Antioxidants from the class of polyphenols, NAC and Vitamin C act differently on amyloid fibril formation by two human cystatins: stefin B and cystatin C

We compare the effect on amyloid fibril formation by two homologous proteins from the family of cystatins, human stefin B (stB) and cystatin C (cysC) in presence of 3 polyphenols: curcumin, resveratrol and quercetin and 2 non-phenolic anti-oxidants: vitamin C (VitC) and N-acetyl cystein (NAC). Some of the experimental data have already been presented, here we compare, further discuss and highlight the results. The amyloid fibril formation was followed by ThT fluorescence and transmission electron microscopy. Inhibitory effects on amyloid fibrillation reaction depended on anti-oxidant class and concentration. The fact that different effect of polyphenols was observed with the two cystatins; Cur acted inhibitory on stB but not on cysC fibril formation, could be explained if the 3 polyphenols would not bind to the same binding site in the fibrils core. Other differences are pointed out and discussed. Synergistic effects of VitC and chosen polyphenols on amyloid fibrilllation of human stB have been explored and are reported here for the first time. Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 13 October 2020 doi:10.20944/preprints202010.0274.v1 © 2020 by the author(s). Distributed under a Creative Commons CC BY license.


INTRODUCTION
Neurodegenerative diseases, among them most prevalent Alzheimer's and Parkinson's disease, stem from protein misfolding and aggregation. Misfolded proteins form deposits extracellularly and inclusions inside cells, including neurons. The soluble fractions of these aggregates are more toxic to the cells than fibrils and it is believed that they interact with membranes and even perforate them. Many indirect membrane oligomers interacton studies confirm this notion, among them in silico and structural studies 1 as well as biophysical lipid protein interaction studies 2 . However, it remains to see amyloid pores in living cells.
As a model for folding and later mis-folding and aggregation to amyloid fibrils we have studied human stefin B in comparison to a more stable stefin A and numerous of its mutants and chimeric variants. Solving 3D structure of the tetramer of stefin B by both X-rays and by NMR in solution, we have shown that the oligomers are domain-swapped and that the tetramer is composed of two domain-swapped dimers formed in a process of loops exchange 3 . Cystatin C also forms domain swapped dimers and their structure has been determined by X-ray and NMR in solution, respectively 4,5 . In the mechanism of stefin B amyloid fibril formation, we have studied the role of oligomers in the process, wher the lower oligomers might be off-pathway 6,7 . Interaction of stefin B oligomers with membranes was demonstrated in vitro. It was shown that the size, morphology and interaction of the oligomers with lipid membranes correlate with toxicity 8-10 , resembling pore forming toxins 11 .
Under normal physiological conditions, human stefin B (alternative name is cystatin B), as a housekeeping gene, acts as intracellular cathepsins inhibitor. It is localized in the cytosol but also in the nucleus 12 . Stefin B protein was found multimeric in cells 13,14 , which suggests it may have alternative functions than cysteine proteases inhibition. Its oligomers were shown to interact with cytoskeletal proteins 13 and to bind Aβ 15 , suggesting a chaperone-like function 16,17 . Recently, the Italian group has demonstrated that stefin B (cystatin B) is important in physiology of the synapse 18 and is essential for proliferation and interneuron migration 19 . When cystatin B gene is mutated and either loses activity or aggregates or both, it causes a rare progressive myoclonus epilepsy -EPM1 20,21 , spred in the Baltics and some parts of the Mediteranian.
Human cystatin C mutation L68Q also causes a severe disease, hereditary cystatin C amyloid angiopathy (HCCAA). HCCAA is a rare autosomal dominant genetic disease observed in a small part of population of Iceland, only. The amyloid deposits of the L68Q mutated CysC, which accumulate in the brain arteries, weaken arterial walls, which leads to repeated brain hemorrhages (mini-strokes) and may result in dementia and paralysis 22 . Cystatin C also binds Aβ [23][24][25] and was found neuroprotective, loaded in extracellular vesicles 26 . Not at least, cystatins are constituents of amyloid plaques 19 .
As we are studying antioxidant effects on protein aggregation it is worth to mention that stefin B was shown to bind Cu2+ in contrast to stefin A 27 . The binding of the metal influenced protein aggregation, it inhibited amyloid fibrils but increased granular and amorphous aggregates 27 . It is known that deposits of amyloid fibrils interact with metals, such as Cu2+ and Fe3+ and raise free radicals in the Fenton's reaction, generating H2O2. Externally applied antioxidant substances can help the body to fight free radicals. Sometimes antioxidants can become pro-oxidant under certain conditions, such as transition metals presence, generating by themselves H2O2 as is the case with vitamin C (ascorbic acid) at higher concentrations 28 . Indirectly, some substances could also act as antioxidants if they would reduce protein aggregation. The idea that polyphenolic compounds reduce protein aggregation by direct binding to amyloid core structure is not new 29 .
Here, we highlight and discuss our own studies of the influence of chosen anti-oxidants on protein aggregation to amyloid fibrils, comparing 2 different human proteins. We studied amyloid fibril formation in presence of polyphenolic compounds, vitC and NAC of human stefin B (stB) and cystatin C (cysC) from the family of cystatins and found different behavior. We try to explain these observations by subtle differences in protein sequence and by several binding sites to different antioxidant substances.

RESULTS AND DISCUSSION:
1. Effect of curcumin on stefin B fibrillation; kinetics and TEM measurements.
We followed the kinetics of amyloid fibril formation of stefin B in presence of various concentrations of curcumin (Cur) and without this substance -as a control, by using ThT fluorescence. Figure 1 shows the fibrillation reaction of stefin B as a function of time for various concentrations of Cur.
Stefin B alone was made to fibrillate at 25 o C, pH 4.8, adding 10% 2,2,2, trifluoro-ethanol (TFE) . as described before 30 and is given in more details under Methods. Inhibitory action on amyloid fibrils formation is most expressed at lower concentrations (1 µM Cur and 4 µM Cur), where the lag phase as compared to stB alone is prolonged to 250 hours. The lag phase prolongation at presence of 10 µM Cur is 150 hours, while at higher Cur concentrations (20 µM Cur and 50 µM Cur) it is even less, only 50 hours.
The intensity of ThT fluorescence is substantially lower at all Cur concentrations apart from the lowest (1 µM Cur), which might indicate inhibition of fibril growth, i.e., that the amount of the amyloid fibrils by stB is less in presence of Cur as compared to stB alone. However, one should be aware that quenching effects might be present on ThT fluorescence by aromatic and polyphenolic compounds 32 , therefore, another method is needed to check the inhibitory effect.
To check the morphology and to some degree also amount of the fibrils we conducted transmission electron microscopy (TEM) experiments. The samples of stB with Cur, which showed the biggest decrease in ThT intensity (50 µM Cur) as compared to stB alone in Figure 2A are shown in Figure 2B.
It can be seen that there are less fibrils and that they are shorter yet thicker. At the lowest Cur concentration (1 µM Cur) fibrils are very scarce, instead, granular aggregates are abundant ( Figure   2C), which agrees with the longest lag phase and the strongest inhibition. Intensity of ThT fluorescence of stB amyloid fibrils at the plateau (Figure 1), therefore, does not reflect their true amount ( Figure 2B); at the 2 highest Cur concentration fluorescence intensity gets lowered due to quenching phenomenon. 2. The effect of resveratrol on stefin B fibrillation; kinetics and TEM measurements.

Effect of quercetin on stefin Bfibrillation
As studied by Hasanbasić et al. 31

Comparison of the effect of 3 polyphenols to stefin B and cystatin C amyloid fibril formation
As shown by Hasanbasic et al. 31 , representative data are presented again in Figures 1 and 3  Regardless of the similarly structured amyloid core with the cross-β structure, this may be explained by an assumption that sequence details and precise arrangement of the subunits in the protofilaments influence aromatic-aromatic π stacking interactions 34 . Indeed, precise arrangement of protofilaments can influence π stacking 35, 36 and the same protein can form polymorphic amyloid fibrils, which have specific chemical and biological properties 37, 38 .
5. The effect of of N-acetylcystein (NAC) on stefin B fibrillation; kinetics and TEM measurements. That NAC does not change the amount of amyloid fibrils, it even seems to increase the rate of their formation at higher than 0.2 mM concentration, was confirmed by TEM. Samples of stB at the plateau of the reaction in presence of 4mM NAC were measured 33 . The fibrils were abundant and did not differ from the control (not shown). .

Comparison of the effects of VitC and NAC on kinetics of stB and cysC fibril formation
ThT measuremnts of cysC in presence of vit C show that at the concentration of 0.5 mM VitC completely inhibits amyloid fibril formation, while it does not prolong the lag phase (not shown). This was confirmed by TEM measurements where nearly no fibrils of cysC were seen at 0.5 mM VitC concentration. At 1 mM concentration also much less fibrils could be detected by TEM in comparison to cysC alone but there was more amorphous material seen 33 . The effect of vitC on CysC is thus similar than the effect on stB, namely, it acts inhibitory at lower concentrations while at higher the effect gets reversed. We explain different effects at lower and higher concentrations of vitC by dual, contrasting effects. The first mechanism is interference with cross-β structure of amyloid, the second is pro-oxidative effect, due to change in the redox state of the redox sensitive amino-acids, such as cysteine and methionine.
In distinction to stB, NAC at the higher concentrations, such as 0.5 mM and 1 mMalso completely inhibits amyloid fibril formation of cysC (Fig.7), while at 0.2 mM concentration it significantly prolongs the lag phase.  In conclusion, cysC amyloid fibrils formation is inhibited both by vit C and NAC. Especially significant is the effect of NAC. This was not observed for stB and is explained by possible reduction by NAC of at least 1 disupfide bond in cysC. Using scanning fluorimetry -DSF 33   Results of the time course of ThT fluorescence along stB fibrillation reaction show that a complete inhibition is observed at all Cur concentrations (1 µM Cur + 0,5 mM VitC, 4 µM Cur + 0,5 mM VitC, 10 µM Cur + 0,5 mM VitC, 20 µM Cur + 0,5 mM VitC) apart the highest 50 µM Cur + 0,5 mM VitC. Even here the lag phase is significantly prolonged by cca 100 hours than stB with 50 µM Cur. One can judge (comparing Figs 6 and 9) that Vit C and Cur act synergistically and strongly inhibit stB amyloid fibril formation. TEM analysis (Fig. 10B) shows that even at 50 µM Cur + 0,5 mM VitC the amount of amyloid fibrils is low in comparison to stB alone (Fig. 10A). Detailed viewing of the fibrils showed that they are of different morphology (shorter and thicker).

Acknowledgements:
This work was funded by ARRS -the Slovenian research agency, programme Proteolysis and its regulation P1-0140 (led by B. Turk).