Disulfide-crosslink scanning reveals prion-induced conformational changes and prion strain-specific structures of the pathological prion protein PrPSc

Abstract

Prions are composed solely of the pathological isoform (PrP^Sc) of the normal cellular prion protein (PrP^C). Identification of different PrP^Sc structures is crucially important for understanding prion biology because the pathogenic properties of prions are hypothesized to be encoded in the structures of PrP^Sc However, these structures remain yet to be identified, because of the incompatibility of PrP^Sc with conventional high-resolution structural analysis methods. Previously, we reported that the region between the first and the second α-helix (H1∼H2) of PrP^C might cooperate with the more C-terminal side region for efficient interactions with PrP^Sc From this starting point, we created a series of PrP variants with two cysteine substitutions (C;C-PrP) forming a disulfide-crosslink between H1∼H2 and the distal region of the third helix (Ctrm). We then assessed the conversion capabilities of the C;C-PrP variants in N2a cells infected with mouse-adapted scrapie prions (22L-ScN2a). Specifically, Cys substitutions at residues 165, 166, or 168 in H1∼H2 were combined with cysteine scanning along Ctrm residues 220-229. We found that C;C-PrPs are expressed normally with glycosylation patterns and subcellular localization similar to WT PrP, albeit differing in expression levels. Interestingly, some C;C-PrPs converted to protease-resistant isoforms in the 22L-ScN2a cells, but not in Fukuoka1 prion-infected cells. Crosslink patterns of convertible C;C-PrPs indicated a positional change of H1∼H2 toward Ctrm in PrP^Sc-induced conformational conversion. Given the properties of the C;C-PrPs reported here, we propose that these PrP variants may be useful tools for investigating prion strain-specific structures and structure-phenotype relationships of PrP^Sc.

Keywords: prion; prion conversion; prion disease; prion protein; protein conformation; protein crosslinking; protein misfolding; protein structure.

Figure 1.

Design and expression of mutant PrPs with two cysteine (Cys) substitutions, 165C;C- and 166C;C-series. A, a schematic illustration of the secondary-structure components of mouse PrP and positions of the substituted cysteines (Cys). MoPrP, sequences of WT mouse PrP. B1 and B2, the first and the second β strands, respectively. H1, H2, and H3, the first, second, and third α helices (46). 165C or 166C was combined with another Cys scanning the distal H3 (Ctrm) from residue 220 to 229 (165C;C- or 166C;C-series, respectively; substituted Cys are underlined). The solid and broken lines with -S-S- represent the native and the newly introduced disulfide-crosslink, respectively. B and C, expression levels and PrP-banding patterns of the 165C;C- and 166C;C-series. Immunoblots were developed with anti-PrP mAb 3F4 from whole-cell lysates of transiently transfected N2a cells. Each two-Cys construct is named after the last digits of the residue numbers of the substituted Cys (in mouse numbering). 165C or 166C, mutants with Cys substitution only at position 165 or 166, respectively. Square brackets indicate positions of dimeric forms of the mutant PrPs. The upper and lower panels of blots represent images of short- and long-exposure of the same PVDF membranes, respectively. Note that all constructs have similar banding patterns as (3F4)MoPrP, indicating complex-type N-linked glycosylation. Two-Cys mutants show trace amounts of dimmers, whereas 165C and 166C have discernible dimers despite their low expression levels. Bottom panels are graphs showing expression levels of each series quantified by densitometry. Each bar represents mean ± S.D. from three independent experiments. D, single-Cys PrPs have substantial levels of PrP dimer formation that disappear after DTT treatment. Immunoblots were probed with 3F4 mAb showing banding patterns of mutant PrPs with only a single Cys substitution either in H1∼H2 or Ctrm. DTT (+) and (−) samples, with or without DTT in the sample buffer. The square bracket indicates the position of the dimeric forms. E, disruption of the native disulfide bond drastically changes the banding pattern, suggesting that the substituted Cys in the 165C;C- and 166C;C-series do not affect the native disulfide bond formation. Left panel, scheme illustrating the positions of alanine substitutions for the native Cys. The solid or broken lines with -S-S- represent putative disulfide-crosslinks of 6C;C213A or 6C;C178A, respectively. Right panel, immunoblots with 3F4 mAb showing PrP banding patterns of the mutant PrPs combining 166C with the alanine substitution. Note that the banding patterns of 6C;C178A and 6C;C213A are very different from that of 166C, presumably due to high-mannose-type N-glycosylation, unlike the banding patterns of 165C;C- or 166C;C-series constructs.

Figure 2.

Confocal microscopy analysis of relative surface expression levels of representative 166C;C-PrPs correlated to CtxB surface staining. CtxB live cell-stained transfected cells were fixed and permeabilized, and then analyzed for 3F4-PrP expression. 3F4-PrP expression is shown in red, CtxB staining in green, and the right panels show merged signals. The upper panels show (3F4)MoPrP-transfected cells as positive control, and lower panels show mock-transfected cells as negative control. Overlaying the CtxB signals with 3F4-PrP expression was used to quantify the relative surface expression levels of C;C-PrPs (lower graph). Constructs were detected at similar rates at the cell surface, excluding major differences in subcellular localization.

Figure 3.

Substituted Cys of 166C;C-constructs form an intramolecular disulfide-crosslink: assessment of fragment patterns after V8-protease digestion. A, a schematic illustrating the position of the FLAG-tagged 166C;C-constructs along with putative cleavage sites by V8-protease (vertical lines). B, V8-digested fragment profiles are similar between FLAG-tagged WT ((3F4)Mo-FL) and 166C-FLAG (166C-FL), but those of 6C;3C-FLAG and 6C;9C-FLAG (6C;3C-FL and 6C;9C-FL, respectively) are different. Immunoblots were probed with anti-FLAG polyclonal antibody, or 3F4-mAb, showing nondigested and V8-digested PrPs, with or without DTT in the sample buffer. The upper and middle panel images were obtained from the same PVDF membrane with shorter and longer exposure, respectively. The bottom panel shows immunoblots re-probed with 3F4 mAb. Arrowhead indicates full-length FLAG-tagged PrP molecules, curved brackets are positions of the intermediate-size fragments that diminish by DTT, and square brackets indicate the smallest fragments.

Figure 4.

165C;C- and 166C;C-series mutants convert to PK-resistant PrP forms (PK-res) in 22L-ScN2a cells. The upper and lower panels represent images of short- and long-exposure of the same PVDF membranes, respectively. The square brackets indicate the position of the dimeric form. A and B, PK-res PrP^Sc of 165C;C- and 166C;C-series, respectively. Immunoblots probed with 3F4 mAb demonstrating the levels of PK-res PrP^Sc. Just as for PrP^C isoforms, PK-res PrP^Sc of the C;C-series lacks the dimeric forms. C, PK-res PrP of 166C;C-series maintain the intramolecular disulfide-crosslinks throughout conversion. Immunoblots were developed with 3F4 mAb showing PK-res PrP^Sc of 166C;C-series and mutant PrPs with a single substituted Cys, either at 166C or in Ctrm. Note that all the single-Cys constructs have substantial levels of dimeric forms, in contrast to 166C;C-series. D, the dimeric forms of PK-res PrP^Sc of single-Cys mutants disappear after DTT treatment, proving intermolecular disulfide-crosslinks. Immunoblots with 3F4 mAb comparing single-Cys PrP, 166C and 229C, and a double-Cys PrP, 6C;9C. DTT (+) and (-) indicate samples prepared with or without DTT in the sample buffer. E, PK-res PrP formation of 6C;9C is PrP^Sc-dependent. Immunoblots with 3F4 mAb demonstrating PK resistance in the lysates from 22L prion-infected and noninfected N2a cells, transiently-transfected with (3F4)MoPrP or C;9C. 22L (+) or (−) indicates samples from 22L-infected or noninfected N2a cells. PK (+) or (−) indicates samples with or without PK digestion. Note that there is no detectable PK-resistant PrP in lysates from noninfected N2a cells. F, conversion-incompetent C;C-PrPs can interact with PrP^Sc. Immunoblots with 3F4 mAb showing efficient dominant-negative inhibition effects on co-transfected (3F4)MoPrP by the conversion-incompetent C;C-PrPs, namely 6C;0–6C;5C. Δ159, a deletion mutant PrP lacking residue 159 as a control (22), was on the same membrane. Note that unnecessary lanes were eliminated. G, the conversion reaction of 6C;9C is relatively resistant to the inhibitory effects of the Q218K substitution. Immunoblots with 3F4 mAb comparing PK-res PrP of WT or 6C;9C and their Q218K counterparts. Decrease of PK-res formation by Q218K is smaller in 6C;9C than that in WT.

Figure 5.

A model to explain the discrepancy between high PrP expression and efficient PK-res conversion of convertible mutants. A, hypothetical positional changes of H1∼H2 in the PrP^Sc-dependent conversion reaction. B, a crosslink between residues 165 (or 166) and the distal portion of Ctrm deforms the conformation of PrP^C, but does not severely affect the conversion because the position of H1∼H2 is suitable for conversion. C, a crosslink between residues 165 (or 166) and the proximal portion of Ctrm, e.g. 220C, might inhibit PK-res conversion by hampering the positional changes of H1∼H2. D, could there be a disulfide-crosslink connecting the more C-terminal H1∼H2 and the more proximal Ctrm, which would not hamper the conversion into PK-res?

Figure 6.

Convertible mutants of the 168C;C-series support the hypothetical positional change of H1∼H2 during conversion to PK-res. The upper and lower panels of blots represent images of short- and long-exposure to the same PVDF membranes, respectively. A, a schematic illustrating positions of substituted Cys residues of 167C;C-, 168C;C-, and 169C;C-series. B, PrP^C forms of 168C;C-series show similar banding patterns as WT PrP. Immunoblots with 3F4 mAb showing expression levels and PrP banding patterns of 168C;C-series. Square brackets denote positions of dimeric forms of mutant PrPs. C, conversion capacities of 168C;C-series. Immunoblots with 3F4 mAb demonstrating the levels of PK-res formation of the 168C;C-series. The square brackets indicate the position of the dimeric form. Like PrP^C forms, PK-res of the 168C;C-series also lack the dimeric forms. D, PK-res PrP formation of 8C;5C is PrP^Sc dependent, such as that of 6C;9C PrP. Immunoblots probed with 3F4 mAb comparing samples prepared from 22L-ScN2a and noninfected N2a cells, transfected with 8C;5C or 6C;9C, and digested with different concentrations of PK. Left and right panels show samples from cells transfected with 8C;5C and 6C;9C, respectively. Note that PK-res PrP of 8C;5C is only found in 22L-ScN2a cells, similar to 6C;9C. PK-digested and nondigested samples were on the same membrane and unnecessary lanes were removed. E, conversion of 8C;5C into the PK-res isoform is prion strain-dependent. Immunoblots were probed with 3F4 mAb comparing PK-res of (3F4)MoPrP and 8C;5C from transiently transfected Fukuoka1-infected or RML-infected N2a58 cells. Note that 8C;5C does not convert to PK-res PrP in Fukuoka1-infected cells while it converts in RML-infected cells.