Reduced translocation of nascent prion protein during ER stress contributes to neurodegeneration

Abstract

During acute stress in the endoplasmic reticulum (ER), mammalian prion protein (PrP) is temporarily prevented from translocation into the ER and instead routed directly for cytosolic degradation. This "pre-emptive" quality control (pQC) system benefits cells by minimizing PrP aggregation in the secretory pathway during ER stress. However, the potential toxicity of cytosolic PrP raised the possibility that persistent pQC of PrP contributes to neurodegeneration in prion diseases. Here, we find evidence of ER stress and decreased translocation of nascent PrP during prion infection. Transgenic mice expressing a PrP variant with reduced translocation at levels expected during ER stress was sufficient to cause several mild age-dependent clinical and histological manifestations of PrP-mediated neurodegeneration. Thus, an ordinarily adaptive quality-control pathway can be contextually detrimental over long time periods. We propose that one mechanism of prion-mediated neurodegeneration involves an indirect ER stress-dependent effect on nascent PrP biosynthesis and metabolism.

Fig. 1. PrP^Sc accumulation induces ER stress and reduces PrP translocation into the ER

(A) Total brain homogenate from normal and PrP^Sc-infected hamsters (‘normal’ and ‘scrapie’) were analyzed by staining for total proteins or immunoblotted for the indicated antigens. Asterisk indicates trace IgG heavy chain that occassionally contaminates tissue homogenates from residual blood. (B) Analysis of PrP from the samples in panel A for resistance to digestion by proteinase K (PK). No protease-resistant PrP was detected in normal tissue even upon gross overexposure of the blot (not shown). (C) ER microsomes from normal and PrP^Sc-infected hamsters were used for in vitro translocation assays for PrP. After synthesis with ³⁵S-methionine, the samples were treated with PK to digest non-translocated products, and the protease-protected PrP (indicative of its successful translocation into the ER microsomes) was recovered by immunoprecipitation. Shown are autoradiographs of the translocated PrP. Note that in the absence of membranes (‘-’), full length PrP is not protected. In samples from animals 1 and 4 weeks post-inoculation (before PrP^Sc accumulation), no difference is observed in normal and scrapie microsomes. By contrast, at a time when PrP^Sc accumulation is high (7 and 10 weeks; see Sup. Fig. 1), translocation is significantly lower in infected microsomes relative to the uninfected control. Note that each pair of microsomes at every time point were isolated and analyzed in parallel; however, comparisons may not be valid between time points, so the apparent increase in translocation in normal microsomes over time may not be meaningful.

Fig. 2. Design and mechanistic analysis of a constitutive pQC variant of PrP

(A) Schematic diagram of cytosolic quality control (cQC), pre-emptive quality control (pQC), and ER-associated degradation (ERAD). Of these, only the pQC pathway has been demonstrated to be utilized by PrP during ER stress (Kang et al., 2006; Orsi et al., 2006). (B) The signal sequences and cleavage site (arrowhead) for constructs used in this study. Lysine residues used for crosslinking analyses are in bold. (C) Crosslinking to cytosolic proteins of ribosome-associated nascent chains (RNCs) synthesized up to PrP residue 150. PrP (P), Ifn-PrP (I), Opn-PrP (O), and ΔSS-PrP (D) are analyzed. The arrowhead indicates crosslinks to SRP54, confirmed by immunoprecipitation (right panel). Asterisk indicates the position of uncrosslinked nascent chains. (D) Crosslinking to ER proteins of RNCs synthesized to PrP residue 150. After crosslinking, the products were fractionated into membrane-associated and lumenal proteins, shown in the left and middle panels, respectively. Open arrow indicates crosslinks to the translocon component Sec61α (verified by immunoprecipitation; not shown), and closed arrow indicates crosslinks to the lumenal chaperone PDI, identified by immunoprecipitation in the right panel. (E) PrP was synthesized in the absence or presence of ER-derived rough microsomes (RM) in a lysate supplemented with His-tagged ubiquitin. Ubiquitin-conjugated products were captured on immoblized Co⁺². The positions of PrP species representing precursor (pre), signal-cleaved (s.c.), glycosylated (glyc) and ubiquitinated (Ub) products are indicated. Also shown are the Ubiquitin-conjugated products for ΔSS-PrP, illustrating its relatively poor ubiquitination. (F) Ubiquitination analysis (as in panel E) of PrP, Ifn-PrP, and Opn-PrP in the absence and presence of RMs. The lower panel shows the total products and the upper panel the Ubiquitin-conjugated species captured via the His-tagged ubiquitin.

Fig. 3. Ifn-PrP metabolism is distinct from ^CtmPrP and ΔSS-PrP

(A) Wild type PrP, Ifn-PrP, and Ifn-PrP(A120L) were analyzed by in vitro translation and translocation assays. An inhibitor of glycosylation was included in all reactions to simplify the banding pattern. Half of each sample was analyzed directly, while the remainder was digested with PK. The positions of full length (FL) PrP, and the proteolytic fragments corresponding to ^CtmPrP and ^NtmPrP are indicated. Note that Ifn-PrP makes comparable amounts of ^CtmPrP as wild type, while Ifn-PrP(A120L) makes substantially more. (B) Wild type PrP, Ifn-PrP, PrP(A117V) and PrP(AV3) were expressed in N2a cells, and microsomes isolated from these cells were subjected to analysis for ^CtmPrP by limited PK digestion. Shown are different relative amounts of undigested sample, as well as the products after digestion under ‘mild’ and ‘harsh’ conditions (see Hegde et al., 1998). The PK-digested samples were deglycosylated with PNGase before analysis. In this assay, PK digestion under mild conditions generates an ∼18 kD fragment corresponding to ^CtmPrP (indicated by asterisk). A smaller band corresponding to the C-terminal globular domain of PrP is indicated by the arrowheads. Note that Ifn-PrP levels are very low due to its constitutive degradation (see Fig. 4B), even though its rate of expression was verified to be comparable to wild type PrP by pulse-labeling experiments as in Fig. 4A (data not shown). A band at ∼14 kD seen in the Ifn-PrP samples appears to be a degradation intermediate that is sometimes observed. (C) PrP and ΔSS-PrP were synthesized in vitro in the absence of ER membranes and analyzed by sucrose gradient sedimentation. An aliquot of the total translation products is also shown. Note that PrP is ubiquitinated significantly more efficiently than ΔSS-PrP, and that the two proteins have different sedimentation profiles indicative of associations with different complexes.

Fig. 4. Ifn-PrP mimics pQC in vivo in the absence of ER stress

(A) ER translocation of the indicated PrP constructs in transiently transfected Hela cells subjected to acute ER stress (15 min) by Ca⁺² depletion using thapsigargin (Tg). Translocation was quantified using relative glycosylation efficiency and is indicated below the respective lanes. The positions of unglycosylated (-CHO) and glycosylated (+CHO) species of PrP are indicated. Note that protein synthesis is reduced in stressed cells due to PERK-mediated phosphorylation of eIF2α. (B) N2a cells transiently transfected with Ifn-PrP were treated with proteasome inhibitor (10 uM MG132) for 0, 2, or 4 h as indicated and analyzed by immunoblotting. Samples were separated into detergent-soluble (S) and insoluble (P) fractions before analysis. ‘4+20’ indicates samples from cells treated with inhibitor for 4 hours, and cultured in the absence of inhibitor for an additional 20 h. The last lane is a marker for mature PrP from cells expressing wild type PrP. (C) N2a cells transiently transfected with ΔSS-PrP were treated with proteasome inhibitor (10 uM MG132) for 4 h as indicated, and either harvested immediately, or cultured for an additional 4 to 24 h in the absence of inhibitor. All samples were analyzed for ΔSS-PrP by immunoblotting with 3F4 antibody. (D) N2a cells transiently transfected with Ifn-PrP or empty vector were separated into detergent-soluble (S) and insoluble (P) fractions before analysis by immunoblot using a PrP antibody that detects both endogenous PrP and Ifn-PrP. Note the lack of changes to endogenous PrP in cells expressing Ifn-PrP (most of which is found in the insoluble fraction as unglycosylated species).

Fig. 5. Phenotype of Ifn-PrP transgenic mice with constitutive pQC of PrP

(A) Lifespans of Ifn-PrP and Opn-PrP mice. All causes of death are included in the analysis except those mice that were sacrificed prematurely for analysis. (B) Representative Opn-PrP and Ifn-PrP mice at ∼2 months. Note rough hair coat and smaller size of Ifn-PrP mouse. (C) Representative Ifn-PrP mouse at ∼2 years. Note kyphosis (arrow), rough hair coat, and abnormal gait (Sup. Movie 2, 3, and 6).

Fig. 6. Histologic analysis of Ifn-PrP mice reveals mild neurodegeneration

(A-D) H&E stained saggital brain sections show no alterations to gross brain morphology or development in Ifn-PrP or Opn-PrP mice at either ∼2 months or ∼2 years of age. Green boxes indicate regions shown in greater detail in panels B-D. (E) Immunostaining for calbindin to visualize Purkinje cells (brown). Note that neither the Purkinje cells, granular layer (left) or molecular layer (right) are affected grossly in Ifn-PrP mice. (F) GFAP staining of Ifn-PrP or Opn-PrP mice at either ∼2 months or ∼2 years of age. Shown is a region of hippocampus where age-dependent increase in reactive astrocytes is observed in Ifn-PrP mice beyond that seen in old Opn-PrP mice. (G) Fluoro-Jade C staining of Ifn-PrP or Opn-PrP mice at ∼2 years of age. Shown are regions where increased staining is observed in Ifn-PrP mice. Note that no staining was observed in young mice of either genotype (data not shown).

Fig. 7. Quantitation of pQC in Ifn-PrP transgenic mice

(A) Expression of Ifn-PrP in mixed cortical cell cultures prepared from newborn transgenic and non-transgenic mice after treatment with proteasome inhibitor (10 μM MG132) for the indicated times. For comparison, PrP expression in normal hamster brain is shown. Detection was with the 3F4 monoclonal antibody selective to hamster (and not mouse) PrP. Two exposures of the blot are shown to illustrate the very low level steady state expression of Ifn-PrP, and the selective increase in the unglycosylated cytosolic form of PrP upon proteasome inhibition. (B) Cortical cultures as in panel A were pre-treated with MG132 as indicated, pulse-labeled for 1 h with ³⁵S-methionine in the absence or presence of MG132, and immunoprecipitated with either 3F4 (to selectively recover the transgenically expressed Ifn-PrP) or a pan-PrP antibody to recover both endogenous and transgenic PrPs. The white arrow indicates the position of unglycosylated (and non-translocated) Ifn-PrP, seen selectively when the proteasome is inhibited. This is also seen in the total PrP immunoprecipitates, where it represents ∼10% of total PrP synthesized.