Mammalian prions

Abstract

Upon prion infection, abnormal prion protein (PrP^Sc) self-perpetuate by conformational conversion of α-helix-rich PrP^C into β sheet enriched form, leading to formation and deposition of PrP^Sc aggregates in affected brains. However the process remains poorly understood at the molecular level and the regions of PrP critical for conversion are still debated. Minimal amino acid substitutions can impair prion replication at many places in PrP. Conversely, we recently showed that bona fide prions could be generated after introduction of eight and up to 16 additional amino acids in the H2-H3 inter-helix loop of PrP. Prion replication also accommodated the insertions of an octapeptide at different places in the last turns of H2. This reverse genetic approach reveals an unexpected tolerance of prions to substantial sequence changes in the protease-resistant part which is associated with infectivity. It also demonstrates that conversion does not require the presence of a specific sequence in the middle of the H2-H3 area. We discuss the implications of our findings according to different structural models proposed for PrP^Sc and questioned the postulated existence of an N- or C-terminal prion domain in the protease-resistant region.

Keywords: :

• prion
• PrP protein
• insertion
• mutagenesis
• protein structure

This article refers to:

Self-Propagation of a Pathogenic Protein: Conversion of PrP into Different Prion Structures or Assemblies

Prions are the etiologic agents of transmissible fatal neurodegenerative diseases affecting both men and animals.¹ They are mainly if not solely composed of assemblies of PrP^Sc, a conformationally altered isoform of the host-encoded cellular prion protein PrP^C. This PrP^C is a glycoprotein tethered at the cell surface by a GPI-anchor. The N-terminal region of the protein is unstructured while the C-terminal moiety is a globular domain containing three α helices and two small anti-parallel β strands.^2-4 In contrast, PrP^Sc is enriched in β structure, insoluble and tends to aggregate.^5-7 Upon infection, exogenous PrP^Sc seeds are thought to self-template host PrP^C, leading to further aggregation and deposition mainly in the nervous tissue. PrP^C expression is essential for prion replication as shown by the resistance of PrP knockout mice to prion infection and the restoration of their susceptibility after introduction of a PrP transgene.^8,9 Within the same host species, different prion strains can be propagated; those can be differentiated on the basis of the incubation time to disease, the pathology and PrP^Sc biochemical signature. Strain properties are assumed to be enciphered within differences in PrP^Sc conformation, at the level of the tertiary and/or quaternary structure.^10,11

PrP^Sc was initially differentiated from PrP^C by its insolubility and resistance to proteases such as proteinase K (PK).¹ While a strain-dependent, variable proportion of PrP^Sc is now recognized as PK-sensitive,^12-15 its relative degree of infectivity is still debated. For most prion strains, the upstream N-terminal part of PrP^Sc is truncated following proteinase K (PK) treatment while the segment encompassing the 140 to 150 C-terminal amino acids shows high PK resistance. Similarly, in vivo, part of PrP^Sc is internalized and subjected to an N-terminal truncation mediated by endo-lysosomal enzymes.^16,17

Natural Variation and Experimental Modification in ihe Sequence of Prions

The susceptibility of many mammalian species to experimental or ‘natural’ prion infection indicates that distinct, yet conserved PrP sequences are competent for prion conversion. Aligning the amino acid sequence of PrP PK-resistant core from mouse, hamster, human and ovine species (Fig. 1) shows that differences can occur all along this segment. While the last 4 or 5 C-terminal amino acids preceding the GPI anchor attachment site are less conserved among species, they do not appear to be essential for prion replication.^18,19 Apart from this C-terminus, the majority of the differences shown in Figure 1 are considered as equivalent replacement by comparison softwares such as BLAST. However, their combination is sufficient to potentially establish a species barrier to foreign prions.^20,21 Oppositely, prions can cross the species barrier when adequate mutations are introduced in the PrP sequence of a given host-species. For example, substitutions in the N-terminal moiety of mouse PK-resistant segment, by residues found at the same place in hamster, human or sheep PrP render the transgenic mice expressing these chimeric proteins permissive to hamster, human and sheep prions respectively.^22-24 These findings, along the pivotal role of the Met/Val polymorphism at PrP codon 129 in the human susceptibility to prions,²⁵ suggested that the region upstream from the last two helices is a major domain involved in the conformational change. The successful, cell-free amplification of a PK-resistant entity from the short PrP segment corresponding to the disease-associated human mutation Y145Stop recently further highlighted the potential role of this region.²⁶ However many familial forms of prion diseases are associated with mutations downstream position 145 of human PrP sequence. The C-terminal part of the protein may thus also play an important role in the conformational change.

Figure 1. Sequence alignment of the human, mouse, hamster and sheep PK-resistant PrP^Sc peptide. Numbering was according to the human sequence. Letters in blue correspond to amino acids present in the main PK fragment of 21 kDa strains (or Type 1) but not in 19 kDa (or Type 2) ones. Letters highlighted in red indicate difference with the human sequence and those highlighted in yellow indicate amino acids considered as similar (or positive) using the blast comparison software.

While PrP with deletions within the protease resistant domain may induce non-transmissible neuropathology in transgenic mice,^27-29 those are not permissive to prions. The only exception was observed with the so called miniprion PrP106, i.e., PrP with two deletions (23–88 and 141–176).³⁰ The 141–176 region was nevertheless suggested to participate in the structural changes leading to PrP^Sc.^31,32 Recently PrPs with myc or tetracysteine tags inserted at the end of the PK-resistant domain were shown to be convertible.^19,33,34 To further gain insight into the critical region(s) of the PrP sequence involved in the conformational change leading to PrP^Sc, we generated a panel of PrP mutants with an insertion inside the PK-resistant core and tested their ability of conversion into prions in cell culture.

Prions with Insertions in the Middle of PK-Resistant Core

We have recently shown that bona fide prions could be generated following insertion of ectopic peptides in H2 end or in the H2-H3 inter-helix loop of ovine PrP^C.³⁵ These mutant proteins were stably expressed in RK13 cells, a rabbit kidney cell line permissive to 127S ovine prions, upon expression of ovine PrP (V¹³⁶R¹⁵⁴Q¹⁷¹ haplotype).^36,37 The mutant PrP were converted by 127S prions and self-propagated PrP^Sc through multiple cell passages. The mutant cells with the original insert (Table1, first lane) were as infectious as their wild-type counterparts, for recipient cells expressing either the homologous insertion mutant or the wild-type PrP and for transgenic mice expressing ovine PrP^VRQ. Other inserts were introduced at the same place. They were eight to 16 amino acids long and some contained a polyglycine stretch, a His-tag, or a FLAG-tag (Fig. 2 and Table 1).³⁵ All accommodated 127S prion replication, indicating that prion conversion did not require a specified size, a specific sequence, or a defined amino acid composition in the H2-H3 loop. While this tolerance to sequence change was exceptional, it has limits as introduction of a tetracysteine-tag renders PrP resistant to conversion (Table 1). This finding might be connected with a previous study, where mouse PrP with a tetracysteine tag introduced into the same loop did not convert into fibrils in vitro.³⁸ F and T residues conservation in front of H3 might be necessary to efficient conversion as these amino acids were kept in our initial study, so as to minimize the eventual impacts on the formation of the last helix. Their removal from an otherwise well tolerated insert indeed impaired prion replication (Table 1, compare lanes 1 and 8 or lanes 5 and 9). Accordingly, substitution in wild-type PrP of the F or the T residues of the NFT glycosylation site was not always compatible with prion conversion³⁶ (see also Table 1). Another salient finding was that prion conversion also accommodated insertion of as much as 8 additional amino acids in the C-terminal region of H2 (Fig. 3). The octapeptide was introduced upstream of the third residue from the C-terminus of H2 and further upstream with gradual amino acid increment. Four consecutive mutants PrP were converted into PrP^Sc upon 127S infection. Nonetheless insertion before the sixth residue from H2 end dramatically reduced conversion efficiency and insertions further upstream were not converted anymore (Fig. 3).

Table 1. Changes introduced in ovine PrPC sequence and their compatibility with prion conversion

Modifications	Amino acids	PrP^res
Insertions (position 203)	AQQGGGFT^1	+
	GGGGGGFT^1	+
	HHHHHHFT^1	+
	AQQGDYKDDDDKGGFT^1	+
	DYKDDDDKFT	+
	CCPGCCFT	−
	CCRECCFT	−
	AQQGGG	−
	GDYKDDDDKG	−
	HHHHHHHHHHFT	−
	AQQGGGAQQGGGFT	−
	AQQGDYEDRYYRENGGFT	−
Substitutions	N200 by either D, Q, S, T, H^1	+
	F201S^1	−
	F201V^1	+
	T202A	−
	GENFT (199–202) by GSGSG	−

¹According to Salamat (refs. 35 and 36). Other modifications where not previously reported.

Figure 2. Insertion site of peptides compatible with PrP^Sc generation. The 3D structure of sheep PrP is shown with the sequence of peptides inserted in the inter-helix loop, i.e., in between the NFT glycosylation site and the beginning of helix α 3. Dark letters correspond to amino acid of the original insert isolated, red ones to engineered modifications and blue ones to FLAG-tag sequence introduced either inside the original peptide as published,³⁵ or directly inserted at position 203 (lane below, not previously published).

Figure 3. Displacement of the octapeptide insert (AQQGGFT) in H2-H3. (A) Segments corresponding to H2 and H3 are in blue and in green, respectively. Cysteines of the disulfide bridge are identified in red. Arrows show the position of the insertions, red arrows indicate compatibility with prion replication and purple ones incompatibility, as indicated by positive or negative production of self-propagating PK-resistant PrP^Sc (PrPres) as shown on the line below the sequence. (B) western blot analysis of PK-digested samples of PrP-ins195 from cell culture passages two to eight following 127S strain infection.

To determine whether such insertions modified PrP^C structure, the same changes were introduced in PrP expressed in E. coli, assuming that the 3D structures of these recombinant proteins would be equivalent to those of the cell-membrane anchored glycoproteins and would provide relevant information.³⁹ Circular dichroism analysis did not reveal significant modification of the secondary structure of PrP with an octapeptide inserted either at position 193 (convertible, see Fig. 3) or two amino acids upstream (unconvertible). Because a slight decrease in the proportion of alpha-helical content may be difficult to detect with such technique, more detailed structural analysis have been initiated by NMR. HSQC recordings indicate that the insertions did not alter PrP global arrangement. It might suggest that some of the amino acids inserted have replaced those of the helix.

Together our findings reveal the high tolerance of prions to major sequence modifications in the C-terminal part of H2 and adjacent loop of PrP. This conclusion was further sustained by the convertibility of a double mutant PrP with two octapeptide inserts, one at the end of H2 and the other in the inter helix loop, right in front of H3.³⁵

Insertions in the H2-H3 Domain and Structural Models of Prp^sc

While high-resolution structure of PrP^Sc is lacking,⁴⁰ several models have been proposed in which the participation of H2-H3 domain is debated.^32,41-44 The β helix^31,32 and the in-register models⁴⁴ have in common to propose the existence of a “prion domain” (as in yeast prions)⁴⁵ shorter than the whole PK-resistant segment, but they conflictingly locate it in the N or C-terminal part, respectively. The tolerance of prions to sequence change in the middle of H2-H3 region, either in the inter-helix segment or in the C-terminal part of H2, tends to exclude the middle part of H2-H3 as the main prion conversion site. Our results also strongly suggest that the inter-helix segment remains essentially unstructured in PrP^Sc because insertion of peptide not prone to form secondary structures, such as the polyglycine stretch can be tolerated at that place. However they do not exclude that the conversion process transform H3 and most of H2 into β strands. Indeed insertions in H3, or upstream to the last turns of H2 were not compatible with prion conversion, while these mutant PrP were normally expressed at the cell surface.

The existence of a small prion domain might itself be questioned. How the conformationally unchanged part outside such domain would resist protease digestion? It was long believed that there were 15% to 30% alpha-helical structures in PrP^Sc, suggesting that the C-terminal part of PrP remained essentially unmodified in the PK-resistant core while the N-terminal half underwent a conformational change. Recent FT-IR spectroscopic data and re-interpretation of earlier works might challenge this view, as they lead to the conclusion that there is no more α-helix in PrP^Sc.^5,46 On the other hand, a putative C-terminal prion domain emerged from structural analysis of recombinant PrP polymers.^42,44 To accommodate these data, it might worth considering that the prion domain is congruent with the whole protease-resistant domain. In this respect some theoretical models such as formation of both an N and a C-terminal left hand β helix or domain swapping of the molecules, were also proposed.^41,43 Establishment of the 3D structure of PrP^Sc arrangement as well as definition of the landscape of sequence elements critical for conversion and those susceptible to be modified remains a major challenge to fully understand the conversion process and identify drug targets.

Abbreviations:
PrP	=	prion protein
PrPC	=	normal cellular PrP
PrPSc	=	scrapie associated PrP
PK	=	proteinase K
PrPres	=	PK-resistant PrPSc
GPI	=	glycophosphatidylinositol

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Analisa Pastore (NIMR London, UK) and Stephanie Prigent (INRA Jouy-en-Josas, France) for their help with recombinant PrP structural analysis. We also thank Jerôme Chapuis for help with mutagenesis and Laetitia Hertzog and Emilie Jaumain for analysis of infected mice. This work was supported by INRA. M. K. F. Salamat was supported by the Higher Education Commission of Pakistan and by INRA Animal Health division.