Search

Resistance to proteolytic digestion has long been considered a defining trait of prions in tissues of organisms suffering from transmissible spongiform encephalopathies. Detection of proteinase K-resistant prion protein (PrPSc) still represents the diagnostic gold standard for prion diseases in humans, sheep and cattle. However, it has become increasingly apparent that the accumulation of PrPSc does not always accompany prion infections: high titers of prion infectivity can be reached also in the absence of protease resistant PrPSc. Here, we describe a structural basis for the phenomenon of protease-sensitive prion infectivity. We studied the effect on proteinase K (PK) resistance of the amino acid substitution Y169F, which removes a single oxygen atom from the β2-α2 loop of the cellular prion protein (PrPC). When infected with RML or the 263K strain of prions, transgenic mice lacking wild-type (wt) PrPC but expressing MoPrP169F generated prion infectivity at levels comparable to wt mice. The newly generated MoPrP169F prions were biologically indistinguishable from those recovered from prion-infected wt mice, and elicited similar pathologies in vivo. Surprisingly, MoPrP169F prions showed greatly reduced PK resistance and density gradient analyses showed a significant reduction in high-density aggregates. Passage of MoPrP169F prions into mice expressing wt MoPrP led to full recovery of protease resistance, indicating that no strain shift had taken place. We conclude that a subtle structural variation in the β2-α2 loop of PrPC affects the sensitivity of PrPSc to protease but does not impact prion replication and infectivity. With these findings a specific structural feature of PrPC can be linked to a physicochemical property of the corresponding PrPSc.

The cellular prion protein (PrPC) consists of a flexible N-terminal tail (FT, aa 23–128) hinged to a membrane-anchored globular domain (GD, aa 129–231). Ligation of the GD with antibodies induces rapid neurodegeneration, which is prevented by deletion or functional inactivation of the FT. Therefore, the FT is an allosteric effector of neurotoxicity. To explore its mechanism of action, we generated transgenic mice expressing the FT fused to a GPI anchor, but lacking the GD (PrPΔ141–225, or “FTgpi”). Here we report that FTgpi mice develop a progressive, inexorably lethal neurodegeneration morphologically and biochemically similar to that triggered by anti-GD antibodies. FTgpi was mostly retained in the endoplasmic reticulum, where it triggered a conspicuous unfolded protein response specifically activating the PERK pathway leading to phosphorylation of eIF2α and upregulation of CHOP ultimately leading to neurodegeration similar to what was observed in prion infection.

Prions induce lethal neurodegeneration and consist of PrPSc, an aggregated conformer of the cellular prion protein PrPC. Antibody-derived ligands to the globular domain of PrPC (collectively termed GDL) are also neurotoxic. Here we show that GDL and prion infections activate the same pathways. Firstly, both GDL and prion infection of cerebellar organotypic cultured slices (COCS) induced the production of reactive oxygen species (ROS). Accordingly, ROS scavenging, which counteracts GDL toxicity in vitro and in vivo, prolonged the lifespan of prion-infected mice and protected prion-infected COCS from neurodegeneration. Instead, neither glutamate receptor antagonists nor inhibitors of endoplasmic reticulum calcium channels abolished neurotoxicity in either model. Secondly, antibodies against the flexible tail (FT) of PrPC reduced neurotoxicity in both GDL-exposed and prion-infected COCS, suggesting that the FT executes toxicity in both paradigms. Thirdly, the PERK pathway of the unfolded protein response was activated in both models. Finally, 80% of transcriptionally downregulated genes overlapped between prion-infected and GDL-treated COCS. We conclude that GDL mimic the interaction of PrPSc with PrPC, thereby triggering the downstream events characteristic of prion infection.

The enzyme thiopurine S-methyltransferase (TPMT) is involved in the metabolism of thiopurine drugs used to treat acute lymphoblastic leukemia and inflammatory bowel disease. Thus far, at least 29 variants of the TPMT gene have been described, many of which encode proteins that have low enzyme activity and in some cases become more prone to aggregation and degradation. Here, the two naturally occurring variants, TPMT*2 (Ala80 → Pro) and TPMT*5 (Leu49 → Ser), were cloned and expressed in Escherichia coli. Far-UV circular dichroism spectroscopy showed that TPMT*2 was substantially destabilized whereas TPMT*5 showed much greater stability comparable to that of wild-type TPMT (TPMTwt). The extrinsic fluorescent molecule anilinonaphthalene sulfonate (ANS) was used to probe the tertiary structure during thermal denaturation. In contrast to TPMTwt, neither of the variants bound ANS to a large extent. To explore the morphology of the TPMT aggregates, we performed luminescent conjugated oligothiophene staining and showed fibril formation for TPMT*2 and TPMT*5. The differences in the flexibility of TPMTwt, TPMT*2, and TPMT*5 were evaluated in a limited proteolysis experiment to pinpoint stable regions. Even though there is only one amino acid difference between the analyzed TPMT variants, a clear disparity in the cleavage patterns was observed. TPMT*2 displays a protected region in the C-terminus, which differs from TPMTwt, whereas the protected regions in TPMT*5 are located mainly in the N-terminus close to the active site. In conclusion, this in vitro study, conducted to probe structural changes during unfolding of TPMT*2 and TPMT*5, demonstrates that the various causes of the low enzyme activity in vivo could be explained on a molecular level., funding agencies|Swedish Childhood Cancer Foundation||Swedish Research Council||EMBO