Delineating the Aggregation-Prone Hotspot Regions (Peptides) in the Human Cu/Zn Superoxide Dismutase 1.

Amyotrophic lateral sclerosis (ALS) is a fatal, incurable neurodegenerative disease described by progressive degeneration of motor neurons. The most common familial form of ALS (fALS) has been associated with mutations in the Cu/Zn superoxide dismutase (SOD1) gene. Mutation-induced misfolding and aggregation of SOD1 is often found in ALS patients. In this work, we probe the aggregation properties of peptides derived from the SOD1. To examine the source of SOD1 aggregation, we have employed a computational algorithm to identify four peptides from the SOD1 protein sequence that aggregates into a fibril. Aided by computational algorithms, we identified four peptides likely involved in SOD1 fibrillization. These four aggregation-prone peptides were 14VQGIINFE21, 30KVWGSIKGL38, 101DSVISLS107, and 147GVIGIAQ153. In addition, the formation of fibril propensities from the identified peptides was investigated through different biophysical techniques. The atomic structures of two fibril-forming peptides from the C-terminal SOD1 showed that the steric zippers formed by 101DSVISLS107 and 147GVIGIAQ153 vary in their arrangement. We also discovered that fALS mutations in the peptide 147GVIGIAQ153 increased the fibril-forming propensity and altered the steric zipper's packing. Thus, our results suggested that the C-terminal peptides of SOD1 have a central role in amyloid formation and might be involved in forming the structural core of SOD1 aggregation observed in vivo.

Introduction

Amyotrophic lateral sclerosis (ALS) is a rapidly developing neurodegenerative disease that causes a slow degeneration and death of motor neurons, leading to death within 2 to 5 years.[1−3] Approximately 10% of ALS cases are familial (fALS), and of all fALS cases, 20% are caused due to mutations in Cu/Zn superoxide dismutase 1 (SOD1) gene. This ubiquitous 32 kDa homodimeric protein plays a key role in cellular defense against reactive oxygen species.[4] More than 180 ALS-related mutations have been described in SOD1 in the Amyotrophic Lateral Sclerosis Online Database [ALSoD (https://alsod.ac.uk/)]. Moreover, wild-type SOD1 on exposure to cellular stress has been suggested in ALS sporadic cases.[5,6] The pathogenesis of SOD1-related ALS is primarily associated with the toxic gain-of-function activity of the soluble misfolded SOD1 species,[7−10] resulting in various pathophysiological effects, including axonal degeneration and loss of axonal transport, impaired mitochondrial metabolism, excitotoxicity, proteasomal disturbance, and endoplasmic reticulum stress.[3,11] ALS is manifested in the misfolding and aggregation of SOD1 and accumulation of amyloid fibrils of misfolded SOD1.[12−15] The SOD1 aggregates formed in vitro share common toxic properties with ALS inclusions, such as inducing inflammation[16,17] and activating microglial cells.[18]

Proper folding of protein plays a significant role in neurodegenerative diseases.[19−21] It is primarily known that the exposure of aggregation-prone peptides drives the formation of amyloid fibrils. Documentation and characterization of these short aggregation peptide segments of proteins such as amyloid-beta (Aβ),[22] Tar-DNA binding protein-43 (TDP-43),[23] SOD1,[24,25] human islet amyloid peptide,[26] prion,[27] and p53[28] have significantly increased the understanding of protein aggregation mechanisms and associated toxicity, along with in designing peptide-based aggregation inhibitors.[29,30] The Eisenberg group[24] has identified four short peptides in SOD1, which showed high propensities to form fibrils and a high aggregation rate of these peptides in both wild-type and mutant SOD1. They have also shown that fALS mutations in these peptides modulate the aggregation of SOD1. In extension to the above study, we also examined the amyloid-forming tendency of SOD1 peptides using the Zipper-DB approach. The Zipper-DB program is based on the Rosetta algorithm and reports a steric zipper spine formation in a protein sequence.[31,32] Based on Zipper-DB prediction, SOD1 possesses four short peptides with high amyloid-forming propensities (Table ).

Table 1

Amyloid Forming Peptides Predicted from Zipper-DB in SOD1 Protein

zipper-DB segments	assigned as
14VQGIINFE21	P1
30KVWGSIKGL38	P2
101DSVISLS107	P3
147GVIGIAQ153	P4
147RVIGIAQ153	P4R
147GVTGIAQ153	P4T

These peptides correspond to the residues, 14VQGIINFE21 located in β2 strand (referred to as P1), 30KVWGSIKGL38 in β3 strand (referred to as P2), 101DSVISLS107 located in the Greek key loop (referred to as P3), and 147GVIGIAQ153 located in the C-terminal β8-loop region (referred to as P4) (Figure ). With this information, we next executed in vitro biophysical experiments to examine the amyloid fibril propensities of these peptides.

Figure 1

(A) Three-dimensional structure of SOD1 dimer (PDB code: 2C9V) indicating the fibrillogenic peptides shown in red. (B) Calculated Rosetta energies for each six-residue sequence were obtained from the Zipper-DB program. Peptides predicted to form amyloid fibrils are highlighted in red.

To determine how fALS mutations modulate the fibril-forming propensities, we also studied the aggregation of 147RVIGIAQ153 (G147R, referred to as P4R) and 147GVTGIAQ153 (I149T, referred to as P4T). Our results clearly showed that the C-terminal peptides P3 and P4 instantly form amyloid fibrils upon aqueous dilution, suggesting that these specific peptides might drive the aggregation of wild-type SOD1 and possibly initiate the pathology of ALS.

Results and Discussion

Amyloidogenic Peptides in SOD1

Similar to Eisenberg’s study, we also find four peptides of SOD1 predicted to form amyloid fibrils using the Zipper-DB method.[31] These four peptides are 14VQGIINFE21 (P1), 30KVWGSIKGL38 (P2), 101DSVISLS107 (P3), and 147GVIGIAQ153 (P4) (Figure ). Zipper-DB investigation indicated four key structural hot spots for amyloid formation, two of which are located in N-terminal β2 (P1) and β3 (P2) strands, and the rest two are located in the C-terminal Greek key loop (P3) and β8-loop (P4). Next, AMYLPRED2 predicts three amyloidogenic regions of SOD1 (6–7, 103–121, and 146–153) as amyloidogenic (Figures S1, S2). The program TANGO predicts residues 146–152 as amyloidogenic, whereas the WALTZ program predicts residues 15–23, 96–115, and 146–154 as amyloidogenic residues. Furthermore, prediction of change in the aggregation rate upon mutation through AggreRATE-Pred server[33] indicates that P4R (147RVIGIAQ153) has a higher aggregation propensity than P4T (147GVTGIAQ153).

Structural Changes during Aggregation

CD spectroscopy is used to investigate the secondary structural changes during aggregation.[27,34−36] The P1 peptide (Figure A) displayed a negative minimum at 220 nm that represents an α/β content in the structure. The negative minima started to lose upon incubation, and the peptide P1 lost its native secondary structure within 5 min. As the incubation time increased, P1 peptide showed clearly visible aggregates. P2 peptide showed a negative minimum at ∼205 nm characteristic of a random coil conformation and a positive maximum at ∼222 nm, indicating the presence of triple-helical conformation (Figure B). Upon incubation, P2 displayed a marked decrease in the negative ellipticity, suggesting the loss of native secondary structure. In addition, the positive peak shifted to a lower wavelength with increasing incubation of peptides, suggesting a slight change in the packing of the triple-helical structure. P3 peptide initially showed negative minima at ∼220 nm suggesting the existence of α/β sheet dominant structure. After 5 min of incubation, negative minima at ∼220 nm completely disappeared, and the peptide P3 lost its native structure entirely after 30 min of incubation (Figure C). P4 peptide showed two new negative minima at ∼222 and ∼214 nm and a shoulder at ∼230 nm in the region of far-UV CD spectrum (Figure D).

Figure 2

Secondary structure of SOD1 amyloidogenic peptides were analyzed by far-UV CD spectroscopy before (black line) and after incubation at 37 °C with shaking of 200 rpm for 5 min (red line), 30 min (green line), and 60 min (yellow line). Structural transformation of (A) P1, (B) P2, (C) P3, (D) P4, (E) P4T, and (F) P4R peptides.

Upon incubation for 5 min, P4 peptide showed a loss of native secondary structure and after 30 min, P4 peptide showed a marginal shoulder at ∼230 nm and two positive signals at ∼215 and ∼208 nm, respectively. Remarkably, the redshift was observed from ∼230 to ∼225 nm during incubation. P4T peptide displayed a negative minimum at ∼220 nm, representing an α–β sheet dominant structure (Figure E). Upon incubation for 5 min, P4T peptide showed two new negative minima at ∼218 and ∼208 nm and a redshift from ∼222 to ∼218 nm, suggesting the presence of β-dominated species. After 30 min of incubation, P4T peptide displayed a significant loss of the negative ellipticity at ∼218 nm, suggesting the loss of native structure, leading to aggregation. Besides, the negative peak at ∼208 nm remains unchanged and decreases only after 60 min of incubation.

P4R peptide showed a negative peak at ∼212 nm and the region at ∼220 nm in the CD spectrum, representing that the peptide forms a weak α/β structure (Figure F). Upon incubation for 5 min, P4R peptide showed redshift from ∼212 to ∼208 nm characteristic of β-sheet structures. After 30 min of incubation, the P4R peptide showed two new negative minima at ∼216 and ∼209 nm, respectively, indicating the presence of a fibrillar structure rich in the β sheet. However, after 60 min of incubation, the β sheet-rich structures convert into random coils with a minimum at 204 nm (Figure F).

The SOD1 peptides indicated significant differences in their secondary structures for the duration of aggregation, as observed using far-UV CD over time. The far-UV CD spectra of P3 and P4 indicated a significant decrease in the β-sheet content with time (Figure C,D). On the other hand, the mutations in the P4 peptides also exhibit different structural transitions as the β-sheet content of P4R decreases much quickly compared to P4T, which shows no significant loss of structure till ∼1 h (Figure D,E).

Thioflavin-T Binding to Aggregated Peptides

The amyloid formation was monitored by amyloid-specific dye Thioflavin T (ThT) fluorescence. After 1 h of incubation, all the SOD1 peptides showed very high binding with ThT, as revealed by the increase in the fluorescence intensity at ∼ 485–490 nm, signifying amyloid fibrils (Figure ). The peptides P1 and P2 (Figure A,B) show less ThT binding relative to P3 and P4 (Figure C,D). The maximum increase in ThT binding upon incubation was observed in P3, P4, P4T, and P4R (Figure C,F), indicating the formation of amyloid fibrils during aggregation. P1 and P2 peptides did not show any remarkable spectral change upon incubation at 37 °C.

Figure 3

ThT binding to SOD1 fibrillogenic peptides. ThT fluorescence intensity of (A) P1, (B) P2, (C) P3, (D) P4, (E) P4T, and (F) P4R incubated at 37 °C with shaking of 200 rpm as a function of incubation time for 5 min (black line), 30 min (red line), and 60 min (green line).

The increase in ThT fluorescence intensity upon incubation suggests fibrillar structures or β-sheet structures in these fibril oligomers. ThT fluorescence study indicates that P3, P4, and P4 mutations (P4R and P4T) form ThT positive-fibrils that appeared at 37 °C instantly after mixing with water.

FTIR Spectroscopy

The secondary structure of fibrils was next examined by Fourier transformation infrared (FTIR) spectroscopy. SOD1 peptide fibrils showed comparable FTIR spectra that recommend similar β-sheet dominant structures.[28,37] The amide I band of the FTIR spectrum (1700–1600 cm–1) corresponds to the absorption of the carbonyl peptide bond group of the main protein chain and is sensitive to variations in the secondary structure content protein. In all cases, except P3, a band at ∼1682 cm–1 leads the spectrum, which indicates the existence of an anti-parallel intermolecular β-sheet structure (Figure ). Such β-sheet structures are characteristic signatures of amyloid-forming parts and are well known as steric zippers.[31]

Figure 4

Secondary structure profiles of SOD1 peptide fibrils examined by ATR–FTIR (1000–2000 cm–1) of amyloid fibrils formed in vitro. (A) P1, (B) P2, (C) P3, (D) P4, (E) P4T, and (F) P4R suspension of peptide fibrils.

In contrast, P3 spectra show a dominant band at ∼1625 cm–1, indicative of intermolecular β-sheet structures, in which the self-assembled β-strands adopt a parallel disposition. Additionally, hydrogen bonding in β-sheets of P1 appeared stronger than in fibrils of P4T or P4R, as indicated by the band’s position at 1625 cm–1 compared to 1628 cm–1 and 1601 cm–1 for fibrils. Overall, these findings indicate amyloid fibrils enriched in the β-sheet structure in the case of peptide aggregation.

X-Ray Diffraction of Peptide Fibrils

More validation of the fibrillar properties of the aggregated peptides was obtained by X-ray diffraction (XRD) experiments. XRD patterns of P3 (Figure A) and P4 (Figure B) showed two spherically averaged reflections, at spacings of ∼15.3/6.1 and ∼8.0/6.1 Å, respectively. These reflections are the characteristic of the inter-sheet and hydrogen-bonding distances, respectively, for β-sheet conformation. The XRD studies of P4T show that the amyloid fibrils have a characteristic “cross-β”-like architecture (Figure C). Strictly, a very strong periodicity observed at 6.1 and 9.0 Å indicates the inter-strand and inter-sheet distances of β-sheet arrangements, respectively. In P4R, P1, and P2, a strong reflection at ∼8–10 Å indicates the presence of cross-β-like conformation (Figure D,F). The XRD results of peptides run in duplicate are also shown in Supporting Information (Figure 1). The results showed a similar XRD pattern of fibrillogenic peptides. These structural features are unique signatures observed in many amyloid fibrils where the β-strands are at right angles to the fibril axis and the β-sheets are made straight parallel to the fibril axis.[36,38]

Figure 5

XRD pattern of SOD1 preformed fibrils. (A) P1, (B) P2, (C) P3, (D) P4, (E) P4T, and (F) P4R peptides. The reflection at ∼ 5.0 Å corresponds to the repeat distance of β-strands aligned perpendicularly to the fiber axis, whereas the reflection at ∼10.0 Å is attributed to the repetitive distance between packed β-sheets.

Next, the aggregation study of peptides revealed that all the four peptides form β-sheet-dominated structures but individually in P3 and P4, peptides form amyloid fibrils mainly, as indicated by ThT binding (Figure ) and XRD (Figure ) studies. This result is also corroborated with the findings from the Eisenberg study, which has shown that N-terminal peptides (P1 and P2) do not form the β-sheet spine of the SOD1 fibrils. In contrast, C-terminal peptides (P3 and P4) are involved in fibril nucleation and growth and form the β-sheet spine of the SOD1 fibrils. Additionally, results from the FTIR study of peptide fibril samples at the area of amide I (1500–1700 cm–1) showed the existence of β-sheet-dominated structures (Figure ).

Therefore, the results presented in this study suggest that in both wild-type SOD1 or mutant SOD1, P3 and P4 peptides are common molecular determinants of SOD1 aggregation observed in ALS cases and thus represent an excellent therapeutic target. Moreover, the peptide P1 contains Asn19, and it has been shown that deamidation of asparagine to aspartate resulted in the formation of amyloid fibrils faster than WT SOD1.[39]

Morphology of Aggregates Monitored by Atomic Force Microscopy

The analyses of the morphology of the different aggregates formed by incubation were monitored through atomic force microscopy (AFM). The images of peptide samples as a function of incubation time at 37 °C were measured by AFM in the tapping mode and are shown in Figure . The differences in structures of the peptide aggregates were further reflected in morphological features of the aggregates ranging from small globular aggregates, oligomers, and fibril’s structure along with the presence of irregular aggregates, as revealed by AFM studies.

Figure 6

AFM images of different aggregates formed en route to fibrilization. The area of measurement is 3 × 3 μM, and scale bar represents 1 μM length. The area corresponding to each stage has been magnified for a clearer view. (A–E) P3 incubated at 37 °C and shaking of 200 rpm for 30 min, 1 h, 4 h, 24 h, and 48 h, respectively. (F–J) P4 incubated at 37 °C and shaking of 200 rpm for 30 min, 1 h, 4 h, 24 h, and 48 h, respectively.

Within 30 min of incubation, P3 began to show short protofibrils, representing pre-fibrillar forms (Figure A) which significantly grow into elongated fibrils by 1 h (Figure B). At 4 h, the fibril length increased dramatically along with the lateral growth of individual fibrils. The length and thickness of the fibril increased upon further incubation with the appearance of thinner rods without branch points and junctions (Figure C–E). In the case of P4 peptide, P4 began to show that small oligomers of irregular shape were observed for 30 min of incubation (Figure F), which transformed into a small proto-fibrillar form at 1 h of incubation (Figure G). With further incubation of 4 h, fibrillation was observed with the presence of elongated fibrils where new filaments branched out from the previously formed fibrils (Figure H). Incubation up to 24 h resulted in the formation of long protofilaments, where some of them showed interwinding (Figure I). The average height and thickness of these filaments increased with further incubation of 48 h, resulting in long and thick cord-like fibers (Figure J). The AFM structures of the peptide fibrils indicated that the steric zippers formed by P3 and P4 differ in their arrangement.

It has been shown that the presence of short fibrillogenic peptides alters the aggregation kinetics of the full-length protein by providing a template for fibril nucleation.[40,41] Consistent with these findings, Ivanova et al.(24) also showed that the peptide P4 accelerated the fibril formation of the wild-type apo SOD1 and apoSOD1G93A mutant in a dose-dependent manner and with variable aggregation kinetics. The different aggregation kinetics and morphology of aggregates of the peptides and the corresponding protein have been demonstrated in many studies.[13,24,42,43]

Conclusions

The pathogenesis of ALS is associated with the accumulation of misfolded and fibrillation of SOD1. Many studies have shown several amyloid-forming segments in SOD1. In the present study, we have identified possible amyloid-forming segments in SOD1 by the Zipper-DB profile method. Over 180 mutations have been known in SOD1 which can cause fALS, and many of those mutations promote the aggregation of SOD1 by destabilizing the native structure of the protein. We show that the fALS mutations (G147R and I149T) modulate the amyloid-forming propensity of the aggregation-prone peptide, 147GVIGIAQ153.

Moreover, the structure of the mutant segments P4R (G147R) and P4T (I149T) revealed differences in their molecular arrangements in the steric zipper organization. This aggregation-prone peptide is natively buried in the dimer interface and is thus protected from exposure. The exposure of this segment upon mutation, oxidative stress, or disulfide reduction indicates the destabilization of the β-barrel fold and subsequently misfolding and aggregation. Moreover, previous studies have shown that this peptide accelerates the aggregation of both the wild-type SOD1 and the fALS mutant apoSOD1G93A.[24,44]

Materials and Methods

Prediction of Amyloid-Forming Peptides in SOD1

The aggregation-prone regions in the primary sequence of human SOD1 (Uniprot id: P00441) were predicted first from Zipper-DB program (https://services.mbi.ucla.edu/zipperdb/), which predicts the peptides forming the steric zipper spines of amyloid fibrils. Next, the aggregation regions in the SOD1 sequence are also predicted with the AMYLPRED2 tool (http://aias.biol.uoa.gr/AMYLPRED2/), which utilizes a consensus of 11 different algorithms to predict aggregation regions in the protein.[45]

Peptide Synthesis

The SOD1 peptides, 14VQGIINFE21, 30KVWGSIKGL38, 101DSVISLS107, 147GVIGIAQ153, 147RVIGIAQ153 (P4R), and 147GVTGIAQ153 (P4T), were synthesized by GL Biochem (Shanghai, China). Reverse-phase HPLC and mass spectrometry assessed the purity of the peptides. The concentration of the peptides in the powder form was expressed in dry weight and was stored at −20 °C until use. The working buffers and solutions were prepared the same day before the experiments.

Materials

The deionized water was obtained using ultra filtration apparatus (Milli-Q gradient A101). For all aggregation experiments, type I purified water was used. All reagents, including ThT, were of investigative grade and obtained from Sigma-Aldrich.

Aggregation of Peptides

The aggregation of peptides was initiated by dissolving 10 mg/mL of peptides in Milli-Q water of pH 5.6 with shaking at 200 rpm/min overnight at room temperature to form the fibrils of the peptides.

Circular Dichroism

The changes in the secondary structure of peptides for the aggregation period were monitored through far-UV circular dichroism (CD) spectroscopy performed on a Jasco-1500 CD spectropolarimeter (Jasco, Tokyo, Japan). The CD spectra were obtained with a 1 cm/min scan speed, and data points were collected from 250 to 200 nm at room temperature with a 0.1 cm path length of the quartz cell. Averages of three repetitive scans are the resultant of all data points. The peptides were dissolved in autoclaved distilled water, and the data obtained are presented in milli degrees and plotted against the wavelength (nm).

ThT Fluorescence Assay

ThT stock solution was in 50 μM in phosphate-buffered saline (PBS) (50 mM, pH 7.0). ThT was freshly prepared and adequately covered to avoid its exposure to light. For the fluorescence study, we took 400 μL of aged peptide solution mixed adequately with 200 μL of solution of ThT and the total working volume of 1000 μL with PBS. ThT fluorescence experiments were achieved with a Jasco spectrofluorometer (Model FP-6200) with a 3 mm path length quartz cuvette with excitation and emission at 5 nm bandwidth slits. Samples were excited at 440 nm wavelength, and the emission spectra were recorded between 460 and 560 nm at 37 °C. For reproducibility check, the experiment was performed twice.

Fourier Transformation Infrared Spectroscopy

We used FTIR spectroscopy to characterize the secondary structural and transformation throughout proteins and polypeptide aggregation. The FTIR spectral data of polymers are usually interpreted in terms of the vibrations of a structural repeat. 2 mg of all peptides was solubilized in autoclaved distilled water. All peptides were incubated at room temperature, and 5 μL of each incubated peptide was utilized for the FTIR measurement. The attenuated total reflection (ATR)–FTIR spectra were recorded at a resolution of 4 cm–1 on a Bruker Tensor 27 FTIR spectrometer (Bruker Optik GmbH, Germany).

X-Ray Diffraction

All samples of peptide 1 mg/mL were dissolved at room temperature for 4 days. The droplets (10 μL) of incubated peptides were kept between two siliconized capillaries properly aligned with 2 mm apart and with wax-covered ends. The droplets of each peptide were dried slowly at room temperature to form an oriented fibril of peptides. The XRD experiment was then performed using a Rigaku Ultima-IV X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å) between 2θ of 3 and 20° at 40 kV and 40 mA.

Atomic Force Microscopy

The peptide samples for AFM experiments were prepared on freshly cleaved mica by incubating a 10 μL drop of peptide solution for 30 min. The samples were then washed with Milli-Q water and dried under a nitrogen flush. AFM images were obtained in the tapping mode using a Bioscope Catalyst atomic force microscope (Bruker Corp., MA), attached with a Nanoscope V controller. All images were processed and analyzed using nanoscope analysis, v.1.4 tool. The AFM imaging experiments were performed at room temperature with a room humidity of 40% or less. The images contain 512 × 512 points in 3 × 3 μm area.

Data Availability Statement

All data generated or analyzed during this study are included in this manuscript and Supporting Information attached to this article.