|
||||||||||||||||||
 |
From: TSS (216-119-138-171.ipset18.wt.net)
Nature 425, 717 - 720 (16 October 2003); RNA molecules stimulate prion protein conversion NATHAN R. DELEAULT, RALF W. LUCASSEN & SURACHAI SUPATTAPONE1 Department of Biochemistry, 7200 Vail Building, Dartmouth Medical School, Hanover, New Hampshire 03755, USA Correspondence and requests for materials should be addressed to S.S. (supattapone@dartmouth.edu). Much evidence supports the hypothesis that the infectious agents of prion diseases are devoid of nucleic acid, and instead are composed of a specific infectious protein1. This protein, PrPSc, seems to be generated by template-induced conformational change of a normally expressed glycoprotein, PrPC (ref. 2). Although numerous studies have established the conversion of PrPC to PrPSc as the central pathogenic event of prion disease, it is unknown whether cellular factors other than PrPC might be required to stimulate efficient PrPSc production. We investigated the biochemical amplification of protease-resistant PrPSc-like protein (PrPres) using a modified version3 of the protein-misfolding cyclic amplification method4. Here we report that stoichiometric transformation of PrPC to PrPres in vitro requires specific RNA molecules. Notably, whereas mammalian RNA preparations stimulate in vitro amplification of PrPres, RNA preparations from invertebrate species do not. Our findings suggest that host-encoded stimulatory RNA molecules may have a role in the pathogenesis of prion disease. They also provide a practical approach to improve the sensitivity of diagnostic techniques based on PrPres amplification. We previously showed that PrPres amplification in vitro shares many specific features with the pathogenic process of prion propagation in vivo, including strain and species specificity3. In a typical amplification reaction, diluted prion-infected brain homogenate (0.1% w/v) is mixed with either 5% (w/v) normal brain homogenate or buffer control and incubated overnight at 37 °C. Hamster Sc237 PrPres is amplified about sixfold under these conditions (Fig. 1a, compare M with Sc). While characterizing the biochemical requirements of PrPres amplification reactions, we were surprised to discover that treatment of such reactions with DNase-free, heterogeneous pancreatic RNase abolished PrPres amplification in a dose-dependent manner (Fig. 1a, first panel). In vitro PrPres amplification was also abolished by treatment with purified pancreatic RNase A, RNase T1, micrococcal nuclease, or benzonase (Fig. 1a). In a control experiment, we found that addition of RNase A for 1 h after overnight incubation did not reduce the recovery of PrPres already amplified (Supplementary Fig. S1). Figure 1 Effect of various enzymes on PrPres amplification. Full legend By contrast, PrPres amplification was not affected by addition of RNase V1, which degrades only double-stranded (ds)RNA molecules5, or RNase H, which specifically cleaves RNA:DNA hybrids6 (Fig. 1b, first and second panels). Taken together, these results suggest that single-stranded (ss)RNA is required for PrPres amplification in vitro, but that dsRNA and RNA:DNA hybrids are not. Addition of DNase or the restriction enzyme EcoRI did not decrease PrPres amplification, showing that DNA is not required for the process (Fig. 1b, third and fourth panels). Addition of the enzymes apyrase and heparinase III also had no effect on PrPres amplification, suggesting that neither high-energy nucleotides nor molecules containing heparan sulphate are required for PrPres amplification in vitro (Fig. 1b, fifth and sixth panels). In control experiments, we confirmed the degradation of the target molecules in each of these reaction mixtures by using appropriate analytical assays for these structures (Supplementary Fig. S2). To ensure that the commercial nuclease preparations we used were not contaminated with proteases, we measured the levels of PrPC and PrPres after overnight incubation with various nucleases. These measurements confirmed that levels of PrPC (Fig. 2a) and input PrPres (Fig. 2b) were both unperturbed by addition of enzymes that inhibited PrPres amplification. Figure 2 Nucleases do not cause proteolytic, steric or end-product inhibition of PrPres amplification. Full legend As a control to confirm that abolition of PrPres amplification depends on the stimulatory activity of each inhibitory nuclease, we added benzonase, micrococcal nuclease and RNase A to PrPres amplification reactions in enzymatically inactive states. Both benzonase and micrococcal nuclease require divalent cations for enzymatic activity, so we inactivated these nucleases by addition of 5 mM EDTA. The active site of RNase A contains a critical histidine residue that can be covalently modified by diethyl pyrocarbonate (DEPC). Therefore, we pre-treated RNase A with DEPC to inhibit its RNase activity, and removed excess DEPC by dialysis. Our results show that none of the three nucleases inhibits PrPres amplification in their inactive states, supporting the hypothesis that intact RNA molecules stimulate this process (Fig. 2c). To test whether inhibition of PrPres amplification might be mediated by end products of RNase digestion, we measured directly the effect of cyclic 2',3'-guanidine monophosphate (GMP) and 3' -cytidine monophosphate (CMP) on PrPres amplification. Neither of these nucleotides inhibited PrPres amplification in vitro at concentrations up to 1 mM (Fig. 2d). Our control experiments rule out the possibility that contaminating proteases, steric hindrance, or digestion end-products account for the inhibition of PrPres amplification by specific nucleases. Taken together, these experiments indicate that RNA is required for PrPres amplification in vitro. We next sought to determine whether a preparation of isolated RNA molecules could reconstitute the ability of nuclease-treated normal brain homogenate to amplify PrPres. Remarkably, total RNA isolated from hamster brain successfully reconstituted the ability of benzonase-pre-treated brain homogenate to amplify PrPres in a dose-dependent manner (Fig. 3a). By contrast, purified heparan sulphate proteoglycan (HSPG) failed to reconstitute PrPres amplification (Fig. 3a). Other polyanions, such as ssDNA (Fig. 3b), polyadenylic acid, heparan sulphate, pentosan sulphate and polyglutamic acid (data not shown) also failed to stimulate PrPres amplification. In this and other reconstitution experiments, benzonase-treated control lanes have a greater level of PrPres amplification than the diluted scrapie brain homogenate control samples, indicating that the benzonase pre-treatment reactions were incomplete. Empirically, we found that it was necessary to perform benzonase pre-treatment reactions at 4 °C to avoid denaturing PrPC before the addition of polyanions. Figure 3 Reconstitution of PrPres amplification with RNA. Full legend To estimate the molecular size of the RNA species capable of reconstituting PrPres amplification, we fractionated our preparation of total hamster brain RNA by ultrafiltration through a filter with a relative molecular mass cutoff of approximately 100,000 (Mr 100K). Using agarose gel electrophoresis, we detected all of the ribosomal RNA bands in the retentate and all of the transfer RNA in the filtrate (data not shown). Using these samples, we discovered that the filter retentate was capable of reconstituting PrPres amplification to a level slightly lower than unfractionated total brain RNA. By contrast, the filtrate was not able to reconstitute PrPres amplification (Fig. 3c). These data indicate that most of the reconstitution activity is conferred by RNA molecules >100K in size (>300 nucleotides). There is currently a need to develop more sensitive diagnostic tests for prion disease—this might be achieved by increasing the efficiency of PrPres amplification techniques. We therefore investigated whether addition of total hamster brain RNA could increase the efficiency of PrPres amplification in vitro in brain samples not pre-treated with nuclease. We mixed a more dilute homogenate of prion-infected brain (0.02% w/v) with 5% (w/v) normal brain homogenate overnight without sonication, and measured PrPres amplification. Our results show that addition of total hamster brain RNA to this mixture of intact brain homogenates significantly stimulates PrPres amplification over baseline (Fig. 4a). As a control, we confirmed that addition of RNA did not alter the level of input PrPres or PrPC in these samples (Supplementary Fig. S3). Densitometric measurements indicate that PrPres in 0.02% (w/v) prion-infected brain homogenate samples is amplified about sixfold after overnight incubation, similar to the PrPres amplification level previously reported for 0.1% (w/v) prion-infected brain homogenate3. By contrast, PrPres in samples amplified with RNA is amplified about 24-fold, indicating that addition of RNA increases the efficiency of in vitro PrPres amplification about fourfold. Addition of RNA also increased the efficiency of PrPres amplification of sonicated protein-misfolding cyclic amplification (PMCA) reactions (Supplementary Fig. S4). Figure 4 Stimulation of PrPres amplification with RNA. Full legend To assess the specificity of RNA-mediated stimulation of PrPres amplification, we isolated total RNA from several sources, including Escherichia coli, Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophila melanogaster, and mouse and hamster brain. Agarose gel electrophoresis analysis of these preparations revealed the expected band patterns for each species and confirmed that each preparation contained high-quality, non-degraded RNA (Fig. 4b). Furthermore, each of these preparations was substantially free from contaminants as judged by optical spectroscopy (A260/A280 > 1.9; where A indicates absorbance and subscript numbers indicate wavelength). Notably, among the six preparations of RNA tested, only hamster and mouse brain RNA could stimulate PrPres amplification in vitro (Fig. 4c). This apparent species specificity cannot be attributed to tissue specificity because total hamster liver RNA also stimulated PrPres amplification (data not shown). This argues that mice and hamsters express specific RNA molecules required for PrPres amplification. Additional experiments show that the RNA stimulation activity within the Trizol-extracted hamster brain RNA preparation was irreversibly destroyed by glyoxylation, but not by deproteination, heating to 60 °C, or transient exposure to 50% formamide for 1 h (Supplementary Fig. S5). If PrPres amplification studies accurately model PrPSc formation in vivo, the results presented here represent a significant advance in our understanding of the mechanism of prion conversion. Previously, it has been shown that purified PrPC can be converted into protease-resistant PrPres in vitro in the absence of cellular cofactors7. However, the fact that a 50-fold molar excess of purified PrPres is required to drive conversion of purified PrPC suggests that efficient PrPres formation may depend on the presence of cellular factors other than PrPC (ref. 8). On the basis of the results presented here, we propose the hypothesis that specific RNA molecules are cellular cofactors for PrPSc formation. Consistent with our hypothesis that specific RNA-converting factors stimulate PrPSc formation, nucleic acids bind avidly to and promote conformational change of recombinant PrP (refs 9–14). However, it is important to note that full-length, refolded recombinant PrP lacking post-translational modifications cannot undergo stoichiometric conversion to PrPres (Supplementary Fig. S6), and therefore the results of biophysical studies using recombinant PrP cannot be directly related to the results described here. It has been proposed that PrPSc molecules might bind to specific host RNA molecules to generate strain diversity15. Whether the RNA-converting factors we describe are also involved in generating strain diversity remains to be determined. Finally, it is important to emphasize that the existence of RNA-converting factors is fully consistent with the protein-only hypothesis proposed previously1, because the nucleic acids we describe are host-encoded and not contained within the infectious agent. Methods In vitro PrPres amplification In vitro PrPres amplification3 and PMCA4 were performed as previously described, except that normal brain homogenates were prepared with EDTA-free complete protease inhibitors (Roche) to facilitate experiments involving metal-dependent enzymes. Two millimolar MgCl2 was added to reactions with benzonase and 2 mM CaCl2 was added to reactions with micrococcal nuclease and apyrase. All amplification and control reactions were performed at 37 °C for 16 h. For PrPres detection, protease digestion was performed with 50 µg ml-1 proteinase K for 1 h at 37 °C and immunoblotting was performed with 3F4 monoclonal antibody (Signet). For PrPC detection, samples were not subjected to proteinase K digestion before immunoblotting. All protein electrophoresis experiments shown were performed on 12% SDS polyacrylamide gels and reference Mr for such experiments are shown. Nuclease inactivation Micrococcal nuclease and benzonase were inhibited with 5 mM EDTA. RNase A (50 µg) was incubated with 1% DEPC in 100 µl at 25 °C for 2 h. After incubation, the reaction was dialysed twice against 1 l 10 mM Tris pH 7.2 at 4 °C using a Pierce 3500 MW Slide-A-Lyzer minidialyis unit to remove free DEPC. Control samples containing active RNase A were dialysed in parallel. Protein recovery >90% was confirmed by BCA assay (Pierce). Active and inactivated nucleases were added to amplification reactions at concentrations designated 'X' in Fig. 1. 'No enzyme' control samples were processed in parallel. Reconstitution assays Nuclease digestion before reconstitution was performed by incubating a batch of normal brain homogenate (10% w/v) with benzonase (final concentration of 2.5 U µl-1) and 2 mM MgCl2 for 16 h at 4 °C in the absence of detergents. Benzonase was then inactivated by the addition of 5 mM EDTA before reconstitution with RNA or other polyanions. Preparation and measurement of RNA RNA was isolated from animals <5 min after death using rotor–stator homogenization, extraction with Trizol reagent (Invitrogen) for 5 min at 25 °C, and isopropanol precipitation according to manufacturer's instructions, using RNase-free reagents, containers and equipment. For yeast, cell walls were disrupted during extraction as previously described, using Trizol in place of phenol16. All RNA solutions were alcohol precipitated, washed and resuspended in RNase-free water before use. The concentration and purity of each solution was determined by spectroscopic measurement of absorbance at lambda1/lambda2 = 260/280 nm and confirmed by electrophoresis on 1% agarose gels stained with ethidium bromide. RNA size fractionation Total hamster brain RNA (0.4 mg) was diluted into 0.8 ml RNase-free water, loaded in 0.2-ml batches onto four separate Schleicher and Schuell Centrex UF-05 (100K cutoff) ultrafiltration devices, and centrifuged for 15 min at 3,000g. The devices were then washed with an equal volume of water. The filtrates were pooled and retentate fractions collected by briefly centrifuging the ultrafiltration devices upside down into new microcentrifuge tubes. Parallel samples of denatured retentate were prepared in 50% formamide to disrupt all intra- and intermolecular interactions. Reverse transcriptase polymerase chain reaction RT–PCR was performed using the One Step RNA PCR kit (AMV) from Takara/Fisher following the manufacturer's instructions, using the PrP-specific primers 5'-CGAACCTTGGCTACTGGCTGCTG-3' and 5'-GCTTGATGGTGATATTGACGCAGTC-3', and the following parameters: reverse transcription at 50 °C for 15 min, heat inactivation of reverse transcriptase at 94 °C for 2 min, times25 PCR cycles (94 °C for 30 s, 55 °C for 30 s, 72 °C for 90 s). Products were run on a 1% agarose gel and stained with ethidium bromide. Supplementary information accompanies this paper. Received 4 July 2003;accepted 31 July 2003 References Acknowledgements. The authors thank G. Saborio, C. Soto, V. Ambros, C. Cole and W. Wickner for helpful advice. This work was supported by the Burroughs Wellcome Fund Career Development Award, the Hitchcock Foundation, and an NIH Clinical Investigator Development Award. Competing interests statement. The authors declare that they have no competing financial interests. © 2003 Nature Publishing Group http://www.nature.com/ TSS
|
