Communication pubs.acs.org/JACS
Purple Fibrils: A New Type of Protein Chromophore Tatiana Quiñones-Ruiz,†,§ Manuel F. Rosario-Alomar,†,§ Karina Ruiz-Esteves,‡ Maruda Shanmugasundaram,† Vladimir Grigoryants,† Charles Scholes,† Juan López-Garriga,*,‡ and Igor K. Lednev*,† †
Department of Chemistry, University at Albany, SUNY, Albany, New York 12222, United States Department of Chemistry, University of Puerto Rico at Mayaguez, Mayaguez, Puerto Rico 00693, United States
‡
S Supporting Information *
eye at a high protein concentration. We observed a separation of the sample into two layers, white and purple, after 48 h of incubation (Figure 1A). AFM confirmed the presence of fibrils
ABSTRACT: A purple color is formed during the fibrillation of lysozyme, a well-studied protein lacking a prosthetic group. The application of Raman spectroscopy, electron paramagnetic resonance and UV−vis absorption spectroscopy indicates the formation of a sulfur∴π-bonded radical cation due to the methionine-phenylalanine interaction, which is consistent with a small molecule model reported in the literature. A purple chromophore with characteristic 550 nm absorption is formed due to a specific orientation of the sulfur-centered radical cation and a phenyl ring stabilized by the fibril framework. A specific fibril conformation and the resulting formation of the chromophore are controlled reversibly by varying the pH. This is the first known example of a side chain selfassembled chromophore formed due to protein aggregation.
P
roteins can undergo conformational changes from their native structures to the highly stable and well-organized βsheet-rich aggregates called amyloid fibrils.1 Amyloid fibrils have long been recognized for their association with neurodegenerative diseases such as Alzheimer’s and Parkinson’s, among others.1 So-called functional fibrils are known to play a positive biological role. For example, a variety of peptide and protein hormones are stored in a fibrillary form in the secretory granules of the endocrine system.2 Fibrils’ unique properties such as self-assembly, electrical conduction,3 comparable strength to steel,4 and silk-like mechanical stiffness4 make them attractive candidates for nanomaterial applications.5 Here, we report on the nature of the purple species formed as the result of hen egg white lysozyme (HEWL) fibrillation and investigated using normal and resonance Raman spectroscopy, electron paramagnetic resonance (EPR) and UV−vis absorption spectroscopy. All the obtained results are consistent with the formation of a three-electron bond between a Met sulfurcentered radical cation and Phe π electrons, which has also been reported for a small molecule6 and theoretically predicted for proteins.7 The fibrillation of 25 mg/mL HEWL was conducted in a 20% acetic acid and 0.2 M NaCl solution (pH 2.0) at 58 °C for over 48 h. Low pH and high temperature are typical conditions for HEWL fibrillation in vitro, whereas the presence of acetic acid is less common. HEWL fibril aggregates are insoluble in water and typically form a white precipitate visible by the naked © 2017 American Chemical Society
Figure 1. Purple and white fibrils form simultaneously and show different morphologies. (A) Purple and white layers of a HEWL fibril sample. AFM images of (B) fibrils from the purple layer (PF), (C) fibrils from the white layer (WF), (D) white spherical aggregates formed due to the disintegration of purple fibrils at pH 6 and (E) reassembled purple fibrils at pH 2.0; scale bars are 500 nm.
in both layers (Figure 1B,C). Elongated structures with a thickness of approximately 6 nm dominated the purple phase (Figure 1B). Thicker and shorter fibrils and their bundles were evident in the AFM image of the white phase sample (Figure 1C). The stability of the purple fibrils (PF) was investigated by changing the pH. PF prepared at pH 2.0 were centrifuged out and resuspended in a pH 6.0, 0.2 M NaCl, 5.2 mM acetic acid and 94.8 mM sodium acetate buffer solution. Figure 1D shows that upon increasing the pH to 6 the fibrils were disintegrated to spherical aggregates. Concomitantly, the purple chromophore was disrupted, turning the sample white. The resulting spherical aggregates could be oligomeric species that assembled the fibril in the first place, in agreement with Hill’s report on HEWL fibrils growing by gathering oligomeric species together under mild salt concentration.8 The obtained white spherical aggregates (Figure 1D) were centrifuged out from the pH 6 Received: March 27, 2017 Published: July 10, 2017 9755
DOI: 10.1021/jacs.7b03056 J. Am. Chem. Soc. 2017, 139, 9755−9758
Communication
Journal of the American Chemical Society solution, resuspended in the initial incubation solution (pH 2.0, 20% acetic acid, and 0.2 M NaCl) and kept at room temperature. PF reformed almost instantaneously with a morphology similar to that of the initial fibrils (compare images in Figure 1B,E). The cycle of pH changes was repeated several times, resulting in the same change in morphology and color of the HEWL aggregates. Taking into consideration that (1) the white fibrils (WF) are morphologically different from the PF and that (2) the PF disintegration results in chromophore disruption, we hypothesized that a unique fibril polymorphic structure of HEWL is responsible for the formation of the purple chromophore. This hypothesis was tested using deep ultraviolet resonance Raman (DUVRR) spectroscopy, which is a powerful technique for the structural characterization of amyloid fibrils.9 A typical DUVRR spectrum of proteins is dominated by the amide chromophore bands, whose shape and position are characteristic of the backbone conformation, and the aromatic amino acid bands. Figure 2 shows the DUVRR spectra of PF at pH 2 and the white spherical species resulting from a change to pH 6.
Figure 3. (A) Schematic representation of the S∴π bond model adapted from Chen et al.7 (B) Resonance Raman spectra of purple fibrils (PF) and nonresonance Raman spectra of white fibrils (WF).
shows the 514.5 nm excitation Raman spectra of the PF and WF, and Table S2 shows the peak assignments. The spectrum of the WF resembles a typical nonresonance Raman spectrum of protein fibrils, including the 785 nm excitation Raman spectrum of the PF and WF (Figure S1). These spectra are composed of the amide I (Am I), amide III (Am III) and the aromatic amino acid contributions (Trp, Tyr and Phe).9 Remarkably, the 514.5 nm excitation Raman spectrum of the PF is quite different from both the 514.5 nm Raman spectrum of the WF and the 785 nm spectrum of the PF, indicating a resonance enhancement of the first spectrum. The 514.5 nm spectrum of the PF contains several peaks, which can be assigned to Phe vibrational modes based on comparison with the Raman spectrum of Phe amino acid powder (Figure 3B). These include 1003, 1035, 1183, 1202, 1289, 1316, 1587 and 1605 cm−1 bands, which correlate well with known vibrational modes of Phe.15 The 1451 and 1377 cm−1 Raman bands in the 514.5 nm spectrum of the PF, which are not evident in the Phe spectrum, can be tentatively assigned to CH2 bending modes of Met.16 An intense band at 1668 cm−1 in the 514.5 nm spectrum of the PF can be assigned to the Am I vibrational mode of the peptide backbone, which is dominated by CO stretching vibrations. However, the peak position differs from that at 1673 cm−1 corresponding to the Am I peak from β-sheet structures in the Raman spectrum of the PF obtained using 785 nm excitation (Figure S1). This indicates that the number of groups with the 1668 cm−1 Am I vibrational mode is relatively small and that these carbonyls contribute to the resonance-enhanced Raman signal. Combining the results of resonance and nonresonance Raman spectroscopic studies, one can conclude that the side chains of Phe and Met, as well as a carbonyl group from the polypeptide backbone, are parts of the purple chromophore. This conclusion made us hypothesize that the purple color
Figure 2. DUVRR spectra of purple fibrils (bottom) and white aggregates resulting from purple fibrils disintegration at pH 6(top). The amide I vibrational mode (Am I) is dominated by CO stretching, with minor contributions from CN stretching and NH bending.10 The Am I B2 vibrational mode corresponds to the cross-β sheet core exclusively. Amide II (Am II) and Amide III (Am III) bands involve significant CN stretching, NH bending and CC stretching.11 The CαH bending vibrational mode involves CαH symmetric bending and CCα stretching.12
The DUVRR spectrum of the PF shows sharp intense amides I and II and the presence of the Am I B2 band, which are all characteristic of a well-organized cross-β core.13 The spectrum of white aggregates at pH 6 shows amide features consistent with the conversion of a fibrillary β-sheet to the predominantly unordered protein conformation. These features include a significant decrease in the intensity of the Am I10 and Am I B2 bands, and the broadening and shift of the Am III band to higher frequency.14 Therefore, we can conclude that the changes in pH resulted in protein structural changes and could disrupt a specific conformation required for the chromophore formation. Table S1 shows the DUVRR peak assignments. The UV−vis absorption spectrum of the dispersion of PF shows a band at approximately 550 nm, which is responsible for the observed purple color. No such band is evident in the absorption spectra of the WF and spherical aggregates. We utilized resonance Raman spectroscopy to probe the colored species directly. A 514.5 nm excitation, which is still within the 550 nm absorption band, resulted in resonance enhancement of the Raman scattering by the chromophore moiety. Figure 3B 9756
DOI: 10.1021/jacs.7b03056 J. Am. Chem. Soc. 2017, 139, 9755−9758
Communication
Journal of the American Chemical Society could originate from the Met sulfur-centered radical cation stabilized by the S∴π bond with a phenyl group as shown schematically in Figure 3A. The thioether in the side chain of Met is one of the most easily oxidized groups in proteins.17 The one-electron oxidation of the thioether group yields a radical cation that tends to stabilize through sulfur three electronbonded systems, with characteristic broad absorption bands in the visible region.18 Monney et al. have reported that such a radical cation forms in 6-endomethylthio-2endoarylbicyclo[2.2.1]heptane and exhibits an absorption band at 550 nm upon S∴π bond formation with π electrons from the aromatic group.6 This aromatic group, analogous to the Phe side-chain, has been shown to significantly decrease the oxidation potential of the thioethers in this model system, which becomes a new three-electron S∴π bond upon removal of an electron.6 The formation of a S∴π bond between a sulfurcentered radical cation of Met and the phenyl group of Phe should require a specific protein conformation that is consistent with a strong dependence of the purple color appearance on the fibril structure and morphology. In further justification of our hypothesis, Chen and co-workers have predicted using ab initio-time-dependent density-functional theory (TDDFT) calculations that Met-Phe S∴π bonding could result in the formation of hydrogen bonding between a hydrogen from the terminal methyl group of the Met side-chain and a carbonyl group (CO) from the peptide bond, in a two chain model where a two peptide-unit chain containing Phe interacts with the side chain of Met.7 In agreement with Chen’s prediction, our results show an enhancement of the Am I band at 1668 cm−1 in the 514.5 nm RR spectrum (Figure 3B) of the PF, indicating that the CO group from the polypeptide backbone (Figure 3, red line) is associated with the purple chromophore in the HEWL fibrils. It is also noteworthy that neighboring amide groups can facilitate thioether oxidation.19 The formation of a Met sulfur-centered radical cation stabilized by the S∴π bond assumes the presence of unpaired electrons in the PF. Electron paramagnetic resonance (EPR) spectroscopy is the method of choice for the detection and characterization of unpaired electron species. Figure 4 shows an explicit EPR signal and associated paramagnetism centered at a g value of approximately 1.97 observed exclusively for PF; no signal was evident for WF.
This EPR signal is not typical of carbon-based or nitrogenbased radicals, whose features would appear closer to g = 2.020 and would frequently show proton or nitrogen hyperfine structure. When EPR signals occur at g values significantly different from 2.00, the implication is that there could be a significant contribution from the electron orbital angular momentum or that the overall spin state could be different from 1/2.20 To eliminate the possibility that transition metal ions could be involved, although the observed signal does not show the typical nuclear hyperfine features associated with Mn(55)+2 or Cu(63,65)+2,21 the EPR experiments were repeated in the presence of the chelating agent EDTA. The purple species were also observed in the presence of EDTA, suggesting that metal ions are not involved in the development of the colored species and the EPR signal. The g values in the range 2.002−2.022 have been reported for Met sulfur radical cations produced by gamma irradiation of methionine,22,23 and g values in the range 1.9−2.5 have been reported for cysteine sulfur radicals produced by gamma irradiation of cysteine single crystals.24,25 Sulfur-radical g values greater than 2.00 have been attributed to the angular momentum contribution of sulfur 2p, and sulfur-radical g values less than 2.00 have been attributed to an admixture of sulfur 3d.20,24,25 The broad EPR signal of Figure 4 with the g value of 1.97 obtained for PF is consistent with a sulfur radical. Moreover, our EPR results agree with Hadley and Gordy’s suggestion, where the 3d shell is employed to expand the number of sulfur bonding orbitals to accommodate the additional three-electron bond.24 The UV−vis, Raman and Resonance Raman results reported here indicate a methionine cation radical in particular, stabilized by a S∴π bond with a neighboring Phe. Fibrils are excellent protein scaffolds with a well-organized structure that maintain a particular conformation due to their high rigidity/stiffness,4 which in turn could help to maintain specific positions of the groups involved in the purple chromophore formation. A reported binding energy of 11.2 kcal/mol for the S∴π bond between Met and Phe implies a good conductivity of the electron holes.7 Therefore, changes in the fibril structure (as we observed by changing pH) can cause the separation of Met and Phe residues, resulting in a shift of the electron holes to a different stabilizing group. Reforming the fibrils restores the correct orientation of Met and Phe and the S∴π bond. Our results show that a unique supramolecular structure of a particular HEWL fibril polymorph is suitable for the formation and stabilization of a novel self-assembled chromophore. This opens a new area of research, where some fibril polymorphs could stabilize Met sulfur radical cations, possibly avoiding a more damaging and reactive radical species cascade. For example, it has been proposed that the pathogenesis of Alzheimer’s disease is influenced by the oxidation of Met-35 in amyloid β (Aβ) peptide,26 which results in the reduction of metal ions producing reactive oxygen species when the Aβ peptide aggregates.27 A possible role of Phe side chain proximity to Met in Aβ peptides has been envisioned. It has been shown that Met-35 lies over Phe-19 in Aβ peptide solution,28 and this proximity was suggested to result in a modification of a reaction at the methionine,29 possibly promoting electron transfer.30
Figure 4. X-band EPR spectra of frozen PF and WF. 9757
DOI: 10.1021/jacs.7b03056 J. Am. Chem. Soc. 2017, 139, 9755−9758
Communication
Journal of the American Chemical Society
■
(22) Kominami, S. J. Phys. Chem. 1972, 76, 1729. (23) Werst, D. W. J. Phys. Chem. 1992, 96, 3640. (24) Hadley, J. H.; Gordy, W. Proc. Natl. Acad. Sci. U. S. A. 1977, 74, 216. (25) Matsuki, K.; Hadley, J.; Nelson, W. H.; Yang, C. J. Magn. Reson., Ser. A 1993, 103, 196. (26) Gu, M.; Viles, J. H. Biochim. Biophys. Acta, Proteins Proteomics 2016, 1864, 1260. (27) Butterfield, D. A. Free Radical Res. 2002, 36, 1307. (28) Shao, H.; Jao, S.; Ma, K.; Zagorski, M. G. J. Mol. Biol. 1999, 285, 755. (29) Varadarajan, S.; Yatin, S.; Kanski, J.; Jahanshahi, F.; Butterfield, D. A. Brain Res. Bull. 1999, 50, 133. (30) Chung, W. J.; Ammam, M.; Gruhn, N. E.; Nichol, G. S.; Singh, W. P.; Wilson, G. S.; Glass, R. S. Org. Lett. 2009, 11, 397.
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b03056. Experimental details and characterization data (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*
[email protected] *
[email protected] ORCID
Igor K. Lednev: 0000-0002-6504-531X Author Contributions §
T.Q.-R. and M.F.R.-A. contributed equally.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We are grateful to Professor Sanford Asher for valuable discussion and Dr. Vladimir Ermolenkov for technical assistance. This work was supported by the National Science Foundation under Grant No. CHE-1152752 (I.K.L.).
■
REFERENCES
(1) Stefani, M.; Dobson, C. M. J. Mol. Med. 2003, 81, 678. (2) Maji, S. K.; Perrin, M. H.; Sawaya, M. R.; Jessberger, S.; Vadodaria, K.; Rissman, R. A.; Singru, P. S.; Nilsson, K. P. R.; Simon, R.; Schubert, D.; Eisenberg, D.; Rivier, J.; Sawchenko, P.; Vale, W.; Riek, R. Science 2009, 325, 328. (3) del Mercato, L. L.; Pompa, P. P.; Maruccio, G.; Torre, A. D.; Sabella, S.; Tamburro, A. M.; Cingolani, R.; Rinaldi, R. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 18019. (4) Smith, J. F.; Knowles, T. P. J.; Dobson, C. M.; MacPhee, C. E.; Welland, M. E. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15806. (5) De Santis, E.; Ryadnov, M. G. Chem. Soc. Rev. 2015, 44, 8288. (6) Monney, N. P.; Bally, T.; Bhagavathy, G. S.; Glass, R. S. Org. Lett. 2013, 15, 4932. (7) Chen, X.; Tao, Y.; Li, J.; Dai, H.; Sun, W.; Huang, X.; Wei, Z. J. Phys. Chem. C 2012, 116, 19682. (8) Hill, S. E.; Miti, T.; Richmond, T.; Muschol, M. PLoS One 2011, 6, e18171. (9) Oladepo, S. A.; Xiong, K.; Hong, Z.; Asher, S. A.; Handen, J.; Lednev, I. K. Chem. Rev. 2012, 112, 2604. (10) Copeland, R. A.; Spiro, T. G. Biochemistry 1985, 24, 4960. (11) Myshakina, N. S.; Asher, S. A. J. Phys. Chem. B 2007, 111, 4271. (12) Wang, Y.; Purrello, R.; Jordan, T.; Spiro, T. G. J. Am. Chem. Soc. 1991, 113, 6359. (13) Xu, M.; Shashilov, V.; Lednev, I. K. J. Am. Chem. Soc. 2007, 129, 11002. (14) Mikhonin, A. V.; Bykov, S. V.; Myshakina, N. S.; Asher, S. A. J. Phys. Chem. B 2006, 110, 1928. (15) Castro, J. L.; Ramírez, M. R. L.; Arenas, J. F.; Otero, J. C. J. Mol. Struct. 2005, 744−747, 887. (16) Graff, M.; Bukowska, J. Chem. Phys. Lett. 2011, 509, 58. (17) Vogt, W. Free Radical Biol. Med. 1995, 18, 93. (18) Asmus, K.-D. In Sulfur-Centered Reactive Intermediates in Chemistry and Biology; Chatgilialoglu, C., Asmus, K.-D., Eds.; Springer US: Boston, MA, 1990; p 155. (19) Yamamoto, T.; Dai, J.; Jacobsen, N. E.; Ammam, M.; Hall, G. B.; Mozziconacci, O.; Schöneich, C.; Wilson, G. S.; Glass, R. S. Org. Lett. 2016, 18, 3522. (20) Gordy, W. In Theory and Applications of Electron Spin Resonance; Weissberger, A., Ed.; Wiley: New York, 1980. (21) Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of Transition Ions; Clarendon Press: Oxford, 1970. 9758
DOI: 10.1021/jacs.7b03056 J. Am. Chem. Soc. 2017, 139, 9755−9758
