Affinity-Tagged Miniprion Derivatives Spontaneously Adopt Protease-Resistant Conformations

Abstract

An abridged PrP molecule of 106 amino acids designated PrP106 can form infectious miniprions in transgenic (Tg) mice (29). Addition of six-histidine (His₆) affinity tags to selective sites within PrP106 resulted unexpectedly in new PrP proteins that spontaneously adopted protease-resistant conformations when expressed in neuroblastoma cells and Tg mice. Acquisition of protease resistance depended on the length, charge, and placement of the affinity tag. Introduction of the disease-linked mutation E200K into the sequence of PrP106(140/6His) increased the recovery of protease-resistant PrP fivefold, whereas introduction of the mutations C213A and Δ214–220 did not affect the recovery of protease-resistant PrP. Treatment of cultured cells expressing affinity-tagged PrP106 mutants with polypropyleneimine dendrimer rendered these proteins sensitive to protease digestion in a manner similar to wild-type PrP^Sc. We conclude that certain affinity-tagged PrP106 proteins spontaneously fold into conformations partially resembling, yet distinct from, wild-type PrP^Sc. These proteins might be useful tools in the identification of new disease-causing mutations as well as for screening compounds for therapeutic efficacy.

Prion diseases are disorders of protein conformation in which the cellular form of the prion protein, PrP^C, undergoes a pathogenic conformational change into an infectious isoform, PrP^Sc (20–22). During this conformational change, full-length PrP^C containing ∼40% α-helix and little β-sheet folds into PrP^Sc that is composed of ∼30% α-helix and ∼40% β-sheet (17, 18, 24). Attempts to identify a chemical modification that drives this profound conformational change that is responsible for PrP^Sc formation have been unrewarding (28). Earlier studies showed that the protease-resistant core of PrP^Sc designated PrP 27-30 polymerizes into amyloid fibrils (23). PrP amyloid, like all other known amyloids, possesses a high β-sheet content (3, 7).

Determining the high-resolution three-dimensional molecular structures of PrP^C and PrP^Sc is an important step toward deciphering the mechanism underlying prion diseases. Recombinant PrP molecules derived from Escherichia coli refolded into conformations with a high α-helical content that appear to approximate the structure of PrP^C (8, 9, 14). On the other hand, investigation of PrP^Sc structure has been hindered by the extreme insolubility and molecular heterogeneity of this isoform (23, 32).

To elucidate the tertiary structure of PrP^Sc, we attempted to design smaller infectious PrP^Sc molecules that might be more amenable to structural investigation. We recently reported that a PrP deletion mutant MHM2(Δ23–88,Δ141–176), designated PrP106, successfully forms infectious miniprions in Prnp^0/0 mice (29). PrP106 contains only 106 amino acids, compared to the 208 residues in full-length PrP. In order to facilitate purification of PrP^Sc106, we incorporated an affinity tag consisting of six histidine residues (His₆) at various sites within the PrP106 backbone. Unexpectedly, we found that some of these affinity-tagged derivatives spontaneously adopted conformations with the same level of protease resistance as PrP^Sc106.

MATERIALS AND METHODS

Explanation of nomenclature.

MHM2 is a full-length chimeric construct that differs from wild-type MoPrP at positions 108 and 111 (27). Substitution at these positions with the corresponding residues (109 and 112, respectively) from the Syrian hamster (SHa) PrP sequence creates an epitope for the anti-PrP 3F4 monoclonal antibody (MAb) (12), which does not recognize wild-type MoPrP, and hence facilitates specific detection of the transgene by Western blot.

Mature MHM2 and MoPrP comprise residues 23 to 230 after processing because residues 1 to 22 are removed by a signal peptidase and residues 231 to 254 are removed during the addition of a GPI anchor. PrP106 refers to a truncated MHM2 molecule, in which residues 23 to 88 and residues 141 to 176 have been removed, and can also be designated as MHM2(Δ23–88,Δ141–176). Additional sequences inserted into MHM2 or PrP106 are denoted by the following convention: N-terminal residue of attachment and/or additional sequence in parenthesis. For example, PrP106(140/GASGAS) includes the following residues: 89-140-Gly-Ala-Ser-Gly-Ala-Ser-177-230.

Construction of DNA plasmids and transgenic (Tg) mice.

Most of the new constructs described here were created by standard cassette mutagenesis of pCOMBO3 (Mike Scott) or psp72PrP106 (Tamaki Muramoto), using oligonucleotides obtained from Gibco-BRL. XbaI and AspI were used to insert new sequences between residues 140 and 177. DNA sequencing (Perkin-Elmer) with T7 and SP6 primers was used to verify the sequence of every new insert. The mutagenized PrP inserts were removed from psp72 or pCOMBO plasmids by digestion with BglII/XhoI and subcloned into BamHI/XhoI-digested pSPOX.neo vector (27) to create pSPOX N2a cell expression plasmids. pSPOXPrP106(225/6His) was created by substitution of a BstEII/XhoI-digested pSPOX72MHM2(225/6His) (Kiyatoshi Kaneko) insert into BstEII/XhoI-digested pSPOXPrP106 vector. pSPOXPrP106(140/6His,E200K) was created by substitution of BstEII/XhoI-digested pCOMBO2MHM2 E200K insert into BstEII/XhoI-digested pSPOXPrP106(140/6His) vector. Qiagen Maxiprep columns were used to purify pSPOX expression plasmids for transfection experiments.

CosTetPrP106(140/6His) and CosTetPrP106(225/6His) cosmids were generated in a two-step process from their respective pSPOX plasmids. First, a 400-bp KpnI/XhoI insert from the appropriate pSPOXPrP plasmid was ligated into KpnI/XhoI-digested psp72(SalI)MHM2 vector (27). Second, the SalI/XhoI PrP insert from the modified psp72(SalI)PrP construct was cloned into XhoI-digested CosTet.neo. Microinjection, breeding, and screening of Tg animals was performed as previously described (26).

Expression in neuroblastoma cells.

Stock cultures of N2a and ScN2a cells were maintained in minimal essential medium with Earle's salts plus 10% fetal bovine serum (FBS), 10% Glutamax (Gibco-BRL), 100 U of penicillin per ml, and 100 μg of streptomycin per ml. Cells from a single confluent 100-mm dish were trypsinized and split into 10 separate 60-mm dishes containing Dulbecco's modified Eagle medium plus 10% FBS, 10% Glutamax, 100 U of penicillin per ml, and 100 μg of streptomycin per ml (supplemented DME) 1 day prior to transfection. For each construct, 15 μg of DNA was resuspended in 150 μl of sterile HEPES-buffered saline (HBS) on the day of transfection. The DNA solution was then mixed with an equal volume of 333 μg of DOTAP (Boehringer Mannheim, Indianapolis, Ill.) per ml in HBS in Falcon 2059 tubes and incubated at room temperature for 10 min to allow formation of DNA-lipid complexes. Supplemented DME (2.5 ml) was added to the mixture, and this was then pipetted onto drained cell monolayers. The following day, the medium containing the DNA-lipid complexes was removed and replaced with fresh supplemented DME.

Three days after transfection, cells were harvested by lysis in 1.2 ml of 20 mM Tris (pH 8.0) containing 100 mM NaCl, 0.5% NP-40, and 0.5% sodium deoxycholate (DOC). Nuclei were removed from the lysate by centrifugation at 2,000 rpm for 5 min. This lysate typically had a total protein concentration of 0.5 mg/ml, as measured by the bicinchoninic acid protein assay (Pierce, Rockford, Ill.).

Western blotting.

For samples not treated with proteinase K, 40 μl of whole lysate (representing 20 μg of total protein) was mixed with 40 μl of 2× sodium dodecyl sulfate (SDS) sample buffer. For proteinase K digestion, 1 ml of lysate was incubated with 20 μg of proteinase K per ml (total protein/enzyme ratio = 25:1) for 1 h at 37°C. Proteolytic digestion was terminated by the addition of 8 μl of 0.5 M phenylmethylsulfonyl fluoride in absolute ethanol. Samples were then centrifuged for 75 min in a Beckman TLA-45 rotor (Fullerton, Calif.) at 100,000 × g at 4°C. The pellet was resuspended by repeated pipetting in 80 μl of 1× SDS sample buffer. The entire sample (representing 0.5 mg of total protein before digestion) was boiled for 5 min and cleared by centrifugation for 1 min at 14,000 rpm in a Beckman ultrafuge. For solubility studies, the supernatant fractions of proteinase K-digested samples were precipitated in 10 volumes of MeOH and resuspended in 80 μl of 1× SDS sample buffer. SDS-polyacrylamide gel electrophoresis (PAGE) was carried out in 1.5-mm, 15% polyacrylamide gels (13) or in 16% Tricine gels (Novex) as indicated.

Following electrophoresis, Western blotting was performed as previously described (25). Membranes were blocked with 5% nonfat milk protein in PBST (calcium- and magnesium-free PBS plus 0.1% Tween 20) for 1 h at room temperature. Blocked membranes were incubated with primary 3F4 MAb at a 1:5,000 dilution in PBST overnight at 4°C. Following incubation with primary antibody, membranes were subjected to three 10-min washes in PBST, incubated with horseradish peroxidase-labeled anti-mouse immunoglobulin G secondary antibody (Amersham Life Sciences, Arlington Heights, Ill.) diluted 1:5,000 in PBST for 25 min at room temperature, and subjected to three additional 10-min washes in PBST. After chemiluminescent development with enhanced chemiluminescence (ECL) reagent (Amersham) for 1 to 15 min, blots were sealed in plastic covers and exposed to ECL Hypermax film (Amersham). Films were processed automatically in a Konica film processor.

RESULTS

Protease resistance of affinity-tagged PrP106 derivatives in neuroblastoma cells.

We incorporated His₆ affinity tags at three different sites within the PrP106 backbone. These sites were: (i) the extreme N terminus at residue 89, (ii) the C terminus between residues 225 and 226, and (iii) the junction of the internal deletion bounded by residues 140 and 177. The new constructs were expressed in both scrapie-infected (ScN2a) and uninfected (N2a) neuroblastoma cells to assess protease resistance as previously described (27). PrP106 expressed in either ScN2a or N2a cells was resistant to proteinase K digestion for 30 min at 37°C using a total protein/enzyme ratio of 71:1 (29), but it was largely proteolyzed by more-stringent digestion for 1 h at a protein/enzyme ratio of 25:1 (Fig. 1A, lane 2). Similarly, PrP106(89/6His) was fully proteolyzed under these stringent conditions (data not shown). In contrast, MHM2, which forms full-length PrP^Sc, was resistant to proteinase K digestion for 1 h at a protein/enzyme ratio of 25:1 in ScN2a cells, as expected (Fig. 1A, lane 1). Surprisingly, PrP106(140/6His), PrP106(225/6His), and PrP106(140/6His,225/6His) were all resistant to proteinase K digestion under stringent conditions in both ScN2a (Fig. 1A, lanes 3 to 5) and N2a cells (data not shown).

FIG. 1.

Expression of His₆-labeled PrP106 derivatives in neuroblastoma cells. ScN2a cells were transfected with the following expression constructs: lane 1, MHM2; lane 2, PrP106; lane 3, PrP106(140/6His); lane 4, PrP106(140/6His,225/6His); lane 5, PrP106 (225/6His); lane 6, mock transfection. (A) The minus symbol indicates the whole lysate of an untreated sample; the plus symbol indicates the pellet fraction of a sample digested with 20 μg of proteinase K per ml for 1 h at 37°C. (B) Supernatant fractions of proteinase K-digested samples precipitated in 10 volumes methanol and resuspended in SDS sample buffer. Western immunoblotting was performed with 3F4 MAb as described in Materials and Methods. The apparent molecular masses based on the migration of protein standards are given in kilodaltons.

The protease-resistant fragment of PrP106(140/6His) migrated on SDS-PAGE with an apparent molecular mass of 19 kDa (Fig. 1A, lane 3), whereas the protease-resistant core of PrP106(225/6His) had an apparent molecular mass of 23 kDa (Fig. 1A, lane 5). PrP106(140/6His) could be distinguished from PrP106(225/6His) after proteinase K digestion not only by its mobility on SDS-PAGE but also by its solubility. In 0.5% NP-40–0.5% DOC, approximately 25% of protease-resistant PrP106(140/6His) and 10% of protease-resistant PrP106(140/6His,225/6His) was soluble (Fig. 1B, lanes 3 and 4), whereas PrP106(225/6His) appeared to be largely insoluble (Fig. 1B, lane 5). Taken together, these results suggest that PrP106(140/6His) might adopt a more open conformation than PrP106(225/6His), thereby increasing its solubility and exposing additional regions of the polypeptide chain to proteinase K digestion. Interestingly, protease digestion of PrP106(140/6His,225/6His) yielded both 19- and 23-kDa fragments in a roughly 1:1 ratio (Fig. 1A, lane 4). Prolonged digestion of PrP106(140/6His,225/6His) did not alter the ratio of 19-kDa to 23-kDa bands (data not shown), indicating that the 19-kDa band does not derive from proteolysis of the 23-kDa band. An alternative explanation is that PrP106(140/6His,225/6His) simultaneously adopts two distinct conformations, which are differentially proteolyzed. The ability of PrP molecules to adopt different protease-resistant conformations has been well documented in the case of PrP^Sc molecules derived from different prion strains (2, 4, 31).

Mutagenesis of affinity-tagged PrP106 molecules.

Because the presence of a His₆ affinity tag at the N terminus did not protect PrP106(89/6His) from stringent proteinase K digestion, we concluded that appropriate placement of the affinity tag is one requirement for formation of protease-resistant PrP106 derivatives. To identify other requirements, we systematically altered the charge and length of the tag in PrP106(140/6His) and measured the protease resistance of the resulting molecules expressed in ScN2a cells. Changing the composition of the tag from six histidines to six lysines did not diminish protease resistance (Fig. 2B, compare lanes 1 and 5), whereas substituting six glutamates, six alanines, or the mixed amino acid spacer GASGAS completely abolished protease resistance (Fig. 2B, lanes 4, 6, and 7). The protease resistance of PrP106(140/6Lys) suggests that positive charges in the affinity tag may facilitate the formation of PrP106 protease-resistant fragments. Increasing the length of the polyhistidine tag to seven residues resulted in less-intense protease-resistant bands (Fig. 2B, lane 2), as did substituting alternating alanine residues within the His tag (Fig. 2B, lane 3). Taken together, these results indicate that appropriate insert placement, spacing, and charge are all necessary for the production of protease resistance in affinity-tagged PrP106 derivatives.

FIG. 2.

Expression of PrP106(140/6His) affinity tag mutants in neuroblastoma cells. ScN2a cells were transfected with PrP106 expression constructs in which the following sequences replaced the internal deletion Δ141–176; lane 1, His₆; lane 2, His₇; lane 3, HAHAHA; lane 4, GASGAS; lane 5, Lys₆; lane 6, Glu₆; lane 7, Ala₆; lane 8, mock transfection. (A) Whole lysates of undigested samples. (B) Pellet fractions of samples digested with 20 μg of proteinase K per ml for 1 h at 37°C. Western immunoblotting was performed with 3F4 MAb as described in Materials and Methods. Apparent molecular masses based on the migration of protein standards are given in kilodaltons.

Mutagenesis studies of full-length MHM2 expressed in ScN2a cells have identified specific amino acids that appear to be important for the formation of full-length PrP^Sc. These include residue Q218 (11), as well as cysteine residues 178 and 213, which form a disulfide bridge (16). Substitution mutations of these amino acids prevent the conversion of MHM2 into its scrapie isoform in ScN2a cells. We sought to determine the effect on protease resistance of similar mutations introduced into PrP106(140/6His). The mutations C213A and Δ214–220 (removal of Q218 along with two adjacent turns of α-helix C) each prevented the formation of PrP^Sc when introduced into MHM2 (Fig. 3B, lanes 5 and 6). However, when these mutations were introduced into PrP106(140/6His), the resulting derivatives remained protease-resistant in both N2a and ScN2a cells (Fig. 3B, lanes 2 and 3).

FIG. 3.

Expression of PrP molecules with disruptive mutations in neuroblastoma cells. ScN2a cells were transfected with the following expression constructs. Lane 1, PrP106(140/6His); lane 2, PrP106(140/6His,C213A); lane 3, PrP106(140/6His,Δ214–220); lane 4, MHM2; lane 5, MHM2(C213A); lane 6, MHM2(Δ214–220); lane 7, mock transfection. (A) Whole lysates of undigested samples. (B) Pellet fractions of samples digested with 20 μg of proteinase K per ml for 1 h at 37°C. Western immunoblotting was performed with 3F4 MAb as described in Materials and Methods. Apparent molecular masses based on the migration of protein standards are given in kilodaltons.

Genetic studies have linked a number of PrP mutations to hereditary forms of prion disease (for reviews, see references 6 and 19). We introduced one such mutation, E200K, into PrP106(140/6His) to evaluate the effect on protease resistance. This mutation has been linked to familial Creutzfeldt-Jakob disease among a population of Libyan Jews (5, 10, 15). Whereas introduction of the E200K mutation did not affect the protease resistance of full-length PrP or PrP106 (Fig. 4A, lane 3), it did increase the recovery of protease-resistant PrP106(140/6His) by approximately fivefold (Fig. 4A, lane 5).

FIG. 4.

Expression and clearance of mutant affinity-tagged PrP106 proteins in neuroblastoma cells. (A) ScN2a cells were transfected with the following constructs: lane 1, MHM2; lane 2, PrP106; lane 3, PrP106(E200K); lane 4, PrP106(140/6His); lane 5, PrP106(140/6His,E200K). (B) ScN2a cells were transfected with MHM2 (lanes 1 and 2) PrP106(225/6His) (lanes 3 and 4), and PrP106(140/6His,E200K) (lanes 5 and 6). Cells from even-numbered lanes were treated with 150 μg of PPI generation 4.0 (Aldrich) per ml for 4 h prior to harvest. In both panels, the minus symbol indicates the whole lysate of an untreated sample and the plus symbol indicates the pellet fraction of a sample digested with 20 μg of proteinase K per ml for 1 h at 37°C. Western immunoblotting was performed with 3F4 MAb as described in Materials and Methods. Apparent molecular masses based on the migration of protein standards are given in kilodaltons.

Dendrimer-mediated clearance of protease-resistant affinity-tagged PrP106 proteins.

Branched polyamines such as polypropyleneimine (PPI) dendrimers render wild-type PrP^Sc sensitive to protease digestion in ScN2a cells (30). To investigate whether affinity-tagged PrP106 derivatives might share with PrP^Sc the property of being susceptible to dendrimers, we treated ScN2a cells expressing either full-length MHM2, PrP106(225/6His), or PrP106(140/6His,E200K) with PPI. The results indicate that both PrP106(225/6His) and PrP106(140/6His,E200K) were rendered protease sensitive by PPI (Fig. 4B).

Tg mice expressing affinity-tagged PrP106 derivatives.

We expressed PrP106(225/6His) in Tg mice deficient for wild-type PrP (Prnp^0/0). Three separate lines of Tg[PrP106(225/6His)]Prnp^0/0 mice were generated: two lines expressed the transgene at 32 times (32×) the expression level of normal Syrian hamster (SHa) PrP, and one line expressed the transgene at 8 times (8×) the level of SHaPrP. Uninoculated mice from both lines of 32× expressor lines spontaneously developed a fatal neurological disease at ∼233 ± 8 days of age (Table 1). The clinical signs of disease resembled scrapie and included ataxia, kyphosis, dull coat, masked facies, loss of deep pain sensation, proprioceptive defects, and weight loss. Neuropathological examination revealed no vacuolation or nerve cell loss and only mild gliosis (data not shown). However, after hydrolytic autoclaving, numerous PrP deposits could be seen in nerve cell bodies, as well as throughout the gray matter neuropil and white matter (Fig. 5A and B). Intracellular deposits measured ∼5 μm in diameter at most, while those in the gray and white matter neuropil were as large as 20 μm in diameter. PrP106(225/6His) deposits were more abundant and larger than the 1- to 10-μm PrP106 deposits seen in uninoculated Tg(PrP106) mice (29). Congo red staining of PrP106(225/6His) deposits was negative for amyloid (data not shown). Western blot analysis of the brain homogenates from Tg[PrP106(225/6His)]15961/Prnp^0/0 mice demonstrated the presence of a proteinase K-resistant, insoluble band (Fig. 5C). A premorbid 40-day-old Tg[PrP106(225/6His)]15961/Prnp^0/0 mouse displayed similar levels of proteinase K-resistant PrP as ∼230-day-old, symptomatic counterparts (data not shown). We attempted to transmit the spontaneous neurological disease of Tg[PrP106(225/6His)]Prnp^0/0 mice to both Tg(MoPrP)Prnp^0/0 and Tg(PrP106)Prnp^0/0 mice by intracerebral inoculation of 1% brain homogenate, but all of the animals remained alive and well >300 days after inoculation (Table 1). In contrast to the 32× overexpressors, uninoculated Tg[PrP106(225/6His)]15947/Prnp^0/0 mice expressing the transgene at 8× the level found for SHaPrP in Syrian hamsters survived >215 days without signs of disease (Table 1).

TABLE 1.

Transmission studies in Tg mice

Transgenea	Expression levelb (fold)	Inoculum	Age (days) at disease onset or incubation period (days ± SEM)	n/n0
PrP106(140/6His)	8	None	>250	0/8
PrP106(225/6His)	32	None	233 ± 8	12/12
PrP106(225/6His)	8	None	>215	0/20
MoPrP	4	PrP106(225/6His)c	>300	0/30
PrP106	8	PrP106(225/6His)c	>300	0/30

^aAll transgenes are expressed in Prnp^0/0 mice. 

^bPrP expression level relative to normal Syrian hamsters. 

^cOne percent of brain homogenates prepared from three separate spontaneously sick Tg[PrP106(225/6His)]Prnp^0/0 mice were inoculated intracerebrally into indicator mice (n = 10 per experiment). 

FIG. 5.

Accumulation of PrP in Tg[PrP106(225/6His)]Prnp^0/0 mice. (A) High-magnification view of the neocortex of a spontaneously ill Tg[PrP106(225/6His])15961/Prnp^0/0 mouse at 225 days of age. A section stained by the periodic acid-Schiff (PAS) histochemical method shows multiple PAS-positive deposits in a nerve cell body (smaller arrow) and larger PAS-positive deposits in the neuropil (larger arrow). (B) Lower-magnification view of the same region of the neocortex immunostained for PrP106 by the hydrolytic autoclaving method using 3F4 MAb and showing large numbers of deposits scattered diffusely throughout neuropil. The bar in panel A is 25 μm, and the bar in panel B is 50 μm. (C) Immunoblot of mouse brain homogenates. Lanes: 1, normal Syrian hamster; 2, 60-day-old, uninoculated Tg(PrP106)4290/Prnp^0/0 mouse; 3, 120-day-old, scrapie-affected Tg(PrP106)4290/Prnp^0/0 mouse; 4, 225-day-old, uninoculated, ataxic Tg[PrP106(225/6His)]15961/Prnp^0/0 mouse. Brain homogenates were prepared as previously described (29). The minus symbol indicates the whole lysate of untreated sample; the plus symbol indicates the pellet fraction of sample digested with 20 μg of proteinase K per ml for 1 h at 37°C (total protein/enzyme ratio = 50:1). Western immunoblotting was performed with MAb 3F4 as described in Materials and Methods. Apparent molecular masses based on the migration of protein standards are given in kilodaltons.

We also generated one line of Tg[PrP106(140/6His)]Prnp^0/0 mice with an expression level that is 8× that of SHaPrP. These mice survived >250 days without any signs of spontaneous neurological disease (Table 1). The neuropathological examination of one healthy 60-day-old Tg[PrP106(140/6His)]Prnp^0/0 mouse was unremarkable, and hydrolytic autoclaving revealed only moderate levels of 1- to 10-μm PrP deposits comparable to those previously seen in Tg(PrP106)Prnp^0/0 mice (29) (data not shown).

DISCUSSION

We previously observed that PrP106 expressed either in uninfected N2a cells or in transgenic mice formed a partially protease-resistant fragment and that recombinant PrP106 refolded into a structure that was predominantly a β-sheet (1, 29). These observations raised the possibility that PrP106 might spontaneously adopt a conformation resembling an intermediate on the pathway to PrP^Sc formation (29). As reported above, we unexpectedly discovered that incorporation of His₆ affinity tags into specific locations within the PrP106 backbone produces molecules that spontaneously adopt even more highly protease-resistant conformations resembling PrP^Sc.

How do affinity tags increase the protease resistance of PrP106? These tags may promote the intrinsic folding of PrP106 derivatives into a more compact tertiary structure. Alternatively, they may facilitate intermolecular interactions with other PrP molecules or with cellular factors that promote compact folding of PrP. It is likely that structural studies will be required to distinguish between these possibilities. Location, charge, and spacing all influence the ability of sequence tags to increase the protease resistance of PrP106. Placement of the affinity tag either at the C terminus or in the gap left by the internal deletion Δ141–176 increased the protease resistance of PrP106. However, PrP106(140/6His) was more soluble and displayed a shorter protease-resistant core than PrP^Sc106 or PrP106(225/6His). Placement of the tag at the N terminus did not increase the protease resistance of PrP106 at all. Therefore, inserting the affinity tag at the C terminus generates the derivative most similar to PrP^Sc106. Our mutagenesis studies also demonstrated that building a tag with six positively charged residues is optimal for acquisition of protease resistance. Furthermore, the protease resistance of PrP106(140/6His) was increased by introduction of the disease-associated mutation E200K located outside the tag. In contrast, the protease resistance of PrP106(140/6His) was not affected by mutations in α-helix C that disrupt the conversion of full-length PrP^C to PrP^Sc. Taken together, our mutagenesis studies suggest that PrP deletion mutants become increasingly protease resistant as their structures become progressively destabilized. If this explanation is correct, we predict that the internal and C-terminal His₆ tags are destabilizing elements for PrP106 and that the E200K mutation is a destabilizing element for PrP(140/6His). Elucidating the precise mechanism by which affinity-tagged PrP106 molecules become protease resistant will ultimately require direct determination of their molecular structures.

Although affinity-tagged PrP106 proteins share some properties with PrP^Sc, such as protease resistance, insolubility, and susceptibility to dendrimers, there are several differences that distinguish these PrP molecules. First, affinity-tagged PrP106 molecules can adopt their protease-resistant conformations spontaneously in N2a cells and uninfected transgenic mice. In contrast, the formation of either PrP^Sc106 or full-length PrP^Sc always requires the presence of preexisting infectious prions. Second, mutations that disrupt the formation of PrP^Sc, such as Δ214–220 and C213A, do not alter the protease-resistance of PrP106(140/6His) derivatives. Third, the protease-resistant fragments of affinity-tagged PrP106 molecules are smaller than the protease-resistant core PrP^Sc106. Finally, brains expressing protease-resistant PrP106(225/6His) do not appear to be infectious.

Although affinity-tagged PrP106 molecules are not perfect models of PrP^Sc, they may be potentially useful in several ways. First, using affinity-tagged PrP106 proteins as a starting point, it might be possible to generate spontaneous infectivity by introducing disease-associated mutations such as E200K. Spontaneously infectious PrP106 derivatives might be well suited for structural studies because their structures could be compared with those of noninfectious PrP106 proteins to identify structural elements that generate infectivity. Second, it may be possible to use PrP106(140/6His) to identify pathogenic PrP mutations. This molecule may be partially destabilized in such a way that it is capable of becoming more protease resistant when pathogenic PrP mutations, such as E200K, are introduced into its sequence. Thus, PrP106(140/6His) may be a sentinel molecule that can be used in a simple assay to identify potentially pathogenic mutations. Third, protease-resistant affinity-tagged PrP106 molecules might be convenient substitutes for PrP^Sc in assays identifying novel therapeutic compounds such as dendrimers.

In summary, affinity-tagged PrP106 molecules spontaneously adopt conformations partially resembling PrP^Sc. These molecules may prove to be useful research tools in areas of prion research such as structural analysis, identification of pathogenic mutations, and drug screening.