2127.

  30. [30]

      E. Malito, B. Bursulaya, C. Chen, et al., Proc. Natl. Acad. Sci. U S A, 109 (2012) 5229–5234. doi: 10.1073/pnas.1201964109

  31. [31]

      M. Broker, P. Costantino, L. DeTora, E.D. McIntosh, R. Rappuoli, Biologicals 39 (2011) 195–204.

  32. [32]

      F. Baruffaldi, A.H. Kelcher, M. Laudenbach, et al., Mol. Pharm. 15 (2018) 4947–4962. doi: 10.1021/acs.molpharmaceut.8b00592

  33. [33]

      J.M. Hickey, V.M. Toprani, K. Kaur, et al., J. Pharm. Sci. 107 (2018) 1806–1819.

  34. [34]

      J. Chen, E.J. Garcia, C.R. Merritt, et al., J. Med. Chem. 66 (2023) 9992–10009. doi: 10.1021/acs.jmedchem.3c00908

  35. [35]

      D.A. Polasky, S.M. Dixit, D.D. Vallejo, K.D. Kulju, B.T. Ruotolo, Anal. Chem. 91 (2019) 10407–10412. doi: 10.1021/acs.analchem.9b02650

  36. [36]

      V. Popa, D.A. Trecroce, R.G. McAllister, L. Konermann, J. Phys. Chem. B 120 (2016) 5114–5124. doi: 10.1021/acs.jpcb.6b03035

  37. [37]

      R.E. Bornschein, S. Niu, J. Eschweiler, B.T. Ruotolo, J. Am. Soc. Mass. Spectrom. 27 (2016) 41–49. doi: 10.1007/s13361-015-1250-7

• 

  Figure 1  CRM₁₉₇ unfolding as a function of cofactor NCA. (A) CRM₁₉₇ (PDB ID: 4ae0), (B) CRM₁₉₇—NCA complex (PDB ID: 4ae1), and (C) crystal structure for CRM₁₉₇—NCA complex showing the evidence of noncovalent specific interactions of NCA with catalytic domain D1, forming a large π-π conjugated hydrophobic pocket with surrounding amino acids such as H21, Y65 and F140 (the key amino acids surrounding NCA within a distance of < 7.0 Å). (D) Representative native MS spectrum for CRM₁₉₇ and CRM₁₉₇—NCA complex (CRM: NCA = 1: 50). (E, F) CIU fingerprints for CRM₁₉₇ (15+) and CRM₁₉₇—NCA complex (15+). Theoretical CCS values were derived from both the trajectory method (TJM) and projection approximation (PA) method. (G) CIU₅₀-based conformational stability comparison between CRM₁₉₇ and CRM₁₉₇—NCA complex. CIU₅₀ data are derived from averaging CIU fingerprints of three independent measurements (Student's t-test, P < 0.05, **P < 0.01).

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Capturing sheet-to-helix dynamic transitions during CRM₁₉₇ and CRM₁₉₇—NCA unfolding in the absence of solvent. CCS changing curves for CRM₁₉₇ (A) and CRM₁₉₇—NCA (B) derived from temperature-jumping MD simulation from 300 K to 750 K. (C) Representative models for CRM₁₉₇ unfolding intermediates. (D-F) Quantitative comparison of secondary structures of five representative unfolding features. (G-J) Overall occupancies of α-helix and β-sheet throughout the unfolding process for each domain. For CRM₁₉₇, the duration of gas-phase heating for F1, F2, F3/F4 and F5 was 0–180 ns, 190–420 ns, 420–504 ns and 504–1000 ns, respectively. For CRM₁₉₇—NCA, the duration of gas-phase heating for F1, F2, F3/F4 and F5 was 0–380 ns, 400–470 ns, 470–550 ns and 550–1000 ns, respectively. All data were obtained from three parallel simulations.

  下载: 全尺寸图片 幻灯片
•