Structural studies of PrPSc

Universidad de Santiago de Compostela Facultad de Medicina Departamento de Medicina “Structural studies of PrPSc” Tesis Doctoral / Doctoral Thesis Director: Dr. Jesús Rodríguez Requena Ester Vázquez Fernández Santiago de Compostela, 2012 Este trabajo ha sido financiado a través de la subvención europea FP7 222887 “Priority” Autora: Ester Vázquez Fernández Director: Jesús R. Requena Año: 2012 Facultad de Medicina Departamento de Medicina El Dr. JESÚS RODRÍGUEZ REQUENA, profesor contratado del Departamento de Medicina, de la Universidad de Santiago de Compostela, INFORMA Que la presente memoria titulada “Structural studies of PrPSc”, presentada por Dña. ESTER VÁZQUEZ FERNÁNDEZ, para optar al grado de doctor por la Universidad de Santiago de Compostela, ha sido realizada bajo su dirección en el Departamento de Medicina de la Facultad de Medicina de la Universidad de Santiago de Compostela y reúne los requisitos exigidos por la normativa vigente para ser valorada por el tribunal correspondiente, Y para que así conste a los efectos oportunos, se firma el presente informe. En Santiago de Compostela, a 4 de Septiembre de 2012. Fdo.: Dr. Jesús Rodríguez Requena Agradecimientos “La gratitud no es solo la más grande de las virtudes, sino la madre de todas las demás” Cicerón Cuando un sueño se hace realidad, no siempre hay que atribuirle el mérito al empeño que uno pone en realizarlo. Detrás de cada sueño siempre hay personas que nos apoyan y que creen en nosotros. Son seres que han aparecido en nuestro camino, que nos animan a seguir adelante en nuestros proyectos, brindándonos de diferentes maneras, su solidaridad y conocimientos. A todos vosotros gracias por enseñarme, entenderme, motivarme, hacerme reír, escuchar mis frustraciones y alegrías, no dejar que me rindiera, por vuestros consejos … En definitiva, gracias por hacerme feliz durante esta etapa de mi vida y porque sin vosotros esta tesis no hubiera visto la luz. Estoy en deuda con vosotros y espero que sintáis parte de esta tesis como vuestra. Aut inveniam viam aut faciam “Encontraré un camino, o yo misma lo crearé” Structural studies of PrPSc Resumen en castellano Resumo en galego Summary in English RESUMEN Estudios estructurales de la PrPSc Los priones son agentes responsables de un conjunto diverso de enfermedades neurodegenerativas transmisibles en humanos y animales, que se caracterizan por una acumulación anormal de la proteína priónica (PrP), principalmente en el cerebro. Las lesiones neuropatológicas que se pueden encontrar son atrofia generalizada, pérdida neuronal, vacuolación, cambios espongiformes, gliosis astrocitica y placas amiloides de PrP. Además dichas enfermedades tienen un amplio espectro de manifestaciones clínicas, caracterizadas por demencia, ataxia, insomnio, paraplejía, parestesia y comportamiento anormal. La replicación de los priones se produce mediante la conversión de la PrPC, que es la proteína normal no infecciosa, en PrPSc. Durante el proceso ocurre un cambio conformacional a través de una modificación post-traduccional, sin embargo, dicho proceso todavía no está completamente entendido. La cadena polipeptídica de la PrPC y la PrPSc son idénticas. En ratón, la PrP contiene 209 amino ácidos (numerado como 23-231, después de la escisión del péptidos señal de 22 amino ácidos) y tiene cuatro modificaciones post-traduccionales covalentes: dos aparaginas con glicanos unidos, en los residuos N180 y N196; un puente disulfuro entre los residuos C178-C213; y un glicosilfosfatidilinositol (GPI) de anclaje a la membrana, en la región C-terminal de la proteína (residuo S231). La PrPC es un monómero, mientras la PrPSc es un multímetro heterogéneo. No se han demostrado diferencias covalentes entre ellas, la única diferencia entre la PrPC y la PrPSc es conformacional, son isoformas. La estructura de la PrPC ha sido resuelta mediante resonancia magnética nuclear (RMN) y cristalografía de rayos X, en grandes rasgos contiene tres hélices-! y dos tramos de hojas-" antiparalelas. En contraste, la estructura de la PrPSc todavía no está resuelta, debido a que la insolubilidad de la PrPSc y la incapacidad de cristalizar de los multímetros heterogéneos de PrPSc impiden la aplicación de las mencionadas técnicas analíticas de alta resolución. Sin embargo, una variedad de técnicas instrumentales de baja resolución han proporcionado alguna información sobre la estructura de la PrPSc. A diferencia de la PrPC, la PrPSc es parcialmente resistente a la digestión con la enzima proteinasa K (PK). Estudios de espectroscopía infrarroja de transformada de Fourier (FTIR) han mostrado que la estructura secundaria de la PrPC se compone, en gran parte, de regiones no estructuradas y hélices-!, mientras que la PrPSc está compuesta, en gran medida, por hojas-" con alguna, si hay, hélice-!. La estructura de la PrPSc también ha sido estudiada usando microscopía electrónica basada en el análisis de cristales bidimensionales del núcleo resistente a PK de la PrPSc de hámster sirio (PrP27-30) y por espectrometría de masas (MS) acoplada al análisis de intercambio hidrogeno/deuterio. Durante estos últimos años, varios modelos teóricos para la estructura de la PrPSc han sido propuestos, pero hay insuficientes datos experimentales para alcanzar un consenso definitivo. La elucidación de la estructura de la PrPSc sigue siendo uno de los mayores retos en la investigación sobre priones. Las bases moleculares de la biología de la proteína priónica (PrP), tal como el mecanismo molecular de replicación y agregación, la barrera de especie y la patogénesis de la neurodegeneración, no serán entendidos hasta que la estructura este resuelta. La proteolisis limitada es una técnica muy útil para elucidar características estructurales de la PrPSc. Parámetros conformacionales como la exposición de los amino ácidos a la superficie, flexibilidad e interacciones locales correlacionan muy bien con la proteolisis limitada. Los enlaces peptídicos localizados dentro de hojas-" son resistentes a la digestión proteolítica, mientras que los enlaces peptídicos localizados dentro de lazos, y mas raramente, en hélices-!, pueden ser cortados. Por lo tanto esta técnica proporciona información exhaustiva del enlace peptídico con respecto al plegamiento de la proteína, permitiendo localizar las áreas mas flexibles y las posibles hélices-! y hojas-". En la literatura hay numerosos estudios que utilizan la proteolisis limitada para averiguar información sobre la proteína priónica. El tratamiento con la enzima no específica PK muestra que la PrPC es completamente sensible a proteolisis y la PrPSc es parcialmente resistente, produciendo un núcleo resistente a PK de diferente longitud. La longitud de este núcleo depende de la especie, en hámster Sirio, consiste de los residuos amino acídicos !90-231. Esta región resistente, llamada PrPres o PrP27-30 debido a su peso molecular, continua siendo infecciosa y tiene las características básicas de un prión. Por lo tanto puede ser concluido que la PrPSc consiste de dos dominios: una área N-terminal lábil y desestructurada y una región globular estructurada resistente a proteasas. Otros estudios utilizando esta técnica han demostrado la existencia de una fracción de PrPSc sensible a la digestión con PK, llamada sPrPSc. Esta fracción también es infecciosa y comparte las características básicas estructurales con la PrPres. Otros experimentos basados en proteolisis limitada han visto que la resistencia a la digestión proteolitica es ligeramente diferente en las distintas cepas de la PrPSc. Por lo tanto, es compartida una arquitectura básica entre las diferentes cepas, con diferencias menores. Otros análisis de la PrPSc resistente a PK han informado de otros fragmentos más pequeños resistentes, detectados con anticuerpos que reconocen la región C-terminal de la proteína. Además, en el estudio (Sajnani et al. (2008)) que precede a los experimentos realizados en esta tesis se demostró la utilidad de combinar proteolisis limitada y espectrometría de masas (MS) para obtener información estructural sobre dos cepas de PrPSc de hámster. En este estudio se llegó a la conclusión de que la mitad amino terminal de la PrPSc está caracterizada por una serie de tramos cortos resistente a PK, presumiblemente son hebras-", intercaladas con tramos sensible a PK, probablemente formados por lazos y giros. Desafortunadamente, la información estructural se limitó a la región N-terminal de la proteína; ya que la presencia del glicosilfosfatidilinositol (GPI) de anclaje a la membrana y los heterogéneos azucares unidos a los amino ácidos asparagina, en la zona C-terminal de la molécula, impiden el análisis de espectrometría de masas (MS) en esta región. Con la finalidad de vencer estas dificultades en el análisis de la proteína, se usó la PrPSc de ratones transgénicos que expresan la proteína priónica (PrP) sin el GPI de anclaje a la membrana. Estos animales, producen PrPSc que está desprovista del anclaje GPI y de los carbohidratos, por lo tanto, se elimina la heterogeneidad, permitiendo la detección y la localización de regiones SUMMARY Structural studies of PrPSc Elucidation of the structure of PrPSc continues to be one major challenge in prion research. Molecular basis of the biology of prion protein, such as the molecular mechanism of prion replication and aggregation, the species barrier and the pathogenesis of neurodegeneration will not be understood until the structure is solved. Given that high-resolution techniques such as NMR or X-ray crystallography cannot be used, a number of lower resolution analytical approaches have been attempted. Limited proteolysis has been successfully used to pinpoint flexible regions within prion multimers (PrPSc). However, the presence of covalently attached carbohydrates and glycosylphosphatidylinositol (GPI) membrane anchor makes mass spectrometry-based analysis impractical. In order to surmount these difficulties it was analyzed PrPSc from transgenic mice expressing prion protein (PrP) lacking the GPI membrane anchor. Such animals produce prions that are devoid of the GPI anchor and carbohydrates, and, thereby, permit the detection and location of flexible, proteinase K (PK) susceptible regions by Western blot and mass spectrometry-based analysis. Then, these properties have allowed to obtain, for the first time, a complete survey of the whole PrPSc sequence, regarding its susceptibility to proteolysis. PrPSc samples were digested with PK, subjected to Tricine-SDS-PAGE, and probed with a Cterminal specific antibody (R1). Seven bands were detected by Western blot analysis, with apparent MWs of approximately 17, 14.6, 13, 12, 10.2, 8 and 6.7 kDa. Subsequently, mass spectrometry-based analysis detected 13 different peptides with exact MWs: 17148, 16726, 16371, 13606, 13463, 12173, 12041, 11171, 9687, 9573, 8358, 7436 and 6274, which match quite well with PK-resistant bands identified by Western blot analysis. These peptides correspond to molecules cleaved at positions 81, 85, 89, 116, 118, 133, 134, 141, 152, 153, 162, 169 or 179, respectively. The first 9 peptides (to position 153), match very well with PK cleavage sites previously identified in wild type PrPSc (Sajnani et al. (2008)). Then, the map of PK-susceptible spots, namely 116-118, 133-134, 141, 152-153, 162, 169, and 179, strongly suggests regions corresponding to loops and turns, while nicks at 81, 85, and 89 signal the frontier between the structured C-terminal and unstructured N-terminal domains of PrPSc. Given the high proportion of "-sheet secondary structure derived from FTIR analyses, it is logical to conclude that PK-resistant sequence stretches flanking these spots most likely corresponds to strands of "-sheet. Furthermore, the sizeable C-terminal stretch of PrPSc is highly resistant to PK and therefore perhaps constituted by "-sheet secondary structure. At the same time, taking advantage of the property of prion protein to polymerize into amyloid fibers, different electron microscopy (EM) techniques, from low-resolution to three-dimensional reconstruction were used to study the ultrastructure of the PrPSc. To perform these comprehensive studies, GPI-anchorless PrPSc was used, because this protein has allowed the improvement of the isolation method, so it provides the adequate cleanest sample, without losing the natural properties of the wild-type PrPSc. It was found out that the single fibers have a width of 3-5 nm, no regular helical pitch and are made up for two protofilaments intertwined, with no regular crossover distance. Threedimensional reconstruction of 2D cryomicroscopy (cryo-EM) images of the fibers also shows two protofilaments coiling around a common axis. Moreover, there are 2-2.5 nm repetition densities along the axis, suggesting each of these is a stacked monomer of PrPSc. These axial densities are also evident in reconstructed tomograms obtained by cryo-EM tomography of fiber sample. Another important measure is a 4.8 Å reflection presents in the Fast Fourier Transform (FFT), characteristic of the cross-" structure. A high-resolution image of cryo-EM also shows clearly the 4.8 Å reflection of the spacing between the beta repeat strands that are stacked like a “ladder”. All of the data, obtained during this research, suggest that, each PrPSc monomer must be coiled in a cross-" along the fiber, as to fit approximately 144 residues (~G89-S232) into the 3-5 nm width while maintaining the observed high proportion of "-sheet secondary structure. In order to do so, the PrPSc monomers must necessarily adopt a multi-layer architecture, specifically the results fit perfectly with the four-rung "-strands architecture proposed by Wille H et al. (2009), based on their interpretation of X-ray diffraction patterns. Accordingly, each rung of the GPIanchorless PrPSc monomer would be comprised of ~36-37 residues. In summary, the data support a PrPSc structure consisting of a four-rung solenoid with a central "-strand-rich core; from which a series of highly PK-resistant "-sheet strands intersperse with PK-sensitive short flexible loops and turns. Furthermore, the region comprising ~V179 to the Cterminus of PrPSc is probably composed primarily of "-sheet, as it is highly resistant to PK. The data obtained from this GPI-anchorless PrPSc work is consistent with the previous study of Sajnani et al. (2008) using SHaPrPSc (263K and Dy strains). Besides, the results are consistent with those observed for human CJD PrPSc, which suggests that the myriad human, hamster and mouse prions share a common basic structure. Table of contents 1. Introduction 1 1.1 Historical background 2 1.2 Prion protein structure 5 1.3 Structural insights on the architecture of PrPSc 8 1.3.1 Limited proteolysis 8 1.3.2 Fourier transform infrared spectroscopy (FTIR) 9 1.3.3 Transmission electron microscopy (TEM) 10 1.3.4 Fiber X-ray diffraction 11 1.3.5 Electron crystallography 12 1.3.6 Antibody mapping studies 13 1.3.7 Chemical cross-linking 13 1.3.8 Chemical surface labeling 14 1.3.9 Hydrogen-deuterium exchange (H/D) 15 1.4 Structural models of PrPSc 15 1.4.1 "-helical model 15 1.4.2 Spiral model 16 1.5 Aim of the thesis 18 1.6 Outline of the thesis 18 1.7 References 19 2. Generation of PrPSc 27 2.1 Introduction 28 2.2 Animal model 28 2.2.1 Syrian hamsters 28 2.2.2 GPI-anchorless PrP transgenic mice 29 2.3 Isolation of PrPSc 33 2.3.1 Brain homogenate 33 2.3.2 Isolation of GPI-anchorless PrPSc 33 2.4 Concluding remarks 36 2.5 References 37 3. Limited proteolysis 41 3.1 Introduction 42 3.1.1 Limited proteolysis 42 3.1.2 Tricine-SDS-PAGE 46 3.1.3 Outline of experimental approach 47 3.2 Studies on GPI-anchorless PrPSc 48 3.2.1 Identification of PK cleavage sites by Western blot (WB) 48 3.2.2 Identification of PK cleavage sites by mass spectrometric detection 49 3.2.3 Kinetics of PK digestion 51 3.2.4 PK cleavage analysis after partial unfolding 52 3.3 Studies on Syrian hamster PrPSc (SHaPrPSc) 53 3.3.1 Identification of PK cleavage sites in two strains of SHaPrPSc by WB 53 3.3.2 Identification of PK cleavage sites in PK-sensitive SHaPrPSc by WB 55 3.4 Conclusions 57 3.5 Experimental 60 3.6 References 62 4. Electron Microscopy 67 4.1 Introduction 68 4.1.1 Outline of experimental approach 69 4.2 Transmission electron microscopy (TEM) 70 4.2.1 Results of TEM 71 4.3 Cryo-transmission electron microscopy (Cryo-TEM) 73 4.3.1 Results of cryo-TEM 73 4.4 Cryo-TEM tomography (Cryo-ET) 76 4.4.1 Results of cryo-ET 77 4.5 Helical reconstruction 79 4.5.1 Results of helical reconstruction 80 4.6 Conclusions 81 4.7 Experimental 83 4.8 References 84 5. Discussion 89 6. Appendices 95 Appendix I. Explanation of the GPI-anchorless tg mice sequence (23-232) 96 Appendix II. Calibration of the GPI-anchorless PrPSc MALDI-TOF spectrum 97 Appendix III. List of abbreviations 101 Appendix IV. List of publications / Participation in conferences 103 1 1 Introduction Abstract This chapter is a general introduction and describes the thesis background and outline. A brief description of the prion field is presented and divided into four different blocks: historical background, prion protein structure, structural insights on the architecture of PrPSs and structural models of PrPSc. Due to the complexity of the issues in the study of the structure of PrPSc, most data available in the literature come from lowresolution biophysical techniques; this information has permitted the development of some structural models in the last decade, but there are many pros an cons to each of the models. Therefore, until know, the structure of mammalian prions has remained unknown. With the support of the research group to which I belong, and collaborators from different fields, I have been capable to present new constraints and a new model of the structure of PrPSc. This thesis has numerous implications for understanding the etiology and pathogenesis of prions and other neurodegenerative diseases. Chapter 1 2 1.1 Historical background Prion diseases or transmissible spongiform encephalopathies (TSE) are a wide group of fatal neurodegenerative diseases that affect humans and animals, associated with the accumulation of aggregates of misfolded conformers of the host-encoded cellular prion protein (PrPC), primarily in the brain [1]. Prion diseases have a broad spectrum of clinical manifestations: dementia, ataxia, insomnia, paraplegia, paresthesia and abnormal behaviour. The neuropathologycal findings range from widespread atrophy, neuronal loss, vacuolation, spongiform changes, astrocytic gliosis and PrP amyloid plaques (Figure 1.1) [1]. AB Figure 1.1. Neuropathological features of TSEs. The images represent histopathological and immunohistochemical studies in scrapie-infected mice with RML strain. A. This panel shows spongiform degeneration in the thalamus, the vacuoles are the unstained globular structures (Haematoxyllin eosin staining). Bar 50 !m. B. PrP deposits in the thalamus, stained in brown. Bar 25 !m. The earliest description of scrapie, a natural TSE affecting sheep and goats exhibiting a strong scraping and ataxia, was made in the 18th century. More recently, other animal TSEs were recognized, including chronic wasting disease (CWD) of deer and elk [2] described in captive and wild animals, bovine spongiform encephalopathy (BSE) or “mad cow disease”, in cattle [3] and transmissible mink encephalopathy (TME) [4]. TSEs were also discovered in domestic cats and in other zoo animals [5]. The natural routes of transmission are unclear; some TSEs are endemic, with a low but stable incidence, however it is known that some epizootics have resulted via contaminated feed, like meat and bone meal (MBM) prepared from carcasses. BSE epizootic caused the infectious agent to spread, transmitting the disease to human beings (vCJD). In humans, prion diseases can arise spontaneously, be inherited or acquired through transmission. The first cases of a TSE in humans were described in 1920; it was termed Creutzfeldt-Jakob disease (CJD). It was discovered that CJD could be caused by the three vias, such as the variant CJD that was transmitted to humans by eating food contaminated. Furthermore, an infectious human prion disease, called kuru, was discovered in the highlanders of Papua New Guinea [6]. Other typical diseases are Gerstmann-Sträussler syndrome (GSS) [7] and fatal familial insomnia (FFI), they are mainly associated to a number of mutations in the prion protein gene, although there are sporadic cases of FFI. Introduction 3 Table 1.I. Prion diseases of human and animals. (Modified from Colby DW et al. (2011) [8]) Scrapie was demonstrated to be transmissible by inoculation between sheep (and goats) with characteristic prolonged incubation periods. It was concluded that some kind of virus must be the infectious agent, and it was defined by the term “slow virus”. Kuru, characterized by a progressive ataxia, probably was transmitted during ritual funeral cannibalistic ceremonies. Researchers observed similarities between kuru and scrapie at the neuropathological, clinical and epidemiologic levels, and transmissibility of kuru and CJD was proved by intracerebral inoculation into chimpanzees. More recently, humans have contracted a new variant CJD (vCJD) from prion-tainted beef products [9]. However, the transmissible agent in these diseases was not clear and the concept of “slow virus” was defeated because the specific virus was never found. There was not immunological response and this agent was resistant to treatment for inactivate nucleic acids, such as ultraviolet radiation or treatment with nucleases [10]. Thus, it was suggested that the transmissible agent might be a self-replicating protein [11]. Successive studies with substantially purified preparations of the scrapie agent showed that a protein is required for infectivity and a protease-resistant protein associated with a glycosylphosphatidylinositol (GPI) anchored membrane, termed prion protein (PrP), was isolated from infected brains [12][13]. Then, the term prion: “small proteinaceous infectious particles that resist inactivation by procedures which modify nucleic acids” was established [14]. Chapter 1 4 The protease-resistant form of the PrP was the major constituent of infective brain fractions, extremely resistant to physical and chemical treatments and hydrophobic. Its molecular size is 27 to 30 kDa and was called PrP27-30 or PrPres. But this form derives from a total protein of 3335 kDa named PrPSc (infectious scrapie isoform) (Figure 1.2). It was discovered that PrP27-30 is encoded by single copy chromosomal gene, nevertheless the normal product of PrP gene (Prnp) is a protease-sensible protein called PrPC (normal non-infectious cellular isoform) [15][16]. Therefore scrapie and cellular PrP isoforms are encoded by the same chromosomal gene [17], then the cellular protein is converted into PrPSc through a posttranslational process, consisting of a conformational transformation [18]. This is the core of the “protein only” hypothesis, according to which the infectious agent (PrPSc) is a misfolded form of the cellular prion protein (PrPC) [1]. Figure 1.2. Western blot of the prion protein isoforms. Brain homogenates from uninfected (lane 1 and 2) and prioninfected hamster (lane 3 and 4). Samples in lane 2 and 4 were digested with 50 !g/ml of proteinase K (PK) for 30 min at 37 ºC. PrPC in lane 2 and 4 was completely hydrolyzed, whereas approximately 67 amino acids were digested from the NH2 terminus of PrPSc to generate PrP27-30. The blot was developed with anti-PrP R073 polyclonal rabbit antiserum. (Adapted from Prusiner SB (1998) [1]) Although the prion replication comes about by a self-propagating conversion of PrPC to the pathogenic isoform, the mechanism is still unknown. Two replication models could be considered (Figure 1.3): “template-directed model”, it proposes that monomeric PrPC is the most stable conformer and PrPSc monomer can bind to PrPC molecule generating a heterodimer. Hence PrPSc acts as a template, inducing a conformational transition of PrPC. In this mechanism, the formation of oligomers occurs after the conversion. The other model is the “seeded nucleation model”, it postulates that the PrPC converts spontaneously to PrPSc, but PrPSc is unstable unless it is “protected” or ”stabilized” by being part of an oligomer bigger that a critical size [19]. Both mechanisms are theoretically plausible, however, the most probably is the seeded nucleation model because prion infectivity is associated with small PrPSc oligomers with 14-28 PrP monomers [20][21]. However, recent studies where transmissible prion disease is generated Introduction 11 treatment. GPI-anchorless PrPSc, produced in mice expressing PrP lacking the GPI anchor, is deposited as amyloid plaques into the brain [54], however they look similar to wild-type fibers [50], which are less prone to form deposits. All these studies suggest that PrPSc is capable of fibrilization, a property shared with other known prions [55][56] and with amyloids [57][58]. 1.3.4 Fiber X-ray diffraction The fiber formation property of PrPSc makes possible to determine additional information about the three-dimensional structure of PrP27-30 molecule by fiber X-ray diffraction technique. In early studies, X-ray diffraction pattern of mammalian prions has shown a meridional diffraction at 4.72 Å, that is a measure characteristic of amyloid structure [59], indicating a cross "-structure, and corresponding to the distance between "-sheets parallel to the fiber axis. Furthermore, the intersheet distance was 8.82 Å, meaning a uniform geometric distribution in the H-bonding direction [60]. Figure 1.7. Fiber diffraction patterns. Black arrows indicate cross-" meridional diffraction at close 4.8 Å resolution. The equatorial reflections do not identified are characteristic of lipids present in the preparations. A. SHaPrP27-30. B. RecSHaPrP(90-231) amyloid. White arrow indicates a broad equatorial diffraction at 10.5 Å, absent in A. C. MoPrP27-30 (RML strain). White arrows (also in D) show second and third orders of meridional 19.2 Å diffraction. D. Synthetic prion strain, MoSP1, derived from recPrP(89-230) amyloid, passaged twice through Tg9949 mice. (Modified from Wille H. et al. (2009) [61]) Chapter 1 12 More recent studies [61] have obtained diffraction pattern from infectious prions that show cross-" diffraction (meridional intensity 4.8 Å). In MoPrP27-30 appears a pattern (meridional reflection of 19.2 Å, and its second, third and fourth order: 4.8 Å, 6.4 Å and 9.6 Å) which has been interpreted by the author as indicating the presence of a repeating unit that correspond to four "-strands. Further, it was found that recPrP amyloid differs from highly infectious PrPSc fibers from infected brains; also presents a strong equatorial intensity (approximately 10.5 Å) very typical of an amyloid structure, showing the existence of flat, parallel stacks "-strands separated by this distance. Such equatorial reflection is absent in PrPSc preparations. This difference suggests that PrPSc from brain-derived prions could have a cross-" structure different from the classic amyloid-" conformation; the most plausible structure would be a "-helical architecture. However, there are different equatorial intensities characteristic of diffraction from lipids or lipid-detergent assemblies due to the isolation process, which cast some doubts on the assignment of other reflections to the protein component of preparations (Figure 1.7). 1.3.5 Electron crystallography The ultrastructure of PrPSc has also been analyzed using electron microscopy-based analysis of two-dimensional crystals of PrP27-30 [62][63]. These 2D crystals were discovered in some purified fractions of the PrPSc, showing a hexagonal lattice with dimensions of ! and "= 6.9 nm and !=120º, as determined by electron diffraction. The obtained data were analyzed by digital processing using a single particle analysis approach. Through labelling with a sugar-specific nanogold derivative, the authors were able to localize the sugars in the periphery of the crystal subunit (Figure 1.8). Furthermore, in this study the 2D crystals of PrPSc106 (lacks residues 23-89 and 141-176) were also analyzed, it is the minimum sequence able to sustain conversion to PrPSc and is called miniprion [1]. PrPSc106 was obtained from PrP106-expressing transgenic mice infected with scrapie prions. PrPSc106 2D crystals are isomorphous to PrP27-30 ones and it was observed that the last 36 deleted residues, associated to a "-sheet structure are located at the inside of the oligomer. The authors concluded that PrPSc is formed by a parallel "-helix, since such structure would be the only one to fit to the hexagonal lattice dimensions. They also observed interactions of heavy metal cations with the centre of the hexagons, providing a positive stain; it could be related with the presence of negative charges of the protein in that region. Furthermore, subsequent studies, analyzing the binding and interaction of different heavy metal salts with the protein revealed three different negative densities into the crystal subunit [63]. In essence, threefold symmetry seems to exist, coinciding with the location of the proposed parallel "-helix. The resolution of this technique is about 2 nm, which is just the diameter of the proposed "-helix. Therefore, it is not possible to provide any detail of higher resolution, just that their architecture might correspond to a solenoid. Introduction 13 Figure 1.8. 2D crystals of PrP27-30. A. 2D crystal of PrP27-30 stained with uranyl acetate showing an apparent hexagonal lattice. B. High view of a crystal after contrast transfer function and correlation-mapping and averaging. C. Subtraction map between unlabelled and labelled with nanogold (labelling the N-linked sugars) crystals, showing differences in lighter shades. D. Projection map of PrP27-30 with the sugars labelled outside (yellow). E. Typical rod with an aggregate of crystal subunits at each end. The scale bars are 100 nm. (Wille H. et al (2002) [62]) 1.3.6 Antibody mapping studies With the aim to prove conformational transition in the formation of PrPSc, monoclonal antibodies and Fab fragments to several epitopes of PrPC and PrPSc have been used [64][65]. In these antibody mapping studies, PrPC and PrP27-20 in its native conformation and in denaturing conditions were studied to find out the exposure of the epitopes within the protein. The authors observed that epitopes in the N-terminal region (residues 90-120) were accessible in PrPC, but stayed hidden in native PrP27-30. However, denaturation of PrP27-30 rendered accessible the epitopes of that region. The extreme C-terminal region of the prion protein is exposed on both PrPC and PrPSc. 1.3.7 Chemical cross-linking Chemical cross-linking, in combination with proteolytic digestion and mass spectrometric analysis is a good tool to obtain valuable information regarding the three-dimensional structure of proteins [66]. This technique is based on chemical modification process where a bifunctional Chapter 1 14 cross-linking reagent (cross-linker) reacts with specific groups in the protein, resulting in the formation of a covalent bond. Moreover, the cross-linker has a specific spacer length between its functional groups [66]. Then, the length of the spacer arm provides a maximum distance constraint between the two reacting residues. A cross-linking study probing Syrian hamster PrP27-30 and using bis(sulfosuccinimidyl) suberate (BS3), an agent that reacts with amino groups, showed that PrP27-30 is easily cross-linked with formation of dimers, trimers, and higher-order oligomers. Furthermore the study showed that that BS3 reacted preferentially with G90 and a cross-link involving two G90 was found in crosslinked PrP27-30 dimers. These results indicated that in PrP27-30 aggregate exists a spatial proximity of G90 amino termini. The maximum possible spacer distance of this cross-linker (1.14 nm) between consecutive G90 amino termini suggested a two/three layer architecture in the PrP27-30 (note that one rung is 0.47 nm) [67]. However, four layers could also be a possibility if G90 are not “flush and rigid” but rather in short but flexible mini-tails. 1.3.8 Chemical surface labeling A useful method to study the surface accessibility of the PrPSc is the chemical surface labeling. The relative reactivity of amino acids to a specific chemical, in native conformation of the protein, reflects the surface accessibility of the residues. In a recent study, this approach was applied to find out locations of conformational change. Recombinant Syrian hamster PrP(90-231) and Syrian hamster PrP27-30 were treated with two chemical modifiers, tetranitromethane (TNM) and acetic anhydride (Ac2O), which specifically target accessible tyrosine and lysine residues, respectively. Several differences in chemical reactivity were observed in specific locations within the common sequence of both isoforms of PrP. With TNM, the most conspicuous reactivity difference seen involves peptide E221-R229 (containing Y225 and Y226), which it was much more modified in recSHaPrP(90-231). Instead, peptides Y149-R151, Y157-R164, and R151-Y162 suffered more extensive modifications in SHaPrP2730. Ac2O modified extensively peptide G90-K106 and the amino terminus in both isoforms of PrP. These results suggested that the C-terminal region of the SHaPrP27-30 has lost part of the solvent accessibility of the residue Y225 and Y226, which can means that the two C-terminal helices are tightly packed in this isoform, but, alternatively, that these C-terminal helices have disappeared in PrPSc, transformed into "-sheet. On the other hand, the stretch spanning approximately Y149-R164 is also more accessible in SHaPrP27-30, suggesting rearrangements in !- helix H1 and the short "-sheet of recSHaPrP(90-231). The last conclusion is the N-terminal region of SHaPrP27-30 is very accessible [68]. Besides, this method can be combined with antibody-based detection, to identify accessible amino acids and differentiate between PrP isoforms or even prion strains, without the use of PK [69]. Introduction 15 1.3.9 Hydrogen-deuterium exchange (H/D) H/D exchange is a chemical reaction in which a covalently bonded hydrogen atom is replaced by a deuterium atom. This reaction is the basis of a method that provides information about the solvent accessibility and tertiary structure of proteins. Backbone amide hydrogens located within the unstructured region of PrPSc exchange quickly with solvent deuterium, whereas in those placed in !-helices the hydrogen exchange rate decrease, and in those located in "-sheets it becomes exceedingly slow, as such hydrogen atoms are tightly held in place by bonding to N and O (CO) atoms. Different surveys of H/D exchange coupled to mass spectrometry, in the structure of amyloid fibrils formed by the recombinant PrP, have been performed [70][71]. However, recently it was published a study on GPI-less PrPSc [44]. It was showed that the C-terminal half of GPI-less PrPSc exhibits extremely low rates of H/D exchange, which is typical of stretches with extensive Hbonding ("-sheet) and the whole 90-215 stretch is relatively resistant to exchange. Some stretches exhibiting a somewhat higher exchange rate, suggested to partially overlap with loops/turns, such as 133-148 or 81-118. Therefore, it was concluded that GPI-less PrPSc consists of a series of "-sheet stretches connected by short loops and/or turns. 1.4 Structural models of PrPSc The above data and other structural experiments have been used to develop structural models of PrPSc. In the last decade, two models of a prion protofibril have been suggested: "-helix model [72] and "-spiral model [73]. 1.4.1 !-helical model The discovery of 2D crystals of PrP27–30 and PrPSc106 [62][63] along with threading the PrP sequence through a known "-helix fold have allowed the development of a plausible model termed "-helical model [72]. This model suggests that PrP27-30 displays a central parallel lefthanded "-helix formed by the residues 89-175, the C-terminal region (residues 176-227) retains the !-helical conformation and is located on the outside together with the sugars moieties. Furthermore, the segment encompassing residues 141-176 is a long unstructured loop. The symmetry of this model is based on trimeric arrangement, where three of such "-helical subunits pack tightly through lateral associations to form a trimer. The vertical stacking of these trimers occurs by hydrogen bonds, through head-to-tail-arrangement. Nonetheless, there is also evidence that the "-helix model cannot be fully correct. Three parallel left-handed "-helices trimerize at the level of their "-helix, providing a natural template to assemble three monomers, but the electron microscopy data disagree with this feature, because the width of PrPSc is 3-5 nm and the trimer is wider [50]. Furthermore, from limited proteolysis Chapter 1 16 studies it is possible to observe which residues participate in the "-strand formation, and in currently studies new cleavage points have been found, so that it is possible that PrPSc may have more loops or/and these are located in different places that in the "-helix model [40]. Also, treatment with BS3 [67] and Ac2O [68] show that the amino termini of PrP27-30 are very solvent exposed, in contrast with their internal location in the model. Also, other problem is that the threading, based on the structure of the "-helical part of trimeric carbonic anhydrase from Methanosarcina thermophila is very arbitrary Figure 1.9. "-helical structural model of PrPSc. A. Residues 89-174 of PrP are threaded onto a left-handed "-helical fold (yellow). B. Model of the monomer of PrP27-30. The !-helices (residues 177-227) are shown in red. C. Trimeric model of PrP27-30 built by superimposing three monomeric models. D. Fibrilization. Two trimers of PrP27-30 are assembling through polar backbone interactions between the lower "-helical rung of the top trimer disc and the upper rung of the bottom trimer. The sugars linked to N180 and N196 are shown like blue spheres and they are extended away from the centre of the structure. E. Projection map of 2D crystals of PrP27-30. F. Differences between PrP27-30 and PrPSc106. The differences attributed to the internal deletion of PrPSc106 (residues 141-176) are depicted in red; the differences in glycosylation are shown in blue. G. Superimposition of the trimeric left-handed "-helical model with the electron microscopy map of the PrP27-30 2D crystals. Scale bar is 50 Å. H. The scaled trimeric model was copied onto the neighbouring units of the crystals to show the crystallographic packing suggested by the model. Scale bar is 50 Å. (Adapted from Govaerts et al. (2004) [72]) In summary, while the solenoid-like basic architecture of individual PrPSc monomers seems very reasonable, the trimeric arrangement is very likely to be the result of tight packing to form the 2D crystals from individual PrPSc monomers that are in equilibrium with the fibers. 1.4.2 Spiral model This model derives from a molecular dynamics simulation of the prion protein under amyloidogenic conditions. The spiral model proposes a spiralling core of extended "-sheets formed by three parallel stranded sheets, spanning through the following regions: residues 116- Introduction 17 119, 129-132 and 160-164; and one isolated one, comprising the residues 135-140, although this last one is connected to the three-stranded sheet, forming a continuous four-stranded sheet. Then, a spiraling protofibril with a three-fold symmetry is formed. Furthermore, the glycosylation sites are exposed, the N and C-terminal regions are accessible with a protected core of more tightly packed "-strands [73]. The spiral model is consistent with a wide variety of low-resolution experimental data; but there is a critical constraint that makes it impossible; TEM (this thesis, Chapter 4) and fiber X-ray diffraction data [61] show that PrPSc fibers contain a cross-" stack of "-strands, that means "- strands are completely parallel to the axis of fiber growth. Figure 1.10. Spiral model of the structure of PrPSc. A-B. Building of a protofibril with a three-fold symmetry. The oligomerization occurs between the isolated "-sheet (fourth) and the first "-sheet of the adjacent monomer. C-D. Views of hexameric representation of protofibril showing the maintenance of symmetry on oligomerization and forming a spiralling core of extended sheets. E. Modeled protofibril (hexamer) with the diglycosylated subunits. F. Superimposition of two hexamers rotated 60º from one another around the fiber axis replicates the 6-fold symmetry of the crystals. G. Representation of a superimposing of two hexamers as in figure F. Residues 142-176 (magenta) represent the residues deleted from PrP27-30 to form PrPSc106. !-helices and sugar groups are shown in white. The EM image is a difference map between PrPSc106 and PrP27-30 with differences shown in magenta. H. The same figure as G, but in this case the EM image is a difference map between PrPSc106 and PrP27-30 with differences in the glycosylation. Sugars are shown in cyan colour. (Adapted from DeMarco ML. et al. (2004) [73]) As it can be seen from the above discussion, although theoretical models for PrPSc have been proposed, there is an insufficient amount of experimental data to reach a definitive consensus. It should be noted that there is a structural model that could help to understand the structure of mammalian prions. HET-s is a protein of the filamentous fungus Podospora anserina, which forms a prion that plays a role in heterokaryon incompatibility and it has some similarities with PrPSc. It Chapter 1 18 has a PK-resistant core like PrPSc and its structure, derived from NMR, forms a left-handed "- solenoid, with each molecule forming two helical windings, a triangular hydrophobic core, at least 23 hydrogen bonds, three salt bridges and two asparagine ladders. Furthermore HET-s protofibrils are also approximately 5 nm wide [55]. It is likely that the structure of PrPSc might have a lot of common with this one, as will be discussed later. 1.5 Aim of this thesis Elucidation of the structure of the mammalian prions continues to be one the major and difficult challenges in prion research due to their insoluble nature and heterogeneity. Molecular basis of the biology of PrPSc, such as the molecular mechanism of prion replication and aggregation, the species barrier and the pathogenesis of neurodegeneration will not be understood until the structure is solved. No less crucial is the fact that the knowledge of the structure of PrPSc will provide the essential clues to develop therapies for the treatment of prion diseases. For this reason, the aim of the thesis is focused on the pursuit of conformational constraints through the study of the GPI-anchorless PrPSc, mainly, and wild type PrPSc by different techniques such as limited proteolysis coupled with high-resolution electrophoresis and mass spectrometry, electron microscopy and computational 3D reconstruction, in order to find out the real structure of the scrapie prion protein. 1.6 Outline of the thesis In chapter 1 some background information, and the aim and outline of the thesis have been presented. As described in the introduction of this thesis, because of the insolubility of PrPSc and the failure to crystallize the heterogeneous PrPSc multimers, the application of high-resolution analytical techniques has not been possible, so that only lower resolution instrumental techniques have provided some information about the structure of PrPSc. Furthermore the high heterogeneity derived from the GPI moiety and the carbohydrates contained in its C-terminal region hamper such analysis. To solve the presented problem and extend my studies of the structure of PrPSc, I have used transgenic (tg) mice expressing PrPC lacking the GPI anchor; they were provided by Bruce Chesebro from the Rocky Mountain Laboratories, NIH, Montana, USA [54]. GPI-less PrPSc produced by these mice, when infected with wt prions, is infectious, lacks the GPI anchor, and is largely unglycosylated, which reduces the heterogeneity in the C-terminal portion of the molecule. The description of these mice, the generation of the PrPSc and the method of isolation are given in great detail in chapter 2. In this chapter, it is also described the generation of wildtype PrPSc from Syrian hamster. Introduction 19 Chapter 3 briefly explain the basis of the limited proteolysis technique [41] that it is combined with high-resolution electrophoresis, called Tricine-SDS-PAGE [74] and two different methods of mass spectrometry, MALDI (matrix-assisted laser desorption/ionization)–TOF (time of flight) and nano-LC-ESI (electrospray ionization)-Qq (double quadrupole)-TOF , to allow an accurate detection an location of flexible, proteinase K susceptible regions. The obtained results are presented and discussed meticulously. It is worth noting that I have been able to survey the whole structure of PrPSc thanks to the GPI-less tg mice. Chapter 4 introduces the general principles of the different electron microscopy techniques. These techniques were used to study the ultrastructure of PrPSc. As the previous chapter and continuing the thread of my studies, the experiments were realized with the fibers of GPI-less PrPSc. This protein has allowed the improvement of the isolation method, so I have got the adequate sample, without any artefacts, to be analyzed by microscopy and subsequently to use computational 3D reconstruction. It is very important to mention that these studies were made possible by the collaboration of different research groups from many diverse fields of knowledge: Dr. Matthijn Vos, from the FEI Company, Eindhoven, The Netherlands; and Dr. Howard S. Young, Dr. Ludovic Renault and Dr. Holger Wille, from the University of Alberta, Edmonton, Canada. In the chapter 5 is presented a common discussion, taken together all the results obtaining during this research. Furthermore, it is described an approach to a new model of the PrPSc structure. Four appendices are enclosed in the chapter 6. The first appendix explain an inconsistency, founded during the experiments by mass spectrometry (MS), in the sequence of the GPIanchorless PrP described previously by Chesebro B. et al. [54]. The second appendix shows the different peaks obtained by MS techniques, subsequently used to calibrate the GPI-less PrPSc MALDI-TOF spectrum. Abbreviations used in this document and a compilation of the publications and participation in conferences during the thesis are also annexed in this last chapter. 1.7 References 1. Prusiner SB (1998) Prions. Proc. Natl. Acad. Sci. U.S.A. 95: 13363–13383. 2. Williams ES, Young S (1980) Chronic wasting disease of captive mule deer: a spongiform encephalopathy. J. Wildl. Dis. 16: 89–98. 3. Wells GA, Scott AC, Johnson CT, Gunning RF, Hancock RD, et al. (1987) A novel progressive spongiform encephalopathy in cattle. Vet. Rec. 121: 419–420. 4. Marsh RF, Hadlow WJ (1992) Transmissible mink encephalopathy. Rev. Sci. Tech 11: 539–550. Chapter 1 20 5. Jeffrey M, Wells GAH (1988) Spongiform Encephalopathy in a Nyala (Tragelaphus angusi). Veterinary Pathology 25: 398–399. 6. Gajdusek DC, ZIGAS V (1957) Degenerative disease of the central nervous system in New Guinea; the endemic occurrence of kuru in the native population. N. Engl. J. Med. 257: 974– 978. 7. Masters CL, Gajdusek DC, Gibbs CJ (1981) Creutzfeldt-Jakob disease virus isolations from the Gerstmann-Sträussler syndrome with an analysis of the various forms of amyloid plaque deposition in the virus-induced spongiform encephalopathies. Brain 104: 559–588. 8. Colby DW, Prusiner SB (2011) Prions. Cold Spring Harb Perspect Biol 3: a006833. 9. Will RG, Ironside JW, Zeidler M, Cousens SN, Estibeiro K, et al. (1996) A new variant of Creutzfeldt-Jakob disease in the UK. Lancet 347: 921–925. 10. Alper T, Cramp WA, Haig DA, Clarke MC (1967) Does the agent of scrapie replicate without nucleic acid? Nature 214: 764–766. 11. Griffith JS (1967) Self-replication and scrapie. Nature 215: 1043–1044. 12. Bolton DC, McKinley MP, Prusiner SB (1982) Identification of a protein that purifies with the scrapie prion. Science 218: 1309–1311. 13. McKinley MP, Bolton DC, Prusiner SB (1983) A protease-resistant protein is a structural component of the scrapie prion. Cell 35: 57–62. 14. Prusiner SB (1982) Novel proteinaceous infectious particles cause scrapie. Science 216: 136–144. 15. Oesch B, Westaway D, Wälchli M, McKinley MP, Kent SB, et al. (1985) A cellular gene encodes scrapie PrP 27-30 protein. Cell 40: 735–746. 16. Chesebro B, Race R, Wehrly K, Nishio J, Bloom M, et al. (1985) Identification of scrapie prion protein-specific mRNA in scrapie-infected and uninfected brain. Nature 315: 331–333. 17. Basler K, Oesch B, Scott M, Westaway D, Wälchli M, et al. (1986) Scrapie and cellular PrP isoforms are encoded by the same chromosomal gene. Cell 46: 417–428. 18. Pan KM, Baldwin M, Nguyen J, Gasset M, Serban A, et al. (1993) Conversion of alpha-helices into beta-sheets features in the formation of the scrapie prion proteins. Proc. Natl. Acad. Sci. U.S.A. 90: 10962–10966. 19. Jarrett JT, Lansbury PT (1993) Seeding “one-dimensional crystallization” of amyloid: a pathogenic mechanism in Alzheimer's disease and scrapie? Cell 73: 1055–1058. 20. Silveira JR, Raymond GJ, Hughson AG, Race RE, Sim VL, et al. (2005) The most infectious prion protein particles. Nature 437: 257–261. Generation of PrPSc 29 Hamsters were inoculated intracerebrally or intraperitoneally, depending on the strain to be inoculated. All inocula were prepared to 2 % in phosphate buffered saline (PBS) from brains taken from clinically affected hamsters, infected with Drowsy (Dy) or 263K. The intracerebral incoculation was carried out in animals of 4 weeks of age under anesthesia; the anesthetic solution was Xylazine-Ketamine (1.5 µg/µl-13.5 µg/µl respectively; 10-14 µl/g hamster) via intraperitoneal injection. Then, 20 µl of inoculum were injected in the frontal lobe of brain. Hamsters started to develop clinical signs and were culled within 70-90 days. The intraperitoneal inoculation was in animals with more than one month of age. By this via 100 µl of inoculum were injected, but only 263K strain can be inoculated in this way, because Dy strain does not lead to development of the disease when is inoculated intraperitoneally. Hamsters started to have clinical signs and were euthanized within 90-110 days. The disease has two phases, latency phase where the hamster is incubating the disease, without clinical signs; and the clinical phase when the hamster exhibits different signs such as aggressiveness, ataxia, reactivity, head bobbing and tremor, until dead unless it is humanely culled. In the terminal phase, there is enough protein accumulated in the brain and hamster is killed with CO2 euthanasia method, and dissected for the brain extraction. Hamster brains were collected and kept frozen at -80 ºC until use. 2.2.2 GPI-anchorless PrP transgenic mice To study the GPI (glycosylphosphatidylinositol) function on pathogenesis of prion disease, GPIanchorless PrP transgenic mice were generated. These mice were built using PrP (-/-) mice and their level of PrP mRNA expression in the brain in the heterozygous lines, was half that in C57BL/6 control mice [10]. PrP encoded by the transgene lacks the GPI anchor and contains the amino acids (23-232), like wild-type PrP after removal of the signal peptide and the Cterminal sequence, but with one amino acid serine (S) more in its sequence (see Chapter 6, appendix I). After scrapie infection, heterozygous mice expressing only anchorless PrP developed high titres of prion infectivity, however clinical signs were not seen after more than 600 days of incubation. Homozygous anchorless PrP transgenic mice expressed two-fold more anchorless PrP than heterozygous and scrapie infection induced a fatal disease after 300 to 480 days post-infection [11]. For my study, these transgenic mice expressing PrP lacking the GPI membrane anchor are very interesting because in the absence of the GPI anchor, PrP is not properly glycosylated by the cell machinery and therefore the PrPC protein is mainly in the unglycosylated form. Thereby, I can work without the peptide heterogeneity resulting from the GPI and the carbohydrates, and I can use techniques previously impossible to implement due to this feature of the PrP. Transgenic heterozygous GPI-anchorless PrP mice (tg44+/-) were generously provided by Bruce Chesebro, from the Rocky Mountain Laboratories, NIH, Montana, USA. Unfortunately, due to logistical problems, the females provided did not breed and that as they grew too old, we decided to cross the heterozygous GPI-anchorless PrP males with wild-type PrP (+/+) females Chapter 2 30 on the C57BL/10 background, in order to obtain homozygous GPI-less PrP line. From this crossing, two genotypes were obtained with or without the transgene and the endogenous mouse PrP gene. Then, the mice with the transgene (GPI-less) and the endogenous PrP allele were crossed to obtain homozygous GPI-anchorless mice. These mice express two anchorless PrP transgene alleles and no wild-type mouse PrP allele. Homozygous GPI-less mice were interbred to create additional mice for experimentation and to maintain the transgenic homozygous GP-anchorless line (Figure 2.1 A). Figure 2.1. A. GPI-anchorless transgenic mice family tree. The different generations and crossbreeds to obtain the homozygous GPI-less mice are represented. B. Mice genotyping corresponding to the each different cross, images of the PCR are shown with the three sets of primers. Mice were genotyped using three DNA PCR test on tail DNA. The DNA was extracted by using the Realpure DNA genomic extraction kit (Durviz, s.l.u, Valencia, Spain), according to the manufacturer’s instructions. Primers for GPI-anchorless transgene: upper oligo 624 (5’AACCGTTACCCACCTCAGGGT3’), lower oligo 2037 (5’CAGGGCGCCTCGAGACGC GTCA3’). The set of oligonucleotide primers for PrP null allele specific for a portion of the neomycin resistance gene: upper oligo 1179 (5’GATGGATTGCACGCAGGTTC3’) and lower oligo 1180 (5’TTGAGCCTGGCGAACAGTTC3’). Endogenous mouse PrP was amplified using: upper oligo 2057 (5’CCAAGGAGGGGGTACCCAT3’), lower oligo 2038 (5’TCCCACGATCAGGAAGATGAG3’) [10]. The PCR reactions were performed following to the manufacturer’s instructions, in a volume of 10 µl containing 0.1 units of PrimeSTAR HS DNA Polymerase (Takara Bio Inc., Otsu, Japan) and 100 ng of DNA. Amplification was carried out by cycling 27 times for 1 min at 94 ºC, 30 sec at 60 ºC and 1 min at 72 ºC. The PCR products were separated in agarose gels to 1.2 % in 1X TAE buffer (Figure 2.1 B). It is worth noting that I have had overlap problems between primers, as can be seen in figure 2.1 B; the primer set of null allele (ko) is able to amplify the GPI-anchorless transgene (PrnpGpi-). I have not found a satisfactory explanation for this and I have tried to solve that problem without Generation of PrPSc 31 success. However, due to the kind of crossbreeds, after the F0 generation it is not necessary to differentiate the null allele and the GPI-less trangene in the same mouse, on any occasion. Then, I decided to continue the safe crossings without confusion. Only female mice were chosen for inoculation to avoid fighting problems leading to premature deaths. Mice at the age of six weeks, were anesthetized, via intraperitoneal injection, with Avertin (2,2,2 tribromoethanol 99 % (12.7 mg/ml) and tert-amyl alcohol (7.8 µl/ml); 1517 µl/g mouse) and subsequently intracerebrally inoculated in the right temporal lobe, with 20 µl of 2 % RML-infected brain homogenate (RML [12] is a mouse-adapted scrapie isolate which was passaged in Swiss CD-1 mice obtained from Charles River Laboratories), kindly provided by Juan Maria Torres, CISA, Madrid, Spain. At 365 days post inoculation, mice, which did not show any scrapie-associated signs, were euthanized and their brains immediately were harvested and stored at -80 ºC until use. Figure 2.2. Accumulation of PrPSc in the GPI-anchorless mice. A. Western blot of brain homogenate from scrapieinfected GPI-anchorless PrP mouse before and after digestion with proteinase K (25 !g/ml.). Samples were detected with a SAF83 antibody, which epitope is unknown. The two bands in the PrPSc treated with PK correspond with the mono and noglycosylated form (from higher to lower size). B. Mortality curves of wild-type mice (C57BL/6) inoculated with 2 % of infected GPI-less brain homogenate (green line) and a negative control of wild-type mice inoculated with PBS (red line). Due to lack of scrapie signs, which on the other hand agrees with what has been previously described [10][11], I decided to determine whether PrPSc had been accumulated in the mice brains. Brain homogenate to 10 % of infected GPI-less mouse was analyzed by inmunoblotting, after digestion with 25 µg/ml of proteinase K (PK), for 1 h at 37 ºC; and PrPres was detected in these transgenic mice inoculated with RML scrapie (Figure 2.2 A). At the same time I tried to detect PrPres depositions in brain tissue. Thus immediately after sacrifice of infected GPI-less mouse, the brain was extracted and formalin fixed. Fixed brains were shipped to CRESA, Barcelona, and a histopathology study performed by Enric Vidal. Transversal sections of the brain were obtained at the level of the optic chiasm, piriform cortex and medulla oblongata. Samples were dehydrated through increasing alcohol concentrations, through xylene and then Chapter 2 32 paraffin-embedded. Four micrometer sections were obtained and haematoxylin-eosin stained for morphological evaluation. Further slides were mounted in 3-trietoxysilil-propilaminecoated glass slides for immunohistochemical (IHC) studies against PrPSc. Briefly, deparaffinised sections were immersed in formic acid and boiled at low pH in a pressure cooker, with endogenous peroxidases blocked. After pre-treatment with PK, the sections were incubated overnight with the primary antibody, which was the anti-PrP mAb 6H4 (1:2000, kindly provided by Prionics AG), and, finally, developed using DAKO EnVision system and 3,3’diaminobenzidine as the chromogen substrate. Then, hyaline deposits were observed arranged radially around blood vessels. Those deposits were strongly immunoreactive to PrP monoclonal 6H4 antibody. Deposits were also located submeningeally, subventricularly and scattered in the neuropil (Figure 2.3 A). Figure 2.3. Histopathological and immunohistochemical analyses. Sections are from the brain at the level of the hippocampus and each section was stained with haematoxylin and eosin (HE) or prepared for immunohistochemical analysis using the 6H4 antibody (PrP). Scale bars 50 !m. A. GPI-less mice inoculatedwith RML strain. In the HE hyaline deposits were observed, arranged radially around blood vessels. The deposits showed strong immunoreactivity to 6H4 antibody against prion protein. Most of the PrP immunoreactive deposits were perivascular although in some of them no central vessel was seen. B. Wild-type PrP mice (C57BL/6) inoculated with GPI-less PrPSc. In the HE, the image shows a moderate spongiosis. In immunohistochemical analysis there are deposits of PrPres with diffuse granular appearance. Amyloid plaques are not observed. C. Negative control. Wild-type PrP mice (C57BL/6) inoculated with PBS. HE slides do not present lesions. There is not immunoreactivity to 6H4 antibody against prion protein. It is possible to observed the blood vessel. Furthermore, to be absolutely sure that the material was an infectious protein, I tested whether these GPI-anchorless mice could replicate the scrapie agent. Then, ten wild-type PrP (C57BL/6) mice were inoculated with 20 µl of 2 % infected GPI-less brain homogenate. All of Generation of PrPSc 33 them developed clinical scrapie and died within 154±15 days post-inoculation (Figure 2.2 B). Infected GPI-less brain homogenate inoculum behaves in a similar way to RML inoculum [12], which has typical incubation times of !150 days in wt mice. Further, PrPSc was detected in brains of these second-passage wt animals and histopathology, performed by Enric Vidal (as seen above), showed typical RML lesion profile (Figure 2.3 B), a rather generalised spongyform appearance caused by vacuolation and the absence of amyloid plaques. 2.3 Isolation of PrPSc A variety of studies were carried out with the PrPSc. Depending on the technique to be used, PrPSc was obtained by two different ways. 2.3.1 Brain homogenate In some experiments, working with the purified protein is not crucial or even ideal; we can study directly the PrPSc from the brain. Therefore, for convenience and whenever the technique allows it, brain homogenate (unpurified protein) was used. To prepare the brain homogenate, hamster and mouse infected brains were homogenized using a dounce homogenizer (Wheaton Industries Inc, NJ, USA), to 10 % (w/v) in phosphate buffered saline (PBS) and 5 % of sarkosyl. Furthermore, to clarify the homogenate it was treated with one pulse of sonication with a probe ultrasonic homogenizer (Cole Parmer Instrument CO., Chicago IL, USA). 2.3.2 Isolation of GPI-anchorless PrPSc To analyze the protein with high-resolution techniques it is essential to isolate the PrPSc, because it is necessary a high grade of purity. The isolates, lacking glycans and GPI anchor (see above), were obtained from brains of infected GPI-anchorless mice. GPI-anchorless PrPSc was isolated using a slightly modified version of the method of Baron GS. et al. [13] (Figure 2.5) based on the special property that anchorless PrPres is accumulated in large amyloid deposits. 10 % brain homogenate (w/v) in PBS was adjusted, to a final volume of 12 ml, to contain 50 mM Tris-HCl (pH 8.0), 137 mM NaCl and 2 % sarkosyl (TBSS) and incubated for 20 min at room temperature (RT). Nucleic acids were digested by treatment with Benzonase (25 U/ml) at 37 ºC for 30 min. The sample was centrifuged at 18,300 rpm, for 24 min at 20 ºC, in a Beckman Type 50.2 rotor. The supernatant was discarded; PrPSc and other detergent-insoluble material were in the pellet. After gently rinsing twice with TBSS, the pellet was resuspended, using the vortex, in a total volume of 5.6 ml of TBSS. Then, the material is digested with 10 µg/ml of PK at 37 ºC for 1 h. Protease digestion is terminated by addition of Pefabloc to 1 mM and incubated on ice for 10 min. Insoluble material, settled by gravity or visible in suspension, was removed by manual aspiration, with the end of a pipette tip, at any Chapter 2 34 stage of the procedure from herein. EDTA to 30 mM was added and incubated for 5 min at RT, after that NaCl to 1.7 M final concentration was added. Subsequently, the preparation was layered onto a sucrose cushion (1 M sucrose, 100 mM NaCl, 0.5 % sulfobetaine 3-14, 10 mM Tris-HCl (pH 7.4) at 20 ºC) of 1/3 the volume of the sample and PrPSc was retrieved by centrifugation at 18,300 rpm, for 24 min at 4 ºC, in a Beckman Type 50.2 rotor. After careful and complete removal of the supernatant and cleaning with cotton swabs the walls of the centrifugation tube, the pellet was resuspended in a total volume of 0.5 ml of 0.5 % sulfobetaine 3-14 in PBS. To wash the sample, it was briefly homogenized with a 0.5 ml dounce homogenizer (Wheaton Industries Inc, NJ, USA) and was recovered by centrifugation at 14,000 rpm (22,000 g) for 20 min at 4 ºC in a microfuge, this washing step is repeated once more. Depending on the kind of the study, the final GPI-anchorless PrPSc pellet was resuspended in 100 µl of deionised water or in 20 µl of 6 M guanidine solution. The stock suspension thus prepared was stored at 4 ºC. Its purity was assessed by SDS-PAGE with Coomassie blue staining and estimated to be 90 % (Figure 2.4). The yield of anchorless PrPSc was ~35 µg per brain, it was determined by BCA protein assay [14]. Figure 2.4. Characterization of isolated GPI-anchorless PrPSc. 10 !l of sample were loaded and separated in a 15 % gel by SDS-PAGE. The purity was assessed by Coomassie blue staining. The molecular weight of the GPI-less PrP27-30 is !16750 Da. For electron microscopy studies, I had to improve the GPI-anchorless PrPSc isolated method. Achieving a complete degradation of lipids was necessary because of lipids settle like a blanket covering the fibers; therefore disturbing the analysis of the PrPSc fibers by electron microscopy. To solve the problem, a final step was added to reach the complete degradation of lipids (Figure 2.5). The final pellet of purified PrPSc was resuspended in 100 µl of deionised water and was treated with lipase at 1 µg/ml, 2 h at 37 ºC. To trap the fatty acids released by the lipase during the incubation, bovine serum albumin (BSA) was added to a final concentration of 10 mg/ml. As fatty acid laden-BSA is soluble, the sample was centrifuged at full-speed (22,000 g) for 20 min and the pellet, containing PrPSc fibers, was resuspended in 100 µl of deionised water. The sample was centrifuged again as above to eliminate completely the BSA. Finally, the pellet was resuspended in 100 µl of deionised water and sonicated three pulses at amplitude of 50 % with a probe ultrasonic homogenizer (Cole Parmer Instrument CO., Generation of PrPSc 35 Chicago IL, USA). With these steps, ultra clean and homogeneous GPI-less PrPSc fibers were obtained for optimal analysis (Figure 2.6). Figure 2.5. Diagram of the isolation of GPI-anchorless PrPSc. Depending on the kind of study, the purification may proceed in three different ways until to have the isolated GPI-anchorless PrPSc. Chapter 2 36 Figure 2.6: Transmission electron microscopy (TEM) images, comparing different ways to prepare anchorless scrapie fibers at the end of the isolation procedure. A. The isolated GPI-anchorless PrPSc was resuspended in 100 !l of deionised water. There are lot of fibers, but they are very crowded together and covered with what I believe to be lipidic material. B. The same sample was sonicated two pulses with a probe ultrasonic homogenizer. They are more scattered but still partially covered with gooey, putatively lipidic, material. C. The isolated GPI-anchorless PrPSc was treated with lipase and BSA. The image show a ball made up of fibers. D. The sample was sonicated three pulses to break up the ball. A lot of small very clean and homogeneous fibers were obtained. 2.4 Concluding remarks Currently, for the study of the prion structure it is necessary to induce the disease in laboratory animals, mainly in mice and hamsters, to obtain enough protein to work with. There is a considerable interest in the field to obtain artificially infectious prion proteins that mimic the same properties of natural prion strains, in large amounts. This objective has been encouraged since the publication of the article of Wang et al. [8]. Hence, despite the difficulty we are working to generate high-quality synthetic prions from recombinant protein. Generation of PrPSc 37 In the meantime, transgenic mouse models have provided valuable insight on the prion diseases. GPI-anchorless PrP transgenic mice, expressing only anchorless PrP molecules, are very useful to study the role of the glycosylphosphatidylinositol in prion pathogenesis. Furthermore, due to the minimal amount of carbohydrates and the absence of the GPI, these mice are very valuable for the analyses of the effect of glycans and GPI anchor on PrPres structure [13] and to study the structure of prion itself [15]. Nevertheless there are very few studies that use the GPI-less animal model to find out more information about this atypical protein. Different GPI-anchorless PrPSc isolation techniques have been described in the literature [13][16]. After several purification attempts to find the best-adapted method to our studies, I decided to use an adaptation of Baron et al. method. It was necessary to improve this purification system because in our microscopy studies, fibers have to be spotless, we cannot have any artefact disturbing the analytical process. I tried different strategies to clean the fibers without success, like washing with methanol/chloroform or heating. Finally, I decided to treat the isolated GPI-anchorless PrPSc with lipase from porcine pancreas, because it is an enzyme that catalyzes the hydrolysis of lipids. Then, the enzyme degraded all these lipid layers that covered up and interspersed between the fibers, obtaining the perfect sample to analyze by microscopy. It can be foresee that the combination of these different ways to generate prion protein and the improved purification method will allow the application of new techniques and therefore the obtaining of new insights on the structure of PrPSc. 2.5 References 1. Legname G, Baskakov IV, Nguyen H-OB, Riesner D, Cohen FE, et al. (2004) Synthetic mammalian prions. Science 305: 673–676. 2. Colby DW, Giles K, Legname G, Wille H, Baskakov IV, et al. (2009) Design and construction of diverse mammalian prion strains. Proc. Natl. Acad. Sci. U.S.A. 106: 20417–20422. 3. Makarava N, Kovacs GG, Bocharova O, Savtchenko R, Alexeeva I, et al. (2010) Recombinant prion protein induces a new transmissible prion disease in wild-type animals. Acta Neuropathol. 119: 177–187. 4. Colby DW, Wain R, Baskakov IV, Legname G, Palmer CG, et al. (2010) Protease-sensitive synthetic prions. PLoS Pathog. 6: e1000736. 5. Makarava N, Kovacs GG, Savtchenko R, Alexeeva I, Budka H, et al. (2011) Genesis of mammalian prions: from non-infectious amyloid fibrils to a transmissible prion disease. PLoS Pathog. 7: e1002419. 6. Deleault NR, Harris BT, Rees JR, Supattapone S (2007) Formation of native prions from minimal components in vitro. Proc. Natl. Acad. Sci. U.S.A. 104: 9741–9746. Chapter 2 38 7. Barria MA, Mukherjee A, Gonzalez-Romero D, Morales R, Soto C (2009) De novo generation of infectious prions in vitro produces a new disease phenotype. PLoS Pathog. 5: e1000421. 8. Wang F, Wang X, Yuan C-G, Ma J (2010) Generating a prion with bacterially expressed recombinant prion protein. Science 327: 1132–1135. 9. Kimberlin RH, Walker C (1977) Characteristics of a short incubation model of scrapie in the golden hamster. J. Gen. Virol. 34: 295–304. 10. Chesebro B, Trifilo M, Race R, Meade-White K, Teng C, et al. (2005) Anchorless prion protein results in infectious amyloid disease without clinical scrapie. Science 308: 1435–1439. 11. Chesebro B, Race B, Meade-White K, LaCasse R, Race R, et al. (2010) Fatal transmissible amyloid encephalopathy: a new type of prion disease associated with lack of prion protein membrane anchoring. PLoS Pathog. 6: e1000800. 12. Chandler RL (1961) Encephalopathy in mice produced by inoculation with scrapie brain material. Lancet 1: 1378–1379. 13. Baron GS, Hughson AG, Raymond GJ, Offerdahl DK, Barton KA, et al. (2011) Effect of glycans and the glycophosphatidylinositol anchor on strain dependent conformations of scrapie prion protein: improved purifications and infrared spectra. Biochemistry 50: 4479–4490. 14. Raymond, G. J., and Chabry, J. (2004) Purification of the pathological isoform of prion protein (PrPSc or PrPres) from transmissible spongiform encephalopathy-affected brain tissue. In Techniques in Prion Research (Lehmann, S., and Grassi, J., Eds.) pp 16-26, Birkhauser Verlag, Basel, Switzerland. 15. Smirnovas V, Baron GS, Offerdahl DK, Raymond GJ, Caughey B, et al. (2011) Structural organization of brain-derived mammalian prions examined by hydrogen-deuterium exchange. Nat. Struct. Mol. Biol. 18: 504–506. 16. Sim VL, Caughey B (2009) Ultrastructures and strain comparison of under-glycosylated scrapie prion fibrils. Neurobiol. Aging 30: 2031–2042. Limited proteolysis 47 peptide size is approximate, in the Tricine-SDS-PAGE the apparent sizes are more accurate. Further, I used three sets of overlapping MW markers: Peptide Molecular Weight (SigmaAldrich, St. Louis, MO, USA), Kaleidoscope Prestained Standard (BioRad, Hercules, CA, USA) and Novex Sharp Protein Standard (Invitrogen, Carlsbad, CA, USA), so as to be able to make mass assignments as accurate as possible. 3.1.3 Outline of experimental approach Figure 3.5. Experimental design scheme. PrPSc is obtained from infected brains. Protein can be purified or can be used as brain homogenate. The isolated protein is resuspended in denaturing buffer and is analyzed by MALDI-TOF. The limited proteolysis experiment is carried out with the brain homogenate; sample is treated with proteinase K at different ratios or different incubation times, at 37 ºC. Then, sample is deglycosylated and precipitated. Next, Tricine-SDSPAGE technique is used to resolve the small peptides and finally by Western blot the different C-terminal fragments are detected. Chapter 3 48 3.2 Studies on GPI-anchorless PrPSc 3.2.1 Identification of PK cleavage sites by Western blot To resolve the different peptides formed after digestion with PK, in particular small ones, with the best possible resolution, I used a Tricine-SDS-PAGE technique [4]. I found that it was necessary to precipitate the sample before resuspending it in the high-glycerol sample buffer in order to get reproducible separations. GPI-anchorless PrPSc brain homogenate was treated with 25 µg/ml of PK; despite GPI-less PrPSc being mainly in the unglycosylated form, the sample was further deglycosylated with PNGase F, so as to eliminate any interference from glycosylated peptides when assigning molecular weight values to the peptides detected. N-terminal antibody #51, which recognizes epitope G92-K100, detected just one intense, wide band with an approximate MW of 17 kDa, (Figure 3.6, lane 1) which suggests that there are different cleavage points (hence the wide appearance of the band) around residue 89. Very importantly, the fact that only one band is detected by this antibody also means that there are no other Cterminally truncated fragments containing this epitope. In the Western blot probed with the W226 antibody (Figure 3.6, lane 2), which is specific for the W144-N152 region, four bands with approximate MWs: 17, 14.6, 13 and 12 kDa were detected, suggestive of three additional cleavage points in the region between residues ~89 and 145. The C-terminal R1 antibody, which recognizes the epitope Y225-S230 (Figure 3.6, lane 3), detected seven different bands with approximate MWs of 17, 14.6, 13, 12, 10.2, 8 and 6.7 kDa, suggestive of yet three additional cleavages resulting in N-terminal truncation around residues 149, 164 and 175. Figure 3.6. PK-resistant fragments in brain homogenate from scrapie-infected GPI-anchorless tg mouse. Samples were treated with 25 !g/ml of PK and subsequently deglycosylated with PNGase F. Samples were resolved on Tricine-SDSPAGE. PrP was detected with different monoclonal antibodies: lane 1: #51 (epitope 92-100); lane 2: W226 (epitope 144-152); lane 3: (epitope 225-230). Limited proteolysis 49 3.2.2 Identification of PK cleavage sites by mass spectrometric detection To get the exact masses from PK-resistant peptides, and therefore unequivocally indentify them, I used mass spectrometry technique. Then, since it is necessary to isolate the sample to avoid impurities, I needed to be sure that the PK-resistant fragments obtained from isolated PrPSc are similar to those obtained from brain homogenates. It was found to be the case (Figure 3.7), all the bands appear at the same molecular weight. Figure 3.7. PK-resistant fragments in brain homogenate (lane 1) and purified GPI-anchorless PrPSc (lane 2). Both samples were digested with proteinase K, 25 !g/ml and 10 !g/ml, respectively and treated with PNGase F. Samples were resolved on Tricine-SDS-PAGE gels and PrP was detected with the C-terminal R1 antibody, which recognizes the epitope 225-230. Firstly, samples were analyzed by nano-LC-ESI-Qq-TOF MS, which provides high mass accuracy, and two peaks were identified with exact molecular masses: 17148 Da and 16371 Da, which correspond to peptides Gly81-Ser232 and Gly89-Ser232, respectively (Chapter 6 - appendix II). It should be noted that Ser232 is the C-terminus of GPI-less PrPSc, which features a stop codon after the triplet encoding this residue (see appendix I in Chapter 6 and [2]). However, despite several attempts the smaller peptides did not ionize well under the ESI conditions. Therefore, I decided to use MALDI-TOF as an alternative analytical method. Given that the mass range of PK-resistant peptides to be analyzed is very wide, spanning between 17 and 6 kDa, a complex calibration method covering the entire range was developed, and providing acceptable mass accuracy across it. For this purpose, an iterative calibration approach was used. First, insulin (m/z = 5733), ribonuclease A (m/z = 13682) and lysozyme (m/z = 14305) were used mixed with the sample, as internal standards, however, they suppressed some of the less abundant peaks in the spectrum. Therefore, they were used to calibrate a peak with a mass of Chapter 3 50 9573 Da, with which they did not interfere, corresponding to Met153-Ser232 (Chapter 6 - appendix II). Then, the 9573, 16371 and 17148 Da peaks were used to calibrate the rest of peaks in the spectrum. Figure 3.8. MALDI spectrum of PK-treated GPI-anchorless PrPSc isolated from scrapie-infected GPI-anchorless mouse. During the process of isolation the sample was digested with 10 !g/ml of PK for 1 h, at 37 ºC. Purified GPIanchorless PrPSc was resuspended in 20 !l of 6 M GndHCl, to a final concentration of 1.75 !g/!l. The sample was analyzed by MALDI-TOF, mixing 1 !L with 49 !L of sinapinic acid (SA) as matrix. Asterisks mark doubly-charged ions from peptides with m/z 16371 and 17148. Unmarked peaks could not be identified due to low resolution in the >16 kDa region. Figure 3.8 presents the MALDI-TOF spectrum. Thirteen peaks of different masses (m/z) were detected: 17148, 16726, 16371, 13606, 13463, 12173, 12041, 11171, 9687, 9573, 8358, 7436 and 6274. They match quite well with the bands identified by Western blot analysis (Table 3.I). Based on the theoretical molecular weight calculated from the sequence of the mouse PrP, using the GPMAW 6.0 software (Lighthouse, Odense, Denmark), ten of these peaks, correspond to peptides spanning from positions 81, 85, 89, 116, 118, 133, 134, 141, 152 and 153 to the C-terminus (Ser232). Besides, three peaks corresponding to new cleavage points in the C-terminal region, at positions 162, 169 and 179, were detected. No peaks corresponding to C-terminally truncated peptides were found (Figure 3.6), in agreement with the presence of just one band being detected by antibody #51, which recognizes the aminoterminally located epitope G92-K100. Limited proteolysis 51 Table 3.I. PK-resistant fragments in GPI-anchorless PrPSc WESTERN BLOT MALDI-TOF Band kDa Peak (Da) Theoretical mass (Da) Cleavage point Peptide 1 17 17148 17148 81 G81 - S232 16726 16729 85 G85 - S232 16371 16371 89 G89 - S232 2 14.6 13606 13605 116 A116 - S232 13463 13463 118 G118 - S232 3 13 12173 12172 133 M133 - S232 12041 12041 134 S134 - S232 4 12 11171 11172 141 G141 - S232 5 10.2 9687 9688 152 N152 - S232 9573 9574 153 M153 - S232 6 8 8358 8358 162 Y162 - S232 7 6.7 7436 7436 169 S169 - S232 6274 6278 179 V179 - S232 3.2.3 Kinetics of PK digestion Next, I carried out a study to addresses the question of observing the kinetic of appearance of the different PK-resistant bands observed by Western blot. Infected brain homogenate was treated with 25 µg/ml of PK at 37 ºC, for 0, 30, 60, 120, 180, 240, 300 and 360 minutes. Samples were deglycosylated and analyzed by Tricine-SDS-PAGE. Gradual disappearance of all PK-resistant bands was observed; however, the kinetics of disappearance was different for different bands. The intensities of the 17, 14.6, 13, 12, and 6.7 kDa bands decreased steadily over time, with faster relative decreases for 14.6, 13 and 12. Chapter 3 52 On the other hand, the 10.2 and 8 kDa bands, exhibited a much slower decrease in intensity. Thus, by 240 minutes the intensity of the 17 and 10.2 kDa bands equaled and by 360 minutes the intensities of the 17, 10.2 and 8 kDa bands were virtually similar (Figure 3.9). Figure 3.9. Kinetic of PK digestion of PrPSc in scrapie-infected GPI-anchorless tg mouse brain homogenate. Samples were digested with PK (25 !g/ml) and the reaction stopped after 0, 30, 60, 120, 180, 240, 300 and 360 minutes. Samples were treated with PNGase F, subjected to Tricine-SDS-PAGE. After blotting, membrane was probed with R1 antibody (epitope 225-230). 3.2.4 PK cleavage analysis after partial unfolding In an attempt to determine the relationship of the resistant peptides to one another and to extend the map of PK-sensible cleavage sites of PrPSc, brain homogenate containing GPIanchorless PrPSc was treated with different concentrations of guanidine before cleavage analysis. Kocisko et al. have shown that partial unfolding of PrPSc with up to 2.5-3 M guanidine is reversible upon dialysis [10]. GPI-less samples were treated with 0.5 M, 1 M, 2 M, 3 M or 4 M of guanidine for 1 h at 37 °C and then diluted to 0.4 M to reduce guanidine concentration before PK digestion at 25 µg/ml (Figure 3.10). As the concentration of guanidine increases, the conformation of PrPSc is destabilized and the molecule is rendered increasingly susceptible to proteolytic digestion, however, the relative abundance of different PK-resistant bands changed conspicuously after treatment with different concentrations of guanidine. Beyond 1 M, the 10.2 and, to a lesser extent, 12 and 8 kDa bands, corresponding to peptides 152/153-232, 141-232, and 162-232 became predominant. Beyond 3 M guanidine, hardly any PK-resistant material was left (Figure 3.10). Of note, just as seen in the PK kinetic analysis, no new cleavage sites were detected. The apparent higher global intensity of all bands seen after treatment with 0.5 M guanidine, as compared to the 0 M guanidine sample, was reproducibly seen in repeated experiments. I interpret it as an effect of Limited proteolysis 53 guanidine to solubilize peptides that otherwise probably stick to the plastic walls of the eppendorf tube throughout the incubation period. Figure 3.10. Effect of guanidine-induced partial unfolding of GPI-anchorless brain homogenate PrPSc on its digestion by PK. Samples were treated with 0 M, 0.5 M, 1 M, 2 M, 3 M and 4 M of guanidine; then they were diluted to a final concentration of 0.4 M of guanidine and digested with 25 !g/ml of PK for 1 h at 37 ºC, and deglycosylated with PNGase F. Samples were resolved on Tricine-SDS-PAGE and PrP was detected with C-terminal R1 antibody (epitope 225-230). 3.3 Studies on Syrian hamster PrPSc (SHaPrPSc) 3.3.1 Identification of PK cleavage sites in two strains of SHaPrPSc by Western blot In order to extend our previous study [1], studying the C-terminal region of the wild type PrPSc, PK-resistant fragments of SHaPrPSc were resolved by Tricine-SDS-PAGE [4] system, obtaining a better resolution of the bands and predicting new cleavage points and then, new flexible areas. I compared two hamster PrPSc strains, 263K and Drowsy (Dy) that differ in the size of their main PK resistant cores (PrP27-30), 20/21 and 19 kDa respectively, incubation time of the disease, clinical signs, and tropism for specific brain areas. 263K and Dy hamster PrPSc brain homogenates were treated with 50 µg/ml and 30 µg/ml of proteinase K, respectively, this difference in the PK concentrations is because Dy strain is more susceptible to digestion that 263K. Subsequently the samples were deglycosylated with PNGase F and were probed with three different antibodies. N-terminal antibody 3F4, which recognizes epitope 109-112 detected just one intense, wide band in both strains with an approximate MW of 20/21 an Chapter 3 54 19 kDa, respectively (Figure 3.11 A) which suggests that there are different cleavage points around residue 90 and excludes the presence of significant amounts of doubly (N-, and Cterminal) cut bands. The Western blot probed with the W226 antibody (Figure 3.11 B), which is specific for the 145-153 region, presents a different pattern in both strains, although this difference could be attributed to the affinity of the antibody. In 263K two bands were observed, with an approximate MW: 20/21 and 17 kDa, however, three bands were detected in Dy, with an approximate MW of 19, 17 and 15 kDa, suggestive of additional cleavage points in the region between residues ~89 and 153. It is known that the 17 kDa band corresponds with the residue 117-119 [1]. The C-terminal R1 antibody, which recognizes the epitope 226-231 (Figure 3.11 C), detected six different bands with an apparent MW of approximately 20/21 or 19, 17, 15, 14, 10.6 and 6.5 kDa, suggestive of there are several cleavage points in the Cterminal region beyond residue 154. It is worth noting that the pattern is the same in both strains, although the intensities between bands are different. Figure 3.11. PK-resistant fragments in brain homogenate from 263K (lane 1) and Dy (lane 2) scrapie-infected hamster. Samples were treated with 50 !g/ml and 30 !g/ml of PK, respectively, and subsequently deglycosylated with PNGase F. Samples were resolved on Tricine-SDS-PAGE and PrP was detected with different monoclonal antibodies. A. N-terminal 3F4 antibody (epitope 109-112). B. W226 antibody (epitope 145-153) and C. Cterminal R1 antibody (epitope 226-231). Given that 20-19 kDa band corresponds to the predominant G86/90-S231 peptide, whose mean MW is !16.2 kDa, the apparent mass contribution of the glycophosphatidylinositol (GPI) anchor is !3.3 kDa. We assume that all bands correspond to peptides containing GPI, because the C-terminal antibody used (R1) recognizes the residues 226-231. Then, subtracting 3.3 Da to all the bands, masses of bands higher than 10.6 kDa fit with cleavage sites at previously identified flexible stretches [1]: the band of 17 kDa would correspond with the residue 117 Limited proteolysis 55 and/or 119, 15 kDa bands with the residue 135 and/or 139, and the band of 14 kDa with the 142 residue. I also observed another susceptibility C-terminal region make up for the 10.6 and 6.5 kDa bands; subtracting the GPI, their mass would be 7.3 and 3.2 kDa, then these bands would correspond to cleavages around positions N170 and M206, respectively. On the other hand, with respect to the difference between strains, we confirm that Dy PrPSc is overall more susceptible to digestion that 263K. Nevertheless, the same cleavage points appear in both, except for the most amino terminal region. Thus they must have very similar tertiary structures. To sum up, these PK-cleavage points analyzed by WB have to get with cautions because this is not an accurate technique and it is not sure whether these PK-resistant peptides have the GPI moiety. Then, I can only get clues to understand better other results. However, it should be noted that there are several similarities with the results obtained from the GPI-anchorless PrPSc. The cleavage points of the bands higher that 10.6 kDa, in addition to being consistent with those seen in the previous study [1], also fit perfectly with the residues observed in the GPI-less PrPSc (as will be explained in the conclusions). Nevertheless, the two C-terminal cleavage sites are not in agreement with the GPI-anchorless PrPSc flexible stretches, but it is impossible to conclude whether the reason is because some bands in SHaPrPSc do not have the GPI anchor and/or correspond with other different cleavage points. I have to mention that several attempts to resolve the GPI issue were made, unsuccessfully. Until I could work with the GPI-anchorless PrPSc, trying to find out with certainty if the bands had the GPI or not, it was a matter of utmost importance; it was necessary to obtain accurate data and design other sort of experiments. Some of my experiments consisted of different strategies. I tried to remove the GPI with 48 % of aqueous hydrofluoric acid, a solution that has been reported to cut the GPI in the phosphodiester bond [14], which would lead to an easily observable electrophoretic migration difference in the bands previously containing GPI. However, this strategy did not work in my hands. Another attempt was carried out with the Streptococcus agalactiae CAMP factor that binds to GPI-anchored proteins, thereby biotinylated CAMP factor, kindly provided by Michael Palmer from the University of Waterloo, Ontario, Canada, was used to probe the blots, followed by probing with HRP-labelled avidin-peroxidase as a secondary antibody to recognize what bands had the GPI anchor [15]. This approach was also unsuccessful because I obtained an awful background impossible to solve and only nonspecific bands were detected, even in the negative control. 3.3.2 Identification of PK cleavage sites in PK-sensitive SHaPrPSc (sPrPSc) by Western blot In 1998 Safar et al. reported the existence of a subset of PrPSc molecules that are completely degraded by PK, which hence were termed, PK-sensitive PrPSc (sPrPSc) [16]. Tzaban et al. later demonstrated for the first time that prion-infected tissues contain sPrPSc molecules that form low molecular weight aggregates [17]. In the last years, a method of isolation of the PK-sensitive Chapter 3 56 PrPSc fraction by centrifugation was developed by the group in which I work, and it was also shown that sPrPSc has PMCA converting activity [18]. It was concluded that sPrPSc is a conformer made up of smaller multimers, whereas PrPres is formed by larger aggregates. However a question still unresolved was raised: are sPrPSc and rPrPSc two populations with different conformations or simply different sized multimers with the same conformation?. To address this question and help to draw conclusions in this study, SHa sPrPSc, provided by Miguel A. Pastrana from our laboratory [18], was further characterized by limited proteolysis coupled to Tricine-SDS-PAGE technique. SHa sPrPSc was obtained from total SHaPrPSc by differential centrifugation at an intermediate speed. Furthermore it has to be noted that sPrPSc has a partial resistance to lower concentrations of PK (Figure 3.12). Figure 3.12. Partial PK resistance of sPrPSc. sPrPScwas treated with increasing concentrations of PK at 37 ºC for 1 h and analyzed by Wester blot, using 3F4 antibody (epitope 109-112). (Pastrana MA et al. (2006) [18]) There are evident similarities of the PK-cleavage pattern between sensitive and total PrPSc when comparing the digested samples by Western blot assay using a C-terminal antibody, R1 (epitope 226-231) after PK-treatment and deglycosylation (Figure 3.13). Besides the intense band corresponding to cleavages around position 90, 6-7 additional lower molecular weight bands are seen in both cases. Furthermore, the relative intensities of some of the bands vary between sPrPSc and total PrPSc. The main band centered at 19-20 kDa smears a bit more into the 2021 kDa region in the sPrPSc sample, the 17 kDa band is more intense in total PrPSc, bands around 15, and finally, bands at 10 and 6 kDa are more intense in sPrPSc, but the overall pattern is very similar. Additionally, comparing the pattern of total and sensitive PrPSc with the resistant PrPSc of 263K strain (Figure 3.11 C, lane 1), it can be observed that the bands are essentially the same, being the main band and the 10.6 kDa band the most intense of all. Therefore, limited proteolysis studies on the sPrPSc fraction of 263K indicate that, as in total and resistant PrPSc, the sensitive fraction is also composed of alternating PK-resistant stretches interspersed with PK-susceptible stretches. The relatively small variations in intensity of some Limited proteolysis 63 10. Kocisko DA, Lansbury PT, Caughey B (1996) Partial unfolding and refolding of scrapieassociated prion protein: evidence for a critical 16-kDa C-terminal domain. Biochemistry 35: 13434–13442. 11. Bartz JC, Bessen RA, McKenzie D, Marsh RF, Aiken JM (2000) Adaptation and selection of prion protein strain conformations following interspecies transmission of transmissible mink encephalopathy. Journal of Virology 74: 5542–5547. 12. Zou W-Q, Capellari S, Parchi P, Sy M-S, Gambetti P, et al. (2003) Identification of novel proteinase K-resistant C-terminal fragments of PrP in Creutzfeldt-Jakob disease. J. Biol. Chem. 278: 40429–40436. 13. Zanusso G, Farinazzo A, Prelli F, Fiorini M, Gelati M, et al. (2004) Identification of distinct Nterminal truncated forms of prion protein in different Creutzfeldt-Jakob disease subtypes. J. Biol. Chem. 279: 38936–38942. 14. Notari S, Strammiello R, Capellari S, Giese A, Cescatti M, et al. (2008) Characterization of truncated forms of abnormal prion protein in Creutzfeldt-Jakob disease. J. Biol. Chem. 283: 30557–30565. 15. Lang S, Xue J, Guo Z, Palmer M (2007) Streptococcus agalactiae CAMP factor binds to GPIanchored proteins. Med. Microbiol. Immunol. 196: 1–10. 16. Safar J, Wille H, Itri V, Groth D, Serban H, et al. (1998) Eight prion strains have PrP(Sc) molecules with different conformations. Nat. Med. 4: 1157–1165. 17. Tzaban S, Friedlander G, Schonberger O, Horonchik L, Yedidia Y, et al. (2002) Proteasesensitive scrapie prion protein in aggregates of heterogeneous sizes. Biochemistry 41: 12868– 12875. 18. Pastrana MA, Sajnani G, Onisko B, Castilla J, Morales R, et al. (2006) Isolation and characterization of a proteinase K-sensitive PrPSc fraction. Biochemistry 45: 15710–15717. 19. Sajnani G, Silva CJ, Ramos A, Pastrana MA, Onisko BC, et al. (2012) PK-sensitive PrP Is Infectious and Shares Basic Structural Features with PK-resistant PrP. PLoS Pathog. 8: e1002547. 20. Dumpitak C (2003) Untersuchungen zu Struktur und Funktion von Polysacchariden und alterungsassoziierten Proteinmodifikationen bei Prionen. Dissertation. Düsseldorf: HeinrichHeine-Universität. 21. Baron GS, Hughson AG, Raymond GJ, Offerdahl DK, Barton KA, et al. (2011) Effect of glycans and the glycophosphatidylinositol anchor on strain dependent conformations of scrapie prion protein: improved purifications and infrared spectra. Biochemistry 50: 4479–4490. 67 4 Electron microscopy Abstract Prion protein has the property to polymerize into amyloid fibers. Taking advantage of this feature, we used different electron microscopy (EM) techniques, from low-resolution to three-dimensional reconstruction of the PrPSc fibers. As the previous studies of this thesis, GPI-anchorless PrPSc was used, because it provides the cleanest possible sample, without losing the natural properties of the wild-type PrPSc. We have obtained dimensions and morphological information about the prion protein protofilaments. Fibers have a no regular helical pattern and are made up for two protofilaments intertwined, with no regular crossover distance. 3D reconstruction of 2D cryomicroscopy (cryo-EM) images of the fibers also shows two protofilaments coiling around a common axis. Moreover, there are 2-2.5 nm repetition densities along the axis, suggesting each of these is a stacked monomer of PrPSc. These axial densities are also evident in reconstructed tomograms obtained by cryo-EM tomography of fiber samples. Another important measure is a 4.8 Å reflection presents in the Fast Fourier Transform (FFT), characteristic of the cross-ß structure. A high-resolution image of cryoEM also shows clearly the 4.8 Å reflection of the spacing between the beta repeat strands that are stacked like a “ladder”. These measurements of the scrapie fibers should lead to develop a real model of the prion structure. Chapter 4 68 4.1 Introduction Electron microscopy (EM) is a technique developed in the 1930s, based on a beam of electrons to create a magnified image of the specimen rather than a beam o light, achieving higher magnifications and a greater resolution. EM has become in a powerful technique to observe the ultrastructure of a wide range of biological and inorganic samples. Furthermore, currently there are different types of EM methods that use several signals arising from the interaction of the electron beam with the sample to obtain information about structure, morphology and composition. Electron microscopy techniques have been very useful for studying the ultrastructural features of prions. In 1981, using transmission electron microscopy (TEM), abnormal fibrillary structures in subfractions of synaptosomal preparations from scrapie affected mice and hamster were observed. These fibers were designated “scrapie-associated fibrils (SAFs)” from scrapie affected mice and hamster [1]. Later, those fibers were also observed in CJD brain fractions [2]. At the same time, Prusiner SB et al., after the purification process of scrapie prions based on centrifugation, could observe that the purified preparations contained numerous rod-shaped particles, measuring 10-20 nm in diameter and 100-200 nm in length. These rods were aggregates of prions and are not fundamental particles themselves; they were termed “prion rods” [3][4]. Knowing these properties of polymerization and aggregation of the PrPSc, into heterogeneous amyloid fibers, more studies were performed. Nevertheless the presence of fibers or amyloid plaques was not always observed; Godsave et al. demonstrated by cryo-inmunogold EM that brain sections from scrapie infected mice do no contain any significant PrPSc fibers [5]. Then, it was concluded that the fiber formation is dependent of the purification method, which is characterized by consecutives steps of ultracentrifugation, protease and detergent treatment [6][7]. On the other hand, GPI-anchorless PrPSc, produced in mice expressing PrP lacking the GPI anchor, is deposited as amyloid plaques into the brain [8]. A recent study using TEM and atomic force microscopy (AFM) has compared PrP27-30 fibers from GPI-anchorless and wild type mice infected with different scrapie strains and the authors obtained dimensions and morphological information about the protofilaments. They have observed that fibers are generally 100-150 nm in length, 4-5 nm in wide and are intertwined. Furthermore, they concluded there are variations in the different strains, ME7 and 22L fibrils contained thinner protofilaments, 22L fibrils preferred left-handed twists, and 22L fibril periodicities averaged 106 nm per half-turn, compared with 64 and 66 nm for RML and ME7 fibrils, respectively [9]. Despite these comprehensive studies, the structure of PrPSc is still poorly understood. Nonetheless, in the last few years, some great technological breakthroughs have significantly contributed to improve the resolution of the different EM techniques. Besides, the Electron microscopy 69 computational developments in conjunction to emerging technologies in EM instrumentation, and novel sample preparations have allowed us to find out in detail new measurements and insights on the structure of the PrPSc. 4.1.1 Outline of experimental approach In these studies, each technique and each result is related to the previous one. They are independent studies, but only with them together it has been possible to reach a threedimensional reconstruction of the PrPSc. Therefore, this chapter is arranged according to this idea. It should also be noted that in this chapter I am going to write in plural, because of the described studies have been performed by several researchers from different laboratories and several fields of knowledge. Without their help and experience it would have been impossible to approach this project. Figure 4.1. Experimental design scheme. GPI-less PrPSc is obtained from RML infected brains of tg mice expressing the GPI-anchorless PrP. The protein is purified and the quality of the fibers is checked by TEM. When the perfect sample is got, the fibers are analyzed by cryo-TEM. Next, to obtain the 3D reconstruction, the 2D projection images obtained by cryo-EM are subjected to helical reconstruction; alternatively, vitrified samples are subjected to electron tomography, and 3D reconstructions are obtained by weighed back projection technique. Studies were carried out with the GPI-less PrPSc from scrapie infected tg mice expressing GPIanchorless PrP; the used EM techniques require the cleanest and most homogeneous fibers of PrPSc available. For the time being this is achievable with the anchorless PrPSc material, since Chapter 4 70 GPI-less PrPSc fibers preformed in the brain of scrapie-infected mice are easy to isolate using mild centrifugation yielding very pure samples. Please refer to chapter 2 for a more detailed discussion. 4.2 Transmission electron microscopy (TEM) Transmission electron microscopy (TEM) involves a high voltage electron beam emitted by a cathode and focused by magnetic lenses. The electron beam that has been partially transmitted through the very thin specimen carries information about the structure of the specimen. The image is magnified by a series of magnetic lenses and it is recorded by an imaging recording system. Figure 4.2. Transmission electron microscopy. A. Transmission electron microscope Philips CM-12. The images are recorded with a MEGA VIEW-II DOCU camera (University of Santiago de Compostela (USC), Galicia, Spain). B. Layout of optical components in a basic TEM. The components are located in a vacuum system, in which the electrons travel. TEM has a limit of resolution of 2 nm. But in the life sciences, the specimen preparation limits the resolution of what we can see in the electron microscope, rather than the microscope itself. The sample should be negative stained, a method in which the specimen is dried and embedded in a layer of electron-dense heavy metal salts to scatter imaging electrons and thus give contrast between different structures, since most biological materials are nearly "transparent" to the electron beam. Then, transmission electron microscopes produce two-dimensional, black and white images. Electron microscopy 71 4.2.1 Results of TEM In this group of experiments, the first aim was to monitor the appearance and quality of GPI-less PrPSc as different modifications to the isolation method were introduced. An ideal sample is, for our purposes a specimen free of bacteria, lipids, other kind of impurities and/or artefacts that might hamper our studies. Furthermore, the more scattered that fibers are, the better, given that the ideal situation for cryo-EM tomography and helical reconstruction from 2D images is one in which individual fibers are available. This is very difficult to achieve, given the tendency of PrPSc fibers to clump together, particularly under centrifugation, a necessary step to isolate them (Figure 4.3). Figure 4.3. TEM images of GPI-anchorless PrPSc fibers. They were analyzed with the transmission electron microscope Philips CM-12 (University of Santiago de Compostela (USC), Galicia, Spain). A-B. Good preparations. They are clean and homogeneous. The scale bar is 200 nm and 1000 nm, respectively. C-D. Bad samples. Fibers are covered by lipids and junk. The scale bar is 200 nm and 1000 nm, respectively. All samples were stained with uranyl acetate. Chapter 4 72 All isolation methods described in the literature were not good enough for the EM surveys [9][10], because of the isolated protein was still having a lot of lipids covering the fibers. I found out a technique to achieve a complete degradation of lipids during the process of isolation of the GPI-less PrPSc. That step was to treat the sample with lipase; it is an enzyme that catalyzes the hydrolysis of lipids (see Chapter 2 - isolation of GPI-anchorless PrPSc). Then, ultra clean and homogeneous GPI-less PrPSc fibers are obtained for optimal analysis (Figure 4.3 A and B). TEM is the best method to observe whether the fibers have achieved the quality required to be further analyzed by cryo-EM technique, because the preparation of the sample is very simple, the results are obtained very fast and the microscope is easy to use. Nevertheless, there is a disadvantage; the negative staining could cause possible distortions of the molecules resulting from the staining/drying procedure; this is why definitive measurements were not made, and the technique was just used as a quality control. Anyway, I had the possibility to analyze the best samples by a more powerful transmission electron microscope at 200 kV but more complicated to use. It was possible to observe with clarity the fiber morphology and how they are intertwined. Sometimes can be seen two twisted fibers and the tendency to form lateral aggregates (Figure 4.4). Although TEM data are not definitive, some measurements were made; the intertwined fibers have 10 nm width, and each protofilament has a wide of 4-5 nm, in agreement with previous studies [9]. Figure 4.4. TEM images of GPI-anchorless PrPSc fibers. They were analyzed with the transmission electron microscope LIBRA 200 FE OMEGA (University of Santiago de Compostela (USC), Galicia, Spain). A-B. Both are good preparation. The scale bar is 100 nm. All samples were stained with uranyl acetate. Electron microscopy 73 4.3 Cryo-transmission electron microscopy (Cryo-TEM) Cryo-transmission electron microscopy (cryo-TEM) or electron cryomicroscopy is a kind of TEM where the sample is studied at cryogenic temperatures. Inside of an automated vitrification robot, a drop of the sample is pipetted onto a grid coated with a thin carbon film previously glow discharged; the grid is rapidly plunged into liquid ethane to form vitreous ice and then imaging the frozen film using an electron microscope. The specimen is maintained continuously below -170 ºC during storage and also during imaging in the electron microscope to prevent the formation of ice crystals [11]. This technique produces many images of the same specimen in different orientations (Figure 4.5). The cryo-TEM images of a protein molecule have much lower contrast than a negative stain and have a lot of noise. However for 3D reconstructions of macromolecules, vitrification of the unstained specimen is the method of choice, due to fact that this technique preserves the specimen virtually free of artefacts, so the morphology and chemistry of the observed sample bears a close resemblance to its in vivo state. Figure 4.5. Cryo-transmission electron microscopy. A. Vitrification robot “VitrobotTM Mark III” (FEI Co., Eindhoven, The Netherlands). B. Schematic representation of cryo-TEM imaging. 4.3.1 Results of cryo-TEM To continue the studies on the structure of the PrPSc, the best GPI-less PrPSc samples obtained in my laboratory were sent to the Dr. Matthijn Vos, from FEI Company, Eindhoven, The Netherlands. He analyzed the samples by cryo-transmission electron microscopy to find out more ultrastructural information of the scrapie prion protein in native conditions and with more Chapter 4 74 powerful microscopes. Note that with cryo-TEM, we will obtain images that allow us to use computational methods to develop an atomic model of the PrPSc. After the meticulous preparation of the sample embedded in a thin layer of vitreous ice, it was loaded into a electron microscope and the GPI-less PrPSc fibers were analyzed using low electron dose, to avoid the radiation damage; the whole grid was screened for areas with suitable ice. The recorded low-magnification images showed clearly a high amount of prion fibers and the most important feature, that can be observed, is that the sample keeps clean and homogeneous (Figure 4.6). Then, the study can continue with confidence. Figure 4.6. Low magnification cryo-EM images of GPI-less PrPSc fibers. The sample was analyzed with the Titan KriosTM microscope and Falcon I direct electron detector (FEI Co., Eindhoven, The Netherlands). A. Scale bar is 500 nm. B. Scale bar is 1 !m. The visible dots are fiduciary gold that serves as references for later studies. In high-resolution cryo-TEM images of the GPI-less PrPSc, after closer examination was observed that the fibers are made up of two thinner protofilaments. Furthermore, their Fast Fourier Transform (FFT), automatically calculated along the length of a fiber, that can be used to determine repeating features in an image, showed a specific 4.8 Å reflection, which indicates a !-sheet stacking along the fiber direction (Figure 4.7). Electron microscopy 81 Figure 4.12. 3D reconstruction of a GPI-less PrPSc fiber. The red surface images are different views of the structure. The bottom images are the top (left) and the bottom (right) views. The scale bar is 100 Å. 4.6 Conclusions In this study, we used different electron microscopy techniques to acquire information on GPIless PrPSc fibers global architecture, protofilament substructure and even location of the cross-! structure within fiber. Our current interpretation, after thorough analysis of the results is that the basic fiber is made up of two intertwined protofilaments (Figures 4.10 and 4.12), each !3-5 nm of wide. This range of width is due to the diameter of each protofilament is variable, as it can be see in the Chapter 4 82 helical reconstruction (Figure 4.12, bottom images). The shape is not circular, it is like a tear of about 2 nm at its narrowest and about 5 nm at its widest. There are repetitions of !2 nm along the axis (Figures 4.10 B and 4.12), this could be interpreted that each density forms one monomer of PrPSc, so 2-2.5 nm axial densities might be individual stacked PrPSc subunits. Furthermore, it is possible to observe that there is no regular crossover distance. Cryo-TEM studies have provided a significant result, the presence of 4.8 Å reflection in FFT, indicating a !- sheet stack along the fiber direction (Figure 4.7); even due to the chance of two fibers were together in the same focal plane, we can see that these 0.5 nm repetitions are stacked like a “ladder” (Figure 4.8). However, no other reflections are observed indicating a head to tail uniform stacking of the basic units. Figure 4.13. Representation of all data together, in the 3D helical reconstruction of GPI-less PrPSc fiber. Scale bar 100 Å. The mammalian prion turns out to be very similar as HET-s prion protein from the filamentous fungus Podospora anserina, as both show densities along the backbone [18]. The HET-s densities are spaced ~1 nm apart while the PrPSc densities are a double length (2-2.5 nm). Moreover HET-s prion stacked monomers consist of solenoids featuring 2 rungs made up of !-strands. In turn, PrPSc has probably to be made up out of 4 rungs (see discussion in next chapter). Looking at the density profile the loop of the prion basic fiber is flatter in structure than the HET-s loop. The HET-s fiber is simple, while the PrPSc almost exclusively bundles 2 protofilaments. Just like HET-S the individual fiber is torqued in the form of a twisted ribbon. For the PrPSc structure two of these ribbons are intertwined. Electron microscopy 83 4.7 Experimental Isolation of GPI-anchorless PrPSc - GPI-anchorless PrPSc was isolated using a modified version of the method of Baron GS et al. [10]. During the purification, total PrPSc was treated with 10 µg/ml of proteinase K. The final pellet of purified PrPSc was resuspended in 100 µl of deionised water and was treated with lipase at 1 µg/ml and bovine serum albumin (BSA) to a final concentration of 10 mg/ml, 2 h at 37 ºC. The sample was centrifuged (in the microfuge) at full-speed for 20 min and the pellet was resuspended in 100 µl of deionised water. The sample was centrifuged again as above to wash completely the BSA. Finally, the pellet was resuspended in 100 µl of deionised water and sonicated three pulses at amplitude of 50 % with a probe ultrasonic homogenizer (Cole Parmer Instrument CO., Chicago, IL, USA). The stock suspension thus prepared was stored at 4 ºC. In some cases to avoid bacteria, 0.1 % sodium azide is added to the sample (for more information see Chapter 2). Transmission electron microscopy - 10 µl of sample was applied to carbon-coated 400 mesh cupper (Ted Pella Inc., Redding, CA, USA), for 2 min. After, the grid with the sample was washed in two different drops of distilled water. Sample was then stained for 2 min with 2 % uranyl acetate (Electron Microscopy Science, Hatfield, PA, USA) and most of the liquid is removed from the drop by blotting with a piece of filter paper. Samples were viewed at 100 kV in a Philips CM-12 transmission electron microscope (FEI Co., Eindhoven, The Netherlands); digital images were taken with an MEGA VIEW-II DOCU camera system and were analyzed with the software SIS NT DOCU. In parallel, a selected number of samples were analyzed using LIBRA 200 FE OMEGA electron microscope (Carl Zeiss Inc.); the sample was view at 200 kV. Cryo-transmission electron microscopy - Samples (3 µl) were mixed with a gold fiducial solution (15 nm gold particle size), for feature tracking during the reconstruction procedure and applied to glow-discharged Quantifoil grids (R2/2 Quantifoil Jena) within the VitrobotTM Mark IV (FEI Company, Eindhoven, The Netherlands) where the environmental chamber was kept at relative humidity of 100 % and temperature of 22 ºC. Excess liquid was blotted away with filter paper using an automatic blotting device within the Vitroblot and immediately the grids were plunge-frozen in liquid ethane. The grids were transferred to a Titan KriosTM transmission electron microscope equipped with Falcon direct electron detector (FEI Co., Eindhoven, The Netherlands), operating at 300 kV, under low-dose conditions (20 electrons per Å2) at liquid nitrogen temperatures. Cryo-TEM tomography - The samples prepared in the above section were also used for lowdose cryo-TEM tomography. Tomograms of selected areas were obtained from -70 to +70 degrees with a 1.5 degree tilt increment. The Titan KriosTM microscope (FEI Co., Eindhoven, The Netherlands) operated at 300 kV, under low-dose conditions (50 electrons per Å2) at liquid nitrogen temperatures. The tilted images were aligned and reconstructed with the FEI inspect3D software (FEI Co., Eindhoven, The Netherlands). Chapter 4 84 Helical reconstruction - Dr. Howard S. Young, Dr. Ludovic Renault and Dr. Holger Wille, from the University of Alberta, Edmonton, Canada, carried out this technique. Results are still preliminary. 4.8 References 1. Merz PA, Somerville RA, Wisniewski HM, Iqbal K (1981) Abnormal fibrils from scrapieinfected brain. Acta Neuropathol. 54: 63–74. 2. Merz PA, Somerville RA, Wisniewski HM, Manuelidis L, Manuelidis EE (1983) Scrapieassociated fibrils in Creutzfeldt-Jakob disease. Nature 306: 474–476. 3. Prusiner SB, McKinley MP, Bowman KA, Bolton DC, Bendheim PE, et al. (1983) Scrapie prions aggregate to form amyloid-like birefringent rods. Cell 35: 349–358. 4. McKinley MP, Braunfeld MB, Bellinger CG, Prusiner SB (1986) Molecular characteristics of prion rods purified from scrapie-infected hamster brains. J. Infect. Dis. 154: 110–120. 5. Godsave SF, Wille H, Kujala P, Latawiec D, DeArmond SJ, et al. (2008) Cryo-immunogold electron microscopy for prions: toward identification of a conversion site. J. Neurosci. 28: 12489–12499. 6. Meyer RK, McKinley MP, Bowman KA, Braunfeld MB, Barry RA, et al. (1986) Separation and properties of cellular and scrapie prion proteins. Proc. Natl. Acad. Sci. U.S.A. 83: 2310–2314. 7. McKinley MP, Meyer RK, Kenaga L, Rahbar F, Cotter R, et al. (1991) Scrapie prion rod formation in vitro requires both detergent extraction and limited proteolysis. Journal of Virology 65: 1340–1351. 8. Chesebro B, Trifilo M, Race R, Meade-White K, Teng C, et al. (2005) Anchorless prion protein results in infectious amyloid disease without clinical scrapie. Science 308: 1435–1439. 9. Sim VL, Caughey B (2009) Ultrastructures and strain comparison of under-glycosylated scrapie prion fibrils. Neurobiol. Aging 30: 2031–2042. 10. Baron GS, Hughson AG, Raymond GJ, Offerdahl DK, Barton KA, et al. (2011) Effect of glycans and the glycophosphatidylinositol anchor on strain dependent conformations of scrapie prion protein: improved purifications and infrared spectra. Biochemistry 50: 4479–4490. 11. Wang L, Sigworth FJ (2006) Cryo-EM and single particles. Physiology (Bethesda) 21: 13–18. 12. Requena JR (2009) Structure of mammalian prions. Future Virology 4: 295–307. 13. Lucic V, Förster F, Baumeister W (2005) Structural studies by electron tomography: from cells to molecules. Annu. Rev. Biochem. 74: 833–865. Electron microscopy 85 14. Fernandez J-J (2012) Computational methods for electron tomography. Micron 43: 1010–1030. 15. Baumeister W, Steven AC (2000) Macromolecular electron microscopy in the era of structural genomics. Trends Biochem. Sci. 25: 624–631. 16. Meng X, Zhao G, Zhang P (2011) Structure of HIV-1 capsid assemblies by cryo-electron microscopy and iterative helical real-space reconstruction. J Vis Exp. 17. Egelman EH (2007) The iterative helical real space reconstruction method: surmounting the problems posed by real polymers. Journal of Structural Biology 157: 83–94. 18. Mizuno N, Baxa U, Steven AC (2011) Structural dependence of HET-s amyloid fibril infectivity assessed by cryoelectron microscopy. Proc. Natl. Acad. Sci. U.S.A. 108: 3252–3257. 89 5 Discussion Abstract In this chapter, I have taken all the data obtained in this thesis to make a final common discussion. It is very interesting to analyze all results as a whole, and observe how each piece of the puzzle fits perfectly, in order to develop a new structural model of the PrPSc. I believe that the results presented in this thesis are an advance in the knowledge of prion protein. Our future perspectives are to refine the results obtained by the 3D reconstruction methods and attempt to get more data with other electron microscopy techniques, to generate a three-dimensional model with the highest possible resolution hitherto achieved. Chapter 5 90 Despite the difficulty to obtain substantial information on the structure of the PrPSc, in this thesis a large amount of experimental data were obtained, allowing the development of a new structural model. GPI-anchorless PrPSc fibrils have a width of 3-5 nm ([1] and Chapter 4 of this thesis). During our electron microscopy studies we have observed !2-2.5 nm densities along the fibril axis, showing each monomer of PrPSc. Also a 4.8 Å reflection was detected, indicating a cross-! along the fiber direction. These constraints mean that each PrPSc monomer must be coiled in such a way as to fit approximately 144 residues (~G89-S232) into this width while maintaining the observed high proportion of !-sheet secondary structure [2][3][4]. In order to do so, the PrPSc monomers must necessarily adopt a multilayer architecture. Such architecture is observed in SH3 fibers [5] or the HET-s fungal prion domain [6]. The HET-s prion domain packs 72 residues (218-289) into two windings of three !-strands each, alternating with turns and loops [6]. Wille et al. have concluded that PrPSc fibrils are composed of four rungs of !-strands, based on their interpretation of fiber X-ray diffraction patterns [7]. Accordingly, each rung of the GPI-anchorless PrPSc monomer would be comprised of ~36-37 residues. Figure discussion. Cartoon of the structure of PrPSc (the shape and connection of the elements are arbitrary). A. Monomer of PrPSc is represented. Electron microscopy studies suggest that each monomer has a length of !2-2.5 nm, and it is made up of four-rung separated !0.5 nm, indicating a cross-! motif. Limited proteolysis reveals a number of cleavage sites, corresponding to loops connecting !-strands. B. Protofilament cores adopt a cross-! structure to form the fibers, stacking the monomers of PrPSc. Association between fibrils is the most common conformation of the fibers. Discussion 91 The information emerging from limited proteolysis indicate that, positions N152-M153 lie near the middle of the G85-S232 sequence, and, therefore, it is tempting to speculate that they might be located at an exposed position at the border between rungs. This might explain why the N152-S232 and/or M153-S232 fragment emerges as the most conspicuous PK-resistant fragment after prolonged treatment with PK or partial unfolding with guanidine (Figures 3.9 and 3.10). Positions A116-G118 might be the border between the two most amino-terminal rungs (approximately G85-A115 and A119-E151). On the other hand, the results presented are partially inconsistent with the specific location assigned by Govaerts et al., using threading algorithms, to residues K100-P104 and E145-R163, placed in loops and not rungs [8]. Our experimental data suggest that the stretches formed by residues K100-P104, N142-E151, and Y154-Y161, are PK-resistant, i.e., likely part of a !-strand rung that is stabilized by extensive H-bonding (Figure 3.6 and Table 3.I). However, due to the limitation of the techniques it is not possible to describe how many !- sheets are forming each rung. Still, the most likely it is that each rung is made up minimum of three !-sheets, but there could be more or even impossible to detect other cleavage points because of the conformation between rungs. In summary, our data support a PrPSc structure consisting of a four-rung solenoid with a central !-strand-rich core; from which a series of highly PK-resistant !-sheet strands intersperse with PK-sensitive short flexible loops and turns. Furthermore, the region comprising ~V179 to the Cterminus of PrPSc is probably composed primarily of !-sheet, as it is highly resistant to PK. The data obtained from this GPI-anchorless PrPSc work is consistent with our previous results (263K and Dy strains) and those of other researchers using SHaPrPSc. Furthermore our results are consistent with those observed for human CJD PrPSc, which suggests that the myriad human, hamster and mouse prions share a common basic structure. References 1. Sim VL, Caughey B (2009) Ultrastructures and strain comparison of under-glycosylated scrapie prion fibrils. Neurobiol. Aging 30: 2031–2042. 2. Caughey BW, Dong A, Bhat KS, Ernst D, Hayes SF, et al. (1991) Secondary structure analysis of the scrapie-associated protein PrP 27-30 in water by infrared spectroscopy. Biochemistry 30: 7672– 7680. 3. Baron GS, Hughson AG, Raymond GJ, Offerdahl DK, Barton KA, et al. (2011) Effect of glycans and the glycophosphatidylinositol anchor on strain dependent conformations of scrapie prion protein: improved purifications and infrared spectra. Biochemistry 50: 4479–4490. 4. Smirnovas V, Baron GS, Offerdahl DK, Raymond GJ, Caughey B, et al. (2011) Structural organization of brain-derived mammalian prions examined by hydrogen-deuterium exchange. Nat. Struct. Mol. Biol. 18: 504–506. 5. Jiménez JL, Guijarro JI, Orlova E, Zurdo J, Dobson CM, et al. (1999) Cryo-electron microscopy structure of an SH3 amyloid fibril and model of the molecular packing. EMBO J. 18: 815–821. Chapter 5 92 6. Wasmer C, Lange A, Van Melckebeke H, Siemer AB, Riek R, et al. (2008) Amyloid fibrils of the HET-s(218-289) prion form a beta solenoid with a triangular hydrophobic core. Science 319: 1523–1526. 7. Wille H, Bian W, McDonald M, Kendall A, Colby DW, et al. (2009) Natural and synthetic prion structure from X-ray fiber diffraction. Proc. Natl. Acad. Sci. U.S.A. 106: 16990–16995. 8. Govaerts C, Wille H, Prusiner SB, Cohen FE (2004) Evidence for assembly of prions with lefthanded beta-helices into trimers. Proc. Natl. Acad. Sci. U.S.A. 101: 8342–8347. Appendix III 101 APPENDIX III List of abbreviations ! 2D: two-dimensional. ! 3D: three-dimensional. ! Ab: antibody. ! ACN: acetonitrile. ! Ac2O: acetic anhydride. ! AFM: atomic force microscopy. ! BS3: bis(sulfosuccinimidyl) suberate. ! BSA: bovine serum albumin. ! BSE: bovine spongiform encephalopathy. ! CFT: contrast transfer function. ! CJD: Creutzfeldt-Jakob disease. ! vCJD: variant CJD. ! Cryo-TEM: cryo-transmission electron microscopy or electron cryomicroscopy. ! Cryo-ET: cryo-electron tomography. ! CWD: chronic wasting disease. ! Dy: drowsy strain. ! EDTA: ethylenediaminetetraacetic acid. ! EM: electron microscopy. ! ET: electron tomography. ! FFI: fatal familial insomnia. ! FFT: fast fourier transform. ! FSE: feline spongiform encephalopathy. ! FTIR: fourier transform infrared spectroscopy. ! GndHCl: guanidine hydrochloride. ! GPI: glycosylphosphatidylinositol. ! GPI-less: GPI-anchorless. ! GSS: Gerstmann-Sträussler syndrome. ! HE: haematoxylin and eosin. ! HGH: human growth hormone. ! H/D: hydrogen-deuterium exchange. ! IHC: immunohistochemical. ! IR: infrared spectroscopy. ! KO: knock out. ! MALDI-TOF: matrix-assisted laser desorption/ionization-time of flight. ! MBM: meat and bone meal. ! MeOH: methanol. ! MS: mass spectrometry. ! MW: molecular weight. ! nano-LC-ESI-Qq-TOF: electrospray ionization-double quadrupole-time of flight. Appendix III 102 ! NMR: nuclear magnetic resonance. ! NTCB: 2-nitro-5-thiocyanatobenzoic acid. ! PBS: phosphate buffered saline. ! PCR: polymerase chain reaction. ! PK: proteinase K. ! PMCA: protein misfolding cyclic amplification. ! PNGase F: peptide-N-glycosidase F. ! Prnp: prion protein gene. ! PrP: prion protein. ! bPrP: bovine prion protein. ! hPrP: human prion protein. ! MoPrP: mouse prion protein. ! PrPC: cellular prion protein. ! PrPSc: scrapie prion protein. ! PrPSc106: miniprion. ! recPrP: recombinant prion protein. ! sPrPSc: protease-sensitive prion protein. ! PrP27-30 or PrPres: protease-resistant prion protein. ! RML: strain of mouse. Rocky Mountain laboratory. ! SA: sinapinic acid. ! SAFs: scrapie associated fibrils. ! SHaPrP: hamster prion protein. ! SDS-PAGE: sodium dodecyl sulfate-polyacrylamide gel electrophoresis. ! SP: signal peptide. ! TEM: transmission electron microscopy. ! TFA: trifluoroacetic acid. ! tg: transgenic. ! tg44+/-: transgenic heterozygous GPI-anchorless PrP mice. ! tg44 -/-: transgenic homozygous GPI-anchorless PrP mice. ! TME: transmissible mink encephalopathy. ! TNM: tetranitromethane. ! TSE: transmissible spongiform encephalopathy. ! WB: Western blot. ! WBP: weighted backprojection. ! Wt: wild-type. Appendix IV 103 APPENDIX IV List of publications/Participation in conferences Articles submitted Vázquez-Fernández E, Alonso J, Pastrana MA, Ramos A, Stitz L, Vidal E, Dynin I, Petsch B, Silva CJ and Requena JR. Structural organization of mammalian prions as probed by limited proteolysis. Under revision, PLoS One (2012). Articles published Sajnani G, Silva CJ, Ramos A, Pastrana MA, Onisko BC, Erickson ML, Antaki EM, Dynin I, Vázquez-Fernández E, Sigurdson CJ, Carter JM, Requena JR. PK-sensitive PrP is infectious and shares basic structural features with PK-resistant PrP. PLoS Pathog. 2012;8(3):e1002547. Gong B, Ramos A, Vázquez-Fernández E, Silva CJ, Alonso J, Liu Z, Requena JR. Probing structural differences between PrP(C) and PrP(Sc) by surface nitration and acetylation: evidence of conformational change in the C-terminus. Biochemistry. 2011 Jun 7; 50(22):4963-72. Studies selected for oral presentation Fernández-Borges N, Vázquez-Fernández E, Rodríguez-Elezgarai S, Parra B, Alonso J, Di Bari MA, Sánchez-Martín M, Eraña H, Harrathi C, Gayosso M, Vidal E, Pumarola M, Agrimi U, Mayoral T, Nonno R, Andreoletti O, Requena JR and Castilla J. Highly infectious recombinant prions - A new challenge for understanding how prions work. International Conference “Prion 2012”, Amsterdam, The Netherlands. Vázquez-Fernández E, Ramos A, Alonso J & Requena JR. Probing the structure of GPI-less PrPSc by limited proteolysis. II National Prion Conference (2011), Madrid, Spain. Fernández-Borges N, Parra B, Rodríguez-Elezgarai S, Vidal E, Sánchez-Martín M, de Castro J, Vázquez-Fernández E, Pumarola M, Requena JR, Mayoral T and Castilla J. Lessons from nature: prion resistant species and their mechanisms. II National Prion Conference (2011), Madrid, Spain. Sajnani G, Pastrana MA, Ramos A, Dynin I, Vázquez-Fernández E, Onisko B, Requena JR. Limited proteolysis of PK-sensitive PrPSc further supports a structure featuring beta sheet stretches interspersed with loops/turns for PrPSc. International Conference “Prion 2008”, Madrid, Spain. Appendix IV 104 Poster communications Vázquez-Fernández E, Alonso J, Pastrana MA, Ramos A, Stitz L, Vidal E, Dynin I, Silva CJ and Requena JR. Probing the structure of GPI-anchorless PrPSc by limited proteolysis. International Conference “Prion 2012”, Amsterdam, The Netherlands. Harrathi C, Eraña H, Rodríguez-Elezgarai S, Vázquez-Fernández E, Fernández-Borges N, Requena JR and Castilla J. Landing in a new recombinant prion world. II National Prion Conference (2011), Madrid, Spain Fernández-Borges N, Vázquez-Fernández E, Rodríguez-Elezgarai S, Gong B, Parra B, Alonso J, Requena JR and Castilla J. In vitro prion transmission studies. A switch to a recombinant PrPSc world. International Conference “Prion 2011”, Montreal, Canada. Vázquez-Fernández E, Sajnani G, Ramos A and Requena JR. Further characterization of flexible regions of PrPSc by limited proteolysis. I National Prion Conference (2010), Bilbao, Spain. Vázquez-Fernández E, Sajnani G, Ramos A and Requena JR. Further characterization of flexible regions of PrPSc by limited proteolysis. International Conference “Prion 2010”, Salzburg, Austria.