Yeast strains VL1, a [psi⁻][PIN⁺]; VL2, a [psi⁻][pin⁻]; VL3, a weak [PSI⁺][pin⁻]; VL4, a strong [PSI⁺][pin⁻] in 74-D694 background (ade1-14 ura3-52 leu2-3, 112 trp1-289, his3-200) were a kind gift from Dr. Susan Liebman. L2723, a SUP35 deletion strain (sup35Δ:LEU2) episomally expressing full length (N-MRF) Sup35 (pRS313-N-MRF, HIS) in 74-D694 was also a gift from Dr. Liebman's lab (Bagriantsev and Liebman, 2006). As described by them, a control strain was generated by replacing the pRS313-N-MRF with plasmid pRS316-MRF expressing only the C-domain of Sup35 (MRF). W303, a wild type (leu2-3,112 trp1-1 can1-100 ura3-1 ade2-1 his3-11,15) yeast strain was also used.

Yeast cells were grown in standard media. Rich media contained either 1 or 0.25% yeast extract, 2% peptone and 2% dextrose. Synthetic minimal media contained all amino acids except those used for selection and 2% dextrose (SD). For overexpression constructs dextrose was replaced with raffinose (2%) and galactose (3%) (SRaf+Gal) in minimal media. To examine the effect of pH, media pH was adjusted to 4.2 with HCl and 6.5 with NaOH. Yeast transformations were done by lithium acetate protocol (Gietz and Woods, 2002).

The plasmid, pRS316-MRF, expressing only the C-domain of Sup35 and another plasmid pNM-RFP for overexpression of prion domain (NM) of Sup35 fused to RFP were again a gift from Dr. Susan Liebman. The wild type TTR-GFP fusion construct was generated by amplifying the full length human transthyretin cDNA from pDNR-LIB clone and inserting it upstream of GFP in a centromeric plasmid, pRS316CG. WT-TTR-GFP fusion fragment was then amplified and cloned at BamHI site in pRS424-GAL1, a 2 micron expression vector to generate pWT-TTR-GFP. Plasmid pM-TTR-GFP was obtained by introducing two point mutations simultaneously; methionine substitution at phenylalanine 87 position (F87M) and methionine substitution at leucine 110 position (L110M) in WT-TTR-GFP using QuickChange site-directed mutagenesis procedure from Stratagene (Cat# 200516). A vector control (pGFP) was created by digesting pWT-TTR-GFP with SacII enzyme and subcloning the GFP fragment at SacII site in pRS424-GAL1 vector. All the clones were verified by restriction digestion and DNA sequencing.

Different yeast strains were transformed with either TTR-GFP overexpression constructs alone or co-transformed with plasmids expressing N-MRF or MRF. Cells were grown in synthetic glucose selection media and reinnoculated in inducing media. (SRaf+Gal) TTR aggregates were analyzed after 72 h using a Nikon Ti-E inverted fluorescent microscope. For each transformant, six or more microscopic visual fields were randomly selected and cells with aggregates and diffused GFP expression were manually counted. More than 600 cells were counted per transformant and three independent transformants were analyzed for each construct.

Yeast cells overexpressing WT-TTR-GFP or M-TTR-GFP were harvested after 24 h of incubation in the inducing media and lysed using 1X lysis buffer (50 mM TrisCl, 50 mM KCl, 10 mM MgCl₂, 5% glycerol). Cell lysates were normalized for total protein and incubated at 37 and 95°C in non-denaturing sample buffer without β-mercaptoethanol and resolved on 12% SDS-PAGE. The blot was probed using anti-GFP antibody (Cat# G6795, Sigma).

Cell lysates were normalized for total protein at 100 μg/μl and 1 mg of total protein was centrifuged at 40,000 rpm for 3 h at 4°C. Supernatant fraction was aspirated and the pellet fraction was resuspended in the same volume (100 μl) of 1X lysis buffer as the supernatant. Equal volumes of supernatant and pellet fractions were loaded. Total protein loaded was one-tenth of the normalized protein for both the samples. The total, supernatant and pellet fractions were resolved on 10% SDS-PAGE, immunoblotted and probed using anti GFP antibody.

Effect of over-expression of WT-TTR-GFP and M-TTR-GFP was measured by spotting serial dilutions of exponentially growing liquid cultures. Yeast cells with WT-TTR-GFP or M-TTR-GFP were grown in raffinose media till early log phase (0.3–0.4). The cultures were normalized to 0.2 OD and 5-fold serial dilutions were prepared (5⁻¹, 5⁻², 5⁻³, and 5⁻⁴). 5 μl of each dilution was spotted on glucose (uninduced media) and galactose (inducing media) selection plates and incubated at 30°C for 2–3 days.

Yeast cells were grown in SD media maintained at different pH at 30°C for 24 and 36 h. Cells were harvested, washed and incubated with 500 nM of Lysotracker Red DND-99 (Cat# L7528) at 30°C for 30 min in their respective media. Cells were washed and visualized using confocal laser scanning microscope Leica TCS SP8. Images were processed using LAS X 3.1.1 software.

Protein levels of the WT-TTR-GFP and M-TTR-GFP in yeast were determined by lysing the cells, normalizing the total protein and immunoblotting using anti-GFP antibody under the defined experimental conditions (different pH, yeast strain variants). GAPDH was used as the endogenous loading control.

Yeast cells expressing M-TTR-GFP aggregates were stained with, 4,6-diamidino-2-phenylindole (DAPI), a nuclear DNA-binding dye and FM4-64, an endocytic vesicle and vacuolar membrane dye. For nuclear staining, cells were fixed with 10% formaldehyde for 2 h, washed with phosphate-buffered saline (PBS), permeabilized by incubation in 70% ethanol for 30 min at room temperature. Cells were washed again and incubated with DAPI (1 μg/ml) for 10 min and visualized. For vacuolar staining, cells were incubated with 8 μM FM4-64 in YPD for 15 min at 30°C. Cells were washed and further incubated for 60 min in fresh YPD and then visualized under fluorescent microscope.

Colocalization of M-TTR-GFP aggregates and Sup35 aggregates were monitored by co-transforming M-TTR-GFP and prion domain (NM) of Sup35 fused with RFP (NM-RFP). Colocalization was evaluated visually, quantitatively and statistically. Cells from independent transformants were analyzed under fluorescent microscope (Nikon Ti, 100X with oil immersion) for aggregates after 72 h under FITC (GFP) and TRITC (RFP) channels and images were captured separately for both the channels. Colocalization was analyzed by merging the images from two channels using NIS-Elements AR 3.2. Cells with M-TTR-GFP, NM-RFP or both M-TTR-GFP/NM-RFP aggregates were manually counted to calculate the percentage of cells showing colocalized aggregates. The degree of colocalization between the fluorophore conjugated proteins (M-TTR-GFP and NM-RFP) was quantified by calculating Pearson's correlation coefficient using Coloc2 Fiji pluggin of the ImageJ software.

To determine the effect of M-TTR on aggregation of Sup35, M-TTR-GFP construct was transformed in Δ sup35 strain maintained by either N-MRF (expressing full length Sup35 where N is the prion domain) or MRF (expressing only functional domain of Sup35). This strain also has an ade1 mutation which encodes for a premature stop codon. Cells were grown in SRaf+Gal selection media, normalized at 0.2 OD and 5 μl of each culture was spotted on 1/4 YPD agar plates and adenine deficient (-Ade) plates. Growth on -Ade plates was monitored after incubating at 30°C for 3 days. The color saturation on YPD medium was observed after incubation at 30°C for 2 days and then at 4°C for 3 days. If Sup35 protein which is a translation termination factor is functional, it terminates the translation of ADE1 gene at the premature stop codon which results in accumulation of a red pigment giving red coloration to yeast cells on rich media and reduced or no growth on media lacking adenine. However, if Sup35 is in the aggregated form and its function is impaired, it leads to read-through of the premature stop codon and synthesis of full length Ade1 protein. There is no accumulation of red pigment, yeast cells appear white on rich media and grow on media lacking adenine.

Transformants expressing M-TTR-GFP in different variants of yeast prions: [psi⁻], weak [PSI⁺], strong [PSI⁺], [pin⁻], and [PIN⁺] were cured of the prions by subculturing the cells 4 to 5 times on YPD plate containing 5 mM guanidine hydrochloride (GuHCl) at 30°C till the colonies turned red from pink or white on YPD plates.

The imunoprecipitation was performed according to the Catch and Release kit protocol (Cat#17-500). Briefly, the protein was isolated from weak [PSI⁺] cells over-expressing either M-TTR-GFP fusion protein or GFP alone by standard glass bead method using 1X lysis buffer (100 mM Tris Cl, 100 mM KCl, 20 mM MgCl₂ and 10% glycerol) and total protein was estimated by BCA pierce kit (Cat# 23225). 1 mg of total lysate was incubated with 10 μl of anti-GFP antibodies (Cat# G1544) on column containing resin for overnight at 4°C in continuous rotation. Following incubation, the resin in the column was washed twice with 1X wash buffer to remove non-specifically bound proteins. Immunoprecipitated protein complexes were eluted with hot (95°C) 1X denaturing elution buffer, resolved by 10% SDS-polyacrylamide gels, and analyzed by immunoblotting using antibodies against the Sup35 (BE4, a polyclonal antibody, was a kind gift from Susan Liebman) and GFP (monoclonal antibody, Sigma Cat# 6795).

Statistical analysis was performed using SPSS software for Windows, version 17.0 (SPSS, Chicago, Illinois). Data was checked for normality and two-tailed t-test was applied to determine the statistical significance and p-value less than 0.05 was considered to be significant.

To investigate TTR aggregation, wild type TTR was fused to GFP (WT-TTR-GFP) and overexpressed in yeast as described in Materials and Methods section. On overexpression of WT-TTR-GFP, most of the cells showed diffused fluorescence and only 5–7% of cells showed dot-like visible aggregates (Figures 1A,B). Since destabilization of TTR tetramer to monomer is important for aggregation, the native state of WT-TTR-GFP was analyzed. WT-TTR-GFP appeared to migrate as a higher molecular weight species at ~120 KDa (Figure 1C, lane 3). It is possible that in yeast WT-TTR-GFP exists as either trimer or tetramer that dissociates into trimer under our experimental conditions (2% SDS without β-mercaptoethanol at 37°C). These higher molecular weight species dissociated into monomer when incubated at 95°C (Figure 1C, lane 5). Another possibility is that these higher molecular weight species occur due to interaction of TTR with other cellular components. However, we reasoned it to be a trimer because trimer species have been shown to populate in the serum (Cubedo et al., 2012). Moreover, WT-TTR tetramers are formed by sequential addition of monomers; and trimer intermediates have been observed in vitro (Wiseman et al., 2005).

We then introduced two previously (Jiang et al., 2001) defined point mutations, F87M at the monomer-monomer and L110M at the dimer-dimer interface of WT-TTR, which destabilizes the TTR tetramer. As expected, this engineered double mutant, M-TTR-GFP, existed as monomer in yeast (Figure 1C, lane 2) and did not migrate at higher molecular weight. On overexpression of M-TTR-GFP, no aggregation of M-TTR-GFP was observed at 24 h and very minute dot-like aggregates start appearing between 36 and 48 h (Figure S1). The aggregates became clearly visible at ~72 h of incubation in inducing media under fluorescent microscope. A significant (p < 0.05) and approximately 5-fold increase in percentage of cells with dot-like visible aggregates compared to WT-TTR-GFP in two different (74D-694 and W303) yeast backgrounds (Figures 1A,B) was observed at 72 h. It is important to point out that more WT-TTR-GFP protein was present in the insoluble pellet fraction compared to supernatant suggesting presence of non-visible higher molecular weight species on overexpression of WT-TTR-GFP. These higher molecular weight species are the biochemical aggregates which are not visible using microscopic techniques and may include aggregation intermediates. However, M-TTR-GFP fractionated more than 2-fold in the pellet as compared to the WT-TTR-GFP (Figure 1D). It was further ensured that the increase in aggregation was not due to any difference in the expression levels of WT-TTR-GFP and M-TTR-GFP (Figure 1B, inset). The effect of aggregates on viability of yeast was monitored by dilution spotting and 7-AAD staining. No effect on growth or viability of yeast cells was observed on either episomal or stable overexpression of WT-TTR-GFP or M-TTR-GFP (data not shown). Thus, TTR-GFP aggregates appeared to be non-toxic in yeast.

Further, to visualize the cellular localization of M-TTR-GFP aggregates, cells overexpressing M-TTR-GFP were stained with FM4-64 (vacuolar staining dye) and DAPI (nucleolar staining dye). Similar to several other amyloid proteins, bigger M-TTR-GFP aggregates were observed at periphery of vacuole and smaller aggregates near the nucleus (Braun et al., 2010) (Figure 1E).

The pH has been previously shown to be a factor that triggers tetramer dissociation, partial unfolding and aggregation of TTR in vitro (Jiang et al., 2001; Lim et al., 2013; Xue et al., 2014). It has been earlier contested and demonstrated that intracellular pH of yeast is altered by the external pH (Slavík, 1983; Thevelein et al., 1987). We also analyzed the affect of pH of media on the intracellular pH of yeast cells using Lysotracker Red probe which is an acidotropic probe and commonly used to stain acidic organelles. At external pH 7.4, vacuoles, the acidic compartment of yeast, were stained in all the cells and no cytoplasmic staining was observed. At pH 6.5, vacuoles were stained in more than 80% of the cells and cytoplasmic staining was observed in approximately 20% of the cells. Interestingly, at pH 4.2, more than 90% of the cells showed cytoplasmic staining suggesting the acidic nature of the cell interior (Figure 2C). There was no difference in Lysotracker staining after culturing for 24 or 36 h in media with different pH.

To monitor whether pH of the media has any effect on TTR aggregation in yeast, cells expressing either WT-TTR-GFP or M-TTR-GFP constructs were grown in inducing media with different pH. The percentage of cells exhibiting M-TTR-GFP aggregates was significantly reduced by 4-fold at higher pH compared to acidic pH (Figure 2A). Even, the WT-TTR-GFP aggregates were completely abolished at pH 6.5. It was also ascertained that the difference in aggregation at two different pHs was not due to different expression levels (Figure 2A, inset) or cell death as monitored by dilution spotting (Figure 2B) and 7-AAD staining (data not shown). Thus, WT-TTR-GFP or M-TTR-GFP aggregation in yeast is susceptible to acidic pH and recapitulates the in vitro scenario.

Endogenous yeast prions have been suggested to interact with other amyloid proteins. [PSI⁺] is a phenotypic trait of endogenous prion form of Sup35p and exists as different variants in yeast (Uptain et al., 2001) whereas [psi⁻] represents the non-prion form of Sup35p. In order to determine the interaction of TTR with yeast prion protein Sup35, we examined the effect of variants of [PSI⁺] on M-TTR-GFP aggregation. A significant (p < 0.05) 2-fold increase in cells with visible M-TTR-GFP aggregates was observed in weak [PSI⁺] and more than 1.5-folds in strong [PSI⁺] as compared to [psi⁻] strain (Figure 3A, dark gray bars). Again, the difference in aggregation was not due to any significant change in expression levels of M-TTR-GFP in the different strains (Figure 3A, left inset).

To confirm that the enhanced aggregation of TTR is solely due to presence of [PSI⁺], endogenous prion was cured from these transformants by passaging 8–10 times on guanidine hydrochloride (GuHCl) plates. Curing of [PSI⁺] leads to the elimination of prion and was monitored by change in coloration from white/pink ([PSI⁺]) to red as in ([psi⁻]) strain (Figure 3A right inset). After curing of the prion, the expression of M-TTR-GFP was re-induced in these [psi⁻] cells. Interestingly, M-TTR-GFP aggregation was remarkably reduced on losing the prion form of Sup35. (Figure 3A, light gray bars). This data indicates that yeast prion, [PSI⁺] is able to promote M-TTR-GFP aggregation.

Additionally, we overexpressed M-TTR-GFP in another [PSI⁺] sup35Δ ade1-14 strain maintained by full length Sup35 expressed on a plasmid (N-MRF). The control non-prion form of this strain was maintained by expressing only the functional domain (MRF) but lacking the prion (N) domain of Sup35. Even in this strain background, a significant (p < 0.05) 2-fold increase in cells with visible aggregates of M-TTR-GFP was observed in the presence of aggregated N-MRF [PSI⁺] as compared to soluble MRF strain (Figure 3B, upper panel). This again supported the interaction amongst these two heterologous amyloid proteins lacking any sequence similarity. We further exploited this experimental set up to examine if overexpression of M-TTR-GFP cures [PSI⁺]. On overexpressing GFP alone, N-MRF [PSI⁺] strain is white or pink in color on YPD media and grows on -Ade media whereas cells expressing the functional, non-aggregated MRF domain of Sup35 are red on YPD and do not grow on -Ade media (Figure 3B, lower panel). On overexpression of M-TTR-GFP, however, no change in the color of the N-MRF expressing cells on rich media or growth on -Ade media (Figure 3B, lower panel) was observed. Thus, it appears that M-TTR-GFP is not able to cure the yeast prion, [PSI⁺].

To further verify whether these two heterologous amyloid aggregates are merely sharing the common cellular space or they exhibit any direct interaction, M-TTR-GFP fusion protein was overexpressed in a weak [PSI⁺] strain and M-TTR-GFP aggregates were pulled down using anti-GFP antibody. Sup35 protein co-immunoprecipated with M-TTR-GFP aggregates but was not pulled down from the cells overexpressing GFP alone as the control (Figure 3C). This observation suggests that Sup35 directly interacts with M-TTR aggregates.

In addition to [PSI⁺], another yeast prion [PIN⁺] was also examined for its effect on M-TTR aggregation. [PIN⁺] is the prion form of Rnq1 and has been shown to induce [PSI⁺] and Htt aggregation in yeast. However, we did not observe any significant difference in the percentage of cells with M-TTR-GFP aggregates in [PIN⁺] and [pin⁻] cells (Figure 3D).

To verify the inducing effect of Sup35 on M-TTR aggregation, we co-overexpressed M-TTR-GFP and prion (NM) domain of Sup35 fused to RFP (NM-RFP) in a [psi⁻] [PIN⁺] strain. A 2-fold increase in M-TTR-GFP aggregation was observed in the presence of overexpressed NM-RFP (~21%) as compared to RFP (~10%). On analyzing the cells for aggregates, single, two and multiple dots were observed for M-TTR-GFP and peripheral rings, lines and dots for NM-RFP as tabulated in Table 1. In about 8% of the cells, both M-TTR-GFP and NM-RFP aggregates were present and more than 90% of these cells showed colocalization. The M-TTR-GFP dots colocalized with dot-, line- (at the ends) and peripheral ring/mesh-like aggregates of NM-RFP (Figure 4). The Pearson's correlation coefficient (R²) ranged from 0.5 (for rings and lines) to 0.95 (for dots) and suggested a positive correlation of colocalization. No correlation was found between M-TTR-GFP and cytoplasmic expression of RFP alone (Figure 4). This co-localization again supports the probability of interaction between the two amyloid proteins. However, we again did not notice any significant increase in aggregation of NM-RFP (data not shown) due to the overexpression of M-TTR-GFP compared to GFP. This finding is consistent with the earlier report where addition of wild type TTR fibrils did not accelerate aggregation of Sup35-NM-YFP (Derkatch et al., 2004). Thus, the interaction of M-TTR and Sup35-NM appears to be unidirectional where overexpression of NM-RFP enhances M-TTR-GFP aggregation but not vice-a versa.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.