The G127V variant of the prion protein interferes with dimer formation in vitro but not in cellulo

Scrapie prion, PrPSc, formation is the central event of all types of transmissible spongiform encephalopathies (TSEs), while the pathway with possible intermediates and their mechanism of formation from the normal isoform of prion (PrP), remains not fully understood. Recently, the G127V variant of the human PrP is reported to render the protein refractory to transmission of TSEs, via a yet unknown mechanism. Molecular dynamics studies suggested that this mutation interferes with the formation of PrP dimers. Here we analyze the dimerization of 127G and 127VPrP, in both in vitro and a mammalian cell culture system. Our results show that while molecular dynamics may capture the features affecting dimerization in vitro, G127V inhibiting dimer formation of PrP, these are not evidenced in a more complex cellular system.


Scientific Reports
| (2021) 11:3116 | https://doi.org/10.1038/s41598-021-82647-w www.nature.com/scientificreports/ providing some insights to the process [21][22][23][24][25][26][27][28][29][30][31][32][33][34][35] . In vitro spontaneous conversion of PrP is generally attempted by applying chemicals and/or varying the pH or temperature in order to enhance the process and to obtain forms of PrP that are at least in some aspects reminiscent to PrP Sc24, [36][37][38][39][40][41][42] . In this line, the reduction of the single disulfide bond present in the PrP's structure, destabilizes the protein 43,44 , while also promotes its conversion to oligomeric forms 39,44,45 . These misfolded oligomers are β-sheet rich and exhibit low-level PK-resistance, however, they are likely of different structures than those of PrP Sc derived from brain, which possess an intact intramolecular disulfide bond 46,47 . However, templated in vitro conversion is much more successful in generating authentic PrP Sc material 23,28 although infectious PrP Sc formation is reported in non-templated conversion as well 22,23,29,48 . The effect of the G127V mutation has been studied on the fibril formation kinetics of the mouse prion protein (mPrP), using an in vitro conversion assay, and important differences had been revealed between the wild type and the G126V mutant mPrP (correspondent of the G127V HuPrP), such as the critical concentration being higher, the lag phase being longer and the initial effective rate constant of fibril growth being slower in the case of the mutant 49 .
While these results provide a potential biophysical explanation for the observed protecting effect of G127V against prion diseases, a structural explanation is still missing. Recently, the structural and dynamic features of G127V variant were determined by NMR 50 , which together with molecular dynamics simulations suggested that HuPrP(G127V) prevents the formation of PrP dimers, offering a potential explanation for the inhibition of PrP Sc formation by the G127V mutation 50,51 .
The role of dimer formation in PrP Sc is not fully elucidated. A number of studies employing various approaches suggested that in in vivo conditions, at least some fraction of PrP C is existent as alpha helical dimers 52-55 evoking an importance of dimerization in PrP C 's physiological functions, importantly also in the cytoprotectivity played by PrP C56 . It has also been demonstrated by using N2a cells that the hydrophobic domain (aa 112 through 133) plays an essential role in the ability of the mouse PrP to dimerize 54 . The fragment aa 90 through 231 of the recombinant Syrian hamster PrP [PrP(90-231)] was also found to form alpha helical dimers in the presence of low SDS concentrations 57 , although, it was not confirmed whether this would represent the native form or if it would already be an intermediate or a species in the conversion pathway of PrP 58,59 . It can be envisaged that dimer formation is critical at the beginning of PrP's misfolding and this may or may not involve at first the partial unfolding of PrP.
To test the effect of G127V on the dimerization of the prion protein we employed two approaches: a mammalian cell culture based approach that follows the formation of mPrP dimer through co-immunoprecipitation of the dimers in presence and absence of an intermolecular disulfide bond 60 and an in vitro system that employs recombinant mouse prion protein containing a genetically encoded crosslinkable pBpa residue 61 . Our in vitro results are in line with the contention that was put forward in molecular dynamics studies 50,51 that the G127V mutation inhibits the formation of PrP dimers, however the cell culture experiments do not seem to display similar effects/observations.

Results
In vitro crosslinking. mPrP variants with site specifically inserted pBpa can be used to study heterodimerization. The formation of a dimer by the bacterially expressed recombinant mouse prion protein has been demonstrated and the dimer interface has been mapped in our earlier work 61 by using a genetically incorporated unnatural amino acid, pBpa, that crosslinks to C-H groups of the protein backbone and side chains of nearby proteins situated within a distance of less than 3.1 Å, when irradiated by UV light. Here we used this system to investigate the effect of the G126V mutation (the mouse equivalent of the human G127V mutation) on the formation of a prion dimer. The pBpa is incorporated either into the wild type (WT) or into the G126V mutant prion protein's sequence and to specifically monitor heterodimerization, an mCherry-tag is fused to one of the mPrPs in order to be able to distinguish the interacting proteins ( Fig. 1), i.e. WT and the mutant form, by SDS-PAGE. Effective crosslinking of mPrP dimers was obtained earlier with 127pBpa positional mutation 61 , accordingly, to monitor the formation of the heterodimers we placed a 127pBpa mutation into the sequence of the untagged mPrP in order to mediate covalent crosslinking to mPrP-mCherry fusion proteins (Fig. 2).
The configuration where the pBpa is incorporated into the mPrP-mCherry fusion protein was not used for heterodimerization studies because upon UV irradiation a band appeared at the expected position of a heterodimer on the SDS-gel of the protein alone, mPrP(Y127pBpa)-mCherry, without the presence of untagged mPrP as its binding partner.
Gly126Val mutation diminishes PrP dimerization. The mPrP(Y127pBpa) effectively crosslinks mPrP-mCherry when the position 126 contains the WT Gly residue (Figs. 3 and 4). However, the presence of a Val in position 126 in mPrP or in mPrP-mCherry or in both of them at the same time, diminishes the appearance of heterodimeric mPrP(Y127pBpa)-mPrP-mCherry crosslinked product, suggesting an inhibition upon heterodimer formation (Figs. 3 and 4). Next, we placed the pBpa residue to another position and performed crosslinking experiments with mPrP(S131pBpa) using the same experimental setup ( Supplementary Fig. S1). pBpa in the 131st position mediated significant, but less efficient crosslinkage between mPrP and mPrP-mCherry with Gly126. Nevertheless, Val126 placed to either mPrP or to mPrP-mCherry diminished the formation of the heterodimers ( Fig. 4 and Supplementary Fig. S1) crosslinked by this positional mutant.
We also examined the effect of the G126V mutation on the homodimerization of mPrP (Fig. 5, Supplementary Fig. S1). We employed three single pBpa mutants, the Y127pBpa, M128pBpa and S131pBpa prion variants, however, the diminishing effect of Val126 on homodimeric PrP formation reached a statistically significant level only in case of crosslinking via mPrP(Y127pBpa), given the relative size of the scatter of the data of 3 parallel experiments with the other two mutants (Fig. 6).  50,54 . After transient transfection of HeLa cells, the Western blot analysis reveals an additional signal at the size of PrP dimers in absence of the reducing agent β-mercaptoethanol (ME) (Fig. 7a, middle panel). A quantification indicated that V127 does not interfere with the formation of PrP homodimers (Fig. 7a, right panel). Under the conditions tested, about 60% of both G127-and V127PrP formed homodimers (Fig. 7a,  The major steps of the procedure employed to investigate the effect of G126V mutation on the heterodimer and homodimer formation of mPrP in in vitro conditions. DNA plasmids were constructed for E. coli expression of wild type, G126V and pBpa mutants mPrPs with or without fusion with mCherry (Step 1). The purified proteins (Step 2) were selected in pairs with one possessing an mCherry tag and another an untagged mPrP that usually contained the pBpa insertion to study heterodimer formation by crosslinking (Step 3). Single protein variants with pBpa were also selected to study homodimer formation (Step 4). Crosslinked hetero-and homodimers were assessed by SDS-PAGE and densitometry analysis (Step 5).  60 . Similarly, V127PrP is located at the plasma membrane and is complex glycosylated (Fig. 7b,c). Next, we wanted to address the possibility that the G to V substitution interferes with the formation of G127PrP/V127PrP heterodimers. To this end we inserted an HA epitope tag into G127PrP and a V5 epitope tag into V127PrP. To allow intermolecular disulfide bond formation both proteins contain a cysteine at position 132 ( Fig. 8a). After co-expressing G127PrP-HA and V127PrP-V5 in HeLa cells the lysates were subjected to an immunoprecipitation with anti-HA antibodies under non-reducing conditions. The immunopellet was then analyzed by Western blotting using anti-V5 antibodies. Indeed, V127PrP-V5 co-purified with G127PrP-HA, indicative of the formation of PrP heterodimers (Fig. 8b). To assess the dimerization under physiological conditions, we performed native immunoprecipitation assays with HA-and V5-tagged PrP constructs, which lacks the Ser-to-Cys mutation (Fig. 9). Cells were co-transfected with an HAtogether with a V5-tagged PrP construct and cell lysates were processed to an immunoprecipitation under native conditions with anti-HA antibody conjugated beads. A Western blot analysis of the immunopellet using anti-V5 antibodies verified the formation of homodimers of G127PrP and V127PrP as well of heterodimers under physiological conditions (Fig. 9a). Quantification of dimer formation as ratio to the input control showed no significant interference of the V127 mutation in the formation of PrP homo-or heterodimers (Fig. 9b).
In sum, the cell culture experiments indicate that the G to V substitution does not interfere with the dimerization of complex glycosylated and GPI-anchored PrP in cells.

Discussion
Molecular dynamics studies suggest that Tyr127 (128 in the human prion protein) has a key role in dimer formation and altering the 126Gly to Val changes the orientation of the Tyr side chain, while diminishing PrP dimerization. We found earlier that position 127 of recombinant mPrP when harboring a photo-crosslinkable amino acid, pBpa, and after UV irradiation results in the highest crosslinking for the Y127pBpa variant among 24 single pBpa-mutant variants tested 61 , confirming its key role in the formation of mPrP dimer interface and indicating at the same time that a Tyr to pBpa mutation does not diminish dimer formation. Since the pBpa substitution here is exactly in this aforementioned position, the G126V mutation in the mPrP(Tyr127pBpa) protein may affect directly the orientation of the pBpa residue, hence its crosslinking, without affecting the dimerization itself. However, the effect of Val126 is apparent in the crosslinking experiments of mPrP(Y127pBpa) with mPrP(G126V)-mCherry where the Val126 mutation is in the partner protein and can not directly affect the orientation of the pBpa side chain.
Former studies have pointed to the similarity between dimers of the full-length prion proteins formed in vitro in a test tube and those formed under more physiological conditions of mammalian cell culture experiments 52,[54][55][56][57][58]61 . Here the two systems seem to provide opposing results. While the presence of G126V diminishes dimer formation under in vitro conditions, this propensity of the G126V mutation is not apparent in cell culture experiments. The discrepancy observed may be due to the confinement of the GPI-anchored prion protein to a two-dimensional membrane surface in the cell culture experiment, in contrast to the recombinant protein that is freely moving in a solution in the test tube. Indeed, transgenic mouse models indicated that the C-terminal GPI anchor has a major impact on folding of PrP C . Expression of a secreted full-length PrP mutant (S231X) induces neurodegeneration in transgenic mice 62 . Moreover, brain extracts of both human patients expressing Y226X and mice expressing S231X contain infectious prions and transmit disease 62,63 . These findings  65 , which may constrain its conformation and thus may be responsible also for the apparent absence of the destabilizing effect of the V127 mutation on PrP dimers in cellulo. Whatever the correct explanation may be, the CO-IP data clearly show that the effect of the Val126 mutation on dimerization of the full length mPrP evidenced by in vitro and molecular dynamics studies, is not becoming apparent within the complexity of the cellular context.
In conclusion, our study emphasizes the important role of the flexible N-terminal and the GPI anchor in the conformational (mis)folding of PrP and the limitations of in vitro and in silico approaches to study misfolding of anchorless PrP as a model for the formation of infectious prions.

Materials and methods
All chemicals were purchased from Merck/Sigma-Aldrich unless indicated otherwise.

DNA plasmids used for bacterial protein expression and purification.
To produce the untagged and mCherry tagged wild type mPrP and the pBpa-mutant variants Y127pBpa, M128pBpa and S131pBpa of the untagged mPrP, the same DNA-plasmids were used as in a previous work 61 , named as pRSETB-mPrP-Cys, pRSETB-mPrP-mCh-6xHis-Cys, pRSETB-mPrP(Y127x)-Cys, pRSETB-mPrP(M128x)-Cys and pRSETB-mPrP(S131x)-Cys, respectively. Each of the plasmid constructs encoding for a G126V mutant mPrP (that corresponds to the G127V mutation of the human PrP sequence) was generated from the corresponding pRSETB plasmids of either untagged mPrP or the mCherry tagged mPrP encoding plasmids by using standard molecular biology techniques. Briefly, the plasmids were digested with NgoMIV and AfeI enzymes (Thermo Fisher Scientific) and the following oligonucleotides were inserted as indicated in Table 1.

Expression and purification of recombinant mPrPs.
Wild type, G126V-and pBpa-mutant variants of mPrP with and without an mCherry fusion tag, were expressed and purified as described 61 , briefly as follows. The plasmids encoding proteins lacking pBpa [pRSETB-mPrP-Cys, pRSETB-mPrP-mCh-6xHis-Cys and pRSETB-mPrP(G126V)-mCh-6xHis-Cys] were transfected into E. coli BL-21(DE3) competent cells, while those aiming for insertion of pBpa [pRSETB-mPrP(Y127x)-Cys, pRSETB-mPrP(M128x)-Cys, pRSETB-mPrP(S131x)-Cys, pRSETB-mPrP(G126V, Y127x)-Cys, pRSETB-mPrP(G126V, M128x)-Cys and pRSETB-mPrP(G126V, S131x)-Cys] into pre-transformed competent E. coli BL-21(DE3) cells comprising already a pEVOL-pBpF plasmid 66 . Single colonies of transfected cells were used to inoculate starter cultures in Luria-Bertani (LB) media containing the corresponding antibiotics and were grown for 16 h at 37 °C. 1% inoculum was added to 800 ml LB media to yield a starting OD 600 of ~ 0.05, and cultures were grown in presence of corresponding antibiotics. Protein expressions were induced at OD 600 of 0.6-0.8 by adding 1 mM IPTG alone for mPrP, mPrP-mCh and mPrP(G126V)-mCh or together with 0.02% arabinose and addition of 1 mM pBpa, for the pBpa-mutant PrPs. After induction, the cells were cultured for an additional 16 h, at 37 °C in case of the untagged mPrPs (WT and mutant proteins), while in case of mPrP-mCherry fusion proteins, for 4 h, at 37 °C. The bacterial cells were harvested and proteins were purified as described in 61 . The purified proteins eluted from the Ni-NTA columns were dialysed in three steps: first in 20 mM sodium acetate, pH 5.5, 1 mM EDTA and 1 mM EGTA, at a protein to buffer ratio of 1:1000 vol:vol, at 4 °C for 6 h, and at the second and third steps in 20 mM sodium acetate, pH 5.5, at 1:1000 vol:vol sample to dialysate ratio at 4 °C for 6 and 12 h, respectively. This buffer was used to store the proteins, either on ice or frozen at − 20 °C, until further use. The purity of the dialyzed proteins was tested on 10% SDS-PA gels with RAMA staining 67 and protein concentrations were measured by Bradford method 68 .

Photo-crosslinking of heterodimers and homodimers of mPrP.
To study the hetero-or homodimerization of mPrP, photo-crosslinking was performed using protein variants containing pBpa mutation, either alone or as being one of the interacting pairs, respectively. In order to examine the heterodimerization, mPrP-mCh or mPrP(G126V)-mCh was used and crosslinked with one of the following proteins: mPrP(Y127pBpa), mPrP(G126V, Y127pBpa), mPrP(S131pBpa) or mPrP(G126V, S131pBpa). Homodimerization was examined by crosslinking of single variants: mPrP, mPrP(Y127pBpa), mPrP(G126V, Y127pBpa), mPrP(M128pBpa), mPrP(G126V, M128pBpa), mPrP(S131pBpa) and mPrP(G126V, S131pBpa). Photo-crosslinking and evaluation of the crosslinked products were done as described earlier in 61 . Briefly, the reaction mixtures of heterogeneous or single proteins were irradiated at 6 µM of total protein in presence of either 0.06% SDS or 2% SDS (to favor dimerization or to assess the nonspecific background crosslinking, respectively) in phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 6 mM Na 2 HPO 4 .2H 2 O, 1.4 mM KH 2 PO 4 , pH 7.4) for 2 h by 365 nm UV light. Reaction volumes of 100 µl in 1.5 ml microfuge tubes were placed on ice at a Figure 6. Homodimer formation is diminished for G126V mutants of mPrPs. The homodimers were detected by photo-crosslinking single-type pBpa-mutant mPrPs: mPrP(Y127pBpa) or mPrP(S131pBpa) either comprising (grey color bars) or not (black bars) a 126Val mutation. mPrP containing no pBpa was used as the negative control (white bar). Percentages were calculated based on gel densitometry analysis as described in the Materials and methods. Statistical significance of differences was calculated by unpaired t-test. Significantly different pairs and the level of significance are indicated on the graph, stars correspond to p levels **p ≤ 0.01. "(X)" stands for pBpa in the notation of Y-axis category names referring to the pBpa-mutation at the indicated positions.  www.nature.com/scientificreports/ distance of 5 cm from the UV-tubes for crosslinking, while similar samples were kept at dark as controls. Samples were analyzed on 10% SDS-PA gels. The efficiency of heterodimer or homodimer formation of mPrP variants were measured densitometrically as described previously 61 , briefly as follows. The gels were scanned and the images were analyzed for band intensities by the ImageJ software version 1.52a 69 . To determine the percentages of the crosslinked heterodimer PrPs, the area corresponding to the heterodimer band on the intensity plot was divided by the total area corresponding to the sum of areas of all protein bands in the lane. The percentage of the homodimers formed was calculated as the area corresponding to the homodimer band divided by the sum of the monomer and the dimer area. All crosslinking experiments were performed three times, using independent protein samples (n = 3) obtained from different expressions. For data representations the mean ± SD of these values

DNA plasmids used for mammalian cell transfection. Plasmids were amplified in Escherichia coli
TOP10 (Thermo Fisher Scientific). The murine prion protein (GenBank accession number M18070) was modified by two amino acid exchanges (L108M/V111M) 70 to enable detection by the monoclonal antibody 3F4 71 .
The amino acid numbering that is indicated here refers to the human prion protein sequence. Mutations of the mouse prion protein were generated by standard PCR to insert an exchange from glycine at position 126, which corresponds to position 127 in the human prion sequence, to a valine (denoted hereafter G127 or V127PrP).   Immunofluorescence analysis was performed as described earlier 60 . HeLa cells were seeded on glass coverslips and transiently transfected with the indicated constructs. After 24 h cells were fixed with 4% paraformaldehyde for 10 min. One set was permeabilized with 0.2% Triton in PBS and the other set was left unpermeabilized. All cells were blocked with blocking solution (5% normal goat serum in PBS) for 1 h and incubated with the anti-3F4 antibody in blocking solution overnight at 4 °C. Cells were washed with PBS and incubated with the Cy3-conjugated anti-mouse secondary antibody (Alexa Fluor 488) for 1 h. After mounting with Fluoromount-G (Thermo Fisher Scientific), cells were visualized by fluorescence microscopy (Zeiss ELYRA PS.1 and LSM 880).

Co-immunoprecipitation.
To analyze formation of dimers, co-immunoprecipitation was performed as described earlier 60 , briefly as follows. HeLa cells were co-transfected with the indicated constructs (V5 or HA-tagged). Transiently transfected cells were lysed in a detergent containing buffer (0.5% Triton X-100, 0.5% sodium deoxycholate in PBS). Post nuclear supernatants were generated (14,000g, 10 min) and collected, and were incubated with anti-HA-agarose beads (overnight; 4 °C; rotating). The immunocomplex was washed twice with the lysis buffer and once with PBS before proceeding for analysis by Western blotting.
Western blotting. Western blot analyses were performed as described earlier 60 . Briefly, cell lysates were boiled in Laemmli sample buffer with or without 4% β-mercaptoethanol. After protein separation using SDS-PAGE proteins were transferred to nitrocellulose by electroblotting. Membranes were incubated with blocking solution (TBS with 0.1% Tween 20 and 5% skimmed milk) for 1 h at RT and incubated with primary antibody in blocking solution for 18 h at 4 °C. The Western blots were washed with TBS-T (TBS with 0.1% Tween 20) and incubated with respective secondary antibody (IRDye-Infrared Technology, LI-COR Biosciences or horseradish peroxidase in TBS-T) for 1 h at room temperature. Protein signals were detected via the peroxidase activity and enhanced chemiluminescence using ECL Western Blotting Substrate (Promega) or by an ODYSSEY 9120 Scanner (LI-COR Biosciences).
Quantification. Dimerization efficiency of the prion protein constructs was measured densitometrically (ImageJ software 69 or Image Studio Lite from LI-COR Biosciences) and visualized as percentage of the dimer signal to the sum of monomer and dimer signal. In the case of immunoprecipitation the dimerization efficiency was calculated as ratio of the dimer fraction to the HA-tagged PrP amount in the input control normalized to the WT/WT homodimers ratio. Data represent mean ± SD (standard deviation). Statistical differences between two conditions were determined by using a Student´s t-test (*p < 0.05; ns: not significant).