DNA aptamers detecting generic amyloid epitopes

Abstract

Amyloids are fibrillar protein aggregates resulting from non-covalent autocatalytic polymerization of various structurally and functionally unrelated proteins. Previously we have selected DNA aptamers, which bind specifically to the in vitro assembled amyloid fibrils of the yeast prionogenic protein Sup35. Here we show that such DNA aptamers can be used to detect SDS-insoluble amyloid aggregates of the Sup35 protein, and of some other amyloidogenic proteins, including mouse PrP, formed in yeast cells. The obtained data suggest that these aggregates and the Sup35 amyloid fibrils assembled in vitro possess common conformational epitopes recognizable by aptamers. The described DNA aptamers may be used for detection of various amyloid aggregates in yeast and, presumably, other organisms.

Keywords: :

• Saccharomyces cerevisiae
• Sup35/eRF3
• [PSI⁺]
• amyloid
• aptamer
• huntingtin
• polyglutamine
• prion

Introduction

Amyloids were originally described as insoluble protein aggregates causing a number of human diseases, such as Alzheimer, Parkinson, Huntington and prion diseases. Remarkably, unlike the proteins which form them, all amyloids are structurally similar, being composed of fibers with a specific cross-β structure (reviewed refs. 1 and 2). Amyloids were also found in lower eukaryotes, in particular, in the yeast Saccharomyces cerevisiae, where they may manifest themselves as non-chromosomal heritable genetic determinants (prions).

The best studied of yeast prions is [PSI⁺], which reflects the aggregated prion state of the translation termination factor eRF3, also called the Sup35 protein. Aggregation reduces the Sup35 function in translation and causes the nonsense suppressor phenotype.^3,4 Aggregation of Sup35 is mediated by its N-terminal (N) domain, which is mainly responsible for the prion properties. The C-terminal (C) domain is required for the translation termination and viability. The middle (M) domain is likely to serve as a spacer between the two other domains and also contains a 20-residues sequence important for Sup35 interaction with the Hsp104 chaperone and prion polymer fragmentation.⁵ Both M and C domains are inessential for the Sup35 conversion into the prion state.^6,7 [PSI⁺] can appear de novo spontaneously with very low frequency,⁸ which may be significantly increased by transient overproduction of full-length Sup35 or its N-terminal fragments^9,10 in the presence of another prion, [PIN⁺]^11,12 related to the prion state of the Rnq1 protein.¹³ [PSI⁺] has variants (also called strains) differing in the strength of their nonsense suppressor phenotype. Strong [PSI⁺] variants show efficient nonsense suppression, while its weak variants are less efficient in nonsense suppression.¹⁰ Variation in phenotypic manifestation of [PSI⁺] is related to heritable differences in the structure of Sup35 prion aggregates.^14,15 When Sup35 is constantly overproduced in the presence of [PIN⁺], it mainly forms aggregates, which, in contrast to prions, cannot propagate due to poor fragmentation by the Hsp104 chaperone.¹⁶ In vitro the Sup35 protein or its N-terminal fragments, isolated from Escherichia coli, form amyloid fibrils. These fibrils can transform non-prion [psi^-] yeast cells to the prion [PSI⁺] state,^15,17 thus demonstrating a direct relation between such fibrils and the prion form of Sup35 propagated in vivo. Though the same relation to in vitro formed fibrils was shown for the yeast prion Ure2 and Rnq1,^18,19 the ability of other amyloidogenic proteins to form amyloid aggregates in yeast was never proved, and the only indication of amyloid nature of such aggregates is their insolubility in the presence of strong ionic detergents, such as SDS or sarcosyl, which dissolve almost all other macromolecular complexes in yeast cells.^20,21 Detection of amyloids formed in vivo requires availability of reagents distinguishing them from non-amyloid protein aggregates and complexes, which may represent aptamers.

Aptamers are single-stranded DNA or RNA molecules able to bind to various molecular targets. They may have a wide range of applications in therapeutics, biosensing, diagnostics and research (reviewed ref. 22). Aptamers can be obtained from a large library of random oligonucleotides through an in vitro selection procedure called SELEX (Systematic Evolution of Ligands by an EXponential enrichment).²³ Earlier we used this procedure to generate a set of DNA aptamers against amyloid fibrils formed in vitro by the Sup35 NM fragment.²⁴ Here we show that such DNA aptamers are able to interact with SDS-resistant polymers of the yeast prionogenic protein Sup35, as well as with that of several other amyloidogenic proteins, thus demonstrating their suitability for amyloid detection.

Results

DNA aptamers interact with SDS-insoluble protein aggregates of [PSI⁺] [PIN⁺] cells

Earlier, using the SELEX procedure, we have obtained ten DNA aptamers to in vitro-generated Sup35 fibrils, which could recognize these fibrils in an ELISA-like assay.²⁴ Here, we studied the ability of these and 11 additional aptamers, obtained through a similar procedure, to interact with ex vivo detergent-insoluble aggregates of Sup35 and of some other proteins in the dot-blot and western blot procedures. The aptamers (Table 1) were chemically synthesized and 5′-biotinylated to allow an aptamer-target binding assay. These aptamers lacked ten terminal nucleotides of constant regions at both sides, which did not impair the interaction of aptamers with Sup35 fibrils. More extensive truncation was not tried, since complete removal of 25 nucleotides from flanking regions in aptamer S5 hindered its binding to Sup35NM fibrils.²⁴

Table 1. Nucleotide sequences of variable regions of selected DNA aptamers

Aptamer ID	Variable sequences
S1	CCGACGCTAGCGACCTATCCCATAAACGAGACACGGGTCC
S2	CCGACCAACGTTGTTACCTTCCCTATAACAACAGGAGCCG
S3	CCCACTGCACACGTCTCTAAACACACGCCCCGTCATCCGG
S4	CCCCGGGAACCACCCTATAACGTACCGCTAAGTACGCGCG
S5	CCACCCTGAGCTCCGTAGACATTTATCGCTCTACCCGCCC
S6	CGCGGGAGGGGGAGGGGAAGAGCCTGGGGGAGACGGCAGG
S7	CCCGACAACCAACGGCTTTAGGGTGCAGCCAACCCGCCTG
S8	CCCGACTTTCCGGTATACGCATTCCACAACACTGTCCTGG
S9	CCACCAGTGATAACGGAAGGAACAGTAGAAATCCGTCCG
S10	CGGCACCGAACGCCCCACAGGAAACAGCGTCTGCGCGCG
A2	CCCTGGGGTTCCAGAGACGAGTAGTAGAACGCATTCCCGG
A5	CCCCTCAACAGCGTGTTGCACCCAATCAGTACCTGCCGTG
A8	CCATGGAGACCCGCGTATTTGTAGTGTTTATGCCCGCG
A18	CGCCCGTGAATAAGCTACCATCCAGAACGACGCCTGGC
A21	CCGGGACAGCGTATTGAAGCTATGAGGCGGCCCTTCCCCA
A22	CGACCGTGAAAAGCTAGGTAATCTTACCCATTCCGCGCGCGC
A24	CGACCCAGTCAAGTTGTAGCTATTGTAACGCCCAAACCGC
A26	CCGCAGCCGACTAGGGCGGAGGATTCAGACCCCGACGCG
A31	CCCCGACGCCGTGGACTAGGTAACACCCTATCTTACTGC
A36	CCAGGCGCCAGAGCCGCGTGTTCGTTTCCTCACGCTTTGC
A37	CCCGGGGCGCCAGAGACGAGTAGGTGCAGCGGTTGTCATTGGG

The variable regions shown are flanked with constant sequences: 5′-agcactaatctatgc … accttatgttgtagc-3′. Except for the A2-A37 pair, which has 90% similarity over 24 nucleotides (indicated in bold), aptamers show no significant sequence similarity of their variable regions neither with each other, nor with DNA aptamers against the prion form of hamster PrP.³⁹

All selected aptamers were tested in a dot-blot assay for their ability to bind to SDS-insoluble aggregates purified from lysates of yeast cells containing [PSI⁺] variant with strong suppressor phenotype ([PSI⁺]s) and [PIN⁺]. Most of the aptamers were able to bind to these aggregates, albeit with different efficiency (Fig. 1). Three DNA aptamers were removed from further study, since two of them (A31 and A37) stained the aggregates faintly and one (A24) weakly stained the [psi^-] [pin^-] control.

Figure 1. DNA aptamers detect prion aggregates of Sup35 partially purified from cells of 5V-H19 [PSI⁺]s [PIN⁺]. The [psi^-] [pin^-] derivative of this strain was used as a control. Aptamer IDs are presented on the left. Samples of protein aggregate fraction were serially diluted in 2-fold increments, as indicated on the top.

DNA aptamers interact with detergent-insoluble aggregates of different amyloidogenic proteins

To further characterize the aptamers, we tested their binding to various detergent-insoluble aggregates. These were isolated from [PIN⁺] cells either lacking [PSI⁺] or containing its weak ([PSI⁺]w) and strong variants. Besides prions, we also studied interaction of aptamers with non-heritable aggregates of overproduced Sup35 or detergent-insoluble aggregates of mouse PrP_90–231 and of the model 103Q-GFP protein composed of expanded polyglutamine (polyQ) domain of human huntingtin fused with the green fluorescent protein (GFP). The PrP and huntingtin proteins are related to transmissible encephalopathies (reviewed refs. 25 and 26) and Huntington disease (reviewed refs. 27 and 28), respectively.

Many of the studied aptamers were able to recognize not only Sup35 aggregates, but also detergent-insoluble aggregates of the other proteins. This allowed us to assign the aptamers to five classes by the specificity of interaction (Table 2 and Fig. 2). Four aptamers recognized Rnq1 aggregates (classes III and V). These aggregates were isolated from cells of the 74-D694 [psi^-] [PIN⁺] strain. Aggregates from the cells containing prion or non-prion polymers of Sup35 were stained by nearly all of the aptamers. Although these cells also contained [PIN⁺], the aptamers of the I, II and IV classes could not interact with the Rnq1 prion polymers and thus detected only Sup35 aggregates (Table 2). The aptamer S5 (class V) was able to interact with all tested targets. Nevertheless, in [PSI⁺] [PIN⁺] cells it primarily detected Sup35 aggregates, as evident from SDD-AGE polymer analysis (Fig. 3). We observed earlier that differential deletion of oligopeptide repeats in the Sup35 prion-forming N-terminal domain results in formation of Sup35 prion polymers of different size.²⁹ Staining of such Sup35 polymers with anti-Sup35 antibody and aptamer S5 revealed a very similar variation of the polymer size. Thus, Sup35 aggregates were the major target of S5. In part, this can be related to Sup35 being much more abundant than Rnq1.

Table 2. DNA aptamers interact differently with various prion and non-prion detergent-insoluble aggregates

Target	Class of DNA aptamers (aptamer ID)
	I	II	III	IV	V
	(S2, S3, S6, S7, S9, S10, A2, A8, A21)	(S1, A18, A22, A36)	(S4, S8, A5)	(A26)	(S5)
^a[PSI^+]s [PIN^+]	+	+	+	+	+
^a[PSI^+]w [PIN^+]	+	+	+	-	+
^b↑Sup35 and [PIN^+]	+	+	+	-	+
^c[PIN^+]	-	-	+	-	+
^d103Q-GFP	-	+	-	-	+
^ePrP90–231	+	+	+	+	+
^f[psi^-] [pin^-]	-	-	-	-	-

Detergent-insoluble protein aggregates were isolated from:^a5V-H19 [PIN⁺] with either [PSI⁺]s or [PSI⁺]w (Rnq1 and Sup35 prion aggregates);^b74-D694 [psi^-] [PIN⁺] overproducing Sup35 (non-prion Sup35 aggregates and Rnq1 prion aggregates);^c74-D694 [psi^-] [PIN⁺] (Rnq1 prion aggregates);^d74-D694ΔS35 [pin^-] overproducing 103Q-GFP (103Q-GFP aggregates);^e74-D694ΔRNQ1 [psi^-] expressing PrP_90–231 (PrP_90–231 aggregates);^f74-D694 [psi^-] [pin^-], control. Aptamers S6, A2 and A21 detected [PSI⁺]s aggregates of Sup35 better than other aptamers of this class. The aptamer S5 stained Rnq1 prion aggregates weaker than aptamers of the class III.

Figure 2. Cross-reactivity of DNA aptamers with partially purified detergent-insoluble aggregates of different proteins. (A) Dot-blot assay. Aptamer representatives of each class are presented. Samples of the protein aggregate fraction were serially diluted in 4-fold increments. (B) SDD-AGE analysis. Only targets stained differently by the aptamer classes are presented. Protein aggregates were partially purified from cells of the following strains: PrP, 74-D694ΔRNQ1 [psi^-] expressing PrP_90–231; 103Q, 74-D694ΔS35 [pin^-] expressing 103Q-GFP and Sup35C; [PIN⁺], 74-D694 [psi^-] [PIN⁺]; NPA Sup35, non-prion aggregates of Sup35 from 74-D694 [psi^-] [PIN⁺] overproducing Sup35; [PSI⁺]s and [PSI⁺]w, 5V-H19 [PIN⁺] carrying either [PSI⁺]s or [PSI⁺]w respectively. Controls: [psi^-], aggregates from 5V-H19 [psi^-] [pin^-] cells and mono, monomers obtained from respective aggregates upon their boiling. The staining of mono 103Q by aptamers A18 and S5 probably relates to incomplete dissolution of 103Q upon boiling.

Figure 3. Interaction of the S5 aptamer with SDS-insoluble aggregates of truncated Sup35. Sup35 aggregates were purified from cells of the strain 22V-H63ΔS35 [PSI⁺] [PIN⁺], transformed with one of the plasmids of the pUKC1512 series, which carried wild-type SUP35, or R1-2, R1-3 and R1-4 deletion alleles, encoding Sup35 with indicated number of oligopeptide repeats in the prion-forming N-terminal domain. Staining with the anti-Sup35 antibody and the S5 aptamer.

Five aptamers stained 103Q-GFP aggregates isolated from cells of the 74-D694ΔS35 [pin^-] strain, which overproduced 103Q-GFP and contained only the non-polymerizing Sup35C protein. The assignment of aptamers specificity in this case was complicated by the observations that 103Q-GFP polymers can seed polymerization of various chromosomally-encoded glutamine/asparagine-rich proteins^30,31 and these polymers could also be detected by aptamers. However, it is unlikely that these polymers contribute significantly to staining, since they should be much less abundant than aggregates of overproduced 103Q-GFP. Otherwise, aptamers would stain all isolated aggregates. However, quite the contrary, among 18 aptamers only five, belonging to the classes II and V, interacted with aggregates isolated from cells expressing 103Q-GFP (Table 2).

All tested aptamers stained aggregates isolated from cell lysates of the strain 74-D694ΔRNQ1 [psi^-] expressing PrP_90–231. The staining was not due to induced aggregation of Sup35, since no detergent-insoluble Sup35 polymers were detected with anti-Sup35 antibodies (Fig. 4). It is surprising that all aptamers stained aggregates of PrP_90–231, which shares no similarity with Sup35, while only few aptamers stained aggregates of Q103-GFP or Rnq1, which, similarly to Sup35, have glutamine-rich polymerization domains. It is known that PrP is capable of nonspecific interaction with nucleic acids³² via its N-terminal region,³³ probably due to the presence of two lysine clusters, one of which is present in PrP_90–231. We observed that this interaction was insufficiently strong to ensure staining of PrP_90–231, since the aptamers did not stain soluble heat-denatured PrP_90–231 (Fig. 2A), while the shortened aptamer S5 lacking 15 nucleotide constant flanks and the 70-nucleotide long DNA aptamer against human pro-urokinase³⁴ did not bind either PrP_90–231 nor Sup35 aggregates (Fig. 4). However, the weak nonspecific affinity of polynucleotides toward PrP could strengthen the specific interaction of aptamers with PrP_90–231 aggregates, resulting in universal binding of all aptamers to these aggregates.

Figure 4. Staining of detergent-insoluble PrP_90–231 aggregates of 74-D694ΔRNQ1 [psi^-] by aptamers is not due to non-specific DNA binding. [PSI⁺] [PIN⁺], 5V-H19 [PSI⁺]s [PIN⁺]; [psi^-] [pin^-], 5V-H19 [psi^-] [pin^-]; PrP, 74-D694ΔRNQ1 [psi^-] expressing PrP_90–231. Sup35 Ab and PrP Ab, antibodies against Sup35 and PrP, respectively; S5, S5-40 and U-1, DNA aptamers S5, S5 without 15 nucleotides from each flank and aptamer against human pro-urokinase, respectively.

Dot-blot assay could not be used for detection of protein aggregates by biotin-labeled aptamers in cell lysates, since lysates contain biotinylated proteins, which are stained by streptavin-peroxidase (data not shown). However, separation of SDS-insoluble aggregates from soluble proteins by SDD-AGE makes purification of such aggregates unnecessary for the study of their interaction with aptamers. Thus, SDD-AGE is a suitable approach for the aptamer-assisted detection of SDS-insoluble polymers of various proteins (Fig. 5).

Figure 5. S5 aptamer interacts with SDS-insoluble aggregates in crude cell lysates. [PIN⁺], 74-D694 [psi^-] [PIN⁺]; 103Q, 70Q and 76QY, 74-D694 [psi^-] [pin^-], expressing 103Q-GFP,³⁷ 70Q-Sup35MC and 76QY-Sup35MC,³⁸ respectively; [PSI⁺], 5V-H19 [PSI⁺]s [PIN⁺]; [psi^-], 5V-H19 [psi^-] [pin^-]. Staining with S5 aptamer or with antibodies against Rnq1, GFP or Sup35.

Discussion

In this work we observed that DNA aptamers obtained against amyloid fibrils of the yeast prionogenic protein Sup35 can bind to detergent-insoluble aggregates formed in yeast cells by various structurally and functionally unrelated proteins. This suggests that cellular detergent-insoluble aggregates and the Sup35 fibrils assembled in vitro possess common structures recognizable by aptamers and, therefore, the interaction with aptamers indicates the amyloid nature of such aggregates.

Earlier, an RNA aptamer to β₂-microglobulin amyloid was generated, which was able to interact with amyloids of lysozyme, though not with three other tested amyloids.³⁵ Similar, but wider, cross-reactivity was shown by DNA aptamers studied here: aptamers against Sup35 amyloid fibrils could interact with prion and non-prion detergent-insoluble aggregates of different proteins isolated from yeast cells. The observed interactions depended on both the aptamer and the target aggregate used in the binding assay. It should be noted that since the amounts of aggregates of different proteins in samples were difficult to equalize correctly, the quantitative comparison of the ability of aptamers to stain different targets was not possible. On the other hand, different aptamers could be compared with each other by the efficiency of interaction with the same target.

Basing on the specificity of interaction, all tested DNA aptamers were assigned to five classes (Table 2). Most of the aptamers detected prion and non-prion aggregates of Sup35. Nevertheless, binding of some aptamers depended on the variations of Sup35 amyloid folding. Three aptamers (S6, A2 and A21; class I) differed from others being especially efficient against [PSI⁺]s, whereas one aptamer (A26; class IV) did not interact with [PSI⁺]w and the non-prion aggregates of Sup35, which should be related to differences in amyloid structure. Aptamers of the classes III and V were able to recognize Rnq1 prion aggregates and class II and V aptamers detected aggregates of 103Q-GFP. Therefore, in vitro formed Sup35 fibrils possess structural determinants, which are shared not only by prion and non-prion aggregates of this protein, but also by detergent-insoluble aggregates of other proteins.

Sup35, Rnq1 and polyQ proteins all have glutamine-rich polymerization domains, while PrP does not share this feature. Nevertheless, PrP_90–231 aggregates were readily recognized by all tested aptamers. The only structural trait common for amyloid conformations of PrP and other studied proteins is the presence of intermolecular β-sheets typical of all amyloids. The presence of common epitopes in amyloid aggregates of polyQ and of proteins, which are not rich in glutamine and asparagine residues, such as Alzheimer Aβ(1–40) peptide, lysozyme and some other proteins, was also observed with the use of specific conformational antibodies.³⁶

The results of this work show that DNA aptamers can be used for the detection of various amyloid aggregates. The higher stability of DNA aptamers gives them an advantage over RNA aptamers and allows using them in complex media containing RNases, for example, for detection of amyloid aggregates of cell lysates, as was shown in this work. This makes described DNA aptamers convenient reagents which can be used for the detection and characterization of amyloid aggregates in yeast and other organisms.

Materials and Methods

Plasmids, strains and growth conditions

The strains and plasmids used in this study are described in Tables 3 and 4, respectively. Yeast were grown at 30°C in rich (YPD, 1% yeast extract, 2% peptone, 2% glucose) or synthetic (SC, 0.67% yeast nitrogen base, 2% glucose supplemented with the required amino acids) media. To induce the synthesis of the 103Q-GFP chimeric protein, transformants with the p103Q-GFP plasmid were incubated in liquid selective media with 2% raffinose as a sole carbon source to mid-log phase. Then, the culture was diluted with an equal volume of the same medium with galactose instead of raffinose and cells were grown for 6 h. The final concentration of galactose in the medium was 2%.

Table 3. Strains

Strain	Genotype	Source
5V-H19	MATa SUQ5 ade2–1 can1–100 leu2–3,112 ura3–52	7
74-D694	MATa ade1–14 his3Δ200 leu3–112 trp1–289 ura3–52	40
74-D694ΔRNQ1 [psi^-] [pin^-]	The same as 74-D694 [psi^-] [pin^-], but with rnq1::HIS3	16
74-D694ΔS35 [psi^-] [pin^-]	The same as 74-D694 [psi^-] [pin^-], but with sup35::TRP1	37
22V-H63ΔS35 [PSI^+] [PIN^+]	MATa ade2–1 SUQ5 kar1 lys1 his3 ura3 leu2 ΔSUP35 [PSI^+] [PIN^+]	29

The [PSI⁺]s [PIN⁺], [PSI⁺]w [PIN⁺] or [psi^-] [pin^-] variants of 5V-H19 and [psi^-] [pin^-] or [psi^-] [PIN⁺] variants of 74-D694 were used. The strain 74-D694ΔRNQ1 [psi^-] does not carry [PIN⁺] due to disruption of the RNQ1 gene. The strain 74-D694ΔS35 [psi^-] [pin^-] with disruption of the chromosomal SUP35 contains pRS315-SUP35C, which supports its viability but cannot support [PSI⁺] propagation. The viability 22V-H63ΔS35 [PSI⁺] [PIN⁺] was supported by one of the plasmids of the pUKC1512 series.

Table 4. Plasmids

Plasmid	Characteristics	Source
YEplac181-SUP35	Multicopy LEU2 plasmid with SUP35	16
pL-moPrP(90–231)	Multicopy LEU2 plasmid, encoding truncated mouse PrP (PrP90–231) under the control of GPD promoter	A. Galkin, unpublished
p70Q-, p76QY-Sup35MC,	Multicopy URA3 plasmids, encoding fusion of 70Q, or 76QY with Sup35MC under the control of SUP35 promoter, respectively	38
p103Q-GFP	Multicopy URA3 plasmid, encoding fusion of expanded polyQ region of huntingtin (103Q) with GFP under the control of GAL1 promoter	37
pUKC1512-SUP35, SUP35R1–2, SUP35R1–3, SUP35R1–4	Centromeric HIS3 plasmids, carrying SUP35 wild-type allele, or SUP35 deletion alleles, encoding Sup35 with two, three or four first oligopeptide repeats of the prion domain, respectively	41
pRS315-SUP35C	Centromeric LEU2 plasmid with SUP35-C	42

Preparation of yeast cell lysates

Yeast cultures grown in liquid selective media were harvested, washed in water and lyzed by glass beads in buffer A: 30 mM TRIS-HCl, pH 7.4, 150 mM NaCl, 10 mM EDTA, 1 mM dithiothreitol and 1% Triton X-100. To prevent proteolytic degradation, Complete protease inhibitor cocktail (Roche Applied Science) and phenylmethylsulfonyl fluoride (final concentration 10 mM) were added. Cell debris was removed by centrifugation at 1,500 g for 4 min.

Centrifugation

Detergent insoluble protein aggregates were isolated as described.²¹ Two to four milliliters of yeast cell lysate were centrifuged at 200,000 g for 1 h at 4°C. Pellet was dissolved in TBS with 3% Sarkosyl and Complete protease inhibitor cocktail (Roche Applied Science), placed in a tube over 20% sucrose solution with 0.3% of Sarkosyl (1/3 of the sample volume) and centrifuged at 200,000 g for 14 h at 6°C. The obtained pellet was dissolved again in TBS with 2% SDS or 3% Sarkosyl in the case of PrP, placed in a tube over 20% sucrose solution with 0.2% SDS or 0.3% Sarkosyl in the case of PrP (1/3 of the sample volume) and centrifuged for 14 h at 240,000 g, 15°C. Control protein aggregate fractions from crude cell lysates of [psi^-] [pin^-] cells were obtained in the same way.

Electrophoresis and blotting

Separation of proteins by semidenaturing electrophoresis in agarose gel (SDD-AGE) was performed as described previously.^20,21 Protein loads were equalized for each gel. For analysis of amyloid polymers we used horizontal 1.8% agarose gels in the Tris-Acetate-EDTA (TAE) buffer with 0.1% SDS. Lysates were incubated in the sample buffer (0.5x TAE, 2% SDS, 5% glycerol and 0.05% Bromophenol Blue) for 5 min at 37°C. After the electrophoresis, proteins were transferred from gels to nitrocellulose membrane sheets (ThermoScientific) by vacuum-assisted capillary blotting for 5 h (agarose gels). Bound antibody was detected using the ECL West Dura system (Thermo Scientific). Rabbit polyclonal antibodies against Sup35 and Rnq1 were used. Anti-GFP mouse monoclonal antibody 3A9 was obtained from Rusbiolink. Anti-PrP rabbit monoclonal antibody EP1802Y was obtained from Abcam.

Aptamer binding assays

Proteins of the total cell lysates or their aggregate fractions were separated by SDD-AGE, transferred to nitrocellulose membrane as described above, and the membrane was incubated for 1 h with blocking buffer (TBS with 0,4% casein and 0.05% tween 20). The aliquot of 5′-biotinilated aptamer was diluted to final 12 pmoles/ml concentration with PSU buffer (500 µl of 5 mM sodium phosphate, pH 6.0, 150 mM NaCl, 0,2 mM urea), boiled for 5 min, incubated at 25C^o for 1 h and diluted 10-fold with mixture of equal volumes of PSU and blocking buffers. The streptavidin peroxidase conjugate was then added at desired concentration; solution was rotated for 20 min at room temperature and then poured out to nitrocellulose membrane. Dot blot assays were used only to detect binding of aptamers with proteins in aggregate fractions isolated from yeast cell lysates. Equal amount of total protein (confirmed by staining the same membranes by Ponceau S, a non-specific protein stain) from each preparation was serially diluted as indicated and applied to nitrocellulose membrane using 96-well sample vacuum microfiltration blotting device (Bio-Rad). Then, the membrane was processed as described above. Bound streptavidin-peroxidase conjugate was detected using the ECL West Dura system (Thermo Scientific).

Abbreviations:
SELEX	=	systematic evolution of ligands by an exponential enrichment
polyQ	=	polyglutamine
GFP	=	the green fluorescent protein

Acknowledgments

The work was supported by the Ministry of Education and Science of the Russian Federation (no. 11.519.11.2001) and the grant from Russian Foundation for Basic Research no. 11-04-00442. Some parts of the study were performed in the Applied Biotechnologies Center for Collective Use, Bach Institute of Biochemistry, Russian Academy of Sciences. Authors are grateful to Alexey Galkin for providing the plasmid pL-moPrP(90-231), Michael Sherman for the plasmid p103Q-GFP, and Mick Tuite for the plasmids of the pUKC1512-SUP35 series.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.