|Home | About | Journals | Submit | Contact Us | Français|
The central pathogenic event of prion disease is the conformational conversion of a host protein, PrPC, into a pathogenic isoform, PrPSc. We previously showed that the protein misfolding cyclic amplification (PMCA) technique can be used to form infectious prion molecules de novo from purified native PrPC molecules in an autocatalytic process requiring accessory polyanions (Deleault, N. R., Harris, B. T., Rees, J. R., and Supattapone, S. (2007) Proc. Natl. Acad. Sci. U. S. A. 104,9741–9746). Here we investigated the molecular mechanism by which polyanionic molecules facilitate infectious prion formation in vitro. In a PMCA reaction lacking PrPSc template seed, synthetic poly(A) RNA molecules induce hamster (Ha)PrPC to adopt a protease-sensitive, detergent-insoluble conformation reactive against antibodies specific for PrPSc. During PMCA, labeled nucleic acids form nuclease-resistant complexes with HaPrP molecules. Strikingly, purified HaPrPC molecules subjected to PMCA selectively incorporate an ~1–2.5-kb subset of [32P]poly(A) RNA molecules from a heterogeneous mixture ranging in size from ~0.1 to >6 kb. Neuropathological analysis of scrapie-infected hamsters using the fluorescent dye acridine orange revealed that RNA molecules co-localize with large extracellular HaPrP aggregates. These findings suggest that polyanionic molecules such as RNA may become selectively incorporated into stable complexes with PrP molecules during the formation of native hamster prions.
Prion diseases are fatal neurodegenerative disorders that occur in inherited, sporadic, and infectious forms (2). In both the sporadic and infectious forms of prion disease, the host-encoded prion protein (PrPC)2 appears to undergo conformational change into a pathogenic isoform (PrPSc), and it has been proposed that prions are unorthodox infectious agents composed of only PrPSc molecules (3). Currently, the molecular mechanism by which native PrPC molecules are converted into PrPSc molecules remains unknown.
A variety of genetic and biochemical studies has suggested that cofactors other than PrPC molecules may be required for prion formation (see Refs. 4–12 and reviewed in Ref. 13). Several classes of negatively charged macromolecules have been shown to bind to and/or influence PrP conformation in vitro, including nucleic acids (14–23), glycosaminoglycans (24–28), phospholipid-rich membranes (29–32), and chaperone proteins (33). However, it is not clear whether similar events occur during the formation of native prions because previously there has been no experimental system available that could produce wild type infectious prions from defined components. Our laboratory has recently used the protein misfolding cyclic amplification (PMCA) technique (34–37) to produce native hamster prions in vitro from purified hamster (Ha) PrPC molecules (1). Interestingly, addition of an accessory polyanion (such as synthetic poly(A) RNA) facilitated the formation of infectious prions in this chemically defined system, as confirmed by bioassays in wild type hamsters. Furthermore, using this polyanion-dependent system, we also demonstrated, for the first time, that wild type prion infectivity could be generated de novo (i.e. without infectious seed in a rigorously prion-free environment) (1). In this study, we characterize the molecular interactions between polyanions and purified HaPrPC substrate during PMCA-induced formation of hamster prions.
Synthetic poly(A), catalog number P9403 (0.2–6 kb by agarose gel electrophoresis), was purchased from Sigma. Synthetic 100-mer oligo(dA), -(dT), -(dC), and -(dG), 100-mer fluorescein-labeled oligo(dT) and oligo(dA), and 100-mer biotin-labeled oligo(dT) were purchased from Operon (Huntsville, AL). Branched oligo(dT) constructs were purchased from Fidelity Systems (Gaithersburg, MD). EasyTides [α-32P]ATP (catalog number BLU503H250UC) was obtained from PerkinElmer Life Sciences. Stock solutions of synthetic polynucleotides were dissolved in 1× TE, pH 8.0, and concentrations were confirmed by A260 prior to use. Dynabeads® M-450 rat anti-mouse IgM (catalog number 110-39D) and Pro-Long® Antifade (catalog number P7481) were obtained from Invitrogen. The anti-PrP 27–33 mAb was obtained from Dr. Peter Morganelli (VA Hospital, White River Junction, VT). The 15B3 mAb, 15B3 coating buffer, immunoprecipitation buffer, and IP wash buffer were obtained from Prionics AG (Schleiren-Zurich, Switzerland). The 89–112 and D13 mAbs were kindly provided by Dr. Anthony Williamson (La Jolla, CA). Benzonase® (catalog number 70664-3) was obtained from Novagen (Madison, WI). The bacterially expressed recombinant mouse PrP was kindly provided by Roland Riek (La Jolla, CA). Anti-actin antibody (catalog number A5060), polyadenylic acid-Sepharose (catalog number P2765), heparin-agarose (catalog number H3025), poly-l-glutamate (catalog number P4886), heparan sulfate (catalog number H7640), polynucleotide phosphorylase (catalog number P2869), acetylated bovine serum albumin (catalog number B8894), hexokinase (catalog number H4502), SP-agarose (catalog number S1799), deoxycholic acid, sodium salt (catalog number D6750), MOPS (catalog number M5789), MES (catalog number M5287), and imidazole (catalog number I-5513) were purchased from Sigma. CNBr-activated Sepharose 4 Fast Flow (catalog number 17-0981-01) was purchased from Amersham Biosciences. RNase-free water and stock solutions of NaCl, EDTA, TE, pH 8.0, and Tris, pH 8.0, were purchased from Ambion (Austin, TX). Other chemical reagents, including protein A-agarose (catalog number PI20333) were purchased at the highest grade available from Fisher. All phosphate-buffered saline (PBS) was without calcium or magnesium and purchased from Mediatech, Inc. (Herndon, VA). Teflon-printed slides (catalog number 63415-15) were obtained from Electron Microscopy Sciences (Hatfield, PA). The hamster scrapie strain Sc237 was kindly provided by Dr. Stanley Prusiner (San Francisco). Sc237 was originally isolated in inbred Syrian hamsters by Richard Marsh (Madison, WI) and subsequently passaged in outbred Syrian hamsters.
All procedures were performed at 4 °C. Six frozen brains from 8- to 12-week-old golden Syrian hamsters of either sex (Harlan Bioproducts, Indianapolis, IN) were suspended in ice-cold PBS plus Complete™ protease inhibitors (Roche Applied Science) to 40 ml total volume and homogenized using a Biospec Products (Bartlesville, OK) Biohomogenizer Mixer at 7000 rpm for 30–60 s. The homogenate was centrifuged at 3200 × g for 20 min, and the pellet was resuspended in a total volume of 40 ml of PBS, 1% deoxycholate, 1% Triton X-100, and Complete™ protease inhibitors using a Kontes (Vineland, NJ) glass Dounce homogenizer (10 strokes with pestle B). The homogenate was solubilized by incubation on ice for 30 min and then centrifuged at 100,000 × g for 30 min. The supernatant was filtered using Millipore (Billerica, MA) 0.2-µm Stericups and poured over a 1-ml ImmunoPure immobilized protein A column (Pierce) to pre-clear tissue-derived immunoglobulins. The flow-through was collected and passed over a Econ-Pac column (Bio-Rad) packed with 1 ml of protein A-agarose resin (Pierce) cross-linked to 3F4 antibody (Signet Laboratories, Dedham, MA) using DMP (Pierce). The column was washed with 15 ml of 20 mm Tris, pH 8.0, 0.5 m NaCl, 5 mm EDTA followed by 10 ml of PBS, 0.5% Triton X-100, and then eluted with 7.2 ml 0.2 m glycine, pH 2.5. The eluate was combined with 800 µl of 1 m Tris, pH 9.0 and applied to Zeba (Pierce) desalting spin columns pre-equilibrated in 20 mm MES, pH 6.4, 0.15 m NaCl, 1% Triton X-100. The resulting 8-ml buffer-exchanged sample was then passed over a 1.5-ml SP-Sepharose ion exchange column (Sigma), and the column was washed with 15 ml of 20 mm MOPS, pH 7.0, 0.25 m NaCl, 1% Triton X-100. The column was eluted with 8 ml of 20 mm MOPS, pH 7.0, 0.5 m NaCl, 1% Triton X-100 and applied to Zeba desalting spin columns pre-equilibrated with 20 mm MOPS, pH 7.5, 0.15 m NaCl, 0.5% Triton X-100 to reduce the salt concentration of the final product. The typical yield from a preparation using six hamster brains is ~ 16 µg of PrPC (~2 µg/ml), based on comparison with known concentrations of Escherichia coli-expressed recombinant PrP, using semi-quantitative Western blots and silver stain gels.
PMCA and serial dilution/propagation experiments were performed as described by Castilla et al. (35), using purified substrates instead of brain homogenates. To avoid cross-contamination, all work was carried out in laminar flow bio-safety cabinets using disposable surfaces and aerosol barrier tips. Typically, we prepared a mixture containing 200 µl of PrPC substrate diluted in 20 mm MOPS, pH 7.5, 0.15 m NaCl, 0.5% Triton X-100 (note that the appropriate dilution of PrPC substrate is empirically determined for each individual preparation using seeded, serial PrPSc propagation assays; the final concentration of PrPC in the mixture is typically between 250 and 500 ng/ml), 40 µl of imidazole diluted in 20 mm MOPS, 0.15 m NaCl, 0.5% Triton, pH 7.0, 4 µl of 0.5 m EDTA, pH 8.0, and 116 µl of TE buffer. From this mixture, multiple 90-µl aliquots were dispensed into 0.5-ml thin walled PCR tubes. Where indicated, the TE buffer contained a polyanionic compound such as poly(A) RNA or synthetic oligonucleotides at a concentration of 68.96 µg/ml (to produce a final concentration of 20 µg/ml in the reaction) except for polyglutamate, heparan sulfate, oligo(dT) construct 1, oligo(dT) construct 2, and oligo(dT) construct 3, which were used at final concentrations of 2 and 50 µg/ml and 800, 80, and 80 nm, respectively. Aliquots were then frozen at −70 °C until use in PMCA reactions. Serial propagation reactions were initiated on day 0 by the addition of 10 µl of PrPSc or PrP27–30 of PrP (or 10 µl of PrPC dilution buffer (20 mm MOPS, pH 7.5, 0.15 m NaCl, 0.5% Triton X-100) for unseeded reactions) to both day 0 and day 1 tubes. Day 0 samples were immediately refrozen without sonication, and day 1 samples were subjected to PMCA for 24 h using a Misonix 3000-MPD programmable sonicator equipped with a microplate horn containing 350 ml of water, set for 30-s bursts every 30 min at output ≤6.0. The appropriate output setting was determined empirically for each individual horn and changed with usage. Temperature was maintained by circulating heated water through a section of aluminum coil to heat the air inside the acoustic chamber (Misonix, Farmingdale, NY), and the microplate horn was covered with a sheet of plastic film to minimize evaporation. Before each burst, the bath temperature measured 41 °C, and after each burst the temperature measured 42 °C. Samples were mounted into a custom-made Plexiglas holder with the lid designed to keep the bottom of the PCR tubes ~3 mm from the horn surface, and the tubes were tightly closed. Samples were generally mounted no further than 4 cm from the center of the circular horn. With appropriate sonicator settings, the temperature inside a thin walled PCR tube filled with 250 µl of water subjected to a single 30-s pulse typically rose 1–3 °C (measured by thermistor readings). Immediately following the last sonication cycle of the first 24-h round, the day 1 PMCA product was removed from the sonicator and centrifuged for 5 s. Pipetting up and down several times resuspended the sample, and 10 µl was used to seed the thawed day 2 aliquot containing a fresh substrate mixture. The remainder of the day 1 sample was stored at −70 °C until analysis, and the day 2 sample was subjected to PMCA for 24 h in the second round of propagation. This process was repeated for multiple propagation rounds, up to 4 days.
All protease-digested (+PK) samples were incubated with 50 µg/ml proteinase K (Roche Applied Science) for 1 h at 37 °C. An equal volume of 2× SDS sample buffer was then added, and samples were boiled for 10 min at 95 °C. SDS-PAGE was performed on 1.5-mm 12% polyacrylamide gels with an acrylamide:bisacrylamide ratio of 29:1 (Bio-Rad). Following electrophoresis, the proteins were transferred to a methanol-charged, buffer-equilibrated polyvinylidene difluoride membrane (Millipore) using a Trans-blot SD semi-dry transfer cell (Bio-Rad) set at 2 mA/cm2 for 30 min.
To visualize PrP signals, Western blot membranes were first treated with 3 m GdnSCN (Roche Applied Science) at room temperature for 30 min. The membranes were then rinsed with TBST (10 mm Tris, pH 7.2, 150 mm NaCl, 0.1% Tween 20) and blocked for 1 h in Hood skim milk (Chelsea, MA) buffered with TBST. The blocked membrane was incubated overnight at 4°C with 3F4, 6D11 or 27–33 mAbs (Signet) diluted 1:5000 or 1:10,000, respectively, in TBST. Following this incubation, the membrane was washed three times for 10 min in TBST and incubated for 1 h at 4 °C with horseradish peroxidase-labeled anti-mouse IgG secondary antibody conjugate (GE Healthcare) diluted 1:5000 in TBST. The membrane was washed again four times for 10 min with TBST. The blot was developed using West Pico or West Femto (Pierce) chemiluminescence substrate, sealed in plastic covers, and either exposed to Fujifilm (Tokyo, Japan) Super RX film or captured digitally using a Fuji (Fujifilm) LAS-3000 chemiluminescence documentation system. Exposed films were developed automatically in a Kodak M35A X-Omat film processor, and digital images were captured using Image Reader version 2.0 and analyzed with Image Gauge version 4.22 software (Fujifilm, Tokyo, Japan). Relative molecular masses were based on migration of pre-stained standards from either Fermentas (Hanover, MD) or Bio-Rad.
Crude hamster and mouse brain homogenates were prepared using a Teflon/glass Potter homogenizer. Brains were homogenized in cold PBS plus Complete™ protease inhibitor mixture to a final brain concentration of 10% (w/v). Crude homogenates were centrifuged for 30 s at 200 × g, and the post-nuclear supernatants were diluted in PBS plus 1% n-octyl β-d-glycopyranoside, 5 mm EDTA to a 1% final brain concentration. Supernatants were solubilized by incubating at 4 °C for 1 h and then centrifuged for 30 min at 20,800 × g at 4°C. 300 µl of solubilized supernatants were incubated with 50 µl of packed polyadenylic acid-Sepharose, heparin-agarose, or polyglutamate-Sepharose (coupled to CNBr-activated Sepharose according to the manufacturer’s instructions) overnight at 4 °C with tilt rotation. Following incubation, the samples were subjected to brief centrifugation, and the sample supernatants were removed. The Sepharose/agarose pellets were washed with PBS plus 1% n-octyl β-d-glycopyranoside, 5 mm EDTA. Pellet-associated proteins were eluted by boiling the pellet with 50 µl of SDS loading buffer. Samples were subjected to SDS-PAGE and Western blot detection with anti-PrP mAb 27–33. Subsequently, each blot was stripped in 3 m guanidine isothiocyanate, 20 mm Tris, pH 8.0, for 1 h at room temperature and re-probed with anti-actin antibody.
Seeded and unseeded 100-µl PMCA reactions were performed in duplicate for each condition. Before addition to each sample, the PrPC substrate was subjected to 100,000 × g centrifugation for 45 min at 4 °C, and the supernatant fraction was then used as the source of PrPC for each sample. After one round of PMCA, 1 ml of PBS, 0.5% Triton X-100 was added to all samples that were then subjected to 100,000 × g centrifugation for 1 h at 4 °C. The sample supernatants were removed; the pellets were washed with 1 ml of PBS, 0.5% Triton X-100, and the centrifugation step was repeated. All but 100 µl of the sample supernatants were removed by aspiration, and the pellets were resuspended by vortexing. An equal volume of 2× SDS loading buffer was added, and each sample was boiled for 10 min at 95 °C. Samples containing the input amounts of PrPC and PrPSc added to the experimental samples were prepared at the same time, and all samples were analyzed together by Western blot with 3F4 mAb.
100-µl PMCA reactions were carried out in duplicate for each condition. Seeded reactions contained PrPSc template generated in a serial PMCA propagation reaction stimulated with synthetic poly(A) RNA. After overnight PMCA, 400 µl of IP buffer (TBS, 3% Tween 20, 3% Nonidet P-40) was added to each sample. 1.5 µg of 89–112 mAb was added to each sample for a final concentration of 3 µg/ml, and the samples were incubated for 2 h at room temperature with slow tilt rotation. 60 µl of a 50% slurry of Immunopure Protein A Plus (Pierce) was added to each sample and allowed to incubate for 40 min at room temperature with slow tilt rotation. Each sample was then centrifuged for 30 s, and the supernatant was removed by aspiration. 750 µl of wash buffer (TBS, 2% Tween 20, 2% Nonidet P-40) was added to each sample that was then mixed briefly by vortexing. The samples were centrifuged again for 30 s, and the wash buffer was removed by aspiration. Each sample was washed three more times in the same manner followed by a final wash with PBS. After aspiration of the PBS, 50 µl of 1× SDS loading buffer was added, and each sample was boiled for 10 min at 95 °C. All samples were analyzed together by Western blot with 3F4 mAb.
Rat anti-mouse IgM Dynabeads® were coated with 15B3 mAb at a concentration of 2 µg of antibody per 100 µl of beads. Beads were incubated with 15B3 in coating buffer for 2 h at room temperature with slow tilt rotation. 15B3-coated beads were washed three times with coating buffer and then resuspended in a volume of coating buffer equivalent to the starting volume of beads. A Dynal MPC™-S Magnetic Particle Concentrator (Invitrogen) was used for all steps involved in 15B3 bead preparation and PrP immunoprecipitation.
For PrP immunoprecipitation, 100 µl of PMCA reactions were carried out for each condition. Seeded samples contained PrPSc template generated in a serial PMCA propagation reaction stimulated with synthetic poly(A) RNA. Each sample was combined with 400 µl of IP buffer and 15 µl of 15B3-coated beads. Samples were immunoprecipitated for 2 h at room temperature with slow tilt rotation. The beads in each sample were washed three times with 1 ml of wash buffer. 30 µl of 2× SDS loading buffer was added to the beads in each sample, which were then boiled for 5 min at 95 °C. All samples were analyzed together by Western blot with 6D11 mAb.
For each sample, a seeded or unseeded 200-µl PMCA reaction was carried out. The polyanion added to the reactions was a synthetic (dT)100 or (dA)100 labeled with a fluorescein group at the 5′ end of the oligonucleotide. The stock oligonucleotide was prepared in 1× TE, pH 8.0, and added to each reaction at a final concentration of 20 µg/ml. After one round of PMCA, each sample was combined with 800 µl of PBS and subjected to 100,000 × g centrifugation for 45 min at 4 °C. All but ~100 µl of the sample supernatants was removed. At this point, MgCl2 was added to a final concentration of 2 mm along with 50 units of Benzonase to those samples undergoing nuclease digestion. Benzonase -treated samples were then incubated for 30 min at 37 °C with shaking at 300 rpm followed by addition of EDTA to a final concentration of 15 mm to inactivate the Benzonase. Following the first centrifugation step, and the incubation step for Benzonase-treated samples, all samples were resuspended in PBS up to a final volume of 1 ml and subjected to 100,000 × g centrifugation for 45 min at 4 °C. All but ~ 150 µl of the sample supernatants were removed. Samples treated with Benzonase were subjected to an additional centrifugation and wash step. The sample pellets were then resuspended by vortexing and applied to 15-mm, single well, Teflon-printed slides that were pre-coated with gelatin subbing solution. To prepare gelatin subbing solution, 6 g of 300 bloom gelatin was dissolved at 58 °C in 500 ml of sterile distilled water. 0.2 g of chromium potassium sulfate was added to the solution after the gelatin dissolved. Gelatin subbing solution cooled to 50 °C was applied to the slides in a volume large enough to cover the entire well and then allowed to dry for at least 2 h at room temperature. Samples were incubated on the slides overnight in a humidified chamber at 4 °C. The following day, the samples were fixed to the slide by addition of 150 µl of PBS, 8% formaldehyde. The samples were allowed to incubate at room temperature for 15 min and then removed by aspiration. Each slide well was washed one time with 150 µl of PBS. 150 µl of mAb D13 diluted 1:250 in PBS + 4% BSA was applied to each slide well and allowed to incubate overnight at 4 °C or for 2 h at room temperature in a humidified chamber. After primary antibody incubation, the slide well was washed three times with 150 µl of PBS + 4% BSA and allowed to incubate for 15 min at room temperature during each wash. 150 µl of secondary antibody, sheep anti-mouse conjugated with Alexa Fluor 568 (Invitrogen), diluted 1:250 in PBS + 4% BSA, was applied to each slide and allowed to incubate for 2 h in the dark at room temperature in a humidified chamber. The slide well was then washed three times with 150 µl of PBS + 4% BSA and allowed to incubate for 15 min at room temperature during each wash. 15 µl of ProLong® antifade solution was added to each sample and number 1.5, 18-mm square glass coverslips (Corning Glass) were mounted on each slide and allowed to dry overnight in the dark in a desiccating chamber. We also prepared and analyzed a control slide coated with only reaction buffer and then stained with D13 mAb and Alexa Fluor 568 secondary antibody to rule out nonspecific binding of either antibody to the slide. A slide coated with only fluorescein-labeled (dT)100 in reaction buffer was prepared to rule out nonspecific binding of the fluorescein-labeled (dT)100 to the slide. Finally, a slide coated with PrPC in reaction buffer that did not contain fluorescein-labeled (dT)100 was prepared to rule out autofluorescence of PrPC in the fluorescein excitation wavelength. Samples were analyzed by fluorescence microscopy using a Leica Confocal model TES SP confocal scanning microscope (Leica Microsystems, Heidelberg, Germany). All images were captured digitally using a 63×, 1.32 aperture, oil immersion objective lens, and Leica confocal system software, version 2.61. Image resolutions are 1024 × 1024 and are averages of three consecutive scans. Alexa Fluor 568 is displayed as magenta to be red-green color-blind-compatible. Images were processed (cropped and re-sized) using Photoshop software.
Radiolabeled poly(A) RNA was synthesized in a 100-µl reaction containing 100 mm Tris-HCl, pH 9.0, 0.4 mm EDTA, pH 7.0, 5 mm MgCl2, 6 mm ATP, 70 mm glucose, 0.1 mg/ml acetylated bovine serum albumin, 2.5 units/ml polynucleotide phosphorylase, 20 units/ml hexokinase, and 2 mCi/ml of 3000 Ci/mmol [α-32P]ATP. All components of the reaction, except polynucleotide phosphorylase, hexokinase, and [α-32P]ATP, were combined and pre-warmed to 37 °C. The enzymes and [α-32P]ATP were then added, and the reaction mixture was incubated at 37 °C for 45 min. Following incubation, the synthesized poly(A) RNA was purified using an RNeasy kit (Qiagen, Valencia, CA) following the manufacturer’s instructions. The concentration of the purified poly(A) RNA was determined by measuring A260
For each condition 300-µl PMCA reactions were prepared on ice. Samples without PrPC contained PrPC dilution buffer (20 mm MOPS, pH 7.5, 0.15 m NaCl, 0.5% Triton X-100) in the place of PrPC. [32P]Poly(A) RNA diluted into 1× TE, pH 8.0, was added to each sample for a final concentration of 20 µg/ml. Samples not subjected to PMCA were frozen at −70 °C. The remaining samples were subjected to PMCA for 24 h. After 24 h the samples subjected to PMCA were frozen at −70 °C for 30 min and then all samples were thawed on ice. 700 µl of PBS, 0.5% Triton X-100 was added to each sample, which was then mixed briefly by vortexing and subjected to 100,000 × g centrifugation for 45 min at 4 °C. All but ~50 µl of the sample supernatants were removed. The sample pellets were washed by addition of 950 µl of PBS, 0.5% Triton X-100, briefly mixed by vortexing, followed by 100,000 × g centrifugation for 45 min at 4 °C. All but ~50 µl of the sample supernatants were removed. 100 µl of Benzonase digestion buffer (50 mm Tris-HCl, pH 9.0, 20 mm NaCl, 2 mm MgCl2) was added to each sample, and the pellets were resuspended by pipetting. 2 µl of 25 units/µl Benzonase was added to nuclease-treated samples. All samples were then incubated at 37 °C for 45 min with 300 rpm shaking (Eppendorf Thermo-mixer, Fisher). 50 µl of 100 mm EDTA, pH 8.0, and 900 µl of PBS, 0.5% Triton X-100 were added to all samples. Each sample was then mixed briefly by vortexing and subjected to 100,000 × g centrifugation for 45 min at 4 °C. All but ~50 µl of the sample supernatants were removed, and 950 µl of PBS, 0.5% Triton X-100 was added to each sample pellet. Each sample was then mixed briefly by vortexing, and subjected to 100,000 × g centrifugation for 45 min at 4 °C. All but ~20 µl of the sample supernatants were removed. An equal volume of denaturation buffer (8 m urea, 10 mm Tris phosphate, pH 8.0, 0.5% Triton X-100) and 2 µl of 10× loading buffer (0.25% bromphenol blue, 50% glycerol) was added to each sample. An input control [32P]poly(A) RNA sample was prepared at the same time, and all samples were heated at 70 °C for 10 min. Denatured samples were subjected to electrophoresis on a 1.2% 1× TBE-agarose gel and then transferred to Biodyne B nylon membrane (Pierce) using 20× SSC transfer buffer and Whatman 3MM chromatography paper (Fisher). After overnight transfer (~17 h) the membrane was exposed to Biomax MS film (Eastman Kodak Co.) with a Hyperscreen intensifier screen (Amersham Bio-sciences). To calculate the stoichiometry of the interaction between nuclease-protected [32P]poly(A) RNA and PrPC, we incubated [32P]poly(A) RNA with PrPC at 37 °C overnight. The sample was then subjected to centrifugation, and the detergent-insoluble fraction was treated with Benzonase. The sample was washed and centrifuged again to re-isolate the insoluble fraction. This fraction was denatured and subjected to agarose gel electrophoresis, transferred to a nylon membrane, and exposed to x-ray film.
PrPSc derived from the fourth round of a seeded PMCA propagation reaction was diluted 1:10 into PBS, 2% Triton X-100. Brain-derived 237 PrP27–30 was prepared as described previously and also diluted 1:10 into PBS, 2% Triton X-100 (9). 20 µl of each PrPSc dilution was combined with 180 µl of PBS, 2% Triton X-100, 4 µl of 100 mm MgCl2, and either 25 or 10 µl of 25 units/µl Benzonase was added to the PrP27–30 preparation and propagated PrPSc samples, respectively. Identical mock-digested samples were prepared except PBS was added in the place of Benzonase. Samples were incubated at 37 °C for 30 min with shaking at 300 rpm. EDTA was added to each sample to a final concentration of 15 mm to inactivate the Benzonase, and samples were subjected to 100,000 × g centrifugation for 1 h at 4 °C. After centrifugation, all but ~15 µl of the sample supernatants were removed by aspiration. Sample pellets were resuspended in 185 µl of PBS, 2% Triton X-100 for a final volume of 200 µl and then sonicated for 30 s in a Misonix 3000-MPD programmable sonicator equipped with a microplate horn containing 350 ml of water and set for output ≤6.0. 10 µl of each PrPSc preparation was used to seed 100-µl PMCA propagation reactions.
To study in situ co-localization of PrP and polyanions, scrapie-infected (n = 3) and wild type (n = 3) hamster brains were fixed in 10% formalin. Tissue sections were pretreated for antigen retrieval using established protocols (75, 82). Briefly, tissue sections were deparaffinized in xylene followed by a 100, 95, 75, and 50% ethanol series. After a water rinse, sections underwent the following: 1) hydrated autoclaving at 121 °C for 10 min; 2) microwave boiling in 10 mm citric acid buffer, pH 6.0, for 7 min; and 3) 88% formic acid for 7 min. Tissue sections stained with acridine orange were first washed with PBS and then stained with PBS, pH 6.0, 0.1% acridine orange (Molecular Probes, Invitrogen) for 15 min, differentiated in 0.1 m CaCl2, and washed with water. For PrP immuno-histochemistry, tissue sections were blocked with 10% normal goat serum in PBS for 30 min at room temperature and incubated overnight at 4 °C with a 1:50 dilution of 3F4 anti-PrP primary mouse IgG (Covance Research Products, Inc). Slides were then washed with PBS and incubated for 2 h at 37 °C with a 1:200 dilution of green fluorescent Alexa Fluor 488-labeled secondary anti-mouse IgG (Molecular Probes, Invitrogen). Samples digested with RNase (Roche Applied Science) were incubated for 1 h at 37 °C with 600 µg/ml of RNase in PBS and then washed with PBS. Tissue section slides treated with heparinase III (Sigma) were digested with 300 µl/slide of 1.6 units/ml heparinase in PBS, 4 mm CaCl2 for 4.5 h at room temperature and then washed with PBS. RNase and heparinase treatment preceded immunohistochemistry, followed by acridine orange staining was performed last. After processing, all tissue sections were air-dried and mounted in Vectashield (Vector Laboratories) with 5 µg/ml 4′,6-diamidino-2-phenyl-indole (Sigma). Images were acquired with a Zeiss Axiophot fluorescence microscope with ×20 air or ×100 oil lenses and Texas Red and fluorescein isothiocyanate filter sets (Chroma). Images were captured with IPlab with manual control, equal and corresponding exposures, and adjusted and merged with ImageJ. Alexa Fluor 488 staining is pseudo-colored green, and acridine orange staining is displayed as magenta to be red-green color-blind-compatible.
To test for Benzonase activity in PMCA reactions containing oligo(dT), 50 ng of biotin-labeled (dT)100 was digested with Benzonase under the same conditions as used in Fig. 3. Control samples containing biotin-labeled (dT)100 not treated with Benzonase were also prepared. After treatment, all samples were transferred by slot blot to Biodyne B membrane (Pierce) pre-soaked in 6× SSC. The membrane was processed using the North2South chemiluminescent detection kit (Pierce) with a streptavidin-horseradish peroxidase secondary antibody. To test for Benzonase activity in PMCA reactions containing RNA, 50 ng (~2000 cpm) of synthetic [32P]poly(A) RNA was digested with Benzonase under the same conditions as used in Fig. 4 and supplemental Fig. S4A. A control sample also containing 50 ng (~2000 cpm) of synthetic [32P]poly(A) RNA was prepared but not digested with Benzonase. After treatment, all samples were transferred by slot blot to Biodyne B membrane (Pierce) pre-soaked in 20× SSC. The membrane was exposed to Kodak BioMax MS film.
Given the ability of PMCA to generate and propagate native infectious prions in vitro from a purified substrate (1), we used this technique to elucidate the mechanism of polyanion-stimulated conversion of HaPrPC to HaPrPSc. To prepare the substrate for these studies, we purified HaPrPC from normal hamster brain, and confirmed its purity by silver stain SDS-PAGE (supplemental Fig. S1). We also performed a serial PMCA propagation reaction to confirm that this preparation of purified HaPrPC is a competent substrate for conversion to HaPrPSc in autocatalytic reactions facilitated by several different polyanionic molecules, including poly(A) RNA, heparan sulfate, and polyglutamate, and some but not all branched single-stranded poly(dT) DNA compounds (supplemental Fig. S2).
In our first series of experiments, we examined the effect of a polyanion on the conformation of purified HaPrPC. One biochemical characteristic of PrPSc is that it is detergent-insoluble under nondenaturing conditions (38). To test whether a polyanion affects the detergent-solubility of HaPrPC, we carried out seeded and unseeded purified PMCA reactions (containing 0.35% Triton X-100), in the presence or absence of synthetic poly(A) RNA (Fig. 1A). Following one round of PMCA, each sample was subjected to ultracentrifugation, and the pellets were analyzed for the presence of detergent-insoluble HaPrP. As expected, detergent-insoluble HaPrP was formed in reactions seeded with HaPrPSc (Fig. 1A, lanes 5 and 6), and HaPrPC alone did not become detergent-insoluble after PMCA (Fig. 1A, lanes 1 and 2). However, detergent-insoluble HaPrPC was formed in unseeded reactions containing HaPrPC and poly(A) RNA without PrPSc template (Fig. 1A, lanes 3 and 4). Only ~25% of the input HaPrPC was recovered in the detergent-insoluble pellet, although this amount is equivalent to the PrP recovered in the reactions seeded with HaPrPSc (Fig. 1A, lanes 3–6 versus lane 7). A typical seeded PMCA reaction results in ~50–75% conversion of HaPrPC to protease-resistant HaPrP, so it is difficult to determine whether incomplete recovery during ultracentrifugation or incomplete conversion during PMCA accounts for this observation. The results of this study show that a polyanionic molecule can induce insolubility of HaPrPC in vitro, possibly by causing it to adopt an altered conformation.
To confirm that poly(A) RNA molecules alter HaPrP conformation, and to test whether the resulting conformation might further resemble that of HaPrPSc, we conducted an immunoprecipitation experiment using 89–112 mAb, a motif-grafted antibody that specifically recognizes the PrPSc conformation (Fig. 1B) (39). HaPrPSc formed in seeded PMCA reactions was successfully immunoprecipitated by 89–112 mAb (Fig. 1B, lanes 5 and 6). Immunoprecipitation of unseeded reactions that lacked poly(A) RNA did not recover HaPrPC (Fig. 1B, lanes 1 and 2). In contrast, unseeded reactions containing HaPrPC and poly(A) RNA resulted in the formation of a PrP isoform that was immunoprecipitated by 89–112 mAb (Fig. 1B, lanes 3 and 4). However, less PrP was immunoprecipitated from the unseeded reaction as compared with the seeded reaction (Fig. 1B, lanes 3 and 4 versus 5 and 6). We obtained qualitatively similar results with a different PrPSc-specific antibody, mAb 15B3, although the immunoprecipitation experiment using this antibody was complicated by background affinity for purified HaPrPC (Fig. 1C) (40). These results indicate that, during PMCA, poly(A) RNA molecules can induce HaPrPC to adopt a conformation that is reactive against antibodies specific for the PrPSc conformation. We also subjected parallel samples to digestion with 50 µg/ml proteinase K (PK) digestion for 1 h at 37 °C (Fig. 1D). This analysis revealed that only the reaction seeded with PrPSc resulted in the formation of protease-resistant HaPrP (Fig. 1D, lane 3). PrP formed during reactions containing only HaPrPC, or HaPrPC plus poly(A) RNA, was not resistant to protease digestion even though HaPrP generated in the latter reaction was reactive against antibodies specific for the PrPSc conformation (Fig. 1D, lanes 1 and 2). These findings indicate that a subset of HaPrP molecules formed during unseeded PMCA reactions containing poly(A) RNA acquire some, but not all, properties of PrPSc. Specifically, they are detergent-insoluble and reactive with antibodies specific for the PrPSc conformation, but they are protease-sensitive. In ongoing experiments, samples containing PrPC plus poly(A) RNA subjected to PMCA for 24 h have not caused disease >200 days after intracerebral inoculation into Syrian hamsters (0/6).
To determine the minimum polyanion length required to facilitate prion formation, we conducted serial PMCA propagation reactions with samples containing synthetic homopolymeric oligonucleotides of defined length (Fig. 2A). In our first experiment we tested the 100-mer polymers (dT)100, (dC)100, (dA)100, and (dG)100. We found that only the oligo-pyrimidines (dT)100 and (dC)100 facilitated prion propagation; (dA)100 and (dG)100 failed to support propagation (Fig. 2A). To determine more precisely the minimum oligonucleotide length required to facilitate serial PMCA propagation reactions, we carried out reactions containing synthetic oligo(dT) ranging in length from 10 to 90 bases (Fig. 2B). In a control reaction lacking oligo(dT), HaPrPSc was generated after the first round of PMCA, but subsequent rounds failed to propagate this HaPrPSc (Fig. 2B, bottom right panel). In contrast, oligo(dT) molecules between 40 and 90 bases in length consistently facilitated efficient propagation of HaPrPSc (Fig. 2B, left and right panels). Oligo(dT) between 10 and 30 bases in length failed to facilitate propagation (Fig. 2B, right panel). The formation of HaPrPSc in these latter samples during the first round of PMCA is attributable to residual nucleic acid molecules present in the HaPrPSc template used to seed the reaction. However, upon serial dilution in subsequent PMCA reactions, the concentration of these seed-associated nucleic acids becomes too low to facilitate the conversion process. These data show that the minimum sized oligo(dT) required for efficient and reproducible in vitro prion propagation is ~40 bases in length.
Our observation that poly(A) RNA molecules can induce a change in HaPrPC conformation led us to investigate whether polyanionic and HaPrPC molecules might form a physical complex during the course of PMCA reactions. For this purpose, we used a modified fluorescence microscopy technique originally developed to study the conformation of synthetic prion fibrils (41). In a pilot experiment, we first identified a fluorescently labeled polyanionic molecule (i.e. (dT)100 labeled with fluorescein at the 5′ end) that was able to facilitate HaPrPSc propagation in a PMCA reaction (Fig. 2A, bottom). We then performed a series of experiments in which PMCA reactions were supplemented with this fluorescein-labeled (dT)100 instead of poly(A) RNA, and the association between HaPrP and fluorescein labeled (dT)100 was analyzed by dual channel confocal microscopy (Fig. 3). In an unseeded PMCA reaction, fluorescein-labeled (dT)100 molecules co-localized with insoluble HaPrP aggregates, indicating that polyanionic molecules can physically associate with HaPrPC molecules during PMCA (Fig. 3, top panels). Fluorescein-labeled (dT)100 molecules also co-localized with the HaPrP aggregates in PMCA reactions seeded with HaPrPSc (Fig. 3, middle panels). In contrast, when PMCA reactions were supplemented with fluorescein-labeled (dA)100, a polyanion that does not support PrPSc propagation (Fig. 2A), fluorescein signals were only detected in a subset of large aggregates (supplemental Fig. S3).
Interestingly, when the products of PMCA reactions containing fluorescein-labeled (dT)100 molecules were treated with nuclease, the co-localized fluorescent signals were not abolished, indicating that fluorescein-labeled (dT) 100 associated with HaPrP is largely protected from nuclease digestion (Fig. 3, bottom panels). Under this condition, however, we also observed some immunostained HaPrP aggregates that lacked fluorescein signal (white arrows). It is possible that the fluorescein-labeled (dT)100 molecules associated with these particular HaPrP molecules were not protected from nuclease digestion, or that it was partially protected but the digested portion of the (dT)100 included the 5′-fluorescein group. We confirmed that Benzonase is active under the conditions used in this assay by digesting dilutions of (dT) 100 under identical digestion conditions (supplemental Fig. S4A, lane 1 versus 2). Taken together, these data indicate that fluorescein-labeled (dT)100 and HaPrP can form a physical complex during PMCA, which renders the nucleic acid largely inaccessible to the activity of a nuclease.
To test whether PrPC can physically interact with polyanionic molecules within a crude homogenate preparation (i.e. in the presence of competing factors), we performed binding assays using agarose-immobilized poly(A) RNA, heparan sulfate, and polyglutamate molecules. These experiments confirmed that both HaPrPC and MoPrPC bind to these polyanionic molecules in the context of a homogenate preparation (supplemental Fig. S5).
To examine more closely the process by which polyanionic molecules are incorporated into HaPrP aggregates, we developed a nuclease protection assay using radiolabeled poly(A) RNA and HaPrPC molecules. To radiolabel poly(A) RNA, we employed a coupled enzymatic synthesis reaction, which generates homopolymeric [32P]poly(A) RNA from a mixture of unlabeled ATP and a trace amount of [α-32P]ATP. This reaction yields a heterogeneous preparation of [32P]poly(A) RNA molecules ranging in length from ~0.1 to >6 kb (Fig. 4, lane 8), which is able to facilitate serial PMCA propagation (supplemental Fig. S2B). During synthesis, [32P]AMP is randomly incorporated into poly(A) strands such that each base has an equal chance of being radiolabeled. This method produces [32P]poly(A) RNA, which is similar but not identical to the commercially available nonradioactive preparations that we routinely use in our PMCA propagation reactions. Specifically, the radiolabeled poly(A) RNA contains some poly(A) strands of greater length than those present in the commercially available poly(A) RNA (data not shown). Although the commercial poly(A) RNA has a uniform size distribution, the radiolabeled poly(A) RNA we used in these experiments was predominantly ~0.7 kb in length, although a widely heterogeneous mixture of poly(A) lengths was present in the preparation (Fig. 4, lanes 7 and 8).
We first studied the effect of PMCA incubation on the interaction between HaPrPC and radiolabeled poly(A) RNA in the absence of a HaPrPSc template seed (Fig. 4). To address the mechanism by which poly(A) RNA renders HaPrPC detergent-insoluble during PMCA (Fig. 1A), we began by characterizing the poly(A) RNA molecules associated with this polyanion-induced insoluble isoform of HaPrP. It should be noted that for the sake of consistency we used the same concentration of [32P]poly(A) RNA molecules (20 µg/ml) in this experiment as we had used previously in our nonradioactive experiments, and therefore the samples contain a vast molar excess of [32P]poly(A) RNA relative to HaPrPC. When a reaction containing [32P]poly(A) RNA and HaPrPC was incubated for ~10 min at 4 °C, we found that [32P]poly(A) RNA is recovered in the pellet after centrifugation, presumably through association with insoluble HaPrPC (Fig. 4, lane 5). The majority of this [32P]poly(A) RNA was about ≤0.2 kb in length, which was smaller than the most abundant [32P]poly(A) RNA molecules present in the input [32P]poly(A) RNA, which are ~0.7 kb in length (Fig. 4, lane 5 compared with lane 7). To test whether the recovered [32P]poly(A) RNA was protected from nuclease digestion, we treated the sample with Benzonase. With this treatment, [32P]poly(A) RNA was not recovered, indicating that the RNA in this sample was not protected from nuclease digestion (Fig. 4, lane 6). We then prepared identical reactions, but we subjected them to PMCA incubation for 16 h at 42 °C (Fig. 4, lanes 3 and 4). Strikingly, in these samples the HaPrP-associated [32P]poly(A) RNA was predominantly ~1–2.5 kb in length, a size much larger than either the HaPrP-associated [32P]poly(A) RNA in samples not subjected to PMCA or the most abundant species of [32P]poly(A) RNA in the input [32P]poly(A) RNA sample (Fig. 4, lane 3 compared with lanes 5 and 7). This observation indicates that over the course of the incubation period, HaPrPC molecules selectively associate with [32P]poly(A) RNA molecules within a specific size range. Additional experiments showed that the shift in selected size also occurs in samples incubated at 37 °C for 16 h without sonication (data not shown). We then tested whether the HaPrP-associated [32P]poly(A) RNA was protected from nuclease digestion after PMCA incubation (Fig. 4, lane 4). Consistent with the results of the microscopy assay, the [32P]poly(A) RNA was partially protected from treatment with Benzonase, suggesting a relatively strong interaction between the [32P]poly(A) RNA and the insoluble HaPrP (Fig. 4, lane 4). Benzonase treatment shifted the size of the predominant [32P]poly(A) RNA species downward to a length between ~0.4 and 1 kb (Fig. 4, lane 4). This mobility shift indicates that the entire length of each [32P]poly(A) RNA molecule was not protected from nuclease digestion and that the Benzonase was active under the tested conditions. We confirmed that Benzonase is active under the conditions used in this assay by digesting a dilution of [32P]poly(A) RNA under identical conditions (supplemental Fig. S4B, lane 1 versus 2).
Densitometric analysis of the x-ray film using Photoshop software allowed us to quantify the amount of protected RNA by comparing it to a known quantity of the input [32P]poly(A) RNA run as an internal control (data not shown). The average size of the protected [32P]poly(A) RNA was estimated by comparing it to a radiolabeled RNA ladder on the same membrane. Using these values, we calculated the molar ratio of detergent-insoluble HaPrP to nuclease-protected [32P]poly(A) RNA to be ~160:1.
Using a serial propagation assay, we confirmed that the autocatalytic activity of HaPrPSc molecules generated in vitro from HaPrPC and poly(A) RNA substrates is not abolished by nuclease digestion (supplemental Fig. S6A). Similarly, nuclease treatment did not destroy the autocatalytic activity of brain-derived HaPrP27–30 molecules (supplemental Fig. S6B). We confirmed that Benzonase is active during treatment of the PrPSc by digesting a dilution of [32P]poly(A) RNA under identical digestion conditions (supplemental Fig. S4B, lane 1 versus 3).
To determine whether polyanions co-localize with HaPrP aggregates in prion disease brains, we conducted neuropathological analysis of scrapie-infected hamsters. First, we performed immunohistochemical staining for HaPrP and histochemical staining with acridine orange (AO), a monomeric cationic fluorescent dye (Fig. 5). HaPrP aggregates were readily detected in the extracellular space of sections from scrapie-infected hamster brains (Fig. 5, top left panel) but not in uninfected brains (supplemental Fig. S7). AO staining co-localized with staining for HaPrP aggregates, which were typically ~5–10 µm in size (Fig. 5, top panels). In some instances, we observed smaller HaPrP aggregates that did not stain with AO (Fig. 5, white arrowheads). AO has been primarily used for detection of RNA, but it can also stain proteoglycans (42–46). To determine the specificity of the AO staining in our samples, we treated brain sections with either RNase or heparinase prior to histochemical staining (Fig. 5). RNase treatment markedly reduced the AO staining associated with the HaPrP aggregates (Fig. 5, middle row). Presumably, the antigen-retrieval pretreatment steps (hydrated autoclaving, boiling in citric acid, and incubation in formic acid) render the majority of HaPrP-associated RNA molecules susceptible to nuclease digestion in these denatured tissue sections. In contrast, heparinase treatment did not significantly decrease the AO staining co-localized to the HaPrP aggregates (Fig. 5, bottom row). Taken together, these findings indicate that RNA molecules are associated with HaPrP aggregates in the brains of scrapie-infected hamsters.
In this study, we report that during the in vitro formation of infectious hamster prions, synthetic nucleic acid molecules induce purified native HaPrPC to adopt a conformation similar to HaPrPSc and, in doing so, become selectively incorporated into a complex that is resistant to nuclease digestion. In addition, we present evidence that RNA molecules co-localize with HaPrP aggregates in scrapie-infected hamster brains.
Folding intermediates have been implicated in the pathogenesis of many protein aggregation diseases (47). Therefore, it is important to identify the intermediate steps of PrP misfolding that lead to the formation of the infectious prion isoform, PrPSc. One recent study identified a detergent-insoluble PrP conformer present in normal human brain, which is partially resistant to PK digestion (48), whereas another study showed that a misfolded PrP isoform was produced by cultured cells subjected to endoplasmic reticulum stress (49). In vitro studies have demonstrated that under chemically destabilizing conditions, both native PrPC and bacterially expressed recombinant PrP (recPrP) can adopt intermediate conformations (50–58). Similarly, nucleic acids have been shown to induce recPrP to adopt β-sheet-rich and amyloid conformations (16, 18, 58). Although these PrP conformers may resemble misfolded precursors to PrPSc, the systems used to generate these conformers have not been shown to produce infectious prions. In this study, using the PMCA system, we found that polyanions can induce native HaPrPC to become detergent-insoluble and reactive against PrPSc-specific antibodies. However, this intermediate HaPrP isoform was not resistant to PK digestion, which indicates that additional conformational re-arrangements are required to generate HaPrPSc. Together, these findings support the hypothesis that polyanionic molecules can facilitate the misfolding of native HaPrPC during the formation of infectious hamster prions, even in the absence of template PrPSc.
Numerous studies have demonstrated that PrP can interact with nucleic acids in vitro (14–23, 59–66). In previous reports, affinity-selected, small highly structured synthetic (shs) RNA molecules were shown to interact in vitro with bacterially expressed recombinant human PrP (hrPrP) as well as brain-derived PrPC molecules (21, 61). In addition, a subset of these shsRNAs formed nucleoprotein complexes with hrPrP, which were partially resistant to RNase A and PK digestion (21). Other shsRNA molecules that lacked affinity for PrP failed to form a nucleoprotein complex or confer PK resistance to PrP (21). This indicates that the shsRNAs required specific structure and sequence, as well as affinity for PrP, for their functional interaction with PrP, an observation that differs from the findings presented in this study. In addition, none of these PrP-nucleic acid preparations have been shown to be infectious.
Previous studies by our group showed that polyanionic molecules facilitated the in vitro PMCA propagation of purified HaPrPSc, as well as the de novo generation of infectious hamster prions in the absence of template PrPSc (1). Utilizing labeled homopolymeric nucleic acids, we found that PrP molecules can form a stable nucleoprotein complex in this system, which protects the nucleic acid components from nuclease digestion. Notably, [32P]poly(A) RNA molecules only became nuclease-resistant when incubated with HaPrP overnight at 37 °C. This finding suggests that the formation of strong intermolecular interactions between polyanionic molecules and HaPrP molecules occurs in a time-dependent manner.
Currently, the molecular composition of native prions has not been fully determined. The most stringent interpretation of the “protein only” hypothesis would be that the infectious unit is composed solely of the misfolded protein, PrPSc (3). Although numerous studies strongly support the contention that PrPSc is the critical entity of infectious prions, the possibility that additional host-encoded component(s) might physically associate with PrPSc and influence its structure cannot yet be excluded. Further studies are required to determine whether native prions contain any essential host-encoded molecules other than PrPSc, such as polyanions, lipids, or metal ions.
Our finding that synthetic homopolymeric nucleic acids facilitate PMCA propagation of HaPrPSc indicates that there is no sequence-specific requirement for functional interaction with HaPrP. However, our finding that only homopolymeric pyrimidine molecules are able to facilitate HaPrPSc propagation indicates that the negatively charged phosphate backbone is not sufficient for the propagation of purified HaPrPSc in vitro. The larger relative size of purine bases may create steric hindrance that prevents the proper physical interaction with HaPrPC required to facilitate the formation of HaPrPSc during PMCA. Our studies with small oligonucleotides revealed that a 40-mer was the shortest length oligo(dT) able to facilitate prion propagation. Based on previous biophysical studies of linear DNA molecules, we estimate the length of linear (dT)40 to be ~28 nm (67). Using data from NMR structural studies of recPrP, we approximate the largest atomic dimension of a full-length HaPrP molecule to be ~7 nm and the smallest dimension to be ~3 nm (68). From these calculations, we estimate that 4–10 HaPrP molecules would be able to interact with a single (dT)40 molecule, depending on the physical arrangement of each HaPrP monomer. Shorter oligo(dT) molecules may not be able to physically accommodate enough molecules of HaPrP to promote the formation of HaPrPSc during PMCA. Using labeled poly(A) RNA we calculated a stoichiometric ratio of ~160:1 between HaPrP molecules and nuclease-protected poly(A) RNA molecules in the nucleoprotein complex formed during PMCA. We estimate that, on average, each nuclease-protected poly(A) RNA molecule could physically interact with ~ 100– 230 HaPrP molecules, depending on the HaPrP monomer arrangement. A stoichiometry of 160:1 falls within this range; each poly(A) RNA molecule could physically bind 160 HaPrP molecules to generate the nucleoprotein complex during PMCA.
Several models could explain how polyanionic molecules promote prion conversion during PMCA. As proposed previously, polyanionic molecules may facilitate prion formation by acting as a physical scaffold that binds to, and promotes intermolecular interaction between, PrP molecules (9,16,24). In the absence of a polyanion, the interaction between template PrPSc molecules and PrPC molecules may occur at a rate too slow to allow for efficient conversion of PrPC to PrPSc during PMCA. In addition, a polyanionic molecule could facilitate prion formation by destabilizing the conformation of PrPC, thereby lowering the energy barrier that must be overcome to generate the infectious isoform PrPSc. Our findings support both of these roles for polyanionic molecules during prion formation. Our data also suggest that the interaction between HaPrP and polyanions is not transient in nature; the polyanion does not appear to dissociate from HaPrP after the conformational conversion of HaPrP into either an insoluble intermediate isoform or HaPrPSc.
During PMCA, we found that HaPrP molecules preferentially associate with poly(A) RNA molecules of specific length. Without PMCA, HaPrP associated with relatively small poly(A) RNA molecules, but this interaction did not protect the RNA from nuclease digestion. These findings suggest a model whereby HaPrP and polyanionic molecules activelyformanuclease-resistantcomplexinatime-and PMCA-dependent process (Fig. 6). PrP molecules may initially associate with shorter poly(A) RNA molecules because of favorable steric effects. However, if these poly(A) RNA molecules are too short to serve as scaffolds for PrP complex assembly, PrP molecules could dissociate and eventually bind with larger poly(A) RNA molecules, leading to the irreversible formation of nuclease-resistant nucleoprotein complexes. Over time, this process would selectively recruit the majority of PrP molecules available into complexes with larger poly(A) RNA molecules.
Prion strains are infectious self-propagating variants possessing distinct clinical and neuropathological phenotypes (69, 70). At present, the molecular mechanism responsible for the existence of different prion strains is unknown. Biophysical studies indicate that fibrils composed of only PrP polypeptide can propagate specific conformations and that the primary amino acid sequence of PrP restricts the number of possible conformations that such fibrils can adopt (71), and elegant studies have shown that prion strain phenotypes in yeast can be completely encoded by the structure of purified proteins (72– 74). However, it is possible that an accessory molecule could influence the cell type-specific strain properties (such as selective neurotropism) of mammalian prions through interaction with PrP molecules. Specific prion strains may preferentially interact with a specific subset of endogenous polyanionic molecules expressed within susceptible host cells.
Our finding that HaPrP selectively interacts with a subset of poly(A) RNA molecules during PMCA supports the hypothesis that prion strains could result from selective interaction with endogenous polyanions in vivo. This “co-factor selection” hypothesis differs from the “co-prion” hypothesis previously proposed by Weissmann (76) in that (i) co-factor molecules would be recruited from pre-existing cellular pools and therefore not need to replicate upon infection, and (ii) nonreplicating polyanionic molecules such as proteoglycans could also serve as co-factors.
Previous studies have shown that cytoplasmic RNA molecules associate with neuro-fibrillary tangles and neuritic plaques in Alzheimer disease brains and Pick bodies in Pick disease brains (42, 77, 78). Similarly, heparan sulfate proteoglycans were found to co-localize with PrP plaques in prion disease brains (79). In this study, we used the cationic dye, acridine orange, to stain for polyanions associated with PrP aggregates in the brains of scrapie-infected hamsters. Interestingly, we found that acridine orange staining, which co-localized with HaPrP aggregates, was primarily due to the presence of RNA molecules. This finding is consistent with previous studies showing that nuclease-protected RNA molecules co-purify with prion infectivity (80, 81), and that endogenous RNA molecules can facilitate the conversion and propagation of hamster prions in vitro (1). It is possible that a diversity of polyanionic molecules, including RNA and heparan sulfate proteoglycans, protein chaperones, and negatively charged lipid membranes, may facilitate the process of conformational conversion during prion formation in vivo, and possibly even promote the differential formation of prion strains. However, further work is needed to identify comprehensively the specific polyanionic molecules associated with PrP aggregates in situ, and to determine fully their roles in the pathogenesis of prion disease.
The confocal microscope used for this study was supported in part by National Science Foundation Grant DBI-9970048 to R. D. Sloboda. We thank Alex Raeber, Francisza Kuhn, and Bruno Oesch (Prionics) for kindly providing the 15B3 immunoprecipitation kit, and Dr. Judy Rees for help editing the manuscript
Supplemental Material can be found at: http://www.jbc.org/cgi/content/full/M704447200/DC1
*This work was supported by the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
2The abbreviations used are: PrPC, cellular prion protein; PrPSc, scrapie PrP; HaPrP, hamster PrP; PMCA, protein misfolding cyclic amplification; mAb, monoclonal antibody; PBS, phosphate-buffered saline; BSA, bovine serum albumin; MES, 4-morpholineethanesulfonic acid; MOPS, 4-morpholinepro-panesulfonic acid; PK, proteinase K; shsRNA, small highly structured RNA; AO, acridine orange; recPrP, recombinant PrP; hrPrP, recombinant human PrP.