It was not too surprising that swa-resistant 22L prions could be selected from a swa-sensitive, cell-adapted 22L population 
, because “natural” brain-derived 22L prions are swa resistant to begin with. It was however unpredictable that RML prions, which are swa sensitive, both when propagated in brain or in PK1 cells, would give rise to swa-resistant and even swa-dependent populations when passaged in the presence of swa.
What is the biochemical basis for swa sensitivity, resistance and dependence? Swa, by inhibiting α-mannosidase II, causes misglycosylation of N-glycosylated proteins, including PrPC
). General misglycosylation could alter the cellular mechanisms involved in PrPSc
synthesis and clearance; however it is more likely that most of the effect is related to the misglycosylation of PrPSc
because different variants of RML prions exhibit a vast difference in their response to swa in the same cell line. How could misglycosylation of PrPSc
lead to these effects? X-ray crystallographic studies on IgG1-Fc have demonstrated that the nature of the N-linked glycans may affect the conformation of a glycoprotein and thereby change its susceptibility to degradation 
. Indeed, we found that swa-sensitive and swa-dependent RML PrPres
have different degradation rates. Moreover, misglycosylation may affect the interaction of PrPSc
with other cellular proteins, such as lectin chaperones, and thereby modulate the rate of synthesis and/or clearance. If, as we propose, prions form quasi-species populations, a conformer present at low levels that is able to overcome a deleterious effect due to swa-mediated misglycosylation could be selected and give rise to swa-resistant or even -dependent PrPres
A further point of interest is that swa and kifu resistance of RML prions do not necessarily go hand in hand: Although both drugs lead to misglycosylation, the resulting glycans are of different nature, namely hybrid-type glycans in the case of swa 
and oligomannose-type glycans in the case of kifu 
. This indicates that different types of misglycosylation might modulate the interactions of PrPC
with proteins involved in PrPres
synthesis or degradation in distinct ways. Castanospermine, an inhibitor of α-glucosidase 1, which removes glucose residues from the initial core and generates a high-mannose glycan, shows an inhibition pattern different from both swa and kifu 
Why was it possible to reproducibly generate swa-resistant prions in AMO10 and AMO18 cells (and at least once in 2E4 cells), but not in PK1 cells, although the latter were able to propagate swa-resistant prions once they were generated in other cell lines and even mediate their conversion to swa dependence? Within the framework of the quasi-species hypothesis, it is possible that the diversity of the prion population in PK1 cells is distinct from, or more restricted than that in AMO10 cells, as suggested by the frequency analysis for swa-resistant prions from RML-infected PK1 and AMO10 cells, so that there are no swa-resistant variants to be selected. This in turn begs the question as to why the diversity is different. We suggest that prions undergo mutations that are based on small, thermally induced conformational changes of short PrPres
, which are stabilized if and when appropriate (“competent”) PrPC
conformers accrete to it (). Thus, as we proposed earlier to explain cell tropism 
, the repertoire of PrPres
variants could depend on the repertoire of PrPC
conformers, which in turn may be determined by the N-linked glycans, of which there is a great variety 
or by association with some cell-derived molecule, for instance a small RNA. The affinity of PrPC
for RNA (reviewed in 
) and the requirement for RNA in PMCA-mediated prion propagation in vitro 
has been documented. Thus, if prion propagation at low concentrations of both swa-resistant seed and competent monomer follows 2nd
order kinetics, the rate of swa-resistant prion formation in PK1 cells may be lower than the replication rate of the cells, thereby preventing emergence of swa-resistant prions. On the other hand, elevated concentrations of swa-resistant seed in the same cells could promote rates of swa-resistant prion formation high enough to allow persistence in dividing cell populations and evolution of prions with increased swa resistance and even “dependence”.
Hypothetical mechanisms for prion mutation and cell tropism.
Several prion strain variants we have described earlier, which resulted from the transfer of brain-derived 22L prions to different cultured cell lines, reverted to the original strain when returned to brain 
, suggesting that the activation energy barriers between the various conformations were low 
. In another instance, brain-derived 139A prions, when passaged through PK1 cells and returned to brain, became indistinguishable from the 79A/RML strain, suggesting that the activation energy barrier between the two strains was high in brain but low in PK1 cells, and that the 79A/RML conformers were preferentially replicated in the cells 
. In the experiments described in this paper, at least one distinct swa-resistant and three swa-dependent RML variants were characterized, derived from AMO10 and from 2E4 cells, which, after passaging in PK1 cells, retained their properties even after propagation in the absence of swa. Upon propagation in brain, the swa-resistant (and kifu-resistant) prions from AMO10 cells reproducibly became swa sensitive, however were still semi-resistant to kifu, showing that a strain distinct from RML had emerged. Whether or not these prions would further change or revert to RML upon continued propagation in mouse brain was not determined. The 2E4-derived prions, when propagated in mouse brain, also became swa sensitive, however the resistance to kifu varied in the three brains tested, ranging from semi-resistance to full susceptibility; this may reflect very slow reversion to the original RML strain, which proceeded at slightly different rates in these brains.
In summary, our experiments have shown that RML prions, when subjected to selective pressure in cultured cells, can develop a variety of novel properties, which are stable in cell culture but are further modified when propagated in brain. The repertoire of possible conformations associated with PrP conformers with the same primary sequence appears to be unexpectedly vast.