Identification of PrP sequences essential for the interaction between the PrP polymers and Aβ peptide in a yeast-based assay

Abstract

Alzheimer disease is associated with the accumulation of oligomeric amyloid β peptide (Aβ), accompanied by synaptic dysfunction and neuronal death. Polymeric form of prion protein (PrP), PrP^Sc, is implicated in transmissible spongiform encephalopathies (TSEs). Recently, it was shown that the monomeric cellular form of PrP (PrP^C), located on the neuron surface, binds Aβ oligomers (and possibly other β-rich conformers) via the PrP_23–27 and PrP_90–110 segments, acting as Aβ receptor. On the other hand, PrP^Sc polymers efficiently bind to Aβ monomers and accelerate their oligomerization. To identify specific PrP sequences that are essential for the interaction between PrP polymers and Aβ peptide, we have co-expressed Aβ and PrP (or its shortened derivatives), fused to different fluorophores, in the yeast cell. Our data show that the 90–110 and 28–89 regions of PrP control the binding of proteinase-resistant PrP polymers to the Aβ peptide, whereas the 23–27 segment of PrP is dispensable for this interaction. This indicates that the set of PrP fragments involved in the interaction with Aβ depends on PrP conformational state.

Keywords: :

• Alzheimer Disease
• amyloid
• prions
• yeast
• protein-protein interactions
• FRET

Introduction

Experimental evidence suggests that amyloid β (Aβ) oligomers play a crucial role in the pathogenesis of Alzheimer disease (AD).^1,2 Accumulation of Aβ oligomers leads to neurotoxicity and impairs synaptic plasticity and long-term memory.^3,4 However, the mechanisms by which Aβ oligomers cause neuropathology remain unclear. One potential lead comes from the ability of Aβ oligomers to interact with certain receptors on the synaptic membrane,^5-7 in particular, with the cellular isoform of prion protein (PrP^C).^8,9 Normal cellular function of PrP^C is still unclear although its polymeric (prion) isoform, PrP^Sc, is a causative agent of transmissible prion diseases, the neurodenerative diseases with ethiology not dissimilar from AD.¹⁰ PrP^C is located on the neuron surface and is shown to interact with β-sheet-rich conformers of various origins, such as heterologous PrP, yeast prion proteins or engineered β-peptides, in addition to Aβ peptide.¹¹ Thus, PrP^C acts as a universal receptor for amyloid-like oligomers. It has been shown that two short segments of PrP^C containing respectively residues 23–27 and 95–110 are critically important for the efficient binding of PrP^C to Aβ oligomers.^8,12 In vitro studies show that the PrP_23–30 and PrP_95–105 peptides efficiently bind to Aβ oligomers and that this interaction is dependent on positively charged residues located in these peptides and also is influenced by the length of the sequence between them.¹³ It is important to note that PrP^C specifically binds to β-conformers but does not interact with Aβ monomer. In contrast, PrP^Sc polymers interact with Aβ monomers and accelerate Aβ aggregation.¹⁴ This suggests that some PrP regions changes their conformations as a result of formation of the polymeric prion, and acquire a capacity of binding Aβ. Specific regions of PrP that are essential for the binding of PrP^Sc to Aβ remain undefined.

We used a yeast-based system to analyze in vivo the interaction of PrP polymers with Aβ. A number of studies have indicated that S. cerevisiae is a convenient model to study the factors that regulate the misfolding and aggregation of mammalian amyloids.^15,16 Various mammalian amyloidogenic proteins, such as α-synuclein, PrP, Aβ and expanded poly-Q region of human Htt, form polymers and/or oligomers in yeast cells that are similar in their biochemical properties to the polymers or oligomers formed in the brains of humans and other animals.^6,17-22 In comparison to mammalian systems, Saccharomyces cerevisiae is a simple, cheap and well-studied genetic model that provides significant advantages for the analysis of amyloid interactions. First, the level of protein expression may be adjusted at experimentator’s will in a wide range using regulated promoters. Second, PrP and Aβ aggregates do not kill yeast cells.

Here, we employed a yeast-based assay for the analysis of in vivo interactions between PrP polymers and the 40 amino acid Aβ peptide, and identified the segments of PrP that are crucial for this interaction.

Results

PrP and Aβ fused to fluorescent proteins form amyloid-like polymers in the yeast cytoplasm

To analyze aggregation of mammalian amyloidogenic proteins in yeast, we have expressed Aβ-YFP (or Aβ-GFP) and various fragments of PrP, fused to CFP (constructed as described in Material and Methods) in the yeast cells. Using confocal fluorescence microscopy, we found that Aβ-YFP forms cytologically detectable aggregates in some cells that appear as “dots” or small “clumps” (Fig. 1). All the PrP-CFP fusions formed aggregates of different shapes, but PrP_23–231-СFP, PrP_28–231-CFP and PrP_90–231-СFP preferentially formed dot- or clump-like aggregates, whereas PrP_110–231-CFP tended to form tape-like aggregates. We have confirmed by using organelle-specific fluorescent dyes that PrP_23–231-GFP and Aβ-GFP aggregates are located in the cytoplasm and do not co-localize with the nucleus, vacuoles, endosome or lipid particles (Fig. 2). To confirm that PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP, PrP_110–231-CFP and Aβ-YFP proteins form insoluble aggregates in yeast cells, the centrifugation analysis has been performed. After centrifugation, Aβ-YFP and fusions of CFP with PrP derivatives were detected both in the soluble and insoluble fractions, whereas monomeric protein CFP was present only in the soluble fraction (Fig. 3A).

Figure 1. Fluorescence microscopy assay of Aβ-YFP and PrP derivatives fused to CFP. Aβ-YFP forms aggregates that appear as “dots” or small “clumps” in some cells and shows diffuse florescence in the other cells. PrP-CFP fusions form aggregates of different shapes: PrP_23–231-СFP, PrP_28–231-CFP and PrP_90–231-СFP preferentially form “dots” or clumps, PrP_110–231-CFP tended to form tape-like aggregates. CFP demonstrates diffuse fluorescence.

Figure 2. Colocalization of PrP_23–231-GFP and Aβ-GFP aggregates with the cell compartments and the lipid particles. The aggregates of PrP_23–231-GFP and Aβ-GFP do not co-localize with vacuoles, nucleus, endosomes and lipid particles. The vacuolar membrane (red) was visualized by staining with the lipophilic dye FM4–64. The endosomes (blue) was visualized by staining with Liso-tracker Blue. Nuclear DNA and mtDNA (blue) were stained with DAPI. The staining of lipid particles (black) was achieved by Sudan Black (see Materials and Methods).

Figure 3. Aβ and full-length PrP or its fragments, fused to fluorescent proteins, exhibit amyloid properties in yeast cells. (A) Analysis of aggregation of Aβ and PrP derivatives fused to YFP and CFP, respectively. Cell lysates were fractionated into soluble (S) and insoluble (I) fractions by centrifugation at 12,000 g, separated on an SDS–PAGE gel, transferred onto a nitrocellulose membrane, and incubated with YFP (CFP)-specific antibody. (B) Analysis of Aβ-YFP and PrP-CFP detergent-resistant polymers by semi-denaturing agarose gel electrophoresis. Cell lysates were treated with 3% sodium lauroyl sarcosinate and analyzed as described using the Ab-13970 antibodies against GFP (Abcam, USA) (see Materials and Methods). (C) Protease K (PK) digestions assay of PrP derivatives and Aβ fused to fluorescent proteins. Samples were digested with 5–100 μg/ml of PK at 37 °C for 1 h. The data presented indicate that all chimeric proteins form aggregates that are highly resistant to proteolysis. The Aβ-YFP protein was detected using the antibodies Ab-13970 against GFP (Abcam, USA) in the sample treated with 40 µg/ml of proteinase K, whereas control GFP completely degraded in the presence of 5 µg/ml proteinase K. Chimeric proteins PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP and PrP_110–231-CFP, produced resistant fragments after the proteinase K digestion with molecular mass approximately 19–21 kDa. The PrP fragments were detected by the Ab-502604 antibodies against PrP (Abcam, USA). (D) The comparative analysis of PK digestions of PrP_23–231-CFP monomers and polymers.

It is known that amyloids exhibit increased resistance to detergents and proteolysis.^10,23,24 To determine, whether aggregates of Aβ-YFP and PrP-CFP form detergent-resistant polymers, cell lysates from respective cultures were treated with 3% Sarkosyl (sodium N-lauroylsarcosine) and analyzed by semi-denaturing detergent agarose gel electrophoresis (SDD-AGE).²³ Data presented in Figure 3B show that all the chimeric proteins formed detergent-resistant polymers. To analyze the resistance of the Aβ-YFP and PrP-CFP to the digestion by proteinase K, we incubated total cell lysates with various concentrations of proteinase K, separated the samples by SDS-PAGE and probed with an YFP-specific or PrP-specific antibody. The data presented in Figure 3C indicate that all chimeric proteins formed aggregates that are highly resistant to proteolysis. The Aβ-YFP protein was still detected in the sample containing 40 µg/ml of proteinase K, whereas control GFP completely degraded in the presence of 5 µg/ml proteinase K. Increased resistance to PK treatment is the specific feature of the polymeric fraction, as monomers of chimeric proteins were essentially undetectable in the supernatant fraction (not shown), possibly due to their degradation in the absense of protease inhibitors.

It was previously shown that PrP_23–231 polymers form in mammalian brains and in yeast cytoplasm a proteinase K-resistant core of approximately 19–21 kDa, which corresponds to amino acids 90 to 231.^19,20 Chimeric proteins PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP and PrP_110–231-CFP, isolated from yeast, produced resistant fragments after the proteinase K digestion with molecular mass approximately 19–21 kDa (Fig. 3C). Protease-resistant products reacted to the PrP-specific antibody but were not detected using by the antibody recognizing CFP (data not shown), confirming that they contain only PrP portion of a chimeric protein. Taken together, these results show that Aβ and full-length PrP or its fragments, fused to fluorescent proteins, exhibit amyloid properties in yeast cells.

The PrP_28–89 and PrP_90–109 segments are involved in the interaction between the PrP polymers and Aβ peptide in yeast

To detect interactions between PrP and Aβ in yeast, we transformed the strain expressing Aβ-YFP with the plasmids expressing PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP or PrP_110–231-CFP. Resulting transformants co-expressing the fusion genes were analyzed by fluorescent microscopy. The PrP_23–231-CFP, PrP_28–231-CFP and PrP_90–231-CFP polymers co-localized with Aβ-YFP with high frequency (from 75 to 85%) (Fig. 4). The frequency of colocalization of PrP_110–231 polymers with Aβ-YFP was approximately five times lower (Fig. 4). The association of PrP aggregates with Aβ was also studied by using a live cell acceptor photobleaching FRET (AΒ FRET) assay. This approach is based on the fact that the energy transfer between the donor (CFP fusion protein) and acceptor (YFP fusion protein) occurs only when the distance between them is not higher than 10 nm.²⁵ Yeast cells co-expressing PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP or PrP_110–231-CFP with Aβ-YFP were analyzed, while cells co-expressing PrP_23–231-CFP with PrP_23–231-YFP or PrP_23–231-CFP with YFP were used as positive or negative controls, respectively. Photobleaching was performed in the regions of the cells containing the PrP-CFP fluorescent clumps. The greatest FRET efficiency (11.5%) was detected for the cells co-expressing PrP_23–231-CFP and PrP_23–231-YFP (Fig. 5). FRET efficiency in the cells co-expressing PrP_23–231-CFP with Aβ-YFP and PrP_28–231-CFP with Aβ-YFP was approximately 6%, whereas the FRET efficiency detected in the cells expressing PrP_90–231-CFP and Aβ-YFP was not significantly different from the negative control (Fig. 5). Thus, the PrP fragment comprising amino acids 28 to 89 plays a role in the molecular association of PrP polymers with the Aβ peptide. FRET was not detected in the cells expressing the PrP_110–231-CFP and Aβ-YFP proteins. Taken together, these data show that, in addition to the PrP_90–109 fragment, the sequence from residues 28 to 89 is also important for the interaction of PrP polymers with Aβ peptide.

Figure 4. Interaction between Aβ and PrP in yeast cells. (A) Analysis of colocalization of PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP or PrP_110–231-CFP aggregates with Aβ-YFP. Efficient colocalization of PrP_23–231-CFP, PrP_28–231-CFP and PrP_90–231-CFP aggregates with Aβ-YFP was observed. (B) The diagram reflecting the frequencies of colocalization of Aβ-YFP with the PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP and PrP_110–231-CFP aggregates. The frequencies of colocalization with error bars are indicated.

Figure 5. The FRET efficiency between PrP-CFP polymers (donors) and Aβ-YFP (acceptor) that reflects the molecular association of proteins analyzed. The FRET efficiency is shown in percent. Error bars are indicated.

Discussion

Our data confirm the previously demonstrated ability of PrP and Aβ-GFP to form aggregates in yeast,^19,26-28 and the similarity in the proteinase K resistance profiles between the PrP aggregates from yeast cells and mammalian PrP^Sc.^19,20 Moreover, we demonstrate that that Aβ-GFP aggregates produced in yeast are also resistant to the digestion by proteinase K. In addition to the previous reported data, were demonstrate that full-length PrP, as well as PrP fragments and Aβ, fused to fluorescent proteins, form detergent-resistant polymers in yeast cells. Therefore, PrP- and Aβ-based aggregates, produced in yeast, exhibit certain amyloid properties. Our results are consistent with recently published data, according to which PrP(90–231) aggregates produced in yeast cells interact with amyloid-specific aptamers.²⁹

We have not detected the colocalization of PrP_23–231-CFP or Aβ-YFP with vacuoles, nucleus, endosomes or lipid particles in yeast cells (Fig. 2). Previously, colocalization of Aβ–GFP with lipid particles has been reported by using staining with Nile Red.²⁷ We used Sudan Black for staining lipid particles, because Nile Red fluorescence is detected with both RFP and GFP filters, even in cells not containing the Aβ-GFP protein. Although we cannot exclude that contradictions with the previously published data may be due to differences between the strains or constructs, our data clearly show that colocalization with lipid particles is not the general pattern of Aβ aggregates in yeast. For the first time, our work demonstrates a colocalization between Aβ-YFP and PrP-derived constructs fused to CFP in the yeast cell. Association between PrP_23–231 polymers and the Aβ peptide is also proven using FRET, confirming direct physical interaction between these proteins. Our findings demonstrate the suitability of the yeast-based experimental assay for studying interactions between mammalian amyloidogenic proteins. Recently, it was shown that the interaction of Aβ oligomers with PrP^C depends on the presence of two short segments in the N-terminal regions of PrP, includes amino acid residues 23–27 and 95–110, respectively.^8,12 However, it remained unclear whether or not the same sequences determine the interaction of PrP polymers with Aβ. We have shown that amyloid-like polymers of PrP_23–231-CFP efficiently co-localize and physically interact with Aβ-YFP in the yeast cell. The segment 23–27 was dispensable for both co-localization and physical interaction detected by FRET, indicating that it is unlikely to play any role in the interaction between PrP polymers and Aβ. The PrP_28–231-CFP and PrP_90–231-CFP polymers both co-localized with Aβ-YFP, however FRET detected a physical interaction with Aβ-YFP only for PrP_28–231-CFP. One possibility could be that PrP_90–231 polymers tend to associate only with aggregated form of Aβ. This decreases the proportion Aβ molecules present in direct contact with PrP, thus decreasing FRET efficiency.

It was demonstrated previously that the positively charged residues in fragments PrP_23–27 and PrP_90–110 are critically important for the effective binding of PrP^C to Aβ oligomers.¹³ However, the PrP_28–89 fragment, which, according our data, regulates the molecular association of PrP polymers with the Aβ peptide, is not enriched with charged amino acids. This suggests that in addition to cationic interactions, other factors influence binding of PrP polymers to Aβ. Further investigation of these factors is made possible by the yeast model developed in our work.

The PrP_110–231-CFP polymers, unlike the PrP_90–231-CFP, PrP_28–231-CFP or PrP_23–231-CFP, polymers, co-localized with Aβ-YFP with a low frequency. In this case, it is probable that we observed a non-specific interaction between heterogeneous amyloids, which has previously been described Derkatch et al.³⁰ According to our results, the absence of the PrP_90–109 sequence affects the morphology of the PrP polymers and prevents their interaction with Aβ. Based on this result and previously published data,^12,13,31 we conclude that the PrP_90–109 sequence is required for the interaction of both monomeric and polymeric PrP with Aβ. Additionally, our data show, for the first time, that the PrP_28–89 fragment is important for the molecular association between PrP polymers and the Aβ peptide in vivo.

Materials and Methods

Strains and growth conditions

The yeast strain BY4742 (MATα his3Δ1 leu2Δ lys2Δ ura3) (Invitrogen) was employed for this study. Standard yeast genetic techniques, media and cultivation conditions were used;^32,33 the yeast cultures were grown at 30 °C. The E. coli strain DH5α³⁴ was grown at 37 °C in Luria-Bertani (LB) broth or on LB agar plates containing ampicillin for plasmid selection.³⁵

Plasmids

The plasmids used in this work are described in Table 1. All the plasmids, with the exception of pcDNA3–1-3F4,³⁶ are shuttle vectors that can propagate in S. cerevisiae and E. coli. The vector pRS425 was described previously by Sikorski and Hieter.³⁷ The pSP-YFP and pSP-CFP centromeric plasmids were kindly provided by Dr S Zadorsky (Saint Petersburg State University). The pcDNA3–1-3F4 plasmid contains the mouse Prnp gene modified to express PrP L108M/V111M for immunostaining with the monoclonal antibody 3F4. The plasmid PGPD-GFP(URA3) was described previously by Rubel et al.³⁸ The plasmid PGPD-PrP-GFP(URA3)³⁸ contains the mouse PrP encoding sequence from residue 90 to 231 aa fused in-frame with the GFP ORF and placed under the control of the GPD promoter. The pGPD-YFP(URA3) plasmid was constructed by replacing the 0.7 kb SacII-SacI fragment containing the GFP ORF from PGPD-GFP(URA3) with the 0.7 kb SacII-SacI fragment of yellow fluorescent protein (YFP) obtained from the plasmid pSP-YFP. The pGPD-GFP(LEU2) plasmid was constructed by inserting the 1.4 kb SalI-SacI fragment containing the GPD promoter and the GFP ORF from PGPD-GFP(URA3) into SalI-SacI-digested pRS425. The pGPD-CFP(LEU2) plasmid was constructed by replacing the 0.7 kb SacII-SacI fragment containing the GFP ORF from pGPD-GFP(LEU2) with the 0.7 kb SacII-SacI fragment containing the CFP ORF from pSP-CFP. The plasmids pGPD-PrP₂₃-CFP(LEU2), pGPD-PrP₉₀-CFP(LEU2), pGPD-PrP₂₈-CFP(LEU2) and pGPD-PrP₁₁₀-CFP(LEU2) were constructed as follows. The BamHI-SacII-flanked sequences of PrP_23–231, PrP_28–231, PrP_90–231 and PrP_110–231 were PCR generated from the plasmid pcDNA3–1-3F4 using the FPrP23, FPrP28, FPrP90, FPrP110 and RPrP231 primers listed in Table 2. These fragments were inserted into BamHI-SacII-digested pGPD-CFP(LEU2) downstream of P_GPD. The pGPD-PrP₂₃-YFP(URA3) plasmid was constructed by inserting the BamHI-SacII fragment containing the PrP₂₃ ORF from pGPD-PrP₂₃-CFP(LEU2) into the BamHI-SacII-digested pGPD-YFP(URA3) plasmid. The PGPD-Aβ-GFP(URA3) plasmid was constructed as follows. The sequence encoding amino acids 1–40 of the human amyloid β peptide (Aβ) was obtained from human brain mRNA by reverse transcription PCR using the FAβ and RAβ primers presented in Table 2. The PCR-generated BamHI-SacII fragment was inserted into the PGPD-PrP-GFP(URA3) plasmid digested with BamHI and SacII. The pGPD-Aβ-YFP(URA3) plasmid was constructed by inserting the 0.7 kb SacII-SacI fragment containing the YFP ORF from pSP-YFP into the SacII-SacI-digested PGPD-Aβ-GFP(URA3) plasmid downstream of P_GPD.

Table 1. The plasmids obtained in this work

Plasmid name	Plasmid marker	Promoter/expression cassette
pGPD-GFP(LEU2)	2μ/LEU2	PGPD-GFP
PGPD-Aβ-GFP(URA3)	2μ/URA3	PGPD-Aβ-GFP
pGPD-Aβ-YFP(URA3)	2μ/LEU2	PGPD-Aβ-YFP
pGPD-PrP23-YFP(URA3)	2μ/LEU2	PGPD-PrP23-231-YFP
pGPD-PrP23-CFP(LEU2)	2μ/LEU2	PGPD-PrP23-231-CFP
pGPD-PrP28-CFP(LEU2)	2μ/LEU2	PGPD-PrP28-231-CFP
pGPD-PrP90-CFP(LEU2)	2μ/LEU2	PGPD-PrP90-231-CFP
pGPD-PrP110-CFP(LEU2)	2μ/LEU2	PGPD-PrP110-231-CFP
pGPD-YFP(URA3)	2μ/URA3	PGPD-YFP
pGPD-CFP(LEU2)	2μ/LEU2	PGPD-CFP

Table 2. Primers used in this study

Primers	Sequence 5′-3′
FPrP23 (BamHI)	CCCGGATCCT ATATGTCTAA AAAGCGGCCA AAGCCTGGAG GGT
FPrP28 (BamHI)	GACTTTGGAT CCTATATGCC TGGAGGGTGG AACAC
FPrP90 (BamHI)	GTTCATGGAT CCTATATGTC TCAAGGAGGG GGTACCCAT
FPrP110 (BamHI)	GAATAGGATC CACAATGCAT ATGGCAGGGG CTGCG
FAβ (BamHI)	GGGTCCACGG ATCCTATATG TCTGATGCAG AATTCCGACAT
RAβ (SacII)	GTTATAAAGG ATCCGACAAC ACCGCCCACC AT
RPrP231(SacII)	CATCCGCGGG CTGGATCTTC TCCCGTCGTA ATAGGCCT

DNA and protein analysis

Standard procedures were used for DNA isolation and plasmid construction (Sambrook et al., 1989).³⁵ The preparation of the yeast lysates and differential centrifugation were performed according to a described protocol Chernoff et al.³⁹ The fractions were separated by 12% SDS-PAGE. PK-resistant PrP was separated using a 4–20% gradient polyacrylamide gel. SDD-AGE of the PrP-GFP and Aβ-GFP polymers was performed as described Kryndushkin et al.,²³ Bagriantsev et al.⁴⁰ with the modifications. The protein lysates (50–100 μg of total protein) were treated with 3% sodium lauroyl sarcosinate in the sample buffer (25 mM Tris, 200 mM glycine, 5% glycerol, and 0.025% bromophenol blue) for a 7 min at room temperature and run on 1.2% agarose gels. After the electrophoresis, proteins were transferred from gels to Immobilon-P PVDF (GE Healthcare) electrophoretically. Western blotting was performed with the chicken primary monoclonal antibodies Ab-13970 against GFP (Abcam) or with the rabbit primary monoclonal antibodies Ab-502604 against PrP (Abcam). The reactions with the secondary anti-chicken or anti-rabbit antibodies conjugated to horseradish peroxidase and chemiluminescent detection were performed using the ECL system (GE Healthcare). Coomassie staining and Bradford technique were used for normalization of the total protein amount.

Protease digestion assay

All the protease K (PK) digestions were performed in a lysis buffer (20 mm Tris-HCl, 150 mm NaCl, and 1 mm EDTA). The yeast lysates were digested with 5–100 μg/ml of PK at 37 °C for 1 h; the reactions were terminated with 2 mm phenylmethylsulfonyl fluoride.

Fluorescence microscopy

The fluorescence microscopy assays of GFP fusion protein were performed using a Leica DM6000B microscope and Leica QWin Standart 3.2.0. software (Leica Microsystems GmBH). The TCS SP5 confocal laser-scanning microscope and “LAS AF Application Wizard Version 1.7.0” software (Leica Microsystems GmBH) were used for the detailed analysis of aggregates, protein colocalization and live cell acceptor photobleaching FRET (AΒ FRET) experiments. For excitation, we used a 488 nm argon laser for the GFP or GFP fusion proteins and a 594 nm helium-neon laser for FM4–64. The emission filters were 505–515 nm for the GFP or GFP fusion proteins and 615–758 nm for FM4–64. DAPI and Liso-tracker Blue-stained cells were detected using an argon-UV laser with 405 nm excitation and 425–475 nm emission. FM4–64 was detected using a helium-neon laser with 594 nm excitation and 615–758 nm emission. For the FRET experiments, a donor (CFP fusion protein) was co-expressed with an acceptor (YFP fusion protein); for excitation, we used 458 nm and 514 nm argon lasers for the CFP or CFP fusion proteins and YFP or YFP fusion proteins, respectively. The emission filters were 461–510 nm for the CFP or CFP fusion proteins and 518–580 nm for the YFP or YFP fusion proteins. FRET measurements were performed using the FRET acceptor photobleaching program of the Leica confocal software. In the photobleaching procedure, the cells were bleached using a 514 nm laser beam at 100% intensity; the bleach time was 10 s.

For staining with FM4–64 (Invitrogen), mid-log-phase cells were incubated with 8 μM dye (5 mM FM4–64 in dimethyl sulfoxide) for 2 h in liquid selective media. The cells were then washed twice with water and observed for vacuolar staining. The cells were stained with 4,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories) for the visualization of nuclear DNA and mtDNA. The cells were first fixed with 4% formaldehyde for 30 min in culture, washed twice with phosphate-buffered saline (PBS) and permeabilized by incubation in 0.1% PBS Triton X-100 for 10 min. The cells were then washed twice with PBS and incubated for 10 min with VECTASHIELD Mounting Medium (Vector Laboratories) with DAPI (1.5 μg/ml). Liso-tracker Blue (Invitrogen) staining was performed according to the manufacturer’s protocol. The staining of lipid particles in the yeast cells was achieved by Nile Red according to Verstrepen et al.⁴¹ or by 0.5% Sudan Black B in ethylene glycol for 60 min according to Thakur et al.⁴²

Statistical analyses

In order to calculate the FRET efficiency, three independent cultures co-expressing Aβ-YFP with PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP or PrP_110–231-CFP were selected in each cases. The cells that showed both PrP-CFP and Aβ-YFP fluorescence were analyzed. Photobleaching was performed in the regions of the cells containing the PrP-CFP fluorescent clumps. The FRET efficiency (FRETeff) was calculated by software according to the following equation: FRETeff = [Dpost−Dpre]/Dpost, where Dpost is the donor fluorescence after photobleaching and Dpre is the donor fluorescence before photobleaching. We applied the Kruskal-Wallis test with the multiple comparison of z-values for the analysis of the FRET efficiency.

In order to measure frequencies of colocalization between the PrP-CFP polymers and Aβ-YFP, four independent cultures co-expressing Aβ-YFP with PrP_23–231-CFP, PrP_28–231-CFP, PrP_90–231-CFP or PrP_110–231-CFP were analyzed in each case. No less than 20 cells containing visible signal of PrP-CFP and Aβ-YFP were analyzed in each culture. The frequencies of colocalization were calculated according to the following equation: [Na/Nt] × 100 where Na – number of PrP-CFP aggregates that co-localize with Aβ-YFP and Nt – total number of PrP-CFP aggregates analyzed. Frequencies were compared by using the Mann-Whitney U-test. Statistical analyses were performed using Statistica v.7.0. (StatSoft Inc., USA) software; P ≤ 0.05 was considered to be statistically significant.

Abbreviations:
AD	=	Alzheimer disease
Aβ	=	amyloid beta peptide
PrPC	=	cellular isoform of prion protein
PrPSc	=	prion isoform of prion protein
AΒ FRET	=	acceptor photobleaching fluorescence resonance energy transfer

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We are grateful to S Zadorskii for kindly providing pSP-YFP and pSP-CFP plasmids, A Saifitdinova and V Korzhova for helping with FRET experiments, N Degtyareva for critical reading of the manuscript. This work was supported by the program of the Presidium of RAS “Basic sciences—medicine” (to APG), the Ministry of Education and Science of Russian Federation (to YOC and AAR); the grant 1 R03 TW006965 01A1 from Fogarty Center (National Institutes of Health, USA) to YOC and APG; the Russian Federation President grant for young PhDs (MK-1152.2013.4) to AAR; the research grant 1.50.2218.2013 from St. Petersburg State University to YOC. The authors acknowledge the Resource Centers “CHROMAS” (SPbSU) for technical support.

10.4161/pri.26867