The Toxicity of Misfolded Protein Oligomers Is Independent of Their

Ladiwala, A. R., Litt, J., Kane, R. S., Aucoin, D. S., Smith, S. O., Ranjan, S., Davis, ...... C. (2013) Lipid rafts mediate amyloid-induced calcium d...
0 downloads 0 Views 2MB Size
Articles Cite This: ACS Chem. Biol. 2019, 14, 1593−1600

pubs.acs.org/acschemicalbiology

The Toxicity of Misfolded Protein Oligomers Is Independent of Their Secondary Structure Mirella Vivoli Vega,† Roberta Cascella,† Serene W Chen,§ Giuliana Fusco,‡ Alfonso De Simone,§ Christopher M. Dobson,‡ Cristina Cecchi,† and Fabrizio Chiti*,† †

Downloaded via UNIV OF SOUTHERN INDIANA on July 20, 2019 at 01:54:55 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Department of Experimental and Clinical Biomedical Sciences, Section of Biochemistry, University of Florence, 50134 Florence, Italy ‡ Centre for Misfolding disease, Department of Chemistry, University of Cambridge, Cambridge CB2 1EW, U.K. § Department of Life Sciences, Imperial College London, London SW7 2AZ, U.K. ABSTRACT: The self-assembly of proteins into structured fibrillar aggregates is associated with a range of neurodegenerative diseases, including Alzheimer’s and Parkinson’s diseases, in which an important cytotoxic role is thought to be played by small soluble oligomers accumulating during the aggregation process or released by mature fibrils. As the structural characteristics of such species and their links with toxicity are still not fully defined, we have compared six examples of preformed misfolded protein oligomers with different β-sheet content, as determined using Fourier transform infrared spectroscopy, and with different toxicity, as determined by three cellular readouts of cell viability. The results show the absence of any measurable correlation between the nature of their secondary structure and their cellular toxicity, both when comparing the six types of oligomers as a group and when comparing species in subgroups characterized by either the same size or the same exposure of hydrophobic moieties.



INTRODUCTION

A series of independent studies has led to an increasing number of stable misfolded oligomers of a variety of proteins, and it has become clear that the different species differ not just in size and exposure of hydrophobic residues, but also in their degree of ordered structure, particularly secondary structure. The characterization of such oligomers suggests that some have a substantial degree of disorder, with no significant content of β-sheet or α-helical structure.14,22,24 On the other hand, other oligomers have been found to have persistent βsheet structure26−28 and are therefore more similar to the amyloid fibrils that are characterized by the presence of cross-β structure. A survey of the different oligomer species shows in particular the variations in the degree of β-sheet structure, with a continuum of values from totally unstructured (ca. 0%) to very highly structured (ca. 50% or more) with many intermediate cases. Variations have also been found in the ability of all these oligomeric species to cause cell dysfunction.14,20,22,24 However, the importance of the secondary structure of these species in determining cellular dysfunction is not yet clear, particularly because it has been challenging to identify pairs or groups of oligomers characterized by the same size and external hydrophobicity but with a different β-sheet content. By contrast, stable oligomers have been generated that differ in hydrophobicity while sharing similar sizes and β-sheet content20,22 or that

The self-assembly of peptide and protein molecules to form highly structured fibrillar aggregates is associated with a range of highly debilitating pathological conditions including Alzheimer’s, Parkinson’s, Creutzfeldt-Jacob’s, and Huntington’s diseases, type II diabetes, systemic amyloidosis, pulmonary alveolar proteinosis, and many others.1 In such diseases, particularly in neurodegenerative conditions, an important cytotoxic role is thought to be played by small soluble oligomers.1,2 These species can form as on- or offpathway intermediates during the process of amyloid fibril formation,3−5 can be released by mature amyloid fibrils acting as a reservoir of misfolded species,6−9 or can form as a consequence of secondary nucleation on the surface of preformed amyloid fibrils.10−13 Given their importance, major efforts have been expended in characterizing the structure of misfolded protein oligomers and to identify the structural features responsible for their ability to cause cellular dysfunction. The size of the oligomers has emerged, from many independent studies, as an important factor associated with oligomer pathogenicity, such that toxicity generally decreases with an increase in oligomer size.1,14−18 The presence of solvent-exposed hydrophobic residues on the surfaces of misfolded oligomers has appeared as another important structural element responsible for oligomer toxicity, particularly as it facilitates inappropriate interactions with cellular components, in particular the lipid bilayer of biological membranes.1,19−25 © 2019 American Chemical Society

Received: April 23, 2019 Accepted: May 10, 2019 Published: May 10, 2019 1593

DOI: 10.1021/acschembio.9b00324 ACS Chem. Biol. 2019, 14, 1593−1600

Articles

ACS Chemical Biology

Figure 1. FTIR spectra (amide I band) of different oligomeric species. (A, D, G) FTIR spectra of the amide I region recorded for the oligomeric species of αS, Aβ42, and HypF-N. (B, C, E, F, H, I) Curve fitting analysis of the FTIR spectra. (J) Percentages of various types of secondary structure estimated from curve fitting analysis of the spectra of the various types of oligomer species.



differ in size while sharing similar hydrophobicity and β-sheet content.16,29,30 In this study, we have compared six distinct preformed oligomeric species having different degrees of β-sheet content and toxicity and formed from three protein systems, namely, αsynuclein (αS), the 42-residue form of the amyloid β-peptide (Aβ42), and the N-terminal domain of the model protein HypF from E. coli (HypF-N), to obtain statistically significant data as to whether the β-sheet content of the oligomers correlates with oligomer toxicity. Subgroups from this set of oligomers contain examples from different sequence sources, with similar sizes and solvent-exposed hydrophobicity and are, therefore, ideal to address this question.

RESULTS

Strategy Employed. In order to investigate the relationship between the secondary structure of misfolded protein oligomers and their ability to cause cellular dysfunction, we selected three protein systems for which stable oligomers can be isolated and characterized, which have been shown to possess different degrees of secondary structure and different biological activities, ranging from highly toxic to totally inert, depending on the solution conditions or the time of incubation in which they are allowed to form. The three protein systems are αS, Aβ42, and HypF-N. The first protein is associated with Parkinson’s disease, dementia with Lewy bodies, Parkinson’s disease with dementia, and multiple system atrophy.1 It has been shown to form, under different solution conditions, stable oligomers denoted 1594

DOI: 10.1021/acschembio.9b00324 ACS Chem. Biol. 2019, 14, 1593−1600

Articles

ACS Chemical Biology

showed low and apparently similar β-sheet contents (Figures 1E,F,J). These small values are in agreement with previous farUV CD spectra, which indicate that the two species have a largely disordered secondary structure.22 The amide I band of the FTIR spectra of the type A and type B species of HypF-N are shown in Figure 1C. Both spectra have a major peak at ca. 1629 cm−1, attributable to β-sheet structure, and a large band or shoulder at ca. 1660 cm−1, consistent with a disordered and β-turn structure. The curve fitting analysis (Figures 1H,I) indicates large and apparently similar contents of β-sheet structure, corresponding to 42% ± 8% and 59% ± 8%, respectively, for these oligomers (Figures 1J), in agreement with the values previously reported.28 Determination of the Toxicity of the Various Oligomers. We then tested the ability of the six oligomeric species to cause cellular dysfunction by adding them individually to the extracellular medium of cultured neuroblastoma SH-SY5Y cells and testing the cell viability with three distinct readouts, namely, measurements of the mitochondrial 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction, the influx of calcium from the extracellular medium to the cytosol and the external release of intracellularly trapped calcein. Although the ability of these six species to impair cell viability has been investigated previously,20,22,24 this was achieved using different methods, cell lines, and laboratories, whereas we have used a uniform analytical setup. The MTT reduction assay revealed that type B* oligomers of αS were highly toxic to SH-SY5Y cells, as the cells treated with these species reduced MTT to ca. 72% ± 2% relative to untreated cells (Figure 2). By contrast, the type A*

type A* and type B*, having disordered secondary structure and persistent β-sheet structure, respectively, and being nontoxic and toxic to cultured human neuroblastoma cells and rat primary neurons, respectively.6,24,27 The second protein system is associated with Alzheimer’s disease and hereditary cerebral hemorrhage with amyloidosis1 and forms, under well-defined solution conditions, type A+ and A− oligomers upon short and long incubation, respectively, both lacking stable secondary structure but acting as toxic and nontoxic species, respectively, to human differentiated rat pheochromocytoma PC12 cells, rat primary neurons, and human neuroblastoma SH-SY5Y cells.22,31 The third protein is not associated with any given disease but has been found to be a very valuable model for studying the fundamentals of protein misfolding diseases, as it forms, under different solution conditions, stable amyloid fibrils or protein oligomers in the form of species that were shown to be toxic and nontoxic to a variety of cell lines and animal models. These oligomers were named type A and type B, respectively, both with large β-sheet structure content.20,28,32,33 For this reason, these six oligomer forms appear to be valuable systems for understanding the importance of secondary structure content in determining oligomer toxicity, as they offer a wide range of both parameters, with differences evident in oligomer types formed from the same protein sequence. Moreover, subgroups of this experimental set, containing oligomers from different sequence sources, share similar sizes, as measured with atomic force microscopy (AFM) and similar levels of solvent-exposed hydrophobicity, as measured with ANS and dye labeling measurements20,22,53 and are, therefore, ideal to address this question (see below). Determination of the Secondary Structure of the Various Oligomers. We first isolated the various forms of oligomers as described in the Materials and Methods. In order to determine their secondary structure, we opted for using Fourier-transform infrared (FTIR) spectroscopy rather than far-ultraviolet circular dichroism (far-UV CD) spectroscopy to avoid all the complications associated with light scattering and differential absorption flattening when studying protein aggregates with this other technique.34,35 Although the secondary structure of the six oligomer species has been previously investigated, this was achieved using somewhat different methods and in different laboratories and often in nonquantitative manner.22,24,28 We therefore decided to reanalyze the six species under the same experimental conditions, techniques, and equipment in a fully quantitative manner. The amide I bands of the FTIR spectra of type A* and type B* of αS are shown in Figure 1A. The first spectrum indicates a major band at ca. 1650 cm−1, attributable largely to disordered or α-helical structure, whereas the second spectrum has two intense bands at ca. 1654 cm−1, which is likely to result from disordered structure, and at ca. 1624 cm−1, and can readily be assigned to β-sheet structure. Curve fitting analysis of the two spectra was carried out to identify the components of the large amide I region (Figure 1B,C) and to assign them to well-defined types of secondary structure (Figure 1J). The results indicate that the β-sheet content of the two species is 7% ± 7% and 38% ± 4%, respectively, in good agreement with estimates obtained in previous work with other techniques.24,27 The FTIR amide I region of type A+ and type A− oligomers of Aβ42 are shown in Figure 1B. Both spectra feature one dominant band at ca. 1650−1660 cm−1, attributable to disordered or α-helical structure. The curve fitting analysis

Figure 2. Cell viability impairment in cells exposed to different oligomeric species. MTT reduction in SH-SY5Y cells treated for 24 h with the different αS, Aβ42, and HypF-N oligomeric species (at 0.3, 1, and 6 μM, respectively). Samples containing oligomers were analyzed by one-way ANOVA followed by Bonferroni’s multiple comparison test relative to untreated cells (***P < 0.001, n = 9). A total of 80000−100000 cells were analyzed per condition. Data represent mean ± SEM of three independent experiments.

species were substantially inert as the cells reduced MTT to a level similar to that of untreated cells, within experimental error (Figure 2). The same difference was found for type A+ and type A− oligomers of Aβ42, with only the former found to be toxic (72% ± 1% MTT reduction) and for type A and type B oligomers of HypF-N, with only the type A species found to be toxic (72% ± 5% MTT reduction). When measuring the influx of Ca2+ from the medium to the cytosol caused by the different oligomeric species in cultured SH-SY5Y cells, we found that only type B* oligomers of αS, 1595

DOI: 10.1021/acschembio.9b00324 ACS Chem. Biol. 2019, 14, 1593−1600

Articles

ACS Chemical Biology type A+ oligomers of Aβ42, and type A oligomers of HypF-N induced a substantial influx of Ca2+ across the cell membrane, as a result of its increased permeability to this ion (Figure 3).

Figure 3. Ca2+ influx across the membranes of cells incubated with the different oligomeric species. Representative confocal microscope images showing the Ca2+-derived fluorescence (green) in SH-SY5Y cells treated for 15 min with the indicated αS, Aβ42, and HypF-N oligomeric species (at 0.3, 1, and 6 μM, respectively). Cells were loaded with the Fluo-4 AM probe that emits fluorescence when bound to Ca2+ ions. The data from the semiquantitative analysis of the Ca2+-derived fluorescence are expressed as the percentages of the values for untreated cells. Samples containing oligomers were analyzed by one-way ANOVA followed by Bonferroni’s multiple comparison test relative to untreated cells (***P < 0.001, n = 6). A total of 100−120 cells were analyzed per condition. Data represent mean ± SEM of three independent experiments.

Figure 4. Disruption of the integrity of cell membranes by the different oligomeric species. Representative confocal microscope images showing SH-SY5Y cells loaded with the calcein-AM probe for 10 min (green) and then treated for 1 h with the indicated αS, Aβ42, and HypF-N oligomeric species (at 0.3, 1, and 6 μM, respectively). The values of the green fluorescence are expressed as the percentages of the values for untreated cells. Experimental errors are SEM. Samples containing oligomers were analyzed by one-way ANOVA followed by Bonferroni’s multiple comparison test relative to untreated cells (***P < 0.001, n = 6). A total of 100−120 cells were analyzed per condition. Data represent mean ± SEM of three independent experiments.

The values were 380% ± 45%, 620% ± 80%, and 555% ± 30% relative to untreated cells. By contrast, type A* oligomers of αS, type A− oligomers of Aβ42, and type B oligomers of HypFN induced a small and nonsignificant influx, amounting to 135% ± 10%, 181% ± 10%, and 177% ± 12% relative to untreated cells (Figure 3). The same scenario was observed when measuring the release of intracellular calcein from the SH-SY5Y cells as a result of membrane disruption (Figure 4), with only the same three species reducing significantly the levels of intracellular calcein (35% ± 4%, 25% ± 3%, and 22% ± 3% of calcein release relative to untreated cells), and the remaining three being substantially inert or only weakly toxic (95% ± 10%, 96% ± 10%, and 86% ± 10% of calcein release relative to untreated cells). The lack of toxicity for αS type A*, Aβ42 type A−, and HypF-N type B oligomers and the inherent toxicity of αS type

B*, Aβ42 type A+, and HypF-N type A oligomers is in agreement with previous reports6,20,22,24,32,36 and emphasizes the reproducibility and robustness of these cell viability results. Absence of Correlation between β-Sheet Content and Toxicity within This Group of Oligomers. The ability to quantify the β-sheet content of the six species investigated here and of their effects on cellular dysfunction provides an opportunity to study in a statistically significant manner whether the β-sheet content is an important determinant of the ability of the oligomers to interfere with cell viability. The plot of MTT reduction versus β-sheet structure content indicates that there is no significant correlation between the two characteristics of the oligomers (Figure 5A, r = 0.06, p ≫ 0.05). In fact, if the content of β-sheet structure were to be an important determinant of oligomer toxicity, a strong negative correlation would be observed. By contrast, if it were a 1596

DOI: 10.1021/acschembio.9b00324 ACS Chem. Biol. 2019, 14, 1593−1600

Articles

ACS Chemical Biology

Figure 5. Correlations between cellular toxicity and β-sheet content for the different oligomeric species. Plots of MTT reduction (A), intracellular Ca2+ levels (B), and intracellular calcein levels (C) in SH-SY5Y cells incubated with the different types of oligomeric species (the y axis in all plots) versus the content of β-sheet structure (the x axis in all plots). The mean values and experimental errors reported in all the plots are those reported in Figure 1D (for β-sheet structure content) and in the histograms of Figures 2, 3, and 4, for MTT reduction, Ca2+ levels, and calcein levels, respectively. The r and p values for each correlation are reported in each plot.

N have very different effects on cellular viability but a similar high β-sheet content.

protective factor against oligomer toxicity, a positive correlation would be observed. Neither of them was found, however. In agreement with the data from the MTT reduction analysis, the β-sheet content was not found to correlate with either of the other two readouts of cell viability, namely, Ca2+ influx (Figure 5B, r = 0.15, p ≫ 0.05) and calcein release (Figure 5C, r = 0.29, p ≫ 0.05). Since the small size and solvent-exposed hydrophobicity of misfolded protein oligomers have been shown to be important factors of their potential to damage cells, we compared oligomer types where these two characteristics are largely similar. The toxic αS type B*, Aβ42 type A+, and HypF-N type A oligomers all have similar sizes, as previously measured with AFM (4.3 ± 0.9, 6.2 ± 0.5, and 2−6 nm, respectively) and all show a high degree of hydrophobic solvent-exposure as determined previously with ANS and other dye labeling measurements.20,22,28,53 The β-sheet content within this subgroup of oligomer species is highly variable, corresponding to 37% ± 5%, 17% ± 5%, and 58% ± 8%, respectively (Figure 1). However, their toxicity is very similar, as revealed by all three assays of cell viability employed here (Figures 2−4). In agreement with this observation, the substantially inert αS type A*, Aβ42 type A−, and HypF-N type B oligomers have sizes within the same order of magnitude (5.1 ± 0.8, 6.1 ± 0.6, and 2−7 nm, respectively) and show a similar but smaller degree of solvent-exposure of hydrophobic surface relative to the toxic subgroup.20,22,28,53 Nevertheless, the β-sheet content within this subgroup of benign oligomers is also very different, corresponding to 10% ± 10%, 17% ± 5% and 54% ± 8%, respectively. A similar conclusion is also evident when comparing two oligomer forms generated from the same protein. The observation that the β-sheet containing oligomers of αS (type B*) are highly toxic, whereas the largely disordered αS oligomers (type A*) are benign, could in principle suggest that an architecture of the oligomers based on β-sheet scaffold is an important factor for misfolded oligomers to induce cellular toxicity. Nevertheless, comparison within the other two pairs of oligomers, each pair having identical sequences, shows that this is not in fact the case, as type A+ and type A− oligomers of Aβ42 have very different toxicities but a similar low β-sheet content and, moreover, type A and type B oligomers of HypF-



DISCUSSION Many of the structurally characterized misfolded protein oligomers with an ability to cause cellular dysfunction contain and are clearly stabilized by a substantial degree of β-sheet structure. These include the A11-positive prefibrillar and OCpositive fibrillar oligomers of Aβ42,26,37 the Aβ40 and Aβ42 protofibrils,38 the type B* oligomers of αS,27 the type A oligomers of HypF-N,28 the short fibrils of huntingtin exon 1,39 the prion examers,40 and several others. The presence of βsheet structure in many of the toxic oligomers so far characterized has led to the assumption that such secondary structure is intimately associated with the toxicity of these aberrant species. However, as structural data on the oligomers from different proteins continue to emerge, it is becoming clear that some of the toxic species are oligomers with little β-sheet structure, such as the type A+ of Aβ42,22 the pentamers and decamers of Aβ42 isolated by Smith and colleagues,41 the oligomers recently characterized for the islet amyloid polypeptide,42 etc. Our data indicate that formation of a β-sheet architecture within the oligomers is not a necessary requirement for their toxicity, because (i) a correlation between toxicity and β-sheet content is not apparent within the group of six oligomers analyzed here, (ii) oligomers from different protein systems but with similar toxicity, solvent-exposed hydrophobicity, and size, as determined previously,20,22,28,53 have variable amounts of β-sheet structure, and (iii) pairs of toxic and nontoxic oligomers from the same protein can have similar contents of β-sheet structure. Analysis of the structural characteristics of the six oligomeric species studied here confirms previous suggestions that the exposure of hydrophobic clusters on the oligomer surface is indeed a major determinant of the ability of these aberrant species to interact with and destabilize the lipid bilayer of cellular membranes, thus causing cellular dysfunction.1,20,22,28 This conclusion was observed in the original reports of the three pairs of toxic and nontoxic oligomers from the same protein sequences, where the higher hydrophobic exposure of the toxic relative to the benign species was noted in each case.20,22,53 Small size has also emerged as an important factor 1597

DOI: 10.1021/acschembio.9b00324 ACS Chem. Biol. 2019, 14, 1593−1600

Articles

ACS Chemical Biology

was used for the FTIR spectra. The resulting samples were deposited on a KBr window in a semipermanent liquid cell using a spacer of 25 μm, and the FTIR spectra were recorded from 4000 to 400 cm−1 at RT using a Jasco FTIR 4200 spectrophotometer (Tokyo, Japan) constantly purged with N2. The spectra were background-subtracted, baseline-corrected, and smoothed. The amide I regions (1600−1700 cm−1) were then analyzed with a procedure of best fitting using Kaleidagraph (Synergy Software, Reading, PA), 3−4 peaks with centers, widths, and heights free to float, and up to 1000 iterations. Residuals between experimental and best fitted spectra were