Báo cáo khoa học: Putative prion protein from Fugu (Takifugu rubripes)

Putative Prion Protein From Fugu (Takifugu Rubripes)

Tác giả: Barbara Christen, Kurt Wüthrich and Simone Hornemann

Lĩnh vực: Institute of Molecular Biology and Biophysics, ETH Zurich, Switzerland

Nội dung tài liệu:
Nghiên cứu này tập trung vào việc xác định cấu trúc 3D của protein dạng prion (PrP) từ cá Fugu (Takifugu rubripes). Mặc dù các gen PrP đã được xác định ở nhiều loài cá, nhưng độ tương đồng trình tự với PrP của động vật có xương sống trên cạn thấp hơn 25%. Các chương trình dự đoán cấu trúc cho thấy một tổ chức 3D tương tự. Tuy nhiên, việc biểu hiện và tinh chế protein dạng tái tổ hợp, bao gồm cả miền C-terminal của Fugu-PrP1, gặp khó khăn. Các phương pháp biểu hiện đồng thời với protein chaperone của vi khuẩn E. coli đã được sử dụng để thu được protein ở dạng hòa tan. Mặc dù phổ CD cho thấy sự hiện diện của cấu trúc bậc hai, phổ NMR không cho thấy dấu hiệu của cấu trúc 3D hình cầu. Kết quả này cho thấy các sản phẩm polypeptide từ các gen được chú thích là gen prnp tương ứng ở các loài không phải cá không thể được chuẩn bị cho các nghiên cứu cấu trúc bằng các phương pháp tương tự đã sử dụng thành công cho PrP từ động vật có vú, chim, bò sát và lưỡng cư.

Mục lục chi tiết:

• Keywords
• Correspondence
• Prion proteins (PrP) of mammals, birds, reptiles and amphibians have been successfully cloned, expressed and purified in sufficient yields to enable 3D structure determination by NMR spectroscopy in solution. More recently, PrP ortholog genes have also been identified in several fish species, based on sequence relationships with tetrapod PrPs. Even though the sequence homology of fish PrPs to tetrapod PrPs is below 25%, structure prediction programs indicate a similar organization of the 3D structure. In this study, we generated recombinant polypeptide constructs that were expected to include the C-terminal folded domain of Fugu-PrP1 and analyzed these proteins using biochemical and biophysical methods. Because soluble expression could not be achieved, and refolding from guanidine-HCl did not result in a properly folded protein, we co-expressed Escherichia coli chaperone proteins in order to obtain the protein in a soluble form. Although CD spectroscopy indicated the presence of some regular secondary structure in the protein thus obtained, there was no evidence for a globular 3D fold in the NMR spectra. We thus conclude that the polypeptide products of the fish genes annotated as corresponding to bona fide prnp genes in non-fish species cannot be prepared for structural studies when using procedures similar to those that were successfully used with PrPs from mammals, birds, reptiles and amphibians.
• Prion diseases, such as scrapie in sheep, bovine spongiform encephalopathy, chronic wasting disease in deer, and Creutzfeldt-Jakob disease in humans, are related to the conversion of the cellular form of the prion protein (PrP) to a protease-resistant β-sheet-rich form (PrPS) [1]. Prion proteins from mammals, birds, reptiles and amphibians all possess the same molecular architecture, consisting of a flexibly extended 100-residue N-terminal tail and a globular C-terminal domain of similar size [2–7]. The C-terminal globular domain is preceded by a highly conserved hydrophobic polypeptide segment (Fig. 1). Its well-defined structure with three α-helices and an antiparallel β-sheet could be identified in all species studied to date [7]. Post-translational modifications such as cleavage of N- and C-terminal signal sequences during the import into the endoplasmic reticulum, formation of a disulfide bond that connects helices α2 and α3, N-linked glycosylation in two sites, and addition of a C-terminal glycosyl-phosphatidylinositol (GPI) anchor are present in all these species, which also contain putative Src homology domain 3- and laminin-α2-receptor binding sites [7,8]. The physiological role in the healthy organisms and the evolutionary origin of PrPs remain controversial [9,10].
• Recently, genes coding for putative prion proteins in fish species such as Japanese pufferfish (Fugu rubripes) [11,12], green spotted pufferfish (Tetraodon nigroviridis) [13], zebrafish (Danio rerio) [13,14], Atlantic salmon (Salmo salar) [12], rainbow trout (Oncorhynchus mykiss) [15], three-spine stickleback (Gasterosteus aculeatus) [8,16], carp (Cyprinus carpio) [8], gilthead seabream (Sparus aurata) [17], Japanese medaka (Oryzias latipes; GenBank: CAL64054), Japanese seabass (Lateolabrax japonicus) and Japanese flounder (Paralichthys olivaceus) [18] have been described and compared (for a sequence alignment, see Rivera-Milla et al. [8]). An early whole-genome duplication that occurred in the evolution of ray-finned fish [19–23] resulted in the presence of two fish PrPs (PrP1 and PrP2), whereas only one PrP has been identified in tetrapod species.
• Comparison of biophysical and structural properties of tetrapod PrPs with fish PrPs might help to improve our understanding of PrP biology, such as structure-function relationships in healthy organisms, and species barriers in transmissible spongiform encephalopathies. In addition, new insights into the evolutionary development of PrPs might be obtained. At the outset of this study, we tried to express and purify putative globular domains of Fugu (Takifugu rubripes) PrP1 (aa 298–423), Fugu PrP2 (aa 215–404), zebrafish (D. rerio) PrP1 (aa 389–581) and zebrafish PrP2 (aa 311–541), using the same protocol as for mammalian PrPs [4,24]. Among these proteins, only Fugu PrP1, spanning residues 298–423, could be obtained in sufficient quantities, and we therefore focused further work on this putative C-terminal domain, which appeared to us to be the most promising candidate for more detailed studies.
• In a first approach, the protein was expressed in inclusion bodies followed by refolding from guanidine-HCl using conventional Ni-affinity chromatography. In a second approach, the protein was obtained in soluble oxidized form by co-expression with Escherichia coli chaperone proteins [25–27], and then purified without the use of denaturants. The proteins thus obtained were studied with CD and NMR spectroscopy.
• Our results show that the putative C-terminal domain of T. rubripes PrP1 does not exhibit a defined 3D fold. We were surprised that fish PrPs could not be handled using the same protocol as for all other natural prion proteins studied in our laboratory, and we therefore conclude that this intriguing negative result should be communicated.
• Identification of the putative C-terminal domain of T. rubripes PrP1
• An alignment of T. rubripes PrP1 and PrP2 with murine PrP is shown in Fig. 1. We determined the polypeptide segment of T. rubripes PrP1 that should correspond to the C-terminal globular domain of tetrapod PrPs on the basis of recently published comparisons of fish and tetrapod PrP sequences [8,12,13,18]. The N-terminus was defined at residue Val298, which is in a hydrophobic segment that has high sequence homology to tetrapod PrPs. The C-terminus could not be identified unambiguously, because the sequence after the predicted α-helix 3 has no homology to non-fish PrPs. The GPI cleavage site could be at either Asn424 or Ser430 [28]. Because no regular secondary structure was predicted for the region between residues 424 and 430, we decided to place the C-terminal end at residue Arg423. In the remainder of this study, the polypeptide fragment of residues 298–423 is referred to as 6x His-tagged Fugu-PrP1(298–423) (tr1-PrP).
• Expression and purification of tr1-PrP
• The His-tagged protein was expressed and purified from inclusion bodies, using the method [4,24] successfully applied to obtain protein samples for 3D NMR structure determinations of a series of recombinant PrPs from mammals, birds, reptiles and amphibians [5,7,29]. Although the far-UV CD spectrum of tr1-PrP indicated the presence of some regular secondary structure, the ¹H-NMR spectrum revealed only small peak dispersion (data not shown), showing that the protein does not exhibit a globular fold and thus indicating possible improper refolding of the protein from the inclusion bodies. In additional experiments, the constructs Fugu-PrP1(298–450)[C426S] and Fugu-PrP1(355–450)[C426S], where Cys426 was replaced by serine, were tested for their folding properties. Fugu-PrP1(298–450)[C426S] was found to have a high tendency to aggregate during purification, whereas the behavior of Fugu-PrP1(355–450)[C426S] was similar to that of tr1-PrP.
• We next used an alternative expression strain with tr1-PrP, E. coli Origami B(DE3), which allows expression of proteins in oxidized soluble form in the cytoplasm of E. coli, and further enables variation of the isopropyl thio-β-D-galactoside (IPTG) concentration used to induce protein expression. In addition, chaperone systems such as Trigger Factor, GroEL/GroES and DnaJ/DnaK/GrpE were co-expressed to assist proper folding of the protein. Co-expression of Trigger Factor was found to yield the highest expression rate of soluble tr1-PrP and the lowest amount of co-purifying protein impurities (Fig. 2), whereas more impurities were observed with the GroEL/GroES system, and with the DnaJ/DnaK/GrpE system no expression of soluble tr1-PrP was obtained.
• In small-scale experiments, the concentrations of the inductors arabinose and IPTG, temperature and expression time were adjusted to maximize the yield of soluble protein. In the final protocol, induction of chaperone pre-expression with (L)-(+)-arabinose (2 g·L⁻¹) for 1 h, a final IPTG concentration of 1 mm, an expression temperature of 25 °C and an expression time of 15 h were used (Fig. 2).
• Soluble tr1-PrP was isolated from cells by sonication and centrifugation in a buffer that did not contain any detergents or denaturants (see Experimental procedures). The protein was purified by Ni-affinity chromatography, using a stepwise imidazole gradient to remove two co-purifying proteins that could be identified by Edman sequencing, MS and a database search as the ribosomal protein S15 and the ferric uptake regulation protein from E. coli (Swiss-Prot accession numbers P0ADZ4 and Q0TK00, respectively). Using this protocol, the yield of soluble oxidized tr1-PrP was 1.8 mg·L⁻¹ in rich medium, and in minimal medium, using ¹⁵N-ammonium chloride as the sole nitrogen source, the yield was 0.4 mg·L⁻¹.
• Characterization of tr1-PrP with CD and NMR spectroscopy
• To compare the conformation of tr1-PrP with that of recombinant mammalian prion proteins, we used CD and NMR spectroscopy. In the far-UV CD spectra, there are indications that tr1-PrP and mPrP(121–231) both contain α-helical secondary structure, but the mean residue ellipticity of tr1-PrP is approximately one-third less negative than that of mPrP(121–231), indicating a lower content of residues located in regular secondary structure elements (Fig. 3).
• In additional CD experiments, the thermal denaturation and the urea-induced unfolding transitions of tr1-PrP and mPrP(121–231) were compared (Fig. 4). Thermal denaturation and urea-induced unfolding of mPrP(121–231) is highly cooperative, as reported previously [30,31], whereas tr1-PrP unfolds in a less-cooperative manner typical of proteins that have no compact globular fold.
• NMR spectroscopy provided further evidence that no conformationally homogeneous sample of tr1-PrP was obtained in our experiments. The presence of peaks with variable line shape and intensity in the 2D [¹⁵N, ¹H]-HSQC spectrum indicates that the protein is prone to aggregation (Fig. 5A). The absence of a globular fold is supported by the small dispersion of the amide proton chemical shifts (Fig. 5). In a 2D [¹H, ¹H]-NOESY spectrum, the region expected to contain NOE-peaks between methyl groups and aromatic rings in globular proteins is empty for tr1-PrP (Fig. 5B).
• Conclusions
• Our investigations indicate that the gene coding for tr1-PrP, which has been annotated as the fish gene corresponding to prnp in mammals [11,12], does not encode a protein that can be isolated and purified with the biochemical methods used for other PrPs. This might be due to the fact that the identification of fish prnp genes was based on the coincidence with characteristic features that had previously been identified in bona fide PrPs, i.e., the N-terminal signal sequence, the Gly-Pro-rich region, the hydrophobic region and the presence of two cysteine residues, two glycosylation sites and the putative C-terminal GPI-anchor site (Fig. 1). The overall sequence homology of the globular C-terminal domain with different tetrapod PrPs is actually only between 15% and 25% [11,12]. Furthermore, the sequence identity is largely concentrated in the segment 114–154 (numeration according to mPrP), which covers a hydrophobic stretch preceding the globular domain, and the regular secondary structures β1 and α1 (Fig. 1). In the remaining part of the putative globular domain with helices α2 and α3, the homology is essentially limited to the alignment of the two Cys residues (Fig. 1).
• On grounds of principle, one cannot a priori exclude that alternative constructs with variable lengths would lead to a folded protein, especially as previous studies with mammalian PrPs have shown that deletions at both the N-terminal and the C-terminal end of the globular domain resulted in destabilization of the 3D structures [32]. However, because the N-terminal part of the fish prion protein studied here includes the highly homologous hydrophobic stretch (Fig. 1), which is unstructured in bona fide prion proteins, it seems unlikely that N-terminal elongation would result in a folded protein. The C-terminal end of the tr1-PrP construct used here was chosen at the proposed GPI-anchor site, and an alternative construct including the natural stop codon (tr1-PrP(298–450)[C426S]) yielded no folded protein either. It thus appears that the absence of a globular domain cannot be rationalized by inappropriate truncation of the tr1-PrP constructs used.
• Overall, we conclude from our data that the Fugu-PrP1 gene annotated as corresponding to bona fide prnp genes in all non-fish species studied to date, does not encode a protein that forms a typical prion protein 3D structure when isolated with the same purification and refolding methods that were successful with the other species. Considering that the sequence homology among fish species is 60% among the PrP1 proteins, 50% among the PrP2 proteins, and 40% between PrP1s and PrP2s [8], one is tempted to hypothesize that with regard to their expression in E. coli and subsequent purification, all fish PrPs might behave differently from tetrapod PrPs.
• Experimental procedures
• Cloning of the proteins
• Expression, purification and refolding of tr1-PrP from inclusion bodies
• Expression and purification of soluble tr1-PrP
• CD spectroscopy
• Thermal unfolding transitions were monitored by following the mean residue ellipticity, [Θ]MRW, at 222 nm between 20 and 90 °C at a constant heating rate of 1 °C·min⁻¹ and protein concentrations of 27 µm tr1-PrP and 19 µm mPrP(121–231), respectively.
• To study the urea-induced unfolding transitions, the mean residue ellipticities at 222 nm were recorded in the presence of different urea concentrations at protein concentrations of 22 µm for tr1-PrP and 33 µm for mPrP, respectively. The mean residue ellipticity was recorded for 30 s and averaged. The data for mPrP(121–231) were analyzed according to a two-state model of folding by using a six-parameter fit [34].
• NMR experiments
• Acknowledgements
• References