DOI QR코드

DOI QR Code

Fragmentation Analysis of rIAPP Monomer, Dimer, and [MrIAPP + MhIAPP]5+ Using Collision-Induced Dissociation with Electrospray Ionization Mass Spectrometry

  • Kim, Jeongmo (Department of Applied Chemistry, Kumoh National Institute of Technology) ;
  • Kim, Ho-Tae (Department of Applied Chemistry, Kumoh National Institute of Technology)
  • Received : 2021.10.28
  • Accepted : 2021.11.25
  • Published : 2021.12.31

Abstract

Collision-induced dissociation (CID) combined with electrospray ionization mass spectrometry (ESI-MS) was used to obtain structural information on rat islet amyloid polypeptide (rIAPP) monomers (M) and dimers (D) observed in the multiply charged state in the MS spectrum. MS/MS analysis indicated that the rIAPP monomers adopt distinct structures depending on the molecular ion charge state. Peptide bond dissociation between L27 and P28 was observed in the MS/MS spectra of rIAPP monomers, regardless of the monomer molecular ion charge state. MS/MS analysis of the dimers indicated that D5+ comprised M2+ and M3+ subunits, and that the peptide bond dissociation process between the L27 and P28 residues of the monomer subunit was also maintained. The observation of (M+ b27)4+ and (M+ y10)3+ fragment ions were deduced to originate from the two different D5+ complex geometries, the N-terminal and C-terminal interaction geometries, respectively. The fragmentation pattern of the [MrIAPP + MhIAPP]5+ MS/MS spectrum showed that the interaction occurred between the two N-terminal regions of MrIAPP and MhIAPP in the heterogeneous dimer (hetero-dimer) D5+ structure.

Keywords

Introduction

Rat islet amyloid polypeptide (rIAPP), a 37-residue peptide differing from human islet amyloid polypeptide (hIAPP) by six residues in the 18-29 residue region, is known not to form amyloid fibrils.1, 2 Of these 18-29 residues, three are proline residues, which are known to disrupt secondary β-sheet structure formation.3 rIAPP is thus an example of a natural “β-breaker inhibitor”, which prevents the formation and propagation of the β-sheet structure.4 The rIAPP-based inhibition of hIAPP amyloidogenicity5-7 or amyloid β aggregation8 were reported by the formation of rIAPP-hIAPP or rIAPP- amyloid β heterocomplexes.

Amyloid plaques containing aggregates of hIAPP, a 37- residue hormone selectively expressed in pancreatic β- cells, are widely associated with the pathology of type-2- diabetes.9-11 These fibrillar aggregates are formed via peptide self-assembly into unbranched elongated structures and present filamentous morphology, a cross-β spine, steric zipper structures, and are cytotoxic.12-14 Recently, there has been much interest in the early oligomerization process of hIAPP because several studies have implicated greater neurotoxicity of smaller prefibrillar hIAPP oligomers over mature fibrils.15-18 Therefore, it is also important to compare the initial aggregation structure of rIAPP, which does not form amyloid, as a control group, in addition to the analysis of hIAPP early aggregation structures growing to ultimately form amyloid plaques.

Accordingly, there have been several studies on early aggregates of rIAPP. In previous IMS-MS experiments, two different rIAPP D5+ conformations were observed based on the arrival time distribution.19 The Radford group confirmed the gas-phase stability of rIAPP oligomers through CID experiments, demonstrating that they are more stable than hIAPP oligomers.20 Using solution NMR spectroscopy, Nanga et al. studied the 3D structure of rIAPP in the presence of a membrane and suggested that the α-helix region plays a crucial role in self-association.1 The structures of rIAPP dimers and oligomers have been studied using molecular dynamics simulations. It was established that the dimer structure of rIAPP is distinct from that of hIAPP because rIAPP displays an α-helix in the N-terminal, with a distortion in the middle of the chain, followed by a second shorter α- helix, and is devoid of β-sheets.21 It was also suggested that rIAPP dimer to pentamer oligomers suffer from marked secondary structural changes due to a significant loss of the C terminal β-sheet and turn conformation.22

Herein, we used CID in conjunction with ESI-MS to obtain structural information on the rIAPP monomer, dimer, and rIAPP-hIAPP hetero-dimer for comparison with the hIAPP dimer structure. The rIAPP homo- or hetero dimer complexes were allowed to form in solution, and electrosprayed onto a quadrupole ion guide. ESI-MS is assumed to produce intact gas-phase dimer complex ions from the dimer complex in solution. The fragmentation patterns for the rIAPP homo- or hetero-dimer structure were investigated in a 50% H2O+50% CH3OH solution. In a previous study, we investigated multiply charged monomers and dimers of hIAPP employing CID-MS/MS.23

Experimental

The MS for protein cation complexes and MS/MS spectra for fragmentation pattern analysis were obtained using a Thermo Finnigan LTQ mass spectrometer (Thermo Electron Corp., San Jose, CA, USA), which is a linear ion trap mass spectrometer equipped with an atmospheric pressure-ionization source.

Mass spectrometry conditions

The protein samples were introduced into the electrospray ionization interface via a direct infusion method using a microsyringe pump (Hamilton, USA) at a flow rate of 1–2 mL min-1. The positive ion MS spectra were acquired over an m/z range of 700–2000 by averaging 1000–2000 scans. The MS/ MS spectra were acquired under the following experimental conditions: activation time of 30 ms, injection time of 100– 200 ms, and an isolation width of 1–1.5 mass units. The parent protein ions were individually and manually selected and then subjected to the CID process. The minimal collision energies were optimized for each MS/MS experiment to be viewed at sufficient signal-to-noise ratios.

Reagents

rIAPP1-37 and hIAPP1-37 (Bachem, Switzerland) were used in the experiments. HPLC-grade H2O (Merck Ltd., Korea) and HPLC-grade CH3OH (Merck Ltd., Korea) were used as solvents. The peptides were dissolved in H2O:CH3OH (1:1, v/v) to prepare 6 × 10-5 M solutions. The solutions were prepared in this manner to achieve sufficient D5+ ion intensity in the CID-MS/MS experiments. The experiments were performed within 24 h of sample preparation.

Results and Discussion

MS Spectra

Under the present ESI experimental conditions, the mass spectrum of rIAPP indicated the presence of multiply charged monomers and oligomers (Figure 1). Monomer peaks were observed at m/z 1960.0, 1307.0, 980.5, and 784.6, indicating that the monomers had multiple proton adduct charges ranging from 2+ to 5+, [M+2H+] to [M+5H+]. The rIAPP sequence contains three basic residues (Lys1, Arg11, and Arg18) and an N-terminal position available for protonation. Depending on the gas-phase basicity, we speculate that protonation occurs at Lys1, Arg11, and Arg18 for M3+ (m/z 1307.0), and at the N-terminal amide for M4+ (m/z 980.5).24 Monomers with charges of 3+ and 4+ gave rise to particularly high-intensity peaks. The signal intensity of the M5+ (m/z 784.6) monomer was lower than that of the other monomer peaks, and it is unclear where the 5th protonation occurs. For the oligomers, peaks were observed at m/z 1568.2, 1120.4, 1680.1, and 1470.2, corresponding to D5+, D7+, T7+, and T8+ (T = trimer), respectively.

E1MPSV_2021_v12n4_179_f0001.png 이미지

Figure 1. ESI-MS spectrum of (a) rIAPP and (b) rIAPP+hIAPP mixed solution. Multiply charged monomers and oligomers are represented as Mz+, Dz+, and Tz+(M = monomer, D = dimer, T = trimer, and z = charge state). rIAPP and hIAPP are colored red and blue, respectively.

Figure 1b shows the MS spectra of a solution containing the two IAPPs, (rIAPP+hIAPP). The MS spectra of amyloidogenic hIAPP can be found in a previous study.23 The insets in Figure 1b show enlarged portions of the spectrum containing the signals for the dimers and trimers. The peaks are labeled with their respective (rIAPP: hIAPP) ratios. D5+(20) represents five positive charges of the homo-dimer complex (two rIAPP: zero hIAPP). In Figure 1b, the D5+(11) notation indicates that the hetero-dimer complex is composed of (one rIAPP: one hIAPP) with five positive charges. The T8+(21) notation means that the hetero-trimer complex is composed of (two rIAPP: one hIAPP) with eight positive charges.

In general, the intensities of the homo-dimer and homooligomer spectral peaks decreased when the second hetero- IAPP was added to enable the formation of hetero-dimer or hetero-oligomer complexes. In addition to the presence of homo-dimers and homo-trimers, the hetero-dimer D5+(11), hetero-trimer T8+(21), and T8+(12) were also observed as D5+(20), D5+(02), T8+(30), and T8+(03), respectively (Figure 1b). Here, homo-dimer D5+(02) or D5+(20) intensities decreased prominently compared to those shown in Figure 1a and in the previously published spectra.23 The observed complexes up to their trimer configuration and their m/z values are listed in the Electronic Supplementary Information Table S1.

MS/MS Spectra

MS/MS spectra of the monomers

CID-MS/MS experiments were conducted to obtain structural information on the parent rIAPP monomer or dimer ions in the early stages of aggregation. The MS/MS spectra of rIAPP monomers are shown in Figure 2. The fragment ions were labeled with various colors and shapes based on the fragment residue regions to compare the fragmentation patterns of each parent ion. The fragment ion assignments for the spectra in Figure 2 are presented in the Electronic Supplementary Information Table S2.

E1MPSV_2021_v12n4_179_f0002.png 이미지

Figure 2. MS/MS spectra of monomers of rIAPP (a) M2+, (b) M3+, (c) M4+, (d) M5+. The bu fragment series peak at u = 23–35 is indicated by a red square at the top of the peak and the yn fragment series peaks at u = 10–14 and u = 27–30 are indicated by red and green triangles, respectively.

In the MS/MS spectrum of rIAPP M2+ (Figure 2a), we observed three fragment ion series, ① singly charged y10– y14 ions, ② doubly charged y27–y30 ions, and ③ doubly charged b23–b35 fragment ions. The MS/MS fragmentation patterns are presented in Table 1. In the MS/MS spectrum of rIAPP M3+ (Fig. 2b), two fragment ion series were observed. Compared to the spectrum of rIAPP M2+, doubly charged y27–y30 ions in residues C7–Q10 were not observed in this case (Figure 2b). The dissociation process of residues C7–Q10 was regarded to be hindered because the Arg11 residue is protonated (H+) depending on the pKa value. The y10–y14 ions were commonly observed as singly charged fragment ions, regardless of the monomer charge states (M2+, M3+, and M4+) (Figure 2a–2c). It is worth noting that the b27 and y10 fragment ions were observed in the spectra shown in Figure 2a–2d as the main dissociation process between L27 and P28.

Table 1. Comparison of MS/MS fragmentation patterns of rIAPP monomers and dimers. The bu and yn ions were observed in the fragmentation of monomers, and bu, yn, (M+ bu), and (M+ yn) were the fragment ions for dimers.

E1MPSV_2021_v12n4_179_t0001.png 이미지

MS/MS spectra of the dimers

In the CID-MS/MS spectrum of rIAPP D5+ (Figure 3a), the fragmentation pattern arising from covalent bond dissociation was observed. Two main fragment ions, [M+ b27]4+ and [M+ y10]3+, were observed. Those ions are indicating that the peptide bond between L27 and P28 is a critical position in the dissociation process of rIAPP D5+. The [M+ y10]3+ is proposed to be a form of [M2++y10 1+]3+, based on the observation of y10 1+ ions (Table 1). The y10 ion was always observed in a singly charged state in the MS/MS spectrum of M2+–M5+ parent ions. Using the same logic, the b27 ion is expected to be in a doubly charged state in the [M+ b27]4+ fragment ion. The formation of M2+ monomer ion was expected to be conserved in the MS/MS dissociation process of rIAPP D5+. In other words, the fragmentation process was expected to occur in the M3+ monomer form in the MS/MS of rIAPP D5+. Therefore, the parent ion rIAPP D5+ is surmised to be composed of M2+ and M3+. As further evidence, in the CID-MS/MS experiments with broader mass isolation widths, it was observed that all of the rIAPP D5+ parent ions dissociated into M2+ (m/z 1960) and M3+ (m/z 1307) (Figure S1).

The two proposed rIAPP D5+ structures are shown in Scheme 1, based on the observation of (M+b27)4+ and (M+y10)3+ fragment ions. We postulate the coexistence of the two proposed configurations of D5+ (Scheme 1a and 1b) in the rIAPP solution. The coexistence of two rIAPP D5+ geometries could be a possible explanation for the observation of two peaks in the IMS-MS spectrum of rIAPP D5+.19

E1MPSV_2021_v12n4_179_f0004.png 이미지

Scheme 1. Schematic diagram of the proposed rIAPP D5+ structures.

The observation of (M+ bu)4+ fragment ion series, u = 23−34, suggests dimer interactions at the 1−22 residue region (Scheme 1a); this spectrum is consistent with dimer structures involving N-terminal helix-helix interactions, as shown in previous hIAPP studies.25 A significant contribution from the L12−L12 and F15−F15 interactions between the monomers has also been reported.26 The observation of (M+bu)4+ fragment ion series, u=23−34, could not provide insight into whether the hIAPP 1–22 region forms a helical coil or not. Nonetheless, the observation of the (M+bu)4+ series (u = 23−34) did indicate the presence of interactions in the 1–22 residue region, where the dissociation process of the dimer complex is expected to be hindered. The assumption of these 1−22 domain interactions between two rIAPP molecules was supported by the observation of the (b27 + b27)4+ fragment ion in the MS/MS/MS spectrum of the (M+ b27)4+ parent ion (Figure S2).

The observation of the (M+ y10)3+ and bu 2+ series (u = 23−27) in Figure 3a is inconsistent with the geometry shown in Scheme 1a. These observations suggest another dimer interaction at the 28−37 region (Scheme 1b), which is incompatible with the geometry shown in Scheme 1a. The dimer interaction in the 28−37 region has been proposed to occur in β-strand interactions between β- hairpin monomers19, 27 and in the disordered loop region of the monomers, 28 in which the C-terminal or 28–37 residue region is not amenable to the CID dissociation process. The inference of these C-terminal or 28–37 domains interaction between two rIAPP molecules was supported by the observation of the bu 2+ series (u = 23−27) and the absence of the bu 2+ ions in the 28–37 residues in the MS/ MS/MS spectrum of the (M+ y10)3+ parent ion (Figure S2).

E1MPSV_2021_v12n4_179_f0003.png 이미지

Figure 3. MS/MS spectrum of (a) rIAPP D5+ dimer and (b) hetero-D5+.[M+ bu]4+ fragment series peaks at u = 23–34 are indicated by a red square at the top of the peak and the (M+ yn)3+ and yn fragment series peaks at n = 10–14 are indicated by red and green triangles, respectively. rIAPP and hIAPP are indicated in red and blue, respectively.

The covalent and non-covalent bond dissociation was observed in the CID-MS/MS spectrum of rIAPP-hIAPP hetero-dimer D5+(11) parent ions (Figure 3b). Three monomer fragment ions (rIAPP2+, rIAPP3+, and hIAPP2+) were observed as products of the non-covalent bond dissociation process of the hetero-dimer D5+(11) complex. It is assumed that the D5+(11) complex comprises two possible hetero-dimer structures, (rIAPP3+ + hIAPP2+) or (rIAPP2++hIAPP3+). The formation of the (rIAPP3++ hIAPP2+) complex was supported by the observation of (hIAPP2++ rIAPP: b23–b27 ions)4+ and (rIAPP: y10–y14 ions)1+ fragment ions in the spectrum of Fig. 3b. The formation of the (rIAPP2++hIAPP3+) structure was supported by the observation of the (rIAPP2++hIAPP: bu)4+ fragment series from u=26 to u=35, which implies an interaction between hIAPP2+ and the 1−22 residue region of rIAPP3+. The fragmentation process was expected to occur in the 3+ monomer charge state, rIAPP3+ in (rIAPP3+ + hIAPP2+), and hIAPP3+ in the (rIAPP2++hIAPP3+) complex. It is worth noting that the absence of (hIAPP2+ + rIAPP: y10)3+ distinguishes the rIAPP-hIAPP hetero-dimer spectrum from the rIAPP homo-dimer MS/MS spectrum in Figure 3a.

Therefore, the (hIAPP2+ + rIAPP3+) hetero-dimer geometry is expected to be similar to that of the Scheme 1a structure, wherein N-terminal domain interaction was proposed for the rIAPP homo-dimer or hIAPP homo-dimer in previous studies.23 The proposed N-terminal domain interaction structure appears to be consistent with the 2D IR experimental results at 24 h after mixing rIAPP and hIAPP.6 Both the observed (rIAPP2+ + hIAPP:b26-35)4+ fragment ions and the absence of (hIAPP2+ + rIAPP: y10)3+ implied that the C terminal domain interaction geometry, which is similar to that of the structure in Scheme 1b, is not favorable in the hetero-dimer 25+(11) complex.

Conclusion

CID-MS/MS experiments were conducted to obtain structural information on the monomers and dimers formed in the early stages of rIAPP aggregation. The MS/MS spectra of the rIAPP monomers showed that there were two different fragmentation patterns, (M2+ fragmentation pattern) and (M3+, M4+ and M5+ fragmentation pattern), indicating that the rIAPP monomer structures depend on the charge state of parent monomer ion. Non-covalent or covalent bond dissociation was observed in the MS/MS spectra of [MrIAPP + MrIAPP]5+ or [MrIAPP + MhIAPP]5+. The fragment ions of [MrIAPP + MhIAPP]5+ for the non-covalent bond dissociation process, M2+ and M3+, indicated that [MrIAPP + MhIAPP]5+ hetero-dimer was composed of M2+ and M3+ monomers. The covalent bond dissociation patterns corresponding to (M+ b27)4+ and (M+ y10)3+ indicated the co-existence of two rIAPP D5+ configurations, proposed in Schemes 1a and 1b, in the rIAPP solution. The observation of (hIAPP2+ + rIAPP: b27)4+ and (rIAPP2+ + hIAPP: b27)4+ in the MS/MS spectrum of [MrIAPP + MhIAPP]5+ hetero dimer indicated that the geometry of Scheme 1a complex was favorable in the hetero-dimer complex structure.

Supplementary Information

Supplementary information is available at https://docs.google.com/document/d/1h9Lcp7xR3woZtmWzfoMJTQ5Q oo80gJ2c/edit?usp=sharing&ouid=1113531400147320509 56&rtpof=true&sd=true.

Acknowledgments

This paper was supported by Research Fund, Kumoh National Institute of Technology (2019-104-061).

Open Access

*Reprint requests to Ho-Tae Kim

https://orcid.org/0000-0002-1541-3081

E-mail: hotaekim@kumoh.ac.kr

All the content in Mass Spectrometry Letters (MSL) is Open Access, meaning it is accessible online to everyone, without fee and authors’ permission. All MSL content is published and distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org / licenses/by/3.0/). Under this license, authors reserve the copyright for their content; however, they permit anyone to unrestrictedly use, distribute, and reproduce the content in any medium as far as the original authors and source are cited. For any reuse, redistribution, or reproduction of a work, users must clarify the license terms under which the work was produced.

References

  1. Nanga, R. P. R.; Brender, J. R.; Xu, J.; Hartman, K.; Subramanian, V.; and Ramamoorthy, A. J. Am. Chem. Soc. 2019, 131, 8252, DOI: 10.1021/ja9010095.
  2. Green, J.; Goldsbury, C.; Mini, T.; Sunderji, S.; Frey, P.; Kistler, J.; Cooper, G.; Aebi, U. J. Mol. Biol. 2003, 326, 1147, DOI: 10.1016/S0022-2836(02)01377-3.
  3. Chiu, C.-C.; de Pablo J. J. AIP Adv. 2015, 5, 092501, DOI: 10.1063/1.4921073.
  4. Wang, H.; Ridgway, Z.; Cao, P.; Ruzsicska, B.; Raleigh, D. P. Biochemistry 2015, 54, 6704, DOI: 10.1021/acs.biochem.5b00567.
  5. Sellin, D.; Yan, L.-M.; Kapurniotu, A.; Winter, R. Biophys. Chem. 2010, 150, 73, DOI: 10.1016/j.bpc.2010.01.006.
  6. Middleton, C. T.; Marek, P.; Cao, P.; Chiu, C.-C.; Singh, S.; Woys, A. M.; de Pablo, J. J.; Raleigh, D. P.; Zanni, M. T. Nat. Chem. 2012, 4, 355, DOI: 10.1038/NCHEM.1293.
  7. Cao, P.; Meng, F.; Abedini, A.; Raleigh, D. P. Biochemistry 2010, 49, 872, DOI: 10.1021/bi901751b.
  8. Bharadwaj, P.; Solomon, T.; Sahoo, B. R.; Ignasiak, K.; Gaskin, S.; Rowles, J.; Verdile, G.; Howard, M. J.; Bond, C. S.; Ramamoorthy, A.; Martins, R. N.; Newsholme, P. Sci. Rep. 2020, 10, 10356, DOI: 10.1038/s41598-020-66602-9.
  9. Cooper, G. J. S.; Willis, A. C.; Clark, A.; Turner, R. C.; Sim, R. B.; Reid, K. B. M. Proc. Natl. Acad. Sci. USA. 1987, 84, 8628, DOI: 10.1073/pnas.84.23.8628.
  10. Hoppener, J. W.; Lips, C. J. Int. J. Biochem. Cell Biol. 2006, 38, 726, DOI: 10.1016/j.biocel.2005.12.009.
  11. Guo, J.; Sun, W.; Li, L.; Liu, F.; Lu, W. RSC Adv. 2017, 7, 43491, DOI: 10.1039/C7RA05742C.
  12. Nelson, R.; Sawaya, M. R.; Balbirnie, M.; Madsen, A. O.; Riekel, C.; Grothe, R.; Eisenberg, D. Nature. 2005, 435, 773, DOI: 10.1038/nature03680.
  13. Sawaya, M. R.; Sambashivan, S.; Nelson, R.; Ivanova, M. I.; Sievers, S. A.; Apostol, M. I.; Thompson, M. J.; Balbirnie, M.;Wiltzius, J. J.; McFarlane, H. T.; Madsen, A.; Riekel, C.; Eisenberg, D. Nature 2007, 447, 453, DOI:10.1038/nature05695.
  14. Jobling, M. F.; Stewart, L. R.; White, A. R.; McLean, C.; Friedhuber, A.; Maher, F.; Beyreuther, K.; Masters, C. L.; Barrow, C. J.; Collins, S. J.; Cappai, R. J. Neurochem. 1999, 73, 1557 DOI: 10.1046/j.1471-4159.1999.0731557.x.
  15. Haataja, L.; Gurlo, T.; Huang, C. J.; Butler, P. C. Endocr. Rev. 2008, 29, 303, DOI: 10.1210/er.2007-0037.
  16. Hebda, J. A.; Miranker, A. D. Annu. Rev. Biophys. 2009, 38, 125, DOI: 10.1146/annurev.biophys.050708.133622.
  17. Lin, C. Y.; Gurlo, T.; Kayed, R.; Butler, A. E.; Haataja, L.; Glabe, C. G.; Butler, P. C. Diabetes 2007, 56, 1324, DOI: 10.2337/db06-1579.
  18. Liu, N.; Duan, M.; Yang, M. Sci. Rep. 2017, 7, 7915, DOI: 10.1038/s41598-017-08504-x.
  19. Dupuis, N. F.; Wu, C.; Shea, J.-E.; Bowers, M. T. J. Am. Chem. Soc. 2011, 133, 7240, DOI: 10.1021/ja1081537.
  20. Young, L. M.; Cao, P.; Raleigh, D. P.; Ashcroft, A. E.; Radford, S. E. J. Am. Chem. Soc. 2014, 136, 660, DOI:10.1021/ja406831n.
  21. Laghaei, R.; Mousseau, N.; Wei, G. H. J. Phys. Chem. B. 2011, 115, 3146, DOI: 10.1021/jp108870q.
  22. Liang, G.; Zhao, J.; Yu, X.; Zheng, J. Biochemistry 2013, 52, 1089, DOI: 10.1021/bi301525e.
  23. Seo, J.-H.; Cha, E.; Kim, H.-T. Chem. Phys. Lett. 2018, 708, 61, DOI: 10.1016/j.cplett.2019.06.044.
  24. Dupuis, N. F.; Wu, C.; Shea, J. E.; Bowers, M. T. J. Am. Chem. Soc. 2009, 131, 18283, DOI: 10.1021/ja903814q.
  25. Wiltzius, J. J. W.; Sievers, S. A.; Sawaya, M. R.; Eisenberg, D. Protein Sci. 2009, 18, 1521, DOI: 10.1002/pro.145.
  26. Zhao, D. S.; Chen, Y. X.; Li, Y. M. Chem. Commun. 2015, 51, 2095, DOI: 10.1039/c4cc06739h.
  27. Mo, Y.; Lei, J.; Sun, Y.; Zhang, Q.; Wei, G. Sci. Rep. 2016, 6, 33076, DOI: 10.1038/srep33076.
  28. Buchanan, L. E.; Dunkelberger, E. B.; Tran, H. Q.; Cheng, P. N.; Chiu, C. C.; Cao, P.; Raleigh, D. P.; de Pablo, J. J.; Nowick, J. S.; Zanni, M. T. Proc. Natl. Acad. Sci. USA 2013, 110, 19285, DOI: 10.1073/pnas.1314481110.