Prion protein—Semisynthetic prion protein (PrP) variants with posttranslational modifications

Deciphering the pathophysiologic events in prion diseases is challenging, and the role of posttranslational modifications (PTMs) such as glypidation and glycosylation remains elusive due to the lack of homogeneous protein preparations. So far, experimental studies have been limited in directly analyzing the earliest events of the conformational change of cellular prion protein (PrPC) into scrapie prion protein (PrPSc) that further propagates PrPC misfolding and aggregation at the cellular membrane, the initial site of prion infection, and PrP misfolding, by a lack of suitably modified PrP variants. PTMs of PrP, especially attachment of the glycosylphosphatidylinositol (GPI) anchor, have been shown to be crucially involved in the PrPSc formation. To this end, semisynthesis offers a unique possibility to understand PrP behavior in vitro and in vivo as it provides access to defined site‐selectively modified PrP variants. This approach relies on the production and chemoselective linkage of peptide segments, amenable to chemical modifications, with recombinantly produced protein segments. In this article, advances in understanding PrP conversion using semisynthesis as a tool to obtain homogeneous posttranslationally modified PrP will be discussed.


| PRION DISEASES
Prion diseases or transmissible spongiform encephalopathies (TSEs) are incurable, neurodegenerative disorders affecting humans and animals. 1 They include scrapie of sheep and goats, bovine spongiform encephalopathy (BSE) of cattle, chronic wasting disease (CWD) of cervids, and several human diseases such as kuru, Creutzfeld-Jakob disease (CJD), Gerstmann-Sträussler-Scheinker syndrome (GSS), and fatal familial insomnia (FFI). The disease progression is accompanied by the loss of cognitive skills and neuronal dysfunction and can be of inherited sporadic or iatrogenic origin. 2,3 The central pathophysiologic event is ascribed to the conformational change of the cellular prion protein (PrP-C ) into scrapie prion protein (PrP Sc ) that then not only propagates further PrP C misfolding in neighboring cells but can also infect other organisms. 4 Identification of the infective pathogen of prion diseases and its proof of transmissibility started in the 1950s. By ending cannibalism within the Fore tribe in Papua New Guinea, the transmission of kuru could be stopped. Experiments with transferring brain samples of these kuru victims into primates induced spongiform encephalopathies. 5 Due to its infective property, the pathogen was first assumed to be of nucleic acid-based, viral nature. However, the application of ultraviolet and ionizing irradiation failed to inactivate the agent, leading to the "protein-only hypothesis" by Griffith in 1967. 6,7 Eventually in 1982 the term "prion" defining a small proteinaceous infectious particle was introduced by Prusiner during the course of discovering the prion protein (PrP). 8,9 PrP 27-30 corresponding to the protease-resistant core of PrP Sc with an apparent molecular mass of 27 to 30 kDa was isolated by enriching fractions from Syrian hamster (SHa) brain for scrapie infectivity. [10][11][12][13] Successful Edman degradation paved the way for subsequent molecular cloning studies of the PrP gene. [14][15][16] The linkage of PrP Sc to prion diseases was recognized as an important feature of the protein, together with its role in transmission and pathogenesis of these illnesses. 17 Thus, the main focus of elucidating prion pathogenicity is assigned to PrP. Understanding the key features in prion diseases can serve as paradigm for other neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS), frontotemporal dementia (FTD), Alzheimer disease (AD), and Parkinson disease (PD), that are characterized by misfolded proteins "prionoids" sharing the aggregation properties but being not strictly infectious. [18][19][20] As it happens, the latter statement might not be entirely correct. Recent prion research reported the discovery of α-synuclein prions 21 in multiple system atrophy (MSA) and iatrogenic AD with evidence of transmissibility of amyloid-β (Aβ), 22 hence highlighting the need to understand prion transmission and toxicity even more.

| Cellular prion protein
High expression levels of PrP C are found in the central nervous system (CNS), but it exists in other cell types and tissues, such as lymphoid organs, as well. [23][24][25][26] Accessing the gene-encoding SHaPrP C , Prnp, 14,27 entailed its further identification in numerous other species and illustrated a highly conserved sequence. 28,29 The entire open reading frame (ORF) is contained within a single exon and primarily translates into a protein composed of 254 amino acids (aas). 30,31 The first 22 aas reflect an N-terminal signal sequence for PrP entering the secretory pathway. Upon its cleavage, glycosylation at asparagine residues and formation of a disulfide bond occur in the endoplasmic reticulum (ER). Lastly, cleavage of the C-terminal signal sequence facilitates the attachment of the glycosylphosphatidylinositol (GPI) anchor, providing mature, posttranslationally modified PrP at the outer leaflet of the cell membrane, typical for glycosylphosphatidylinositol anchored proteins (GPI APs). 32 Interestingly, PrP can be found in three topologic forms at the ER. Apart from the fully translocated PrP, two transmembranal types occur with the N-or C-terminus facing the ER lumen, denoted as Ntm PrP or Ctm PrP, respectively. 33,34 Normally, Ntm PrP and Ctm PrP only comprise a small portion of PrP C , whereas an excess of Ctm PrP induces neurotoxicity. Neuronal cell death is caused in the absence of PrP Sc formation, obviously by an aberrant metabolism of PrP C . PrP C mislocalization represents another mechanistic possibility for prion toxicity next to the alteration of PrP C -mediated signaling and PrPderived oligomeric species. 23 First structural studies on PrP C isolated from brains of SHas demonstrated a predominantly α-helical content. 35 As these measurements agreed well with subsequent spectroscopic data of recombinant PrP, accessible in larger amounts, it was considered an appropriate surrogate in biochemical experiments, [36][37][38][39] as well as in solving nuclear magnetic resonance (NMR) and crystal structures of PrP. 40-47 The PrP structure comprises an unstructured N-terminal (aa 23-120) and a globular C-terminal part (aa 121-231) ( Figure 1).
In more detail, the N-terminal segment consists of a nonapeptide (PQGGGGWGQ) followed by four octapeptide (PHGGGWGQ) repeats (OR) with a high affinity for copper, [48][49][50][51] and other divalent cations, 52 adjacent to a charged cluster (CC) or polybasic region ( Figure 2). Noteworthy, the configuration of the copper binding region in hPrP (aa 23-231) has been determined combining different experimental methods by using synthetic octapeptide and tetraoctapeptide as well as fulllength hPrP. [53][54][55][56] Depending on the concentration of the metal and pH, the OR region is capable to bind up to four copper ions in distinct coordination geometries. 50,57 Current estimates for dissociation constant (K d ) values vary betweeen the micromolar and femtomolar range. 50 The central hydrophobic domain (HD), comprising of aas 113 to 135, serves as a transmembrane domain 33 and includes a palindromic region (AGAAAAGA, aa 113-120) thought to be important in the PrP C -PrP Sc conversion. 58,59 Within the C-terminal region, three α-helices (aa 144-154, 175-193, and 200-219), with two of them connected by a disulfide bond, 60 and a small antiparallel β-sheet (aa 128-131 and 161-164) are present. As posttranslational modifications (PTMs), a Cterminal GPI anchor linked to serine 231 and two N-linked glycosylation sites at asparagines 181 and 197 exist. PrP C can occur in nonglycosylated, monoglycosylated, and diglycosylated forms. 61,62 Variations in glycan structures attached to PrP may be differentially distributed depending on the areas of the CNS. 63 Molecular dynamics simulations indicate that the N-linked oligosaccharides located at two helices within the structured region of PrP contribute to its stabilization in generating a negative electrostatic field covering the helical surface, 64 thus impacting strain diversity and prion infection. [65][66][67][68] The C-terminal GPI anchor tethers PrP to the outer leaflet of the plasma membrane. 69 It has been postulated that mutations in the Prnp gene facilitate the pathogenic process by destabilizing the tertiary structure of PrP C .
More than 30 mutations in Prnp could be linked to inherited prion diseases. 70 In affecting the primary sequence of PrP, concomitant changes in its 3D structure may arise, and not cause, but influence a person's risk of developing a disease. Indeed, thermodynamic measurements of mutated PrP variants indicated destabilizing effects only for some of them. 71 For example by comparing the wild-type variant to the E200K mutant almost identical structures resulted, but major perturbations of the surface electrostatic potential were found. This suggests that these defects cause abnormalities in PrP interactions and should be considered as key determinants in the misfolding process. 72 Moreover, it has been speculated that methionine oxidation in PrP C plays a destabilizing role and supports spontaneous conversion into PrP Sc . Wolschner et al 73 found a strong proaggregation behavior for hPrP C with oxidized methionine residues and a variant with methionine replaced by hydrophilic methoxinine as a stable substitute for oxidation-sensitive methionine. These findings suggest a pivotal role of oxidative stress in PrP conversion.

| Scrapie prion protein
PrP Sc is the toxic, misfolded isoform of PrP. It is, as PrP C , encoded by the Prnp gene and exhibits identical PTMs, but distinct structural, biochemical, and physiological features. 13 Despite a large interest in elucidating the structure of PrP Sc , there are only limited data about its molecular details available. 74 To date, obtaining a high-resolution structure of PrP Sc has been impaired by its insolubility, propensity to aggregate, and heterogeneity. Structural variations, such as differences in the glycosylation patterns, suggested to correlate with biochemical changes, including the extent of the proteinase K (PK) resistance, the electrophoretic mobility of the proteolytic fragments, and the conformational stability, depend on the distinct strains and complicate the determination of PrP Sc structure. 75 Besides, in agreement with discussions from the Prion 2018 round tables, 76 the diversity of PrP assemblies implicates that there may be no single PrP Sc structure. Data generated by biochemical and physical methods, such as spectroscopy analysis, electron microscopy, and limited proteolysis, have led to several 3D structural models. Govaerts and colleagues suggested that left-handed β-helices assembled into trimers, also known as the 4-rung β-solenoid model. 77 Based on electron spin resonance (ESR) measurements, Cobb et al proposed a parallel in-register intermolecular β-sheet (PIRIBS) architecture where PrP Sc consists of βstrands and short turns and/or loops with no residual α-helices. 78 Still, so far, all models display discrepancies with experimental data. 79 Notably, cryo electron microscopy (cryo-EM) is a technique providing high-resolution structures. 80  amyloidosis. 90 To date, cryo-EM studies of tau and AL represent the only structural data of fibrils directly extracted from human tissue under pathologic conditions. For tau-paired helical and straight filaments could be identified with cores made of two identical protofilaments that adopt a combined cross-β/β-helix structure. AL fibrils were found to be helical with a single protofilament showing a cross-β architecture. It is widely accepted that during the PrP C -PrP Sc conversion, the β-strand content increases vastly 91,92 and the PK resistance of the "folded core" (aa approximately 90-231), as well. 9,14 Whereas PrP C is dominated by αhelices, monomeric, soluble, and highly susceptible to proteolytic digestion, PrP Sc contains predominantly β-sheets (>43%), 92 aggregates into amyloid fibrils, 93 is insoluble in detergents and partially resistant to proteolysis. 35,94 These biochemical differences between the PrP isoforms appear to be associated with the changes of the secondary structure in PrP Sc .

| Function of PrP C
Although the relevance of PrP C in TSEs is widely accepted, its physiological function remains enigmatic. Studies with PrP knockout mice have failed on this regard. Transgenic mice lacking PrP were found to develop normally. 95,96 A multitude of functions has been ascribed to PrP in different tissues, cells, and experimental settings, although not always without controversy or questionable reproducibility.
Already 10 years ago, it was suspected that reports on the function of PrP represent just specific aspects of a more complex physiological role of PrP C . 23 Causes for the functional diversity of PrP C might not only be its alternating transient binding partners in different cellular locations but also its proteolytic processing. 124 Despite multiple evidence of PrP in physiological processes, the functional diversity based on its manifold binding partners and proteolytic fragments complicate an exact definition of its physiological function. Yet successful elucidation of pathways and roles of PrP could help to understand its linkage to toxicity in prion diseases and to other neurodegenerative diseases. 136 3.2 | Trafficking of PrP C As the PrP function is closely intertwined with the cellular compartments where the protein is located, having a closer look at trafficking may assist in elucidating its involvement in pathological and physiological processes. PrP C is tethered via its GPI anchor to the outer leaflet of the plasma membrane. 69 In 1993, data by Shyng et al 137 revealed constitutive cycling of PrP C between the cell surface and endocytic compartments on varying times scales dependent on the cell line, as demonstrated in later work. 138 From the cellular membrane, PrP C can enter the cell via multiple pathways, mediated mainly by the unstructured N-terminal domain. 139,140 Evidence for a cooperation between clathrin 138,141,142 and rafts [143][144][145] in the internalization of PrP C was found. 146 Clathrin is a large, oligomeric protein assembling into lattice structures on the inner surface of the plasma membrane. Thereby, it causes the membrane to invaginate and pinch off to form clathrincoated vesicles (CCVs), which can then fuse with other intracellular organelles. 147 Although a clathrin-dependent internalization might appear unusual since PrP lacks a cytoplasmic domain necessary for the direct interaction with clathrin and the adaptor protein, GPI APs can indeed enter the clathrin-dependent pathway upon interaction with transmembrane proteins possessing a clathrin-coated pit internalization signal. 144 Moreover, the endocytosis of PrP C was found to be associated with the low-density lipoprotein receptor-related protein 1 (LRP1) 142,148 that belongs to a receptor family of cell-surface transmembrane proteins capable of binding a variety of ligands and internalizing via clathrin-coated pits. 149,150 As a nonclassical clathrinindependent pathway, the raft-dependent internalization route distinguishes caveolae-dependent and caveolae-independent endocytosis. 151 Caveolae are membrane invaginations, originating from the oligomerization of caveolins, their integral coat proteins, and are considered to be specialized raft domains. 152,153 Due to the presence of PrP C in caveolae-like domains 154,155 and its colocalization with caveolin-1 (cav-1), 143

| MECHANISM OF PrP C -PrP S c CONVERSION
To date, despite considerable knowledge about the characteristics of the infective prion pathogen, its mechanism of replication and the molecular pathways leading to neurodegeneration are largely unknown. There is evidence from invitro and transgenic mouse studies that the conversion to PrP Sc implicates PrP C -PrP Sc interactions. 84,[159][160][161][162][163] The rate of PrP Sc formation and disease progression appears to be directly proportional to the level of PrP C expression, indicated by PrP knockout mice not propagating scrapie infectivity and transgenic mice heterozygous for a disrupted PrP gene requiring prolonged incubation times upon prion inoculation. [164][165][166] In agreement with the "protein-only hypothesis," these findings have raised two models explaining prion replication (Figure 3). The template-directed refolding model by Prusiner 167 proposes that a high-energy barrier prevents the spontaneous PrP C -PrP Sc conversion. Upon interaction, monomeric PrP Sc induces PrP C to convert into PrP Sc . However, until now, there is no experimental evidence for the existence of a stable PrP Sc monomer. 168 PrP Sc seeds in this prion propagation process are not considered essential. Alternatively, in the more accredited seeded nucleation model by Jarrett and Lansbury, 169 a reversible thermodynamic equilibrium between PrP C and PrP Sc is postulated. In the presence of stable oligomeric PrP Sc aggregates, the conversion from PrP C to PrP Sc is favored, thus making PrP Sc aggregates (seeds) inevitable for prion spread. Fragmentation of PrP Sc aggregates increases the number of nuclei capable of recruiting further PrP Sc . In fact, these soluble oligomers produced during the PrP amyloid aggregation have emerged as the primary neurotoxic species, supporting the seeded nucleation model. [170][171][172] Ultimately, evidence for a direct PrP C -PrP Sc interaction in the con- This in vitro PrP res propagation recapitulates the species and strain specificity of prion transmission invivo. 173,174 Mechanistically, it has identified structural factors underlying the species barrier and optimal conditions for the PrP res formation. 66,175,176 The ability to generate PrP res not only from purified but also recombinant protein 177 provides FIGURE 3 Models for the conversion of cellular prion protein (PrP C ) into scrapie prion protein (PrP Sc ). The model for template-directed refolding (top) and seeded nucleation (bottom) are depicted. The figure was modified from Aguzzi and Calella 23 a unique opportunity to study prion propagation. CFC assays can be used as screening experiments as they have the potential to identify compounds directly inhibiting the PrP C -PrP Sc interaction or its subsequent conversion. 178,179 Still, the proportionally large amount of PrP Sc seeds required to drive the CFC assay (PrP Sc :PrP C = 50:1) has prevented it from generating de novo infectivity. 180 A more efficient method for mimicking the autocatalytic replication of PrP Sc was provided by Soto and colleagues 38 in affording a larger than 10-fold increase in PrP res with the usage of a 1:100 ratio of PrP Sc to PrP C . By subjecting scrapie-infected and normal brain homogenate to the so-called protein misfolding cyclic amplification (PMCA) procedure, PrP res is amplified in cycles of sonication and incubation. Successive rounds of PMCAs and fragmentation of PrP Sc rise the available amounts of replication-competent species. [181][182][183] Thus, with automation, this assay offers a promising diagnostic tool in presymptomatic blood screening, 184,185 and eventually, it has facilitated the detection of de novo infectivity in hamsters. 38 However, the levels of infectivity still remain lower than with a similar quantity of brain-derived PrP Sc , and the usage of complex brain homogenate itself represents an obstacle in thoroughly elucidating the conversion and association of infectivity with PrP res . 186 Besides, the distinct efficiency differences between the CFC and PMCA assays applying purified and crude brain-derived PrP Sc proposed that cellular accessory factors are involved in the generation of PrP res . In fact, polyanionic molecules were identified as factors present in the brain homogenate that contribute to the conversion efficiency. 187 Ongoing development of PMCA assays aiming to detect and early diagnose TSEs has led to the quaking-induced conversion (QuIC) method. 188 Sonication is replaced by reproducible and easier controllable shaking during the prion amplification process, which enables the application of standardized protocols. This accomplishment is reflected by the multiple variations currently available, such as standard (S-QuIC), real-time (RT-QuIC), and enhanced QuIC (eQuIC). [189][190][191][192] Apart from the autocatalytic propagation of PrP Sc , another crucial hallmark of the PrP C -PrP Sc conversion is the de novo generation of infectivity. However, when inoculated into animals, PrP fibrillar assemblies can range from being biologically inert to fully infectious, pathogenic, and transmissible in subsequent passages. 37, 76,[193][194][195] Legname and coworkers 196  can easily occur. [200][201][202][203][204] This is supported by the finding that by releasing PrP C from the cell surface or interrupting its transport to the plasma membrane prevents the formation of PrP Sc . [205][206][207] More precisely, both PrP isoforms were found to be associated with rafts. [208][209][210][211] These are defined as highly dynamic microdomains wherein specific lipids stabilize larger lipid platforms and compartmentalize cellular processes at the membrane. 212 Impairing the integrity of the cholesterol-enriched rafts associated with PrP by lowering the intracellular levels of cholesterol reduced the formation of PrP Sc in infected cells. 213 Moreover, PrP Cand PrP Sc -associated rafts were found to have distinct characteristics, as they can be separated from each other by solubilization and flotation on density gradients. 208 According to Campana et al,200 this proposes that either the types of raft or the membrane association of each isoform has different characteristics.
However, Baron et al 209 illustrated that the PrP Sc -PrP C conversion only takes place in the presence of fused PrP Scand PrP C -containing membranes, suggesting that the two PrP isoforms need to be inserted into contiguous membranes. Alternatively, rafts were proposed to stabilize PrP in its conformation via a direct interaction with cholesterol.
Thus, changes in the local lipid environment can mediate PrP conformation. 214 Studies on model lipid bilayers regarding the impact of the PrP-lipid interaction on structure and affinity of PrP support the idea that predominantly α-helical PrP C is stabilized upon binding to raft membranes, whereas binding to negatively charged lipid (nonraft) membranes leads to an increased β-sheet content. 215 Interestingly, the PrP-raft association is mediated by the GPI anchor 213,216 and the N-terminal region of PrP. 217 Unlike for a typical GPI-anchored protein, for PrP, this raft association occurs already earlier in the secretory pathway and appears to be involved in the maturation and folding process of PrP C . 218,219 Alternatively to the plasma membrane, the formation of PrP Sc was suggested to involve additional cellular places.
Immediately after PrP internalization, the PrP C -PrP Sc conversion may occur in the endolysosomal compartment, 205 in the Golgi apparatus and/or the ER following retrograde transport. 220,221 In infected cells, stimulation of retrograde transport towards the ER leads to an increase in PrP Sc formation from PrP C precursor, 222 suggesting that the ER may represent an amplification compartment for PrP Sc . 223

| Impact of GPI anchor on PrP C -PrP Sc conversion
Typically, PrP is attached to membranes by its GPI anchor ( Figure 6A). 69 A better understanding of the interplay between membranes, GPIanchored PrP, and PrP C -PrP Sc conversion is provided by work from Baron and Caughey. 209,210 First, they studied the conditions necessary for PrP res formation of PrP associated with detergent-resistant membranes (DRMs). 209 Based on that, in CFC assays, Baron   threonine, or serine residue of the C-extein to alanine block the splicing process and only allow the initial N → S acyl shift, which enables the generation of a protein α-thioester by addition of an excess of a thiol, such as sodium 2-mercaptoethanesulfonate (MESNA), to trap the protein thioester. 264 PTS is a process that relies on the assembly of two divided segments of inteins, so-called split inteins, to form a functional intein. Upon assembly of the split inteins, PTS occurs and links the N-and C-exteins in a similar sequence of events as described above ( Figure 4B). 265 The generation and biophysical characterization of PrP constructs containing a GPI anchor mimic started more than 10 years ago in the Becker laboratory with work described in Olschewski et al. 228 Two strategies based on the EPL approach provided PrP Palm , an N- to global deprotection provided a synthetic, cysteine-tagged GPI anchor suitable for NCL reactions ( Figure 6B). In a following EPL reaction, PrP with a C-terminal thioester was linked to this synthetic GPI anchor.
Analysis of the secondary structure of PrP attached to the synthetic GPI revealed that the CD curves are indistinguishable from the spectra of PrP and comparable with the spectra of PrP C . Moreover, the CD spectra were found to agree with the spectra of PrP Palm . This observation confirms the successful application of the GPI anchor-mimicking peptides ( Figure 6C) as an alternative to circumvent the elaborate synthesis of a GPI anchor ( Figure 6B). Even though the synthesis of the GPI anchor succeeded, it remains a challenge to provide sufficient amounts for subsequent experiments and extension to other proteins.
Isolating mostly homogeneous, cysteine-carrying GPI anchors from natural sources could help to avoid this problem, and first steps have been made towards this goal by using yeast as an expression system for GPI-anchored proteins, from which the GPI anchor is proteolytically released and purified. 271 GPI-anchored PrP was also found to quantitatively bind to DOPC vesicles. This emphasizes the contribution of GPI anchors in the membrane association of PrP. 257    PTMs in prion pathogenesis has not been forthcoming mainly due to the confusing complexity and heterogeneity of these glycans. 62 A similar question can arise from the cysteine residues introduced during the EPL reactions described above as in our previous approach depicted in Figure 5A. The introduced ligation site cysteine at the C-  Figure 5D). Comparing PrP variants containing a