Encouraged by the production of prion infectivity by polymerizing recMoPrP(89–230) into amyloid fibers 
, we undertook a study aimed at identifying conditions that would shorten incubation times for synthetic prions in Tg mice. We explored an array of variables, including the composition and concentration of denaturant, the number of seeding rounds, and the number of freeze-thaw cycles, none of which modified experimental outcomes. Twenty-five preparations of recMoPrP(89–230) polymerized into amyloid were inoculated into 204 Tg9949 mice. Eighty percent (or 164) of the Tg9949 mice were found to have sPrPSc
and neuropathology typical of experimental prion disease. Three of the amyloid preparations failed to produce measurable sPrPSc
and neuropathology while six other preparations showed incomplete transmissions (Table S3
). Three of the 22 infectious, recMoPrP amyloid preparations were studied in detail; these were designated MoSP2, MoPSP3 and MoSP4. Each of these synthetic prion isolates transmitted disease upon serial passage in Tg9949 mice (). In addition, MoSP2 and two other protease-sensitive synthetic prion isolates transmitted disease to Tg4053 mice overexpressing MoPrP ().
Our creation of these novel protease-sensitive prions challenges the accepted definition of what constitutes a prion. Mammalian prions have been most closely associated with PrP that resists protease digestion 
. Additionally, mammalian prions typically cause disease that shortens the lifespan of the animal. While the novel synthetic prions reported here do not have either of these characteristics, they share four traits common to all mammalian prions: (1) they possess an alternatively folded isoform of PrP (); (2) they cause neurologic dysfunction in animals (); (3) they cause profound neuropathologic changes ( and
); and (4) they are transmissible ( and
). We suggest that these four traits define mammalian prions.
Many prions observed in nature appear to be composed of mixtures of rPrPSc
and sPrPSc 
, though the relationship between the two is unclear. The creation of synthetic prions composed solely of sPrPSc
offers new insight into this relationship and the role of sPrPSc
in disease. Our results demonstrate that sPrPSc
is transmissible and causes neurodegeneration in the absence of rPrPSc
. Our findings also suggest that sPrPSc
does not arise as an off-pathway product during the replication of rPrPSc
. Examples of natural prion diseases that feature sPrPSc
predominantly are rarely reported 
. In the work reported here, it was necessary to use genetically modified lines of mice to make this unusual prion phenotype more readily accessible. Notably, inoculation of wt FVB mice with the amyloid fibers used in these studies did not result in prion disease (Table S7
It is intriguing that MoSP2 remained protease-sensitive even after repeated serial passage. The protease-sensitive prion fraction isolated from Syrian hamsters infected with 263K prions was shown to give rise to rPrPSc
in the protein misfolding cyclic amplification assay 
. Our findings indicate that infection with sPrPSc
does not necessarily lead to rPrPSc
Because some lines of Tg mice overexpressing wt PrP develop spontaneous neurological dysfunction 
, we observed 96 uninoculated, control Tg9949 mice and ic inoculated 78 control Tg9949 mice with BSA in PBS. Unexpectedly, most of these control Tg9949 mice developed late-onset, spontaneous neurological dysfunction. All the ill, control Tg9949 mice showed no neuropathological changes typical of prion disease. Additionally, no sPrPSc
was detected in the brains of these control Tg9949 mice. These studies established the validity and limitations of transmitting prions to Tg9949 mice.
In our initial report of synthetic prions, we described the onset of neurological dysfunction in Tg9949 mice between 380 and 660 days after inoculation 
. Three sets of Tg9949 mice were used as controls. In the first set, 10 of 12 healthy, uninoculated Tg mice were terminated at 574 days of age; the other two Tg9949 mice developed signs of neurological dysfunction at 564 and 576 days of age but had neither rPrPSc
nor neuropathology typical of prion disease. In the second set of control mice, eight Tg9949 mice were inoculated with Syrian hamster Sc237 prions and were healthy at 525 days of age when they were sacrificed. Third, seven Tg9949 mice were inoculated with PBS and remained healthy at 672 days of age when they were sacrificed. In light of the current work, the first and second control groups were terminated too early to observe neurological dysfunction and the third group appears to be an outlier. Our discovery that Tg9949 mice develop late-onset neurological dysfunction does not undermine the key finding of the earlier work 
, which demonstrated that prions could be generated de novo
from recombinant protein, but it does raise the possibility that the incubation period for the initial transmission may have been longer than reported. Incubation periods for some prion strains in Tg9949 mice cannot be determined when they approach or exceed the age of onset of spontaneous neurological dysfunction in these mice.
Despite the observation that uninoculated, control Tg9949 mice were prone to ataxia in old age, we found no evidence of prions in these mice by biochemical means, by histopathology, or by attempted serial transmission of their brain homogenates (). Neuropathological analysis of the brains of these mice excluded that neurologic dysfunction was caused by the spontaneous generation of prions. It is noteworthy that neurological deficits in Tg mice overexpressing PrP are not uncommon and are distinct from those caused by prion infection. Tg mice overexpressing wt MoPrP-B, Syrian hamster PrP, or ovine PrP develop disease featuring hindlimb paralysis, tremors, and ataxia, with mean ages of onset at ~550 days 
. Deletion of specific N-terminal segments of PrP results in fatal ataxia accompanied by degeneration of the cerebellum at 90–275 days of age 
. Deletions of helical regions near the C-terminus result in CNS illnesses similar to neuronal storage diseases 
. Like Tg9949 mice, none of these neurologically compromised mice spontaneously generated prions.
Evidence of prion disease was observed in 22 of 25 amyloid inoculations in Tg9949 mice, but was not observed from any of 7 control inoculations, including PBS, BSA, α-helical recPrP, β-oligomeric recPrP, and 3 uninfected Tg9949 brain homogenates. These results exclude the possibility that the observed neuropathology resulted from contamination of the inocula.
It is possible that a small titer of rPrPSc
that eluded detection is responsible for the disease observed in these studies. Given the extensive neurodegeneration observed in the brains of infected Tg9949 mice (Fig. S4
), this possibility seems unlikely. In fact, the vacuolation profile generated by inoculating the protease-resistant MoSP1 strain into Tg9949 mice was much less severe than that observed for MoSP2 prions, which lack protease-resistance 
. Furthermore, despite its tendency to accumulate, no rPrPSc
could be detected even upon serial passage (). Nonetheless, it is conceivable that some rPrPSc
may be detectable under conditions not yet explored, for example, using alternate proteases. This would not alter our conclusions, however, that such protease-sensitive prions would be overlooked using the standard conditions used to detect prions.
Whereas protease-sensitive prions composed of mutant PrPSc
(P101L) in Tg mice have been studied extensively 
, wt sPrPSc
has been less well investigated. While rPrPSc
is clearly transmissible, it is unknown what role, if any, rPrPSc
plays in the pathogenesis of prion disease. From the studies reported here as well as other investigations, sPrPSc
is clearly pathogenic.
The pathogenicity of sPrPSc
calls into question the adequacy of some terms used to describe different isoforms of PrP, such as PrPres
and PrPsen 
is often equated with PrPSc
, and PrPsen
. From the work presented here, we contend that PrPSc
can be both protease-resistant and protease-sensitive, rendering terms that describe only the protein's response to limited protease digestion as ambiguous. Therefore, the use of terms describing both infectivity and resistance to protease digestion (i.e., sPrPSc
, and PrPC
) is necessary in order to avoid confusion.
While inoculation ic of recMoPrP(89–230) amyloid did not shorten the lives of Tg9949 mice (Table S2
), the amyloid preparations provoked severe neurodegeneration (). Serial transmission of protease-sensitive prions MoSP2, MoSP3, and MoSP4 in Tg9949 mice did not alter the incubation periods (Table S5
), suggesting that these prion isolates encipher long incubation times.
Because the formation of rPrPSc
has been used as an operational assay for the identification of prions, protease resistance has been often viewed as an intrinsic and obligatory feature of prions 
. The results reported here extend our more recent findings that challenge the notion that protease resistance is an obligatory feature of PrPSc
that is required for the transmission of prions 
The production of synthetic prions, which are sensitive to proteolysis but cause transmissible disease, is an important step toward understanding the role of protease-sensitive forms of PrPSc
in the pathogenesis of prion disease. Recent reports suggest that prions with low levels of rPrPSc
occur naturally in sheep 
and humans 
. Our results show the importance of using alternate methods for detecting PrPSc
, rather than employing only the presence of PK-resistant PrP. Exclusive reliance on the detection of rPrPSc
as a surrogate marker for prion infectivity may overlook the contribution of sPrPSc
to prion infectivity and the pathogenesis of prion disease