Biochemical Journal

Research article

Hsp104 as a key modulator of prion-mediated oxidative stress in Saccharomyces cerevisiae

Kuljit Singh, Aliabbas A. Saleh, Ankan K. Bhadra, Ipsita Roy

Abstract

Maintenance of cellular redox homoeostasis forms an important part of the cellular defence mechanism and continued cell viability. Despite extensive studies, the role of the chaperone Hsp104 (heat-shock protein of 102 kDa) in propagation of misfolded protein aggregates in the cell and generation of oxidative stress remains poorly understood. Expression of RNQ1-RFP in Saccharomyces cerevisiae cells led to the generation of the prion form of the protein and increased oxidative stress. In the present study, we show that disruption of Hsp104 in an isogenic yeast strain led to solubilization of RNQ1-RFP. This reduced the oxidative stress generated in the cell. The higher level of oxidative stress in the Hsp104-containing (parental) strain correlated with lower activity of almost all of the intracellular antioxidant enzymes assayed. Surprisingly, this did not correspond with the gene expression analysis data. To compensate for the decrease in protein translation induced by a high level of reactive oxygen species, transcriptional up-regulation takes place. This explains the discrepancy observed between the transcription level and functional enzymatic product. Our results show that in a ΔHsp104 strain, due to lower oxidative stress, no such mismatch is observed, corresponding with higher cell viability. Thus Hsp104 is indirectly responsible for enhancing the oxidative stress in a prion-rich environment.

  • antioxidant enzyme
  • chaperone
  • Hsp104
  • oxidative stress
  • prion
  • RNQ1

INTRODUCTION

In fungal species such as the budding yeast Saccharomyces cerevisiae prions are present as a system of non-Mendelian phenotypic inheritance. Yeast prion is used as a model for studying both the intrinsic properties of the proteins that enable prion propagation and the cellular factors that facilitate prion formation and maintenance [1]. In the non-prion state, these proteins are soluble and are able to perform their molecular function in the cell. Transition to the prion state causes the proteins to aggregate and generally causes a loss-of-function phenotype. Transmission of these aggregates during cell division results in the inheritance of the prion state by daughter cells [13]. Due to the self-propagating nature of the prion structure, the associated phenotypes are always dominant. Yeast prion proteins contain N-terminal or C-terminal regions, termed prion domains, that are required for prion formation and propagation [4]. Mechanisms of amyloid formation in yeast and other fungi are similar to mammalian systems. Therefore yeast prions provide easy and efficient experimental assays for studying the factors and conditions influencing amyloid formation and propagation [1,5].

In the yeast S. cerevisiae, several proteins can exist in prion form. The three most well-studied are: (i) a protein of unknown function, RNQ1; (ii) translational termination factor Sup35 (also called eRF3); and (iii) a regulatory protein in the nitrogen metabolism pathway, Ure2. The prion forms of these proteins are termed [PIN+] (or [RNQ+]), [PSI+] and [URE3] respectively [1,5,6]. The yeast prion [RNQ+]/[PIN+] is formed by the yeast protein RNQ1. The [RNQ+] state facilitates the conversion of other proteins into prions in yeast. RNQ1 possesses a C-terminal glutamine/asparagine-rich prion domain which helps it to assemble into amyloid-like fibrils and induces prion formation when fused with the glutamine/asparagine-rich N-terminal domain of Sup35 [7,8]. The N-terminal non-prion domain of RNQ1 regulates [RNQ+] prion propagation. Interaction between prion aggregates of RNQ1 and soluble Sup35 facilitates [PSI+] initiation. Initiation of [PSI+] is enhanced by the temporary overexpression of Sup35p in [RNQ+] cells [9,10]. Further propagation of [PSI+] does not require [RNQ+] prion.

Yeast prions depend upon molecular chaperones for efficient maintenance and propagation of prion structures. Many Hsps (heat-shock proteins) are present in the yeast cell and have different roles to play in the formation of amyloid-like fibrils. Expression of certain chaperones prevents the yeast cell from having heritable prions. Some chaperones play a role in the conversion of native proteins into the prion conformation [11,12]. Hsp104 encodes an anti-stress chaperone of the Hsp100 gene family in S. cerevisiae. Hsp104 works in conjunction with other co-chaperones, namely Ssa1 and Ydj1, which have roles to play in the disassembly of protein aggregates that could have accumulated due to stress [13]. Hsp104 chaperone is expressed at very low levels under normal conditions and its expression is induced under stress conditions. Transcriptional activators mediate transcriptional up-regulation. The activators bind to three stress-response elements and two heat-shock elements present in the Hsp104 promoter. Hsp104 mutant strains are unable to propagate the yeast prions [PSI+], [PIN+]/[RNQ+] and [URE3] [1,14,15]. Overexpression and inactivation of Hsp104 help in disaggregation of amyloid fibrils in cells containing [PSI+] prions [14]. Overexpressed Hsp104 leads to [PSI+] loss by blocking the incorporation of newly synthesized Sup35p into [PSI+] complexes or by directly disaggregating the complexes of prions into monomers [12]. The role of Hsp104 in the (dis)assembly of [RNQ+] is less well understood.

Oxidative stress results in cells upon an imbalance between the production and elimination of ROS (reactive oxygen species). Instability in the normal redox state of tissues can cause toxic effects through the formation of peroxides and free radical species, which could damage major components of the cell, like DNA, lipids and proteins [16]. Intracellular ROS production changes in acute, adaptive and chronic phases of prion infection. Some cells have a rapid prion protein-dependent increase in intracellular ROS [17]. The production of free radical species can be correlated with increased accumulation of protease-resistant prion proteins. The loss of membrane phosphatidylserine asymmetry is seen in cells with a high number of peroxides resulting from intracellular aggregates contained within prion protein [17]. The loss of activities of enzymes such as peroxidase, catalase and SOD (superoxide dismutase) can lead to increased levels of free radical species in yeast cells. Metal ions such as copper, zinc and iron have a significant role to play in antioxidant activity, as metal ions are sequestered from antioxidant enzymes into prion protein aggregates [18], thus decreasing the activity of antioxidant enzymes. Therefore prion propagation and formation of ROS has a significant role to play in normal functioning of cellular components.

In the present study, we have studied the role of Hsp104 in modulating the cellular ROS level during prion propagation in yeast. In the presence of the disaggregase activity of Hsp104, RNQ1 is expressed in the aggregated prion form. This leads to an increased level of oxidative stress in the yeast cell. We show that the yeast cell tries to combat this toxicity by increased gene expression of scavenger enzymes. However, a concomitant increase in the enzyme activity is not seen, which finally results in reduced cell viability.

EXPERIMENTAL

Materials

S. cerevisiae BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) parental and ΔHsp104 strains were products of Open Biosystems and were purchased from SAF Labs. LB broth, agar, ampicillin, DTNB [5,5′-dithiobis-(2-nitrobenzoic acid)], GSH, GST assay kit, DNPH (2,4-dinitrophenylhydrazine), primers and goat anti-rabbit (FITC-conjugated) antibody were purchased from Sigma–Aldrich. PMSF, SDS, glutathione reductase and sodium azide were products of Fluka and were purchased from Sigma–Aldrich. Succinic acid, dextrose anhydrous, ferrous sulfate, potassium thiocyanate, NBT (Nitro Blue Tetrazolium), NADPH, hydroxylamine hydrochloride, copper sulfate, Triton X-100, GSSG and amino acids for amino acid dropout mixtures were purchased from SRL. Yeast nitrogen base (without amino acids) was purchased from HiMedia Laboratories. Oligo dT(18) primers were purchased from Fermentas. MMLV (Moloney murine leukaemia virus) reverse transcriptase was purchased from Promega. SYBR® Premix Ex Taq™ (Perfect Real Time) kit was purchased from TaKaRa Bio. PCR tubes were purchased from Axygen Scientific. H2O2 was purchased from RFCL. DCFH-DA (dichlorodihydrofluorescein diacetate) was purchased from Cayman Chemical Company. Rabbit anti-2,4-dinitrophenol antibody was obtained from Molecular Probes. All other reagents and chemicals used were of analytical grade or higher.

Construction of pRS315-RFP

pRS315-RFP (empty vector), which lacks the RNQ1 sequence, was constructed by the replacement of the RNQ1 region of pRS315-RNQ1-RFP with a stuffer sequence flanked by restriction sites for BamHI and SacII. The parent vector (pRS315-RNQ1-RFP) was digested with BamHI and SacII to remove RNQ1. The gel-purified linearized vector was ligated to the oligonucleotide sequence containing pre-digested BamHI and SacII restriction sites. The removal of RNQ1 was confirmed by agarose gel electrophoresis.

Transformation and expression of plasmids in S. cerevisiae

S. cerevisiae BY4742 parental and ΔHsp104 strains were transformed with purified pRS315-RNQ1-RFP and pRS315-RFP plasmids following the standard lithium acetate-PEG protocol [19]. The integrity of the inserts was checked by agarose gel electrophoresis.

Expression of RNQ1-RFP in S. cerevisiae cells

Glycerol stocks of transformed strains were inoculated in 50 ml of SC-LEU (synthetic complete medium without leucine) containing 2% (w/v) dextrose and grown at 30°C, 200 rev./min until a D600 of 0.6 was reached. Expression of RNQ1 was induced by the addition of 500 μM CuSO4 and incubated at 30°C, 200 rev./min for 4 h. After the induction period, expression of RNQ1-RFP was studied using fluorescence microscopy with a 100× objective lens (Eclipse E600; Nikon).

Measurement of oxidative stress by DCFH-DA

For quantification of ROS levels, the post-induction cells were counted using Neubauer's chamber. Cells (1×107) were aliquoted into a microcentrifuge tube. DCFH-DA (10 mM, dissolved in DMSO) was added at a final concentration of 10 μM, followed by addition of H2O2 at a final concentration of 1 mM [20]. The final reaction mixture was made up to 1 ml with PBS (10 mM, pH 7.4). The emission intensities of DCF (dichlorofluorescein) (de-esterified and oxidized metabolite of DCFH-DA) were recorded after 60 min of incubation using a spectrofluorimeter (RF5301; Shimadzu) using an excitation wavelength of 504 nm and an emission wavelength of 519 nm. The emission intensity of DCF in the absence of cells was used as the control and was subtracted from each of the readings. For imaging, the cells incubated with DCFH-DA were washed twice with PBS. Conversion into the oxidized product was monitored using fluorescence microscopy with a 100× objective lens (Eclipse E600; Nikon).

Activities of antioxidant enzymes

Yeast cells were lysed using acid-treated glass beads [21]. After lysis, the microcentrifuge tubes containing the lysates were set aside for 1 h for the glass beads and cell debris to settle down. The lysate was then transferred to another centrifuge tube and centrifuged at 800 g for 10 min. The amount of protein in the resulting supernatant was estimated by a dye-binding method [22] using BSA as the standard protein.

Catalase assay

The catalase assay was carried out as described previously [23]. Reagents and buffer were incubated at 0°C in an ice bath, except for 0.3 M sulfuric acid. Absorbance values of the red colour of ferric thiocyanate were recorded in the kinetic mode as a measure of residual H2O2 at 460 nm using a spectrophotometer (UV1700; Shimadzu). Results were expressed as follows [23]: Embedded Image where A1 and A2 are the observed absorbance at two selected time points and t is the time differential between the two points.

SOD assay

The assay was carried out as described previously [24]. The reaction was initiated by the addition of 100 μl of NH2OH.HCl (1 mM) to the reaction mixture containing the yeast lysate and the rate of reduction of NBT following addition of NH2OH·HCl was recorded every 30 s for 10 min at 560 nm using a spectrophotometer (UV1700). The rate of reduction of NBT in the absence of the enzyme source was used as the control. Results were calculated as [24]: Embedded Image (1)

Glutathione peroxidase assay

The assay was carried out as described previously [25]. The rate of oxidation of NADPH was recorded every 15 s for 5 min at 340 nm in a spectrophotometer (UV1700) following addition of H2O2 and yeast lysate to the reaction mixture. The rate of oxidation of NADPH in the absence of reduced glutathione and enzyme source was taken as the reaction control. The activity of the enzyme was expressed as micromoles of NADPH oxidized/min by considering a molar absorption coefficient of 6.22 mM−1·cm−1 [25]. Results were calculated as: Embedded Image (2) where C2 and C1 are the concentrations of NADPH oxidized at two selected time points, calculated as C=A/ϵ, where A is observed absorbance, ϵ is the molar absorption coefficient of NADPH (6.22 mM−1·cm−1) under the reaction conditions and t is the time differential between the two points.

GST assay

The assay was carried out as described previously [26]. The rate of increase in the absorption of GS-DNB (glutathione-dinitrobenzene) conjugate following addition of 1-chloro-2,4-dinitrobenzene and yeast lysate to the reaction mixture was recorded every 30 s for 5 min at 340 nm in a spectrophotometer (UV1700). The rate of increase in the absorption of the conjugate in the absence of enzyme source was taken as the reaction control. Results were calculated using eqn (1) [26].

Glutathione reductase assay

The glutathione reductase assay was carried out as described previously [27]. The rate of oxidation of NADPH in the presence of GSSG and yeast lysate was recorded every 15 s for 5 min at 340 nm in a spectrophotometer (UV1700). The rate of oxidation of NADPH in the absence of the enzyme source was taken as the reaction control. The activity of the enzyme was expressed as micromoles of NADPH oxidized/min and was calculated using a molar absorption coefficient of 6.22 mM−1·cm−1 [27]. Results were calculated using eqn (2).

Thioredoxin reductase assay

The thioredoxin reductase assay was carried out as described previously [28]. The reaction was initiated by the addition of DTNB to the reaction mixture and the rate of reduction of DTNB in the presence of yeast lysate was recorded every 15 s for 5 min at 412 nm using a spectrophotometer (UV1700). The rate of reduction of DTNB in the absence of enzyme source was considered to be the control. The activity of the enzyme was expressed as micromoles of DTNB reduced/min and was calculated using a molar absorption coefficient of DTNB as 14.150 mM−1·cm−1 [28]. Results were calculated using eqn (2).

Western blotting

Yeast lysates, obtained as described above, were mixed with denaturing gel loading buffer and proteins were separated electrophoretically on a 12% cross-linked denaturing polyacrylamide gel [29]. The gel was then equilibrated in electrotransfer buffer [Tris-glycine and 10% (v/v) methanol] and electroblotted on to a precharged (in 100% methanol) PVDF membrane. The membrane was immersed in 100% methanol and dried at room temperature (27°C). For derivatization, the membrane was incubated in a solution of DNPH (1 mM, 2 M HCl) for exactly 10 min and washed three times with 2 M HCl [30,31]. The membrane was then incubated for 5 min in TBS. Non-specific protein binding was blocked by incubating the membrane for 1 h in 2.5% (w/v) non-fat dried skimmed milk powder in TBS. The membrane was washed with TBS-T (TBS containing 0.05% Tween 20) and TBS and incubated with rabbit anti-2,4-dinitrophenol antibody (1:2500 dilution) overnight at 4°C. The membrane was then washed as above and incubated with goat anti-rabbit (FITC-conjugated) secondary antibody (1:500 dilution) for 2 h at room temperature. The membrane was washed thoroughly with TBS-T and TBS, dried and scanned on an image scanner (Typhoon Trio; GE Healthcare) in the fluorescence mode, at an excitation wavelength of 488 nm and an emission wavelength of 526 nm.

For immunoblotting of SOD and thioredoxin, blotting was carried out as described above, but using a nitrocellulose membrane (0.45 μm). After blocking for 1 h with 2.5% (w/v) non-fat dried skimmed milk powder, the membrane was incubated with rabbit anti-SOD1 (1:1000 dilution) or anti-thioredoxin (1:400 dilution) antibody for 4 h and overnight respectively, at 4°C. The rest of the protocol was the same as described above.

Gene expression analysis

The level of expression of various genes in parental and ΔHsp104 strains transformed with pRS315-RNQ-RFP and pRS315-RFP was measured by real-time reverse transcription–PCR. For this, total RNA was isolated from yeast cells using the hot phenol method [32]. Genomic DNA contamination was removed by treatment with DNase I. This was followed by reverse transcription using oligo dT(18) primer and MMLV reverse transcriptase according to the manufacturer's protocol. The cDNA obtained was used to study the changes in expression profiles of various genes in yeast cells. Primers were designed for the desired genes using Primer3 software (v. 0.4.0) [33]. A 1:10 dilution of cDNA was used with the SYBR® Premix Ex Taq™ (Perfect Real Time) kit according to the manufacturer's protocol. Cycles were performed on an Eppendorf Mastercycler® ep realplex Thermal Cycler using SYBR Green detection. Cycling parameters were as follows: (A) initial denaturation at 95°C, 20 s; (B) 35 cycles of (i) denaturation at 95°C, 5 s, (ii) primer annealing at 52°C, 25 s, (iii) amplification at 72°C, 15 s. Melting curves were generated by: (i) incubating the amplicons at 95°C, 15 s; (ii) reducing the temperature to 51°C, 15 s; and (iii) increasing the temperature to 95°C over a dissociation time of 15 s. Data were analysed using realplex 2.2 software (Eppendorf) to calculate cycle threshold (CT) values and melting curve to estimate optimal melting temperatures for all reactions. The actin1 gene was taken as the housekeeping gene (internal control). Relative fold change in gene expression was calculated by the comparative CT method (also known as the 2−ΔΔCT method) using the following equation: Embedded Image

This form of the equation is used to compare the gene expression in two different samples (e.g. treated against untreated), where expression of each gene is related to an internal control gene [34].

Measurement of cytotoxicity

Transformed parental and ΔHsp104 strains were allowed to grow in SC-LEU containing 2% dextrose until D600 reached 0.6. Cells were counted using a Neubauer chamber. Cell suspension was diluted separately, in such a way that cell suspension contained 4500 cells/μl, which was further diluted 3-fold serially. Cell suspensions (2 μl each) were plated on to SC-LEU+2% dextrose agar plates (with and without CuSO4). The growth of colonies was monitored at 30°C for 2–3 days.

RESULTS

Hsp104 is necessary for prion propagation

Plasmids pRS315-RNQ1-RFP and pRS315-RFP (empty vector) were isolated from transformed Escherichia coli DH5α cells. Double digestion of pRS315-RNQ1-RFP (Figure 1A) and pRS315-RFP with BamHI and SacII yielded bands at ~2000 bp (Figure 1B) and ~700 bp (Figure 1C) respectively, which corresponded with the lengths of the inserts RNQ1-RFP and RFP in the recombinant vectors. This also confirmed the removal of RNQ1 to generate the empty vector (pRS315-RFP).

Figure 1 Expression of RNQ1-RFP in S. cerevisiae BY4742 parental and ΔHsp104 strains

(A) Partial vector map of pRS315-RNQ1-RFP showing the relevant restriction sites used in the present study. Restriction digestion patterns of (B) pRS315-RNQ1-RFP and (C) pRS315-RFP by BamHI and SacII enzymes. (B) Lane 1, DNA ladder; lane 2, double-digested plasmid; lane 3, undigested plasmid. (C) Lane 1, DNA ladder; lane 2, undigested plasmid; lane 3, plasmid digested with SacII alone; lane 4, double-digested plasmid. The gels were scanned using an image scanner (Typhoon Trio) in the fluorescence mode. (DG) Expression of RNQ1-RFP in parental (D and E) and ΔHsp104 (F and G) strains. Images for red fluorescence (D and F) were recorded with an excitation of 558 nm and emission of 578 nm (E600 Eclipse microscope, Nikon), with an objective lens of 100×. Corresponding DIC images (E and G) are also shown.

In the parental strain, RNQ1 was expressed in the aggregated form as seen by dotted RFP fluorescence (Figure 1D) whereas in the ∆Hsp104 strain, RNQ1 was seen to exist in the soluble form (Figure 1F). The role of Hsp104 in prion partitioning and subsequent transmission to daughter cells is still not clear. Due to its disaggregase activity [14], Hsp104 accelerates the formation of amyloid-like fibrils by increasing the concentration of prion templates, which act as seeding agents for rapid prion replication in the daughter cells. ∆Hsp104 strains are unable to propagate yeast prions [15]. The results obtained in the present study confirm that the presence of the chaperone Hsp104 is essential to the propagation of prion protein and agree well with what is reported in the literature [1,14,15].

[RNQ+] leads to enhanced oxidative stress in cells

DCFH-DA is a cell-permeable fluorogenic probe that measures the level of intracellular ROS. The increase in the fluorescence intensity due to the oxidation of DCFH (the de-esterified form of DCFH-DA after it enters inside the cell) to highly fluorescent DCF, indicates the level of ROS [20]. In the present case, DCFH-DA was used to measure the change in the level of ROS during the expression of RNQ1 in parental and ∆Hsp104 strains.

Expression of RNQ1-RFP in the aggregated form (Figure 2A, fluorescent puncta) led to a higher level of DCF fluorescence in parental cells (Figure 2B). In the ΔHsp104 strain where RNQ1-RFP was expressed in the soluble form (Figure 2D), the fluorescence signal due to conversion of DCFH-DA into DCF was almost negligible (Figure 2E). The higher level of DCF fluorescence in the parental strain indicated a higher level of ROS in the strain expressing RNQ1 in the aggregated form as compared with the ∆Hsp104 strain which expressed soluble RNQ1.

Figure 2 Detection of oxidative stress in transformed S. cerevisiae BY4742 parental and ΔHsp104 strains

Images for expression of RNQ1-RFP in (A) parental and (D) ΔHsp104 strains were recorded as described in Figure 1. Images for DCF fluorescence in (B) parental and (E) ΔHsp104 strains were recorded with an excitation of 488 nm and emission of 526 nm (E600 Eclipse microscope), with an objective lens of 100×. Corresponding DIC images (C and F) are also shown.

Quantitative estimation of ROS in the transformed parental strain induced with CuSO4 showed higher fluorescence intensity of DCF (Figure 3A), which correlated well with the image shown in Figure 2(B). The presence of aggregated RNQ1 in the parental strain caused a significant increase in the oxidative stress level when compared with the uninduced parental strain. Untransformed parental strain was also grown in the presence of 500 μM CuSO4 to confirm whether the increased stress level in the transformed parental strain was due to aggregated RNQ1 or the presence of CuSO4 alone. The results shown in Figure 3(A) confirm that the presence of CuSO4 had a negligible effect on the ROS level in the untransformed parental cell. Similarly, expression of RNQ1 resulted in an increase in the level of ROS in the ∆Hsp104 strain as compared with the untransformed or uninduced cells (Figure 3A). In the presence of CuSO4 alone, untransformed ∆Hsp104 cells did not exhibit any significant increase in oxidative stress, confirming that the oxidative stress was generated due to the expression of RNQ1.

Figure 3 Determination of the level of ROS using the DCFH-DA assay (λex=504 nm, λem=519 nm) in S. cerevisiae BY4742 parental (empty bars) and ΔHsp104 (filled bars) strains

(A) Untransformed cells are denoted as −,−; transformed uninduced cells are denoted as +,−; and transformed induced cells are denoted as +,+. The effect of addition of CuSO4 alone was monitored by adding 500 μM Cu2+ to the growth medium of untransformed cells (−,+). ***P<0.001 against transformed uninduced parental cells; ##P<0.01 against transformed uninduced ΔHsp104 cells; and $$P<0.01 against transformed parental strain. (B) Effect of the presence of the vector alone was monitored by transforming the cells with the empty vector pRS315-RFP. ***P<0.001 against parental strain transformed with the empty vector; aaaP<0.001 against untransformed parental strain; ##P<0.01 against ΔHsp104 strain transformed with the empty vector; bbP<0.01 against untransformed ΔHsp104 strain; and N.S., non-significant. All experiments were carried out in triplicate, mean±S.E.M. are shown.

Comparison between the two strains showed that the transformed mutant strain induced with CuSO4 had a lower ROS level (Figure 3A) which correlated well with the image shown in Figure 2(E). RNQ1 is expressed as a soluble protein in the strain where Hsp104 has been disrupted. Expression of RNQ1 in the transformed parental strain showed a significant increase in oxidative stress level (Figures 2B and 3A). In order to confirm whether the presence of the vector backbone and/or RFP was responsible for the increased ROS level, both strains were transformed with the empty vector (pRS315-RFP). As can be seen, no significant difference could be observed between the untransformed strain and the same strain transformed with empty vector using either the parental or ΔHsp104 strain (Figure 3B). Comparison of parental and ΔHsp104 strains transformed with the empty vector showed that there was no difference between the level of ROS in both strains (Figure 3B). Thus the difference in the level of oxidative stress observed in these strains was due to the aggregation status of RNQ1 and not due to the absence/presence of Hsp104. Comparison with cells transformed with pRS315-RNQ1-RFP showed that the ROS levels in these cells were significantly higher as compared with either untransformed cells or cells containing empty vector. Thus the increase in ROS levels observed in the transformed cells was because of the expression of RNQ1 only and not due to any other factor. The higher level of ROS observed in the parental strain was because of aggregation of RNQ1 as compared with the strain which lacked Hsp104. In the latter case, propagation of RNQ1 did not occur in the prion form, which lowered the oxidative stress in these cells.

Antioxidant enzymes play a crucial role in Hsp104-mediated oxidative stress

Since ROS levels were significantly enhanced in both strains following expression of RNQ1, the cells were lysed and the activities of antioxidant enzymes were measured in the cell lysates.

The activities of all enzymes measured (i.e. catalase, SOD, glutathione peroxidase, GST, glutathione reductase and thioredoxin reductase) in the transformed parental cells were found to be significantly lower than in the uninduced cells or those transformed with the empty vector. However, in the case of the transformed ΔHsp104 strain, a significant increase in catalase activity and a significant decrease in the activity of thioredoxin reductase were observed, as compared with the cells carrying the empty vector. The activities of other enzymes (SOD, glutathione peroxidase, GST and glutathione reductase) remained unaffected (results not shown). Expression of RNQ1 aggregates in the transformed parental strain showed a significant decrease in catalase activity level compared with the Hsp104-deleted strain (Figure 4A). A similar reduction in enzyme activities was also seen in the case of other enzymes, i.e. SOD (Figure 4B), glutathione peroxidase (Figure 4C), GST (Figure 4D), glutathione reductase (Figure 4E) and thioredoxin reductase (Figure 4F), for the parental strain. Yeast glutathione peroxidases do not exhibit classical peroxidase activity. They have been shown to demonstrate greater similarity with PHGPxs (phospholipid hydroperoxide glutathione peroxidases) [35]. However, they can still be assayed using the protocol employed in the present study [25]. Lower enzyme activity in the parental strain as compared with the ΔHsp104 strain correlated well with the build-up of ROS in the parental strain, as reported above (Figures 2 and 3).

Figure 4 Assay of antioxidant enzymes in S. cerevisiae parental (empty bars) and ΔHsp104 (filled bars) strains

(A) Measurement of catalase activity. ***P<0.001 against transformed parental strain. (B) Measurement of SOD activity. *P<0.05 against transformed parental strain. (C) Measurement of glutathione peroxidase activity. **P<0.01 against transformed parental strain. (D) Measurement of GST activity. **P<0.01 against transformed parental strain. (E) Measurement of glutathione reductase activity. **P<0.01 against transformed parental strain. (F) Measurement of thioredoxin activity. **P<0.01 against transformed parental strain. Experiments were carried out in triplicate, means±S.E.M. are shown.

Prion propagation leads to oxidative damage to proteins

The presence of RNQ1 in the aggregated form in the parental strain led to greater accumulation of ROS in yeast cells. The presence of free radicals has been reported to generate carbonyl groups on amino acid residues [36,37]. The formation of carbonyl groups is enhanced during oxidative stress and has been taken to be a marker of protein damage in cells [38]. The carbonyl groups on amino acid residues were derivatized using DNPH and detected using an anti-DNPH antibody. Western blotting showed that, in the case of the parental strain expressing RNQ1in the prion form, higher oxidative stress resulted in a higher level of protein carbonylation and greater oxidative damage to intracellular proteins (Figure 5A). Upon expression of RNQ1in the soluble form, no difference in the carbonylation pattern of the proteins could be seen, as compared with uninduced cells or cells transformed with empty vector. This correlated strongly with the significantly lower oxidative stress observed in the ΔHsp104 strain (Figures 2E and 3B) compared with the transformed parental strain (Figures 2B and 3B).

Figure 5 Oxidative damage to proteins

(A) Immunoblotting using an anti-DNPH antibody to detect oxidative damage to intracellular proteins in S. cerevisiae. Lane 1, ∆Hsp104 strain transformed with pRS315-RNQ1-RFP and induced with 500 μM CuSO4; lane 2, uninduced ∆Hsp104 strain transformed with pRS315-RNQ1-RFP; lane 3, ∆Hsp104 strain transformed with pRS315-RFP and induced with 500 μM CuSO4; lane 4, parental strain transformed with pRS315-RNQ1-RFP and induced with 500 μM CuSO4; lane 5, uninduced parental strain transformed with pRS315-RNQ1-RFP; and lane 6, parental strain transformed with pRS315-RFP and induced with 500 μM CuSO4. Protein loading was equal in all lanes. (B) Immunoblotting to determine the expression levels of SOD (upper panel) and thioredoxin (lower panel) using anti-SOD1 and anti-thioredoxin antibodies. Lane M, standard protein markers with molecular masses in kDa to the left-hand side.

Discrepancy between gene expression analysis and enzyme activity assay

The expression of RNQ1 in the aggregated or soluble form resulted in a difference in the activity levels of all antioxidant enzymes assayed (Figure 4). The genes which could have been affected due to RNQ1 toxicity in parental and ΔHsp104 strains were Trx1 (thioredoxin), Glr1 (glutathione reductase), Gtt1 (GST), Gtt2 (GST), Gpx2 (glutathione peroxidase), Gpx3 (glutathione peroxidase), Ctt1 (catalase), Sod1 (SOD) and Sod2 (SOD). The expression of Gpx1 has been reported to be up-regulated under glucose-starved conditions [39] and hence its level was not monitored here. The major glutathione peroxidase activity against oxidative stress is reported to be coded by Gpx3. The expression of Gpx2 is also reported to be up-regulated on exposure to a number of oxidative stress conditions [39]. Hence, the levels of these two glutathione peroxidase-coding genes were determined. Expression of genes coding for all antioxidant enzymes was significantly up-regulated in the parental strain as compared with the ΔHsp104 strain, except for catalase (Table 1). The integrity of amplification was confirmed by analysing the amplified products by agarose gel electrophoresis. Single sharp bands were observed for all genes (results not shown). The positions of the bands matched with the amplicon sizes expected from the parameters used to design the primers.

View this table:
Table 1 Fold change in gene expression in transformed S. cerevisiae parental strain as compared with ΔHsp104 strain

PCRs were carried out in triplicate for three different cDNA preparations. Values are fold change±S.E.M. in gene expression as described in the Experimental section.

The expression of genes that are affected due to RNQ1 toxicity in the parental strain as compared with the parental strain transformed with the empty vector is shown in Table 2. Expression of genes coding for all antioxidant enzymes increased significantly in the parental strains, except Trx1. Increased gene expression levels of prion proteins were observed in the cells expressing RNQ1 aggregates. This confirmed that the altered levels of gene expression were due to the expression of RNQ1 in the aggregated form.

View this table:
Table 2 Fold change in gene expression as a result of expression of an insert in S. cerevisiae

PCRs were carried out in triplicate from three different cDNA preparations. ΔΔCT values indicate fold change in gene expression (RNQ1 transformed−empty vector) as explained in the Experimental section and are means±S.E.M.

The expression of genes in the mutant strain expressing RNQ1 as compared with the mutant strain transformed with the empty vector is shown in Table 2. Expression of Gtt1, Gpx3, Ctt1 and Sod1 was significantly up-regulated in the ΔHsp104 strain. Expression of Trx1, Glr1, Gtt2 and Gpx2, however, was down-regulated in the case of cells expressing RNQ1. No significant difference was observed in the expression levels of genes coding for prion proteins in this case. Overexpression of Trx1 and Trx2 has been reported to enhance the response of wild-type S. cerevisiae to stress generated by peroxides [40]. However, the expression of Trx1 is reported to remain unaffected on exposure to H2O2 or diamide, an oxidative stress-inducing agent. In our case, the expression of Trx1 was uniformly down-regulated in the case of parental cells where the oxidative stress was higher, unlike genes coding for other antioxidant enzymes. In order to confirm whether the decrease in enzyme activity is due to a defect in translation or a post-translational effect, the expression level of one of these enzymes, SOD, was measured. Immunoblotting showed that the expression level of SOD was only marginally lower in the parental strain (Figure 5B), which exhibited a higher level of oxidative stress (Figure 3B). The expression level of thioredoxin, a redox protein, whose gene expression level was up-regulated in the parental strain as compared with the mutant strain (Table 1), was also monitored. As can be seen in this case as well, immunoblotting showed that the protein expression level of thioredoxin was not significantly different in the parental strain as compared with the ΔHsp104 strain (Figure 5B). Thus there is a mismatch between gene and protein expression levels in cells exposed to various levels of oxidative stress, which proved that the lower enzyme activity observed in parental cells was not only due to an error in translation, as has been reported by others [4143], but also due to a post-translational defect leading to the synthesis of a hypoactive enzyme.

Prion propagation leads to decreased cell viability

A toxicity assay was carried out to estimate cytotoxicity due to expression of the RNQ1 protein in parental and ΔHsp104 strains. When cells were grown in dextrose without induction of expression of RNQ1, no difference in the growth pattern of cells could be seen in either case (Figure 6). When cells were grown in the presence of CuSO4, no difference in the growth pattern of ΔHsp104 cells, which express RNQ1 in the soluble form, could be seen. In the case of the parental strain, however, reduced viability was observed.

Figure 6 Assay of cell viability

Toxicity assay for parental (P) and ∆Hsp104 (M) strains was carried out as shown in the Figure. Decreasing order of cell count is indicated with the grey gradient.

As the RNQ1 protein was expressed in the soluble form in the ΔHsp104 strain (Figure 1) and the level of oxidative stress was lower than in the parental strain (Figures 2 and 3), the viability of these cells was higher. No significant difference in the toxicity of the cells was observed in the presence of CuSO4 alone in both parental and ΔHsp104 strains containing the empty vector. This too correlated well with the non-interference of Cu2+ in the generation of ROS in either strain (Figure 3A). Viabilities of parental and mutant empty vector strains were higher as compared with the parental strain, which confirmed that RNQ1 was toxic in its aggregated form, which led to higher oxidative stress in the cell.

DISCUSSION

Catalase is reported to have a significantly lower affinity for the major ROS H2O2 in the cell as compared with glutathione peroxidase [39]. In fact, unlike glutathione peroxidase, many catalases do not recognize higher peroxides as their substrates [39]. This may explain why Ctt1 was not up-regulated in the transformed parental strain expressing RNQ1 in the prion form. Along with their role as anti-scavengers, glutathione peroxidases are known redox sensors in yeast cells [39,44]. The thioredoxin and glutaredoxin (glutathione) (Grx) systems help in maintaining the redox potential of the cell. They thus form an important element of the cellular defence mechanism against oxidative stress. The yeast thioredoxin system consists of the thioredoxins Trx1 and Trx2, and thioredoxin reductase. Trx1 and Trx2 are highly homologous (~74% identity) which probably explains their overlapping functions [45]. They are not required for normal cell growth. Overexpression of Grx1 and Grx2 has been shown to cause only a minor resistance to stress due to H2O2 and none due to higher peroxides [40]. Increased expression of genes for expression of other prion proteins was also seen to occur in the parental strain as compared with ΔHsp104 strain. It is probable that due to the conversion of RNQ1 into the prion form in the parental strain, the other proteins are sequestered, become non-functional and are unable to carry out their cellular activities [7,8]. The cell attempts to overexpress the genes for the corresponding proteins as a compensatory defence mechanism as in the case of antioxidant enzymes, leading to an up-regulation of the corresponding genes.

What is the link between [RNQ1+] and diminished response to oxidative stress? Previous reports have clearly demarcated the fates of misfolded proteins which are substrates of the ubiquitin–proteasome system and those which are not [46,47]. The former, which include most of the cytosolic proteins, are localized in the subcellular compartment proximal to the region participating in ERAD (endoplasmic-reticulum-associated degradation). It is referred to as the JUNQ (juxtanuclear quality control compartment). The second class, which constitutes amyloidogenic and prion proteins, including RNQ1 and Ure2, are partitioned off to a perivacuolar space referred to as the IPOD (insoluble protein deposit). Notably, the subcellular localization of the two classes of proteins can be altered by modifying their ability to be conjugated (or not) to ubiquitin [46]. What is more interesting in the present context is that oxidatively damaged proteins, i.e. those which are carbonylated, are co-localized in the IPOD along with RNQ1 [48]. Both JUNQ and IPOD are retained by the mother cell. The co-localization thus also partially explains why oxidatively damaged proteins are asymmetrically partitioned off to the mother cell during budding. The presence of a ‘sink’, in the form of IPOD, presumably hastens the effect of the proteotoxic insult due to elevated ROS in the cell. The [RNQ1+] cell is unable to compensate for this even after overexpressing the genes for scavenger activities.

Comparison of gene expression levels with enzymatic activity showed discrepancy in almost every case. In general, it would be expected that, as a part of the oxidative stress response of the cell, both the gene expression and the activity of the antioxidant enzymes would increase. However, a significant observation in this regard has been that an increased level of gene expression need not necessarily correspond to an increased expression of the corresponding protein [4143]. Exposure to menadione, a quinone compound that induces oxidative stress in the cell finally leading to cell death, resulted in an increased transcription rate of genes for various antioxidant enzymes and increased mRNA levels in yeast cells [41]. However, neither the expression level nor the activities of the corresponding enzymes showed any change. It has been shown that H2O2 inhibits the initiation of translation via Gcn2 protein kinase, which phosphorylates eukaryotic initiation factor-2 [42]. This results in global inhibition of protein translation inside the cell. Thus, as a result of toxic insult due to increased levels of ROS in the cell, transcription of genes for antioxidant enzymes is up-regulated. However, expression and activity of the enzymes are reduced. Since a higher level of ROS is produced in the parental than in the mutant strain, the mismatch between gene expression data and enzyme activity is higher in the former case. Oxidative stress was shown to lead to changes in the amino acid residues of the proteins (Figure 5). The possibility of post-translational inactivation of antioxidant enzymes cannot be completely ruled out, which is observed with immunoblotting of scavenger and redox proteins. As stated above, studies with menadione have reported lower expression of scavenger enzymes in yeast as well as lower activities [41]. Hence, it can be speculated that the loss of enzymatic antioxidant activity is partially due to error in translation as well as due to inactivation of correctly translated proteins, probably by incorrect folding.

Conclusion

The presence of Hsp104 is necessary for the prion protein RNQ1 to form aggregates. Hsp104 generates prion seeds leading to prion propagation and increased level of oxidative stress, further promoting prion propagation. When expressed in the prion form, RNQ1 leads to an enhanced level of oxidative stress in yeast cells. As is evident from gene expression analysis, when challenged with oxidative stress, the cells overexpress genes coding for antioxidant enzymes. However, presumably due to the adverse effect on the global translation machinery of the yeast cell as well as impaired post-translational efficiency, the corresponding enzymes are synthesized in a functionally inactive form and hence intracellular enzymatic scavenging activity is low. This leads to accumulation of ROS, which proves to be lethal for the cell, as seen in the viability assay. Thus, in prion-containing cells, controlled down-regulation of Hsp104 may be explored as a means to counter the toxic effects of increased oxidative stress.

AUTHOR CONTRIBUTION

Kuljit Singh, Aliabbas Saleh and Ankan Bhadra generated and analysed the data. Ipsita Roy designed the experiments, analysed the data and wrote the paper.

FUNDING

Partial financial support from the Department of Science and Technology is acknowledged. A.A.S. is grateful to the Council for Scientific and Industrial Research (CSIR) for the award of a senior research fellowship. A.K.B. is grateful to the Department of Science & Technology Innovation in Science Pursuit for Inspired Research (DST-INSPIRE) programme for the award of a junior research fellowship.

Acknowledgments

The pRS315-RNQ1-RFP construct was received as a gift from Professor Douglas Cyr, Cell Biology and Physiology, University of North Carolina, Chapel Hill, NC, U.S.A. We thank Professor Thomas O’Halloran, Department of Chemistry, Northwestern University, Evanston, IL, U.S.A., for the gift of the anti-SOD antibody and Professor Chris M. Grant, Faculty of Life Sciences, University of Manchester, Manchester, U.K., for the gift of the anti-thioredoxin antibody.

Abbreviations: DCF, dichlorofluorescein; DCFH-DA, dichlorodihydrofluorescein diacetate; DNPH, 2,4-dinitrophenylhydrazine; DTNB, 5,5′-dithiobis-(2-nitrobenzoic acid); Hsp, heat-shock protein; IPOD, insoluble protein deposit; JUNQ, juxtanuclear quality control compartment; MMLV, Moloney murine leukaemia virus; NBT, Nitro Blue Tetrazolium; ROS, reactive oxygen species; SC-LEU, synthetic complete medium without leucine; SOD, superoxide dismutase

References

View Abstract