Pretreatment of chemically-synthesized Aβ₄₂ affects its biological activity in yeast

Abstract

The tendency of amyloid β (Aβ₄₂) peptide to misfold and aggregate into insoluble amyloid fibrils in Alzheimer's disease (AD) has been well documented. Accumulation of Aβ₄₂ fibrils has been correlated with abnormal apoptosis and unscheduled cell division which can also trigger the death of neuronal cells, while oligomers can also exhibit similar activities. While investigations using chemically-synthesized Aβ₄₂ peptide have become common practice, there appear to be differences in outcomes from different preparations. In order to resolve this inconsistency, we report 2 separate methods of preparing chemically-synthesized Aβ₄₂ and we examined their effects in yeast. Hexafluoroisopropanol pretreatment caused toxicity while, ammonium hydroxide treated Aβ₄₂ induced cell proliferation in both C. glabrata and S. cerevisiae. The hexafluoroisopropanol prepared Aβ₄₂ had greater tendency to form amyloid on yeast cells as determined by thioflavin T staining followed by flow cytometry and microscopy. Both quiescent and non-quiescent cells were analyzed by these methods of peptide preparation. Non-quiescent cells were susceptible to the toxicity of Aβ₄₂ compared with quiescent cells (p < 0.005). These data explain the discrepancy in the previous publications about the effects of chemically-synthesized Aβ₄₂ on yeast cells. The effect of Aβ₄₂ on yeast cells was independent of the size of the peptide aggregates. However, the Aβ₄₂ pretreatment determined whether the molecular conformation of peptide resulted in proliferation or toxicity in yeast based assays.

Keywords:

• Alzheimer's disease
• Candida glabrata
• peptide conformation
• quiescent cells
• Saccharomyces cerevisiae
• synthetic Aβ₄₂

Introduction

Soluble oligomers of amyloid β (Aβ₄₂) peptide have been found to be highly neurotoxic and important in the etiology of AD in comparison to fibril forms.¹ However, it is still unknown which form of the aggregates are more dangerous; oligomers and protofibrils or fibrils and filaments.^2-4 Despite these uncertainties aggregation studies in search for chemo-preventatives that block Aβ₄₂ oligomer and fibril formation using synthetic forms of Aβ₄₂ have become common practice.^5-8

The preparation of Aβ₄₂ peptide can affect its amyloidogenicity⁴ yet pretreatments are required in order to produce monomers.⁹ Lioudyno et al.¹⁰ reported that pretreatment with 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) enhanced the amyloidogenicity of the Aβ₄₂ peptide as well as its interaction with biological membranes.

Complex mechanisms contribute to development of AD. Synaptic loss due to plaque formation¹¹ and neuronal death due to oligomeric forms of Aβ₄₂ are all hallmarks of the disease.¹² It has been reported that different conformations of Aβ₄₂ induce neurotoxicity by distinct mechanisms in human cortical neurons.^13,14 The longer variants have higher propensity to aggregation than the shorter form.¹³ The mechanism of Aβ₄₂ toxicity is still unknown, although it is suggested that amyloidogenic protein aggregates on neuronal cells cause permeabilisation of lipid membranes,^15-17 and increase oxidative stress,^18,19 mitochondrial dysfunction,^20-22 and endoplasmic reticulum stress^23,24 which ultimately leads to apoptosis.

Yeast exhibit some responses to Aβ₄₂ peptide that indicate they are good models for studying the effects of on neuronal cells.²⁵ However, some limitations are expected since yeast lack receptors that have been proposed for Aβ₄₂ binding. Putative receptors include the insulin receptor,²⁶ the alpha7 nicotinic receptor,²⁷ the metabotrophic glutamate receptor,²⁸ and cellular prion protein.²⁹ Chacińska and colleagues³⁰ showed an induction of proliferation in the presence of synthetic Aβ₄₂ in exponentially growing Saccharomyces cerevisiae, while Bharadwaj et al.³¹ showed toxicity of synthetic Aβ₄₂ with a different preparation method on Candida glabrata cells. In this study we report on the preparation of Aβ₄₂ peptide with 2 different methods and treatments of identical yeast cell cultures. The two pretreatments resulted in totally opposite outcomes.

Results

Pretreatment determines the activity of Aβ₄₂ peptide

HFIP pretreatment of Aβ₄₂ was previously shown to cause toxicity to C. glabrata cells.³¹ This study confirms significant toxicity of HFIP pretreated Aβ₄₂ to C. glabrata (Fig. 1A) and to S. cerevisiae (Fig. 1B). There was even more cell killing after 48 h of incubation. Conversely, NH₄OH pretreatment resulted in a significant increase in the number of yeast cells. C. glabrata (Fig. 1A) more than doubled in number (P < 0.001), and S. cerevisiae (Fig. 1B) numbers significantly increased over the time (P < 0.005).

Figure 1. Effect of HFIP and NH₄OH pretreated synthetic Aβ₄₂ on the yeast cells. 2 μM peptide was added to exponential phase C. glabrata (A) and S. cerevisiae (B) cells, suspended in water. Samples were incubated at 30°C for 48 h. Viability was determined as percentage of untreated cells (control). (* P < 0.05, ** P < 0.01, *** P < 0.005, and **** P < 0.001). Data are shown as Mean ± SEM.

Susceptibility of quiescent and non-quiescent cells to toxic Aβ₄₂

Stationary phase fractions of C. glabrata were tested for their susceptibility to 2 μM Aβ₄₂ generated by HFIP pretreatment method. Results indicated that quiescent daughter cells (Fig. 2A) are more resistant to the toxic effect of Aβ₄₂ than older non-quiescent mother cells (Fig. 2B). The older cells lost their viability within 24 h in the presence of amyloid peptide whereas younger cells were affected by the toxicity after 48 h of incubation (see Fig. 2).

Figure 2. Effect of Aβ₄₂ on the quiescent and non-quiescent fractions of C. glabrata. The quiescent daughter cells (A) were more resistant to Aβ₄₂ prepared with HFIP method than non-quiescent mother cells (B) in the first 24 h of incubation. After 48 h of incubation both cell fractions were susceptible to the toxicity effect of Aβ₄₂ (*** P < 0.005).

Amyloid staining on the yeast cell surface and aggregation propensity

In order to examine the presence of amyloidogenic proteins on the surface of yeast cells ThT treated cells were visualised for fluorescence. S. cerevisiae exhibited presence of a high level of ThT stained proteins on their surface (Fig. 3). Yeast treated with HFIP prepared Aβ₄₂ on the other hand have more aggregation propensity, causing the cells to adhere and attach to each other (see Figs. 3G, I). Cell aggregation appears absent in those cells treated with NH₄OH prepared Aβ₄₂ (Figs. 3D, F)_. This result indicates that peptide produced by each method of preparation interacts differently with the yeast cell walls.

Figure 3. Staining of amyloid proteins present on S. cerevisiae cell walls by thioflavin T. First row (A–C) shows untreated S. cerevisiae cells in presence of ThT in comparison with cells treated with 5 μM NH₄OH (D–F) and HFIP (G–I) prepared Aβ₄₂ for 24 h in H₂O. ThT staining comparison of untreated (control), NH₄OH and HFIP pretreated cells are shown in panels A, D and G respectively. The HFIP pretreatment of Aβ₄₂ results in higher fluorescence intensity and adhesion of peptide to the S. cerevisiae yeast cell wall protein. Scale bar = 50 μm.

Flow cytometry for analysis of fluorescence intensity associated with peptide

The fluorescence intensity associated with ThT level on the surface of the yeast cells treated with HFIP- and NH₄OH-prepared amyloid peptide was measured by flow cytometry. Data showed an increase in the fluorescence intensity (FI) in the presence of HFIP pretreated peptide (Fig. 4C) compared with the NH₄OH method (Fig. 4D), and untreated controls (Fig. 4A): fluorescence intensity was measured to be 8.4%, 5.4% and 3.7% respectively (see Figs. 4C, D). It seems the peptide conformation generated from the HFIP method has a greater propensity to induce aggregation and may cause amyloid deposition on the surface of yeast cells.

Figure 4. Flow cytometry analysis of S. cerevisiae treated with HFIP and NH₄OH prepared Aβ₄₂. The fluorescence intensity (FI) of ThT in the experimental samples was determined as a percentage of unstained cells. Yeast cells were grow to exponential phase, washed and resuspended in H₂O. The cells were then treated with HFIP and NH₄OH pretreated Aβ₄₂ and incubated overnight along with control (untreated samples) at 30°C with shaking. The next day, cells were treated with 20 μM ThT and were analyzed by flow cytometry. Pacific blue filter was chosen to measure fluorescence. Unstained control cells (A), ThT stained control cells (B) and those treated with 5 μM of HFIP (C), and NH₄OH pretreated Aβ₄₂ have been shown.

TEM analysis of peptide conformers generated by pretreatment

The Aβ₄₂ prepared for viability counts was analyzed by TEM. The NH₄OH method resulted in more uniform aggregate free peptide solution (Fig. 5A) compared with HFIP (Fig. 5C). The latter showed the presence of protofibrils structures at 0 h. After 24 h incubation at 37°C short fibrils were observed in NH₄OH prepared peptide (Fig. 5B) while, the HFIP method (Fig. 5C) resulted in long and twisted fibrils (Fig. 5E).

Figure 5. TEM analysis of peptide pretreated by NH₄OH and HFIP at 37°C. The NH₄OH method of oligomer formation (A) results in a higher amount of monomeric peptide compared with HFIP method in which still contain long fibrils (B). After 24 h of incubation the NH₄OH method of pretreatment (C) induced a lower tendency form fibrils and those generated were shorter fibrils. The HFIP method (D) on the other hand resulted in a higher level of twisted and longer fibrils. The HFIP twisted fibril is shown (E). Scale bar = 200 nm.

Effect of fibrils on the yeast cells

The toxicity of fibrils, formed after prolonged incubation, was examined. Fibrils generated from HFIP and NH₄OH pretreatment were both toxic to yeast cells (Fig. 6). However, HFIP prepared Aβ₄₂ fibrils exhibited greater toxicity compared to those fibrils prepared with NH₄OH.

Figure 6. Effect of aged peptide on yeast cells. The NH₄OH and HFIP pretreated Aβ₄₂ peptide was incubated for 6 d at 37°C with no shaking. Exponentially growing S. cerevisiae were washed and suspended in H₂O and then treated with 2 μM of these aged peptides and incubated overnight. Viability was determined the next day by transferring 100 μl of treated and untreated samples onto YEPD solid agar. (* P < 0.05 and **** P < 0.001).

Discussion

The current study indicates that different preparations of Aβ₄₂ peptide can lead to different biological activity toward yeast. The recently introduced NH₄OH pretreatment of synthetic Aβ₄₂ reportedly results in formation of higher levels of monomeric peptide³² and has not previously been tested on yeast. Soluble monomers of the Aβ₄₂ prepared by NH₄OH caused cell proliferation and an increase in the number of colonies in both S. cerevisiae and C. glabrata, regardless of the type of solvent (H₂O or dilute NaOH). With prolonged incubation more fibrils were formed and these were toxic to S. cerevisiae cells.

In contrast, the Aβ₄₂ peptide prepared by the HFIP method contained high levels of aggregates and caused cell death within the first 24 h as well as after this time. Older cells were shown to be more susceptible to the effects of Aβ₄₂. The Aβ₄₂ peptide conformation appears to be the major determinant of its effect on yeast cells. Previous studies showed that Aβ₄₂ prepared using HFIP tends to be more potent in the electrophysiological tests and has a greater tendency to modulate biological membrane properties.^4,9,10 To further investigate the different outcomes caused by HFIP, it would be worth examining if HFIP affects the toxicity of additional fibril forming peptides.

The attachment of the Aβ₄₂ peptide to the yeast cell surface has been demonstrated by Bharadwaj (2011) who demonstrated direct binding of fluorescein-labeled Aβ₄₂ to the cell surface by fluorescence microscopy.³³ This finding was further confirmed by electron microscopy and western blotting of sub-cellular fractions.³³ In this study we showed that the Aβ₄₂ peptide caused increased ThT staining of the yeast cell surface. Without Aβ₄₂ peptide treatment, there was some ThT staining which may be attributed to proteins such as Flo1p, Als1p, Als5p, Muc1p and Bgl2p.^34-37 While there is clear evidence for peptides from Flo1p, Als1p, Als5p, and Muc1p to be amyloid forming, the evidence for these proteins forming amyloid on the yeast surface is currently incomplete and limited to the birefringence and ThT staining that they cause.³⁶ While it is tempting to speculate that Aβ₄₂ may be forming amyloid on the yeast cell surface, proof of amyloid formation requires analysis of the resistance to an ionic detergent. This evidence is still lacking for other yeast cell surface proteins.

HFIP-treated Aβ₄₂ resulted in peptide with long strand fibril morphology compared with the NH₄OH method. It is known that HFIP pretreatment method does not efficiently generate aggregate free solutions. This HFIP pretreatment results in peptide that has a high propensity to adhere to the yeast cell wall. In this form the peptide in any size, monomeric, oligomeric and even fibrils are toxic to yeast cells. On the other hand, Aβ42 that is freshly pretreated with NH₄OH caused cell proliferation due to monomeric and oligomeric forms while prolonged incubation led to a fibrillar conformation which was toxic to S. cerevisiae cells.

The data presented here show that different methods for the preparation of Aβ₄₂ can lead to different biological activities in in vitro and in yeast biological assays. This work should serve as a caution in working with and interpreting the outcomes of biological assays.

Materials and Methods

Yeast strains and growth conditions

C. glabrata (ATCC90030) and S. cerevisiae BY4743 (a/α his3Δ/his3Δ leu2Δ/leu2Δ +/lys2Δ met15Δ/+ ura3Δ/ura3Δ) were utilised for experiments and were grown in YEPD (1% yeast extract, 2% peptone, and 2% _D-glucose) with aeration at 30°C. For fractionation of quiescent and non-quiescent cells, growth in the YEPD media continued for 7 d with shaking (200 rpm). Cells were subsequently pelleted and washed for further analysis. For exponential phase of growth, yeast cells were grown overnight and were harvested by centrifugation the next day and resuspended in 20 ml fresh media. Then one ml was inoculated into 80 ml of fresh media and incubated at 30°C with shaking (200 rpm) for another 2-3 h to reach an OD₆₀₀ of 1.0.

Fractionation of stationary phase C. glabrata and S. cerevisiae

Percoll density gradients (GE Healthcare, United Kingdom) were prepared according to the manufacturer's preformed gradient protocol and method described by Allen et al.⁴⁰ with some modification. Briefly, Percoll was diluted 9:1 (v/v) with 1.5 M NaCl. To form a gradient 10 ml of the Percoll solution was added to 15 ml Corex® tubes and centrifuged for 15 min at 20°C in a Beckman Coulter centrifuge (SW41Ti rotor). Stationary phase yeast cells which were grown for 7 d at 30°C with shaking were then pelleted and resuspended in 1 ml Tris buffer (50 mM, pH 7.5). Around 2 ml of cell suspension was overlaid onto preformed gradient, and centrifuged at 400 g for 60 min in a Beckman Coulter centrifuge. Fractions of quiescent (lower fraction) and non-quiescent cells were collected and washed twice in 40 ml Tris buffer to remove any Percoll residue, pelleted and then resuspended in ddH₂O.

Preparation of Aβ₄₂ peptide

Synthetic human Aβ₄₂ peptide was purchased from Keck laboratories (Yale University, New Haven, CT). The peptide was synthesized and purified using tBOC chemistry with DCC and HOBT coupling reagents in the form of lyophilized powder. All solvents used for the preparation of Aβ₄₂ solutions were filtered and centrifuged to minimise the presence of debris that could induce aggregation of the peptide.

Preparation of Aβ₄₂ using NH₄OH

Peptide solutions were prepared according to the protocol of Ryan et al.³² Briefly, 20 mg peptide was dissolved in 40 ml of 10% NH₄OH to final concentration of 0.5 mg/ml solution (w/v). The solution was then left at room temperature for 10 min and then sonicated for 5 min. For yeast treatments, the peptide was prepared immediately prior to use. Each 0.5 mg aliquot of Aβ pellet was dissolved in 60 μl of ddH₂O and then vortexed to promote the dissolving of the peptide. The solution was then sonicated for 5 min and centrifuged at 14,000 g for 5 min to pellet any undissolved peptide. The supernatant was carefully recovered for use in subsequent assays. Peptide concentration was determined from the A_214nm measurement of the solution in a quartz cuvette (BioMate™ Spec, Thermo Scientific, UK).

Preparation of Aβ₄₂ using HFIP

Aβ₄₂ was dissolved in 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) according to the method described by Crescenzi and colleagues.³⁹ Briefly the peptide film was dissolved in HFIP to a concentration of 1 mg/ml, sonicated for 5 min on ice and dried overnight to form a clear film. The film was subsequently dissolved in 60 μl of ddH₂O, sonicated for 5 min and centrifuged to remove insoluble aggregates. Peptide concentrations were determined described above.

Effect of prolonged incubation of Aβ₄₂ on the survival of yeast

The Aβ₄₂ samples from each pretreatment method were incubated at 35°C for 6 d. Exponentially growing S. cerevisiae were washed and suspended in H₂O and then were treated with 5 μM of HFIP and NH₄OH pretreated peptide and incubated overnight. Treatment and viability were determined as follows.

Exponentially growing yeast cells were washed twice and resuspended in ddH₂O. After estimation of cell numbers by a haemocytometer, cells were transferred into 96-well culture plates at a concentration of 10³ cells/ml, treated with 2-5 μM of Aβ₄₂ peptide in a final volume of 200 μl per microtiter well and incubated overnight at 30°C with shaking. The next day 100 μl aliquots of cell suspension were plated onto YEPD plates and incubated for 2 days at 30°C. Colonies were counted for each treatment and the viability was estimated as percentage of untreated cells. All treatments were performed in triplicate.

Flow cytometry analyses after Thioflavin T staining of treated cells

Exponentially growing S. cerevisiae were washed and incubated in H₂O in the presence and absence of pretreated HFIP and NH₄OH peptide. After 16 h of incubation, Thioflavin T (ThT) was added to each sample to a final concentration of 20 μM and incubated for 10 min.⁴⁰ Around 10,000 cells per sample were analyzed by flow cytometry (FACS, Canto II- BD™) using a pacific blue filter with 351 nm excitation and collecting fluorescent emission at 450/65. Data were recorded and saved as FCS3 files. All data were analyzed by WEASEL (WEHI, Melbourne, Australia) software. ThT fluorescence intensity of multiple experimental samples was calculated as percentage of unstained yeast cells.

Confocal microscopy after Thioflavin T staining of treated cells

Confocal fluorescence and differential interference contrast (DIC) microscopy was performed using a Nikon® Eclipse Ti confocal laser-scanning microscope. Treated and untreated yeast cells with Aβ₄₂ peptide in H₂O in a 96 well microtiter plate were stained with ThT to a final concentration of 20 μM. ThT fluorescence was observed using a DAPI filter following excitation of the sample with light of 405 nm wavelength. Images were acquired using a Nikon camera and NIS-Elements imaging software.

Transmission electron microscopy

Freshly prepared aliquots of Aβ₄₂ peptide with both pretreatment methods were used for imaging the starting peptide conformation. The peptide samples were then incubated overnight at 35°C for analysis of peptide conformation after 24 h. Samples from 0 and 24 h were gently agitated before transferring a 4 μl aliquot onto the grids. After 30 sec adsorption time, the excess sample was drawn off using Whatman® 541 filter paper. The grids were stained subsequently with 2% w/v potassium phosphotungstate (pH 7.2) for 10 sec. The grids were then air dried before examination. The samples were examined using a Tecnai™ 12 Transmission Electron Microscope (FEI, Eindhoven, The Netherlands) at an operating voltage of 120 kV. Images were recorded using a Megaview III CCD camera and AnalySIS camera control software (Olympus Australia).

Statistical analysis

Graphs were made using Microsoft Office Excel 2010 software and data were analyzed using PRISM® 6 (Graph Pad Software, Inc.., La Jolla, CA, USA). All data are presented as mean + SEM. Significant differences between samples were determined by comparing the mean score of samples by an unpaired Student's t-test. A P-value of <0.05 is deemed as significant.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This work is supported in part by a grant from the Medical Advances Without Animals Trust (MAWA).