Methylene blue inhibits nucleation and elongation of SOD1 amyloid fibrils

Stock solution preparation

Methylene blue (Sigma-Aldrich) was dissolved in deionized water in an approximate concentration of 10 mM and stored in the dark. The exact concentration was measured using a diluted MB solution with a UV-1800 Shimadzu spectrophotometer in a one cm pathlength cuvette (ε₆₆₄= 74,028 M⁻¹ cm⁻¹). The stock solution was filtered through a 0.2 µm syringe filter (Kinesis) before use.

SOD1 synthesis and purification

Recombinant C-terminally his-tagged SOD1, cloned in pET303 vector, was expressed in E. coli BL21 Star (DE3) strain (Invitrogen). Culture from stock was grown overnight in ZYM-5052 autoinduction media (Studier, 2005) at +37 °C. Cells were harvested by centrifugation for 30 min at 6 000 rpm (HeroLab), 4 °C. The culture was first homogenized mechanically and then sonicated (VS70/T probe, Bandelin) in 50 mM sodium phosphate, 100 mM NaCl pH 7.5 buffer. Cell debris was pelleted by centrifugation for 30 min at 18 000 rpm, 4 °C. The supernatant was mixed with pre-equilibrated, Ni ²⁺ ion loaded IMAC resin (GE Healthcare) and left at 4 °C to equilibrate. The resin was loaded into GE Healthcare HiScale 26/40 column, the column was connected to ÄKTApurifier chromatographic system and target protein eluted with 200 mM imidazole step gradient. ApoSOD1 was obtained as described (Chattopadhyay et al., 2008) with slight modifications. Briefly, SOD1 was dialysed in 100 mM acetate, 50 mM NaCl, 10 mM EDTA pH 3.8 buffer overnight; then in 100 mM acetate, 50 mM NaCl, 50 mM EDTA pH 3.8 buffer for 4 h and two times in 10 mM potassium phosphate pH 7.4 buffer. Dialysed protein was filtered through a 0.22 µm filter (Millipore) and stored at −80 °C.

Development of SOD1 aggregation assay

Wild-type metallated and disulfide-oxidized superoxide dismutase I is known to be an extremely stable enzyme with a melting temperature of 92 °C (Culik et al., 2018) and retains its activity in high concentrations of denaturing agents (Arnesano et al., 2004). Therefore, it is difficult to follow aggregation of wtSOD1 in vitro under physiological conditions in real-time. What is more, fibrillization kinetics of SOD1 are difficult to reproduce due to competing aggregation pathways and process stochasticity (Abdolvahabi et al., 2016). In this study, we have developed an assay allowing one to follow rapid and reproducible SOD1 kinetics at physiological pH (see ‘Methods’). One key component used in the assay is a disulfide reducing agent, since disulfide reduction in SOD1 leads to destabilization of the SOD1 dimer and facilitates fibrillization (Furukawa, Torres & O’Halloran, 2004; Sheng et al., 2013; Culik et al., 2018). Another key component is guanidine hydrochloride (GuHCl), a chaotropic agent that accelerates SOD1 aggregation (Sheng et al., 2013) and in our case, strongly promotes SOD1 fibrillization and prevents “off-pathway” oligomer formation (Fig. S1). In the absence of GuHCl, SOD1 forms round-shaped oligomers (Fig. S1A), whereas in the presence of 0.5 M GuHCl, thread-like amyloid fibers are formed (Fig. S1B). This most likely occurs due to chaotropic agent facilitated SOD1 unfolding and obtainment of a conformation that is prone to aggregation into intermolecular β-sheet rich amyloid fibers. In other words, GuHCl promotes the fibrillization pathway, but not the “off-pathway” oligomer formation. In our study, demetallated and disulfide bond reduced SOD1 is used, since demetallation, as well as disulfide bond reduction, leads to protein destabilization and an increased propensity to aggregate (Furukawa & O’Halloran, 2005; Sheng et al., 2013). To confirm that purified SOD1 is in the apo form, a thermal unfolding assay was performed (see Supplementary Methods section). The measured melting temperature of SOD1 is 50.3 ± 0.1 °C and corresponds to a disulfide-oxidized and demetallated form of the enzyme; this is in agreement with the data presented in literature (Fig. S2, Furukawa & O’Halloran, 2005). These SOD1 destabilizing conditions do not reflect the biological environment that the protein is found in vivo (apart from the pH value), however, theyare necessary to obtain aggregation kinetic data and examine the effect of MB in a reasonable timescale and with minimal stochasticity.

SOD1 aggregate preparation

wtSOD1 aggregates were prepared by diluting 200 µM protein in 10 mM potassium phosphate buffer containing 0.5 M GuHCl and 5 mM DTT. Solutions were incubated at 60 °C under shaking conditions at 600 rpm (Ditabis MHR 23 shaker, Fisher Scientific) for 72 h. Mixtures containing MB and/or preformed seed were prepared by adding 400 µM to the buffer and/or 5% pre-formed SOD1 fibrils respectively.

ThT fluorescence assays

Aggregation experiments were carried out with 200 µM apoSOD1 monomer in 10 mM potassium phosphate, 0.5 M GuHCl, 5 mM DTT buffer, 50 µM of Thioflavin T and 0–400 µM of MB. For seeded aggregation experiments, 5% of ultrasonicated fibrils were added into reaction mixtures and mixed thoroughly. Aliquots of 100 µl from aggregation mixtures were poured into a 96-well low-binding microplate (Corning 3881). ThT fluorescence (excitation at 442 nm, emission at 482 nm) was monitored at 60 °C in continuous shaking conditions every 5 or 10 min in Biotek Synergy H4 Multi-Mode microplate reader. Seeded aggregation experiments were carried out at 60 °C under quiescent conditions in RT-PCR analyzer (Rotor Gene Q).

Dye binding assays

Before any dye-binding assay, 2 ml of SOD1 aggregates were dialyzed 4 times in 1 L of 10 mM potassium phosphate buffer, pH 7.4 at room temperature. In Congo Red binding assay, 10 ul of aggregates are mixed with 1 ml 10 mM potassium phosphate buffer containing 10 µM Congo Red. The sample is incubated for 15 min at room temperature, then light absorption in the 400–700 nm range is measured (Shimadzu UV-1800 spectrophotometer). Nile Red binding assay is performed the same way, except that 10 µM of Nile Red is added to the buffer instead of Congo Red. The dye is excited with 530 nm light and fluorescence emission is measured at 550–800 nm range. In ThT assay, 50 µM of ThT is added instead of Congo Red; fluorescence emission is detected in 450–600 nm range after excitation with 440 nm light beam.

Atomic force microscopy

30 µl of the sample was deposited on freshly cleaved mica and left to adsorb for 5 min; followed by rinsing with distilled water and gentle drying using airflow. AFM images were recorded in the Tapping-in-Air mode at a drive frequency of 300 kHz, with the Dimension Icon microscope (Bruker).

Fourier transform infrared spectroscopy

A buffer of prepared fibers was exchanged into D₂O by a series of centrifugation and rinsing steps. Shortly, fibrils were pelleted in a microcentrifuge tube by spinning down at 12 000 rpm for 20 mins. The supernatant was removed and one mL of D₂O was poured on the pellet. The rinsing cycle was repeated 4–5 times, followed by resuspension of fibrils in 300 µl D₂O during the last step. Fibrils were then sonicated for 1 min at 20% MS72 probe power (Bandelin). FTIR spectra were recorded in a vacuum FTIR spectrophotometer (Bruker). Obtained spectra were processed by subtracting D₂O reference spectrum and then normalized using GRAMS/AI software (Fisher Scientific).

Data normalization

Experimental data were normalized using (1) equation: (1)${\displaystyle {\mathrm{I}}_{\text{norm}}=\frac{\mathrm{I}-{\mathrm{I}}_{\mathrm{a}}}{{\mathrm{I}}_{\mathrm{b}}-{\mathrm{I}}_{\mathrm{a}}}100}$

where I_norm—normalized ThT fluorescence, I—fluorescence intensity at a certain time point, I_a and I_b—minimum and maximum of fluorescence intensity.

Spontaneous SOD1 aggregation

The impact of MB was first investigated on the spontaneous aggregation kinetics of SOD1, using a ThT fluorescence assay (Fig. 1A). According to the data, SOD1 aggregation curves have a typical sigmoidal shape and in the control sample (Fig. 1A) the endpoint ThT fluorescence emission value is significantly higher than the background fluorescence, suggesting that long intermolecular β-sheet structures are being formed. The evidence of fibril formation is shown in an AFM image (Fig. 1B), where several µm-long fibrillar structures are visible. MB shows an inhibitory effect on SOD1 aggregation in a concentration-dependent manner, affecting the nucleation phase (as seen by the increase in lag time). Its effect on the quantity of fibrils formed cannot be determined accurately, as MB is known to partially quench the fluorescence of ThT (Ziaunys, Mikalauskaite & Smirnovas, 2019). Because of the possibility of MB interacting with ThT and affecting both the total fluorescence, as well as aggregation kinetic monitoring, the aggregation reaction was simulateneously tracked by measuring sample absorbance and ThT fluorescence (Fig. S3). Absorbance measurements had a good correlation with the ThT fluorescence assay. Due to this quenching effect, AFM was also employed to confirm aggregate formation and morphology. As the concentration of methylene blue reaches 15 µM, SOD1 is no longer capable to aggregate into amyloid fibrils within the timescale of the experiment (Fig. 1A and Fig. S3). Interestingly, in high enough concentrations of MB, SOD1 no longer aggregates into amyloid fibrils, but rather forms small oligomers/protofibrils, as seen from an AFM image (Fig. 1C). Furthermore, these small aggregates are not capable of binding several widely used dyes for amyloid detection (Figs. 1D–1F). In the presence of these aggregates, the absorbance spectra of Congo Red dye contains a single peak at 490 nm and the spectrum is nearly identical to the spectrum of Congo red in just the buffer solution, indicating the presence of an unbound dye, whereas upon fibril binding, a typical peak at 540 nm appears (Fig. 1D). Nile Red and ThT assays also do not demonstrate an enhanced fluorescence intensity when aggregates are formed in the presence of MB (Figs. 1E, 1F).

Seeded SOD1 aggregation

Seeded aggregation experiments were conducted to find out whether MB is capable of modifying the SOD1 aggregation pathway when pre-formed fibrils (seeds) are present. In the presence of seeds, SOD1 aggregates immediately without a lag time, (Figs. 2A–2F). At MB concentrations up to 50 µM, the process of aggregation reaches a plateau within 150 min (Figs. 2A–2C). At higher MB concentrations, the initial increase is followed by a drop in a ThT fluorescence emission signal (Figs. 2D–2F). The morphology of aggregates formed this way is shown in Fig. 2H. The most abundant species are small, round-shaped aggregates similar to those formed during the process of spontaneous aggregation (shown in Fig. 1C) and look significantly different from fibrils formed in a control sample (Fig. 2G). Several fibers, visible in Fig. 2H, could be traces of the primary fibrillar seed, used to initiate the elongation process. These aggregates, obtained from seeded fibrillization experiments, are also incapable of binding amyloid specific dyes Congo Red, Nile Red and ThT (Fig. 3). Unlike in the spontaneous aggregation experiment, where fibrillization was completely stopped at 15 µM MB, seeds were still capable of elongating at MB concentrations above 15 µM.

Seeding properties of non-fibrillar SOD1 aggregates formed with MB

To find out what are the seeding properties of aggregates formed in the presence of 400 µM MB, the following aggregation kinetics experiments were carried out. Various fractions (1–10%) of the aforementioned aggregates were used as seeds to initiate SOD1 monomer fibrillization (Fig. 4A). In particular, round-shaped aggregates, formed by incubating 200 µM SOD1 with 400 µM MB (shown in Fig. 1C), were mixed with monomeric SOD1 in the aggregation favoring buffer and fibrillization kinetics were followed by a ThT assay. As seen in Fig. 4A, ThT fluorescence remained at the baseline level even in the presence of 10% seed and no fibrils were formed (Fig. 4C), whereas in the control samples (SOD1 aggregation initiated by SOD1 fibrils) an immediate rise in ThT signal occurs and fibrils are observed by AFM (Fig. 4B).

MB effect on pre-formed SOD1 fibrils

In order to test the hypothesis that MB is capable of disassembling SOD1 fibrils, fibers were incubated with MB in a ratio of 1:2 under optimized aggregation conditions. According to AFM data presented in Fig. 5A, SOD1 fibers are present even after 3 days of incubation, indicating that once aggregation occurs, MB does not disassemble fibrils within the experimental timescale.

To determine alterations in the secondary structure of SOD1 fibrils upon incubation with MB, FTIR spectra of the aggregates were recorded (Figs. 5B, 5C). SOD1 fibers have a strong signal at 1,624 cm⁻¹ (with the main minimum of the second derivative at 1,622 cm⁻¹ and a weaker one at 1,638 cm⁻¹), indicating the dominance of an intermolecular β-sheet structure. A second derivative minimum at 1,669 cm⁻¹, corresponding to a polypeptide chain turn motif is also observed. On the contrary, in the sample of non-aggregated SOD1, a clear single peak at 1,633 cm⁻¹ (with the main minimum of the second derivative at 1628 cm⁻¹ and a weaker one at 1,646 cm⁻¹), is observed, indicating the dominance of a weaker intramolecular β-sheet structure (Barth, 2007). In the case of preformed fibrils that were incubated with MB for 3 days, the FTIR spectrum main maximum is at 1,632 cm⁻¹, while the second derivative major minimum is at 1,624 cm⁻¹, with a weaker one at 1,634 cm⁻¹. This implies that the MB-treated fibrils have a secondary structure that is in between the non-aggregated protein and the control fibrils. Interestingly, fibrils incubated with MB were also incapable of binding Congo Red, Nile Red and ThT (Figs. 5D–5F), similarly to previously described non-fibrilllar aggregates.

As MB affects SOD1 fiber dye-binding properties, which may be related to its interaction with the fibril surface, it is possible that MB-incubated fibrils possess altered seeding properties as well. Therefore, this hypothesis was tested by following seeded aggregation kinetics with a ThT fluorescence assay. According to the data (Figs. 6A–6C), MB-incubated SOD1 fibrils elongate at a similar rate as control fibers. These results show that upon incubation with MB, fibril ends remain as active aggregation centers and initiate self-replication of their structure. These ThT-positive aggregates are in the form of elongated fibers (Fig. 6E) and are morphologically similar to fibers formed in a control aggregation mixture (Fig. 6D).