| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Institute of Virology, Technical University of Munich, Munich, Germany;
Department Chemie, Lehrstuhl für Biotechnologie, Technical University of Munich, Garching, Germany; and
Institute of Molecular Immunology, GSF National Research Center for Environment and Health, Munich, Germany
1Correspondence: Institute of Virology, Technical University of Munich, Trogerstr. 30, 81675 Munich, Germany. E-mail: vorberg{at}lrz.tum.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: Key Words: amyloid • protein misfolding • seeding • nucleation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. During the course of infection, PrPC undergoes conformational transition into an abnormally folded aggregate structure that appears to have self-propagating capacities. Prion pathology shares several similarities with other protein-misfolding disorders on the basis of its association with proteinaceous aggregates with common beta-sheet-rich amyloid structures (2
, 3)
. However, prions are unique in that they are not only able to replicate their conformation but are also transmissible. Prion-like phenomena have also been observed in yeast, first identified as epigenetic elements of inheritance (4)
. The yeast prion phenotype [PSI+] is caused by self-propagating aggregates of the translation termination factor Sup35p (5
6
7)
. Unlike mammalian prions, yeast prions are not generally pathogenic (8
, 9)
. [PSI+] can be faithfully inherited by progeny cells in a dominant non-Mendelian manner and can be transmitted by cytoplasmic transfer independently of nucleic acid (reviewed in refs. 10
, 11
). Recent findings demonstrate that self-perpetuating prion conformations might be much more common than currently anticipated even in higher eukaryotes (12)
.
Fundamental aspects of mammalian prion biogenesis have not been elucidated. The mechanism of prion replication, the domains in PrP mediating prion assembly and potential cofactors remain elusive. Several striking similarities between mammalian and yeast prions make the well-characterized yeast prion Sup35p an attractive model to unravel basic mechanisms of prion biogenesis and assembly of aggregation-prone proteins in general (13
14
15
16
17
18)
. Sup35p and PrP display no amino acid (aa) identity, but amino termini of Sup35p and PrP exhibit an unusually high degree of structural flexibility. In Sup35p, the carboxy-terminal domain provides the translational termination activity (reviewed in refs. 14
, 16
), while the prion status of the protein is governed by the amino-terminal NM region. The N region is indispensable for conformational alteration and polymerization and comprises an oligopeptide repeat region and a glutamine/aspargine-rich region. Overexpression of N is sufficient to induce the appearance of the [PSI+] phenotype in yeast (19)
, and this region seems to be solely responsible for formation of aggregation nuclei (20)
. Experiments in yeast have demonstrated that the prion domain confers its ability to aggregate when fused to other proteins (6
, 21
, 22)
. The highly charged M region stabilizes the prion during mitosis and meiosis (23)
and increases the solubility of the protein (24
25
26)
. Unlike for Sup35p, the exact prion-determining elements in PrP remain obscure. Similar to Sup35p, oligopeptide repeats are present in the amino-terminal region of PrP. Interestingly, oligopeptide repeats in Sup35p and PrP can significantly modulate spontaneous prion formation (27
28
29
30)
.
An important difference between yeast and mammalian prion biogenesis is the cellular location of prion formation. The Sup35p prion isoform [PSI+] is present in the yeast cytosol. As PrP is attached to the outer leaflet of the membrane by a GPI moiety, conversion to PrPSc is believed to occur either on the cell surface or along the early endocytic pathway (reviewed in ref. 31
). Elegant experiments have, however, recently demonstrated that GPI anchorage is dispensable for mammalian prion formation (32)
. Furthermore, aggregated PrP has been found in the cytosol of mammalian cells (33
, 34)
and may actually have warrantable relevance in prion diseases (35
36
37
38
39)
. Recently, a prion-like propagation for cytosolic PrP aggregates has been proposed (35)
. Thus, these findings clearly demonstrate that the ability of PrP to aggregate is not restricted to the extracellular space or organelle compartments.
We have employed mammalian cell culture studies to gain insights into structural and environmental requirements for aggregation of Sup35p and cytosolic PrP. Our data demonstrate that Sup35p-NM is unable to spontaneously aggregate at detectable levels when transiently expressed in the cytosol of mammalian cells, while cytosolic PrP readily formed visible aggregates. Interestingly, fusion of N and/or M to the globular domain of PrP significantly altered aggregate appearance and/or size, demonstrating a modulating activity of Sup35p fragments. In vitro studies revealed that both NM and NM-PrP were capable of assembling into fibrils under near-physiological conditions, suggesting that certain cofactors and/or environmental conditions in mammalian cells either specifically promote or inhibit aggregation of prion proteins.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) as a template. The 3F4 epitope allows recognition of transfected mouse PrP using the monoclonal antibody (mAb) 3F4 and discrimination from endogenous wild-type PrP. Sup35p regions N, M, and NM (aa 1–123, 124–250, or 1–250, respectively) were either amplified by standard PCR using the pJC25-NM expression vector as a template (41)
or subcloned from this vector. Insertion of appropriate restriction sites allowed cloning of the respective sequences of mouse PrP and Sup35p into pcDNA3.1 and pET29b(+). A carboxy-terminal HA epitope tag for NM expression in mammalian cells was generated using a primer coding for the peptide sequence of hemagglutinin (YPYDYPDYA) for PCR amplification. Rab4 was a kind gift from M. McCaffrey (University College, Cork, Ireland) and was cloned into pcDNA3.1/NT-GFP-TOPO (G. Lutzny, Institute of Virology). The vector pEGFP-C1-Rab11 was generously provided by D. L. Marks and R. E. Pagano (Mayo Clinic, Rochester, MN, USA).
Cell culture and transient expression of Sup35p/MoPrP constructs
The mouse neuroblastoma cell line N2a (ATCC CCL-131) and the African green monkey kidney fibroblast-like cell line COS-7 (ATCC CRL-1651) were maintained in Opti-MEM or D-MEM medium (Invitrogen), respectively, containing 10% fetal calf serum, antibiotics, and glutamine in a 5% CO2 atmosphere. The medium was changed every 48 h. Transient transfections of constructs were carried out by FuGENE transfection (Roche, Mannheim, Germany), according to the manufacturer’s protocol (Invitrogen). Briefly, cells were plated in 6-cm dishes and transfected 24 h later. Cells were harvested 48 h post-transfection.
Immunoblot analysis and solubility assay
Cell lysis and immunoblotting were performed as described previously (42)
. Briefly, cells were lysed (10 mM Tris/HCl, pH 7.5; 100 mM NaCl; 10 mM EDTA; 0.5% Triton X-100; 0.5% DOC), and lysates were cleared of cell debris (1 min, 1000 g, 4°C). Cell lysates were supplemented with 0.5 mM Pefabloc protease inhibitor (Roche) and insoluble proteins separated (20 min, 20,000 g, 4°C). Soluble proteins in the supernatant were precipitated with four volumes of methanol at –20°C over night. Insoluble fractions (pellet) were resuspended in TNE buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM EDTA), and both soluble and insoluble protein fractions were separated on 12.5% SDS gels. Expression of recombinant proteins was analyzed by Western blot using either mouse anti-HA monoclonal antibody (mAb) F7 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), mouse mAb 3F4 (43)
(Signet Pathology, Dedham, MA, USA) or rat mAb 7H5 of IgG2a subclass raised against a peptide comprising aa 8–27 of Sup35p and ECL plus (GE Healthcare, Buckinghamshire, UK). Signal intensities were quantified using a ScanJet 4100C scanner (HP) and Image Quant TL software (GE Healthcare). The chi-square (
2) test was used for statistical analysis.
Proteinase K treatment
Confluent transfected cells were lysed in lysis buffer (0.5% Triton X-100, and 0.5% DOC in PBS) for 10 min on ice followed by a nucleic acid degradation using benzonase (Sigma, Steinheim, Germany) (75 U, 30 min, 4°C). Lysate aliquots were incubated for 30 min at 37°C with the indicated amount (5–50 µg/ml) of proteinase K (PK). Proteolysis was terminated by addition of protease inhibitor Pefabloc (Roche, Mannheim, Germany). Samples were incubated with 5 M GdnSCN for 1 h at 37°C with agitation, precipitated with four volumes of methanol and analyzed by Western blot analysis.
Confocal laser scanning microscopy
Cells were treated for immunofluorescence as delineated earlier (42)
, with slight modifications. Briefly, cells were plated on poly-L-lysine-coated coverslips and transfected with the respective constructs. 48 h post-transfection, cells were fixed and permeabilized (0.1% Triton X-100). Primary mAbs anti-PrP 3F4, anti-HA F7, anti-Sup35-N 7H5, anti-
-tubulin GTU-88 (Sigma), or anti-LAMP-1 CD107a (BD Pharmingen, Heidelberg, Germany) and polyclonal rabbit antisera antivimentin H-84, anti-Calnexin (C-20), and anti-Rab6 (C-19) (Santa Cruz Biotechnology) or anti-PrP A7 (44)
were diluted in blocking solution (0.2% gelatin). Cy2– or Cy3– secondary antisera (Dianova, Hamburg, Germany) were added in blocking solution. Nuclei were stained with Hoechst DNA staining dye (2 µg/ml in PBS) (Sigma) for 10 min at room temperature (RT). After three additional rinses with PBS, slides were mounted in PermaFluor Aqueous Mounting Medium Liquid (Beckman Coulter, Marseille, France) and kept dry at –20°C. Confocal laser scanning microscopy was carried out using a LSM 510 laser-scanning microscope (Zeiss, Göttingen, Germany). For determination of the ratio of transfected cells displaying visible aggregates to transfected cells without aggregates, 
300 cells were scored for each construct. The graphic data reflect the average of three independent experiments.
Production and purification of recombinant NM, NM-PrP, and PrP-M
All proteins were produced in E. coli BL21(DE3) codon+ (Stratagene, La Jolla, CA, USA) by recombinant expression on induction with 1 mM IPTG (at an OD600 of 0.7) and incubation at 30°C for 4 h. NM was purified via anion exchange chromatography (Q-Sepharose Fast Flow; Amersham, Piscataway, NJ, USA) in the presence of 8 M urea and a subsequent heat precipitation step (1 h at 60°C in PBS; NM remains soluble). PrP-M was purified via SP-Sepharose Fast Flow cation exchange chromatography (Amersham) and a subsequent heat precipitation step (1 h at 60°C in PBS; PrP-M remains soluble). NM-PrP was found in inclusion bodies that could be separated by centrifugation. Subsequently, the protein was purified chromatographically by Q-Sepharose Fast Flow (Amersham, Piscataway, NJ, USA) in presence of 8 M urea. Protein concentrations were determined using the calculated extinction coefficients of 0.99 (NM), 1.70 (PrP-M), and 1.28 (NM-PrP), respectively, for a 1 mg/ml solution in a 1-cm cuvette at 276 nm for NM or 280 nm for PrP-M and NM-PrP (45)
.
Fibril assembly
The in vitro assembly of recombinantly produced proteins into amyloid fibrils was monitored by thioflavin T (ThT) (Sigma) fluorescence with a Spex FluoroMax-3 fluorescence spectrometer (Jobin Yvon, Edison, NJ, USA). Samples of the respective proteins were denatured with small volumes of 8 M GdmCl for at least 2 h. Subsequently proteins were refolded by a hundred-fold rapid dilution in PBSazide (containing 0.02% NaN3). Insoluble protein was removed by centrifugation, and the samples were adjusted to 5 µM of the respective proteins. The protein solutions were aliquoted and incubated under quiescent conditions or rotated (60 rpm) at room temperature. At specific time points, samples were diluted with a ThT solution (5 µM in PBSazide), and fluorescence emission at 480 nm was measured on excitation at 450 nm (25°C). Degradation of the proteins during long-run incubation was monitored by SDS-PAGE analysis (15% wt/vol acrylamide) followed by staining with Coomassie blue R (Serva).
Analysis of fibril morphology
Morphology of amyloid fibrils was investigated by contact mode atomic force microscopy (AFM). Samples were placed on freshly cleaved mica attached to 15-mm AFM sample disks (Ted Pella, Redding, CA, USA). After 3 min of adsorption at 25°C, disks were rinsed several times with Millipore-filtered H2O. The samples were allowed to air dry. Contact-mode imaging was performed on a multimode scanning probe microscope (Veeco, Santa Barbara, CA, USA) by using long, thin-leg standard silicon nitride (Si3N4) probes.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A) in mammalian cells, we established a model system using Sup35p and mouse PrP (MoPrP) fusion proteins transiently expressed in N2a or COS cells (Fig. 1B
). Constructs were generated coding for cytosolic PrP (PrPcyto, PrP aa residues 23–230), Sup35p-NM (Sup35p aa residues 1–250), or chimeric proteins thereof (Fig. 1B
). For N-PrP and NM-PrP, aa residues 1–123 or 1–250 of Sup35p were fused to the murine PrP (aa residues 90–230). In PrP-M, the amino-terminal N region in Sup35p (aa 1–123) was replaced by PrP residues 23–120. For detection purposes, PrPcyto, N-PrP, and NM-PrP were tagged with the PrP epitope for the mAb 3F4 that only detects exogenous PrP. For detection of NM, a HA tag was inserted at the carboxy-terminus (NM-HA). Cells expressing fusion proteins showed no growth defect for the course of the experiment and no obvious cytotoxic effects, except for cytosolic PrP that appeared to be slightly toxic (data not shown). Western blot analysis revealed that proteins were migrating at the expected size of 25 kDa (PrPcyto), 32 kDa (N-PrP), 50 kDa (NM-PrP), 35 kDa (PrP-M), and 40 kDa (NM-HA), respectively (Fig. 2
A).
|
|
On overproduction in yeast cells, the prion domain of Sup35p assembles into aggregates that can form visible foci (46)
. Interestingly, 48 h post-transfection, NM-HA was almost completely soluble in the cytosol of transfected mammalian cells (Figs. 2B
and Supplemental Fig. S1A). Similarly to NM-HA, expression of PrP-M yielded visible aggregates in only very few cells, with the vast majority of PrP-M being dispersed homogeneously throughout the cytoplasm. By contrast, expression of cytosolic PrP (PrPcyto) led to formation of small aggregates (Fig. 2B
) (47)
. Replacement of amino-terminal PrP residues 23–89 with Sup35p-N (N-PrP) did not abolish aggregate formation. Multiple small aggregates sometimes clustered in one area of the cell. Surprisingly, fusion of both N and M to the carboxy-terminal domain of PrP (NM-PrP) led to the spontaneous formation of huge, single aggregates (or foci) that strongly differed from the smaller aggregates observed with PrPcyto or N-PrP (Fig. 2B
). A significant fraction of cells contained 1–4 dot- and ring-shaped aggregates of NM-PrP reminiscent of Sup35p aggregates associated with the [PSI+] status in yeast (46)
. Multiple ring-shaped aggregates were occasionally present within one cell. Collection of sequential scans along the vertical (z) axis revealed a spherical structure of the giant NM-PrP aggregates (data not shown). Thus, the ring-like appearance of the aggregates is likely due to the antibody binding only to the surface of the aggregate. The size of single aggregates ranged from small (PrPcyto and N-PrP, less than 1 µm) to big aggregates (NM-PrP) with aggregates 5 µm in diameter.
Confocal microscopy analysis further demonstrated that foci were present 6 h post-transfection, indicating that newly translated proteins were immediately incorporated into the growing aggregate (data not shown). Comparable aggregate formation was also observed in COS cells, and the PrP-deficient hippocampal cell line HpL3–4 transiently transfected with the respective constructs except that for NM-PrP, some COS cells appeared to harbor more aggregates compared to N2a cells (data not shown). Of note, endogenous PrP was not recruited into the aggregates, and surface levels of PrP remained unaltered due to the expression of chimeric constructs (data not shown).
Comparison of the frequencies of cells containing aggregated proteins revealed very few cells containing visible foci of NM-HA or PrP-M (Fig. 2C
). By contrast, expression of PrPcyto led to aggregate formation in 24 ± 5% of transfected cells. Replacement of the amino-terminal region of PrP with the Sup35p-N or -NM regions significantly increased the number of cells that harbored visible foci to 50 ± 2 and 55 ± 9%, respectively. In conclusion, when transiently expressed in the cytosol of mammalian cells, PrPcyto, N-PrP, and NM-PrP displayed an intrinsic property to aggregate, while NM-HA and PrP-M did not. Interestingly, the frequency of aggregate induction seemed to be influenced by the N and/or M regions, as the number of cells harboring aggregates increased in cells transfected with NM-PrP or N-PrP. By contrast, striking qualitative differences are apparent between PrPcyto or N-PrP and NM-PrP foci, suggesting that the M region alone or in combination with N is capable of modulating aggregate size.
PrPcyto, N-PrP, and NM-PrP aggregates lack characteristic aggresome features and are not subjected to lysosomal degradation
Misfolded and aggregated proteins that are a potential hazard to the cell can be rapidly eliminated by sequestration into intracellular, perinuclear inclusions, so-called aggresomes (48
49
50)
. Aggresome formation is an active cellular process dependent on retrograde protein transport along microtubule tracts leading to protein complex formation around the microtubule organization center. A common feature of aggresomes is the colocalization with centrosome markers (50)
. In our study, inclusions of recombinant proteins did not appear to be formed in the perinuclear region, suggesting that aggregates differed from conventional aggresomes (Figs. 2B
and 3)
. Interestingly, staining of N2a or COS cells did not reveal any colocalization of the recombinant proteins with centrosome marker -tubulin (Fig. 3
A). Disruption of the microtubule filaments by nocodazole for 24 h beginning at the time of transfection had no influence on aggregate formation, demonstrating that inclusions formed independently of an active transport along microtubule tracts (data not shown).
|
Sequestration of misfolded proteins into aggresomes has further been reported to be associated with a collapse of the intermediate filament vimentin forming a cage around the aggresome (48)
. However, costaining for vimentin did not demonstrate any association of recombinant proteins with vimentin (Fig. 3B
).
Cytosolic protein complexes can also actively be captured into bilammellar autophagosomes (51)
that ultimately fuse with lysosomes for acidic hydrolytic degradation (52)
. To analyze whether aggregates formed by PrPcyto and fusion proteins were subjected to lysosomal clearance, cells were costained for lamp-1 that is abundant in lysosomes. None of the aggregated recombinant proteins costained with this lysosomal marker (Fig. 3C
), demonstrating that protein aggregates were not present in this compartment of the endocytotic pathway.
In conclusion, the lack of aggresome markers suggests that PrPcyto, N-PrP, and NM-PrP aggregates do not evoke an active cellular sequestration into aggresomes. Furthermore, none of the protein aggregates is targeted to lysosomal degradation. Further studies revealed that none of the constructs colocalized with markers for Golgi (Rab6), endoplasmatic reticulum (calnexin) or early or late endosomes (Rab4, Rab11) (Supplemental Fig. S2). The observed aggregate formation occurring within 6 h post-transfection indicates an intrinsic property of PrPcyto, N-PrP, and NM-PrP to spontaneously aggregate in the cytosol of mammalian cells.
PrPcyto, NM-PrP, and N-PrP form insoluble complexes in the cytosol of mammalian cells and display increased resistance to proteolysis
Fractionation analysis was performed to determine the ratio of insoluble to soluble recombinant proteins in N2a cells transiently transfected with the respective constructs. Whole cell lysates were subjected to centrifugation to separate soluble from insoluble proteins, and fractions were analyzed by Western blot analysis. Both NM-HA and PrP-M mainly remained soluble (Figs. 4
A and Supplemental Fig. S1B). By contrast, PrPcyto, N-PrP, and NM-PrP were detected in the pellet fraction (Fig. 4A
). Quantification of bands revealed that 54 ± 7% of PrPcyto, 62 ± 13% of N-PrP, and 62 ± 11% of NM-PrP were sequestered into insoluble complexes that were pelleted by centrifugation (Fig. 4B
). Thus, solubility analysis confirmed that NM-PrP, N-PrP, and PrPcyto tend to aggregate into insoluble high-molecular mass complexes.
|
An increased protease resistance is a characteristic feature of Sup35p in [PSI+] cells and abnormal prion protein (PrPSc) in mammalian prion diseases. Analysis of PK resistance revealed a complete PK sensitivity of NM-HA (Fig. 5
). This PK sensitivity of NM-HA was further confirmed using a mAb against the Sup35p N region (Supplemental Fig. S1C). PrPcyto and N-PrP displayed increased protease resistance up to 50 µg/ml tested (Fig. 5)
. Surprisingly, NM-PrP was less PK-resistant. The full-length protein was digested on treatment with 20 µg/ml PK. Interestingly, PK treatment of PrPcyto, N-PrP, NM-PrP, and to a minor extent of PrP-M produced several low-molecular weight fragments, which were detected using the 3F4 antibody. In conclusion, PrPcyto, N-PrP, and NM-PrP share a partial protease resistance reminiscent of the protease resistance of Sup35p in yeast [PSI+] cells and mammalian PrPSc.
|
Aggregate formation is mainly due to the globular domain of PrP but is modulated by Sup35p fragments
Previous experiments demonstrated that PrPcyto, N-PrP, and NM-PrP exhibited different aggregation characteristics, suggesting that Sup35p-N or -NM regions significantly affected aggregate formation. However, PrPcyto still harbored the amino-terminal region of PrP that could potentially account for the observed differences. Thus, two additional constructs, M-PrP (Sup35p-M fused to PrP90–230) and PrP90–230, were generated to assess the influence of Sup35p-M and the amino-terminal region of PrP on aggregate size, induction, solubility, and PK resistance. Western blot analysis revealed expression of the new constructs in N2a cells (Fig. 6
A). PrP90–230 formed aggregates comparable to PrPcyto in size and localization (Fig. 6B
). As sometimes observed for PrPcyto and N-PrP, small PrP90–230 aggregates frequently clustered in one specific area of the cell. Experiments were independently repeated three times, and cells displaying visible aggregates counted. Results were compared to data obtained with PrPcyto (Fig. 2C
). Interestingly, significantly more cells exhibited visible PrP90–230 aggregates compared to cells expressing PrPcyto (45±5 vs. 24±5%, P<0.01, compare Figs. 2C
and 6C
), suggesting that the amino-terminal part of PrP had a negative effect on the nucleation rate of PrP90–230. A sedimentation assay demonstrated comparable solubility and PK-resistance of PrP90–230 and PrPcyto (Fig. 6D, E
, and compare Figs. 6F
and 5
). Thus, the globular domain of PrP promotes aggregate induction and is sufficient to confer PK resistance and partial insolubility.
|
M-PrP formed huge foci that sometimes even outgrew NM-PrP (up to 10 µm in diameter). This confirms that the M region is responsible for the aggregate size effect and increases seeding efficiency (Fig. 6B
). Interestingly, the ratio of cells with visible M-PrP aggregates to total M-PrP expressing cells was significantly lower compared to the ratio of cells with visible PrP90–230(30±3 vs. 45±5%, P<0.05), or NM-PrP foci to total transfected cells (30±3 vs. 55±9%, P<0.001) (compare Figs. 2C
and 6C
). This suggests that M alone exhibited an inhibitory effect on nucleation similar to the amino-terminal region of PrP, but that this activity could be counteracted by N. No significant differences in sedimentation characteristics (Fig. 6D, E
) were observed for NM-PrP and M-PrP. However, M-PrP was drastically more sensitive to PK compared to PrP90–230 (Fig. 6F
), implying that fusion with M altered the aggregate packing, making it more amenable to proteolysis. Thus, resistance to PK appears to be conferred by the globular domain of PrP but can be modulated by the M fragment of Sup35p.
Confocal microscopy analysis of N2a and COS cells expressing NM-PrP revealed an unexpected high number of aggregates in some COS cells compared to N2a cells. To study whether the amount of foci varied dependent on the cell type, 20 cells per N2a and COS cell lines transiently expressing NM-PrP were analyzed for their aggregate contents (Fig. 7
). Because of the small aggregate sizes of N-PrP, PrPcyto, and PrP90–230 it was not possible to determine whether there is a quantitative difference in the number of foci per cell with these constructs. For both N2a and COS cells, the amount of NM-PrP aggregates varied greatly. Surprisingly, however, in N2a cells, aggregate numbers rarely exceeded 14 aggregates per cell, while in COS cells, 14 and more aggregates (up to 183) were detected in at least half of the studied cells (Fig. 7A
). Notably, the observed differences in aggregate amounts per cell were dependent on N, as M-PrP failed to show a drastic cell type-dependent difference in aggregate numbers (Fig. 7A
, B). These data suggest that N influenced nucleation in a cell-type dependent manner. In conclusion, these data confirm the previous findings that the globular domain of PrP is likely the main aggregation-inducing domain, while a dynamic interaction of Sup35p-N and -M regions modulates nucleation and seeding, and, therefore, the aggregate structure reflected by the aberrant PK resistance.
|
NM and NM-PrP both form amyloid fibrils in vitro
It has previously been shown that Sup35p, Sup35p-N and Sup35-NM are capable of assembling into amyloid fibrils under near-physiological conditions in vitro (24
, 41
, 53
, 54)
. To test the capability of Sup35p-PrP fusion proteins to self-assemble into amyloid fibrils in vitro, assembly was monitored by Thioflavin T (ThT) fluorescence. Solutions of the respective recombinantly produced proteins (5 µM each) were incubated under quiescent conditions. On incubation for more than 4 wk, Sup35p-NM formed detectable amyloid structures, indicated by an increase of ThT fluorescence emission at 480 nm (Fig. 8
A). Neither PrP-M nor NM-PrP showed any amyloid-specific increase in ThT fluorescence emission under these experimental conditions. These findings could be confirmed by AFM imaging after 4 wk of incubation (Fig. 8C
). SDS-PAGE analysis of the samples revealed that none of the proteins was degraded during the incubation period (Supplemental Fig. S3). Fibril formation of NM can be accelerated by rotation (60 rpm), increasing turbulence and surface area (55)
. As expected, rotating accelerated NM fibril formation (data not shown) and, interestingly, also NM-PrP was capable of assembling amyloid structures under these conditions (Fig. 8B
). PrP-M did not form amyloid structures at all, as confirmed by AFM (Fig. 8D
). Thus, the fact that both NM and NM-PrP assemble into amyloid fibrils in vitro but exhibit strikingly different aggregation propensities in vivo confirms that factors or environmental conditions in mammalian cells influence aggregation of yeast and mammalian prion proteins.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The extraordinary hallmark that sets prions apart is that they are infectious in the case of mammalian prion diseases or inheritable in the case of yeast prions. This discrimination between typical amyloid proteins and prions, however, has recently been challenged by findings that injection of protein aggregates can at least accelerate aggregation of homologous proteins in some animal models of amyloid diseases (reviewed in ref. 57
). Whether protein aggregates produced in our system are prion-like concerning transmissibility is so far unknown.
Interestingly, Sup35p-NM alone was virtually incapable of acquiring a detectable aggregated state when transiently expressed in the cytosol of mammalian cells. In yeast, de novo appearance of [PSI+] and aggregate formation requires the presence of another prion-like element, termed [PIN+] (19
, 58
, 59)
. Under some circumstances, however, [PIN+] appears to be dispensable for induction of the [PSI+] phenotype, for example, if a short N-terminal fragment of Sup35p is overexpressed (60
, 61)
or if other Q/N-rich proteins or polyglutamine moieties function as seeds (62
, 63)
. Evidence that protein deposits can also coaggregate with heterologous proteins in mammalian cells is based on studies with diverse aggregation-prone proteins (64)
. In light of these findings, factors that are capable of inducing the rapid aggregation of NM might be absent in the cytosol of N2a cells under normal cell culture conditions.
Misfolded and aggregated proteins are a major threat to cell function and viability and key players in many neurodegenerative diseases and other proteinopathies (57
, 65)
. Consequently, cells are equipped with several defense mechanisms against potentially toxic aggregation-prone proteins. Molecular chaperones assist proper folding of nascent proteins, while proteins that are unable to fold are degraded by the ubiquitin-proteasome system. Misfolded proteins that escape the surveillance of these two systems can eventually be sequestered into aggresomes and/or degraded by autophagy (reviewed in ref. 50
). Conflicting results exist concerning aggresome formation of PrP. In prion-infected cell cultures, PrPSc appeared to be sequestered into aggresomes under mild proteasome impairment (66)
. In the same study, however, nondisease-associated, normal PrP present in the cytosol remained soluble. By contrast, it has been reported that transient overexpression of cytosolic PrP leads to the sequestration of misfolded PrPcyto into juxtanuclear aggresome-like structures (47)
. Furthermore, mutant PrPs appeared to accumulate in the cytosol in response to proteasomal inhibition and assemble as classical aggresomes (67)
. In our hands, PrPcyto was neither consistently located in close proximity to the nucleus nor did it colocalize with -tubulin or lead to a collapse of vimentin-intermediate filaments. These findings argue that the observed aggregates in our studies were not aggresomes. Furthermore, it has been suggested that aggresome formation might facilitate autophagosomal degradation of aggregation-prone proteins (50)
. Autophagosomes capturing misfolded protein subsequently mature into lysosomes. However, no costaining of any of the visible aggregates of recombinant proteins was observed with Lamp-1, suggesting that protein aggregates were not destined for proteolytical clearance in lysosomes.
Several lines of evidence argue that aggregate induction of the fusion proteins is mediated by the globular domain of PrP (47)
comprising aa residues 90–230. First, PrP90–230 is readily aggregating. Second, fusion of N, M, or NM to this PrP region rendered aggregation-prone molecules that readily formed visible deposits in mammalian cells, while NM alone appeared to be incapable of forming detectable aggregates, at least during transient expression. Third, replacement of the N region with the amino-terminal part of PrP (aa 23–120) in PrP-M did not lead to a drastic increase in the tendency to spontaneously aggregate. Interestingly, the amino-terminal region of PrP appeared to decrease the nucleation rate of the globular domain of PrP, as deletion of this region leading to PrP90–230 revealed a significant increase in cells displaying visible aggregates (compare Figs. 2C
and 6C
). Thus, PrP23–90 leads to a lower rate of nucleation. So far, no overt cytotoxic effect has been observed for N-PrP, M-PrP, and NM-PrP (data not shown). Further studies will be necessary to elucidate whether replacement of the amino-terminal part of PrP with Sup35p regions can modulate toxicity of cytosolic PrP.
The facts that fusion of N, M, or NM to PrP90–230 influenced frequency of aggregate appearance and/or aggregate size indicate that these fragments could still exert a modulating activity on PrP aggregation. Expression of M- and NM-PrP led to the appearance of large, dotlike or ring-shaped aggregates, suggesting a modulating role of the M region for aggregate size. Thus, M appears to influence the seeding activity of existing nuclei, potentially by facilitating incorporation of soluble chimeric protein into the growing aggregates. A similar phenomenon of ring-shaped and dotlike NM aggregates has previously been reported for yeast overexpressing either NM or wild-type Sup35p (46)
. Experiments in yeast using Sup35p with a deleted middle region demonstrate that this Sup35p variant is mostly insoluble and aggregated (23)
, suggesting a solubilizing tendency of the middle region for Sup35p. The minor resistance of NM-PrP and M-PrP to PK might also be due to the solubilizing effect of Sup35p M region, which could render the protein more sensitive to proteolysis (23)
.
M also lowered the nucleation rate of PrP90–230. The ratio of M-PrP foci-bearing cells to transfected cells was significantly decreased compared to the ratio of PrP90–230 aggregate-containing cells to transfected cells (30±3 vs. 45±5%, respectively) (Fig. 6C
). Furthermore, fewer aggregates per individual cell were detected for M-PrP compared to PrP90–230 (Fig. 6B
). Thus, M exhibited a solubilizing effect, decreasing the rate of nucleation, like the amino-terminal region of PrP, suggesting that both regions share at least some functional role in prion protein polymerization. By contrast, N only marginally influenced the nucleation rate of PrP90–230, as aggregate frequency of PrP90–230, and N-PrP was comparable (45±5 vs. 50±2%, respectively) (Figs. 2C
and 6C)
. A precise determination of the influence of N could not be made in this study, as the very high numbers of aggregates in cells expressing PrP90–230 and N-PrP precluded quantitative analysis. Nevertheless, N was capable of modulating aggregate appearance by counteracting the solubilizing effect of M. When expressed in N2a cells, the ratio of cells harboring visible NM-PrP foci to total transfected cells significantly increased compared to the ratio of M-PrP aggregate-containing cells to total transfected cells (55±9 vs. 30±3%, respectively) (Figs. 2C
and 6C)
. Furthermore, fusion of N to M-PrP increased the number of aggregates per cell, albeit in a cell-type-dependent manner (Fig. 7)
. Thus, a dynamic interaction of N and M is at least partially dependent on cell-specific conditions or factors and modulates rate of nucleation and seeding capacity. The interplay between N and M described here is likely to be important for the [PSI+] phenotype in yeast that is crucially dependent on the dynamics of nucleus formation and aggregate growth for prion inheritance.
Overall, our results suggest that aggregation of PrPcyto, PrP90–230, N-PrP, M-PrP, and NM-PrP in the cytosol of mammalian cells is a spontaneous event independent of an active cellular sequestration of aggregation-prone proteins. The fact that in vitro fibrillization experiments demonstrated that both NM and NM-PrP can form amyloid fibrils, suggests that the cellular environment strongly influences the aggregation capacities of both proteins. Several possible reasons may account for the finding that PrPcyto, PrP90–230, N-PrP, M-PrP, and NM-PrP readily accumulated in the cytosol of mammalian cells, while NM-HA and PrP-M did not (Fig. 9
B). The driving force of aggregation appears to be mainly governed by the globular domain of PrP but can be modulated by the interplay of the Sup35p N and M regions (Fig. 9A
). Either the physiological conditions in the mammalian cytosol per se (e.g., pH) or a PrP90–230-specific aggregation cofactor might promote aggregation of PrPcyto, PrP90–230, N-PrP, M-PrP, and NM-PrP, while the same conditions or factors leave NM and the amino-terminal region of PrP relatively unaffected (Fig. 9B
). Alternatively, in mammalian cells, a seed composed of aggregated heterologous protein might exist that can specifically cross-seed cytosolic PrP90–230. Further investigations are necessary to define the exact regions in PrP for aggregation, as well as to identify potential interaction molecules in the cytosol.
|
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 16, 2007. Accepted for publication September 13, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Krammer, D. Kryndushkin, M. H. Suhre, E. Kremmer, A. Hofmann, A. Pfeifer, T. Scheibel, R. B. Wickner, H. M. Schatzl, and I. Vorberg From the Cover: The yeast Sup35NM domain propagates as a prion in mammalian cells PNAS, January 13, 2009; 106(2): 462 - 467. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
