What could account for a species barrier...
Quasi homo-oligomer interactions?
Experimental selective bias?
The concept of a "species barrier" to transmission of spongiform encephalopathies is a rather loaded term, whose name alone presupposes the outcome of many experiments as yet unperformed. No scientific argument has been advanced as to why there should exist a priori an inter-species barrier and in fact there may exist species in which the recipient is actually much MORE susceptible than various host-to-host. (An example might be sheep-to-oryx; these species differ at only one residue whereas sheep can differ from other sheep breeds in several postions.)
It is important to distinguish natural barriers to exposure from barriers to transmission. A kangaroo, being a vegetarian in Australia, has little exposure to TSE from a grizzly bear in Montana. True, but this tells us nothing about the relative outcomes of intra-cerebral injection of infected material bear-to-bear versus bear-to-kangaroo. What is the molecular reasoning for thinking bear-to-kangaroo is more likely than bear-to-bear? More to the point, we could ask whether cow-to-sheep, cow-to-cat, cow-to-elk, cow-to-people, or cow-to-mink is more likely than cow-to- cow for particular strains, and so on.
It seems a good idea to componentize route of exposure and whole-animal, immune system, and cell biology uptake process contributions from the ultimate molecular mechanisms relevent at the end, when rogue prion has found its way to the CNS and is able to encounter and possibly recruit normal conformer. I don't discuss here issues such as should rogue sheep prion more or less resistant to bovine chymotryptic digestion than rogue cow prion, but ratherI address mainly the molecular components of the "species barrier." The analysis assumes the mainstream version of transmissible spongiform encephalopathies as put forward and documented by Prusiner and many others.
There are two distinct events in trans-species transmission of TSE: primary recruitment events where the trans-species prion is directly involved in conversion of normal host prion [trans recruitment] and subsequent secondary self-recruitment [cis recruitment] whereby newly formed rogue host prions -- which must soon vastly outnumber the original bad prions -- do the converting of normal host prions. Passaging in a constant and normal genetic background eventually stabilizes strain properties. The initial rogue prion is gone, though not necessarily without a trace (strain types), but is no longer contributing directly to the conversion process.
Familial CJD is a signficant variation [endogenous recruitment] on this because the mutant allele is produced in situ from within the cell and does not necessarily have to recruit the normal conformer allele because there is a steady supply of mutant allele gene product. As with trans recruitment, the prion primary sequence is a different covalent molecular species.
Note that the initial cross-species event differs from subsequent iterations in later (third-generation) animals. If BSE crosses to nvCJD only with great difficulty , then nvCJD (rogue conformer of normal human prion) might spread quite easily through, say, blood transfusions or maternally -- so any "species barrier" quickly loses its comfort value as the situation changes to cis recruitment. It is not know how the British BSE epidemic (or for that matter scrapie or kuru) got started (first generation). Sheep with scrapie, cows with BSE, and people with kuru and nvCJD generally have wild-type prions (modulo polymorphic backgrounds); they do not have a mutant allele sufficient to cause disease, as do victims of familial CJD [as concluded from high lod statistics].
I propose replacing "species barrier" with the neutral scientific phrase "species transmission coefficient" or STC. This compares the relative efficiency of transmission intra-species [cis] to inter-species [trans] under a given set of conditions. The STC is defined as log[S*(cis/trans)] where S is a normalization factor. The properties of the logarithm are such that the STC is positive in the case of a species barrier, zero if results are about the same, and negative if transmission is easier cross-species. Division by zero and log of zero are precluded by lower bounds on experimental precision.
Since there are thousands of species of mammals, there are millions of species transmission coefficients (STCs) to determine -- these form a square matrix with species-to-itself comprising the diagonals. In this context, the "species barrier" is a hypothesis that the STC matrix has significant entries only along the diagonal. Note that there is a separate STC matrix for each route of transmission, each strain, passaging, each allele, each significant polymorphic genetic background, and that deliverable threshold dosages, life spans, age of onset, and so on complicate matters. We have experimental data of varying quality on perhaps two dozen STCs; transmission efficiency is difficult to quantitate; we often have to settle for a scale of harder, easier, or about the same. It is no surprise that the STC matrix can be asymmetric, i.e., hamster-to-mouse is not the same as mouse-to-hamster, because the experiment is asymmetric.
Hope has been publicly expressed that scrapie-to-humans and cow (UK BSE)-to-human (val/val) have a high or even infinite STC (absolute barrier), and if not that at least cow-to-human (met/met) is orders of magnitude lower than cow -to-cow. Another common Ministerial assertion is that we have 250 years of data on scrapie-to-humans -- yet the life span up to 1920 was in the 40's and CJD was not even described until 1921 [Jacob's case was familial, D178N: Lancet 344: 130-1 (1994)]. Direct experiments of these two STCs are not possible using human subjects. Trans-genetic knock-in models, even chimeric ones, are no experimental panacea: they have never been established to quantitatively emulate actual cross-species data. However, a very interesting in vitro result in the May 13, 1997 PNAS suggests that an adequate method for rapid semi-quantitative measurements of STCs may be at hand.
Now "infectivity" in prion diseases lacks any deep resemblance to conventional bacterial or viral infectivity -- in particular, the immune system plays no role in TSE (unless microglia are counted here). However, some aspects are so similar that viral theories of TSE remain alive to this day -- the mechanisms are difficult to distinguish. The pig-to-human 1918 influenza epidemic shows that pig-to-human can be far worse than pig-to-pig for this viral disease -- it wasn't 20 million pigs that died. This is actually the usual state of affairs for immunological infectious disease: worse in the cross-species situation. There are additionally valid analogies in terms of barriers to exposure.
In fact, the worse cross-species viral situation is even predictable because the donor develops resistance to low levels of disease agent common in its environment, unlike the recipient. In prion diseases, the opposite is generally assumed without any underlying explanation, i.e., that donor-to-donor is worse. This could ultimately turn out to be a good assumption; my point is no one has brought forward any reason why it should be so. Until such time, there are no theoretical restrictions on the STC matrix.
As noted, experimental data is sparse, marginally quantitative, laborious to come by, and a difficult platform from which to extrapolate. However, a great deal is known about the molecular biology of prion diseases. I examine below various aspects of this with unpromising results. Some argument not considered here or in the scientific literature could provide an eventual basis for a "species barrier." The bottom line for now: there is really nothing on the table intrinsically favoring lower cross-species transmission coefficients from the perspectives of molecular biology:
1. Phylogenetic remoteness. This could be a valid consideration, but it is mooted by the immense phylogenetic distances needed. Prion protein has been extraordinarily conserved during mammalian evolution, especially the toxic core peptide so central to the rogue conformer, which has hardly changed from human to chicken -- and immunological cross-reactivity occurs even in mammal-to-salmon. Primary amino acid sequence can never be adequate to predict delicate 3-D quaternary interactions. Yes, mad nematodes might not be able to bind vertebrate prions, but who can say that a rogue pig conformer cannot recruit a normal chicken conformer? The largest gap validated experimentally to induce disease is primate-rodent. The fact of the matter is that rogue prion proteins are not all that fussy as hydrobic core fragments and the CJD amber mutation tyr 145 stop show.
At the other extreme, great ape prion proteins are so close that they fall within the range of mere neutral polymorphisms. Thus human, under maximal parsimony, exhibit only an E168Q change versus gorilla in a weakly selected loop region. Given the myriad influences of met/val at codon 129, an allele not selected by experimentalists for its properties but merely happening to have high frequency in caucasians, who is to say that E168Q could not have similar impacts? In fact, since chimp are met/met, it is quite likely that some human familial CJD would manifest themselves much more strongly in the chimp than in a human 129 val/val [negative species barrier].
In summary, phylogeny does not even guarantee sequence distance; even if it did, that variation might be in insignificant loop regions; even considerable sequence distance does not affect the hydrophobic core 104-122 that seems to play a key role in conversion and toxicity; and we even have no sound basis for thinking sequence distance is necessarily unfavorable.
2. Natural selection. It is conceivable that susceptibility to cross-species transmission has been deselected as a survival mechanism, i.e., that prion protein in, say, lions has evolved so not to get mad oryx disease from their ungulate prey base. However, there is no evidence that the disease is prevalent in wild African ungulates. While ungulates might have evolved to have high endogenous levels of sporadic CJD as a way of keeping lion populations down, lion populations are already limited by other factors. And why wouldn't lions simply evolve a resistant prion under these circumstances?
When an animal has 65,000 genes, mild selective pressure for a single gene can be dwarfed by cumulative demands on other genes. Age of onset would largely need to precede age of reproduction. Relatively few mammals are carnivores; while cows or mice might gnaw at skeletons for marrow or calcium under some conditions, only one per million of these is expected to have a TSE. This is a key difference between prion disease and ordinary infectious diseases: lack of an amplification or pooling mechanism under normal circumstances.
The better mystery in TSE is why natural selection hasn't weeded out susceptibility to prion disease from within. Prions from different species may well differ in intrinsic susceptibility to conversion but none seem to have an absolute resistance to it. The capacity to undergo conformational shift to the rogue form has not been implicated as a requirement of normal prion function. The answer would be: it is rare, generally has late (so irrelevant) onset, and so is not that much of a target for selection in the broader scheme of things. Indeed, observed changes in prion protein over evolutionary time largely appear adaptively neutral, clustered in loops and transitional domains not contributing particularly to structure/function in other well-studied proteins.
3. Prion protein hetero-oligomers. The evidence to date suggests that quaternary structure is important for prions. This can be taken for purposes of discussion as a dimer . In classical familial CJD, usually one protein molecule is normal, the other carries a point mutation. Just as in sickle cell hemoglobin, the hetero-dimer seems to undergo a cooperative change in 3-dimensional structure to a rogue form that has the undesirable propensity to form insoluble aggregates. It is reasonable to anticipate that two halves of the oligomer probably fit better [bond more strongly because they have evolved to fit if the dimer is functional] in the intra-species situation than the inter-species case, and to guess that on the whole the more primary sequence changes, the weaker the interaction. Could a species barrier [large positive STC] consistently arise in proportion to the quality of this fit?
Yes, a better fit might imply more conformational influence. However, we are not asking for stability of normal conformer here, but instability. Better docking pushes the equilibrium in the direction of stable dimer. We are looking for something structurally unstable that still docks (binds to normal conformer) but not with favorable free energy, so the pair is better off with both chains in the rogue conformation. Primary sequence differences that exist between species might result in protein species that dock unstably across a hetero-oligomer interface -- they can't decide in whose stable dimer they should be in, so end up in rogue conformer.
In the kuru-type situation, where the dimer chains have the same primary structure, it seems that recruitment must take place in a pseudo homo-dimer of rogue and normal. In other words, despite more beta sheet, rogue is not so different in tertiary structure that it cannot dock to normal as a dimer. It does not follow that wild-type prion, in rogue conformation, is consistently better at recruitment than mutant alleles or cross-species polymorphisms: the docking interface was designed for two normal conformers. The termination mutant Y145stop shows that distal and proximal portions of the prion protein appear irrelevant to the conversion process: the key part of the molecule for the disease process is a short peptide that unfortunately, the best conserved domain in the whole macro-molecule.
The other option is that the amyloid fiber uses a second docking surface distinct from the dimer interface. This would be like hemoglobin which has alpha-alpha, beta-beta, and alpha-beta tetramer interfaces. However, here there can be no capping off here of interface sites, the amyloid must grow indefinitely, like nested spoons, with cytoskeleton proteins as a better metaphor. In this scenario, the normal conformer could be destabilized by a abnormal conformer that is only slightly off. The weak dimer pair could either form a new rogue seed or more likely be recruited to a already nucleated growing rogue fiber. And the fiber-fiber lateral sites could allow larger aggregate forms -- bonds here can be weakly positive since there are many of them in a periodic fiber. Fibers are not crystals (17 frieze or 230 space groups don't describe symmetry) but helices (translation and pitch).
4. Experimental selection. The popular impression that species barriers exist may be affected by systematic experimental error. The vast majority of STC determinations are to mouse or hamster. Are the alleles tested from other species representative? No, they mainly get noticed because they have a marked propensity to form rogue conformer on their own and to recruit normal prion in cis. In other words, there is a bias towards a high numerator in the STC but neutrality towards the denominator. This results in an overall trend towards positive STCs in a data set limited by this screen. We do not even test alleles like 129 val trans-species because they doesn't cause TSE in cis. A animal with an allele ineffectual in cis but highly active in trans might be called a (carrier species). Note 85% of human CJD is sporadic: could some of this result from dietary intake of negative STC prion? In other words, what prevents an allele from being so slow-acting in its host that no pathology is evident during the life span, yet being highly efficient in at least some cross-species situations?
REFERENCES Brown,P.; Cervenakova,L.; Boellaard,J.W.; Stavrou,D.; Goldfarb,L.G.; Gajdusek,D.C. Identification of a PRNP gene mutation in Jakob's original Creutzfeldt-Jakob disease family Lancet 1994 Jul 9; 344(8915): 130-1 Creutzfeldt,H.G. On a particular focal disease of the central nervous system Uber eine eigenartige herdformige Erkrankung des Zentralnervensystems Zeitschrift fur die gesamte Neurologie und Psychiatrie 1920; 57: 1-18 Jakob,A.M. TO uber eigenartige Erkrankungen des Zentralnervensystems mit bemerkenswertem anatomischem Befund (Spastische Pseudosklerose - Encephalomyelopathie mit disseminierten Degenerationsherden) Deutsche Zeitschrift fur Nervenheilkunde 1921; 70: 132-46 PNAS Vol. 94, pp. 4931-4936, May 13 1997 Polymorphism barrier article: free fulltext
Species barriers in a model for specific prion protein dimerisation. Warwicker J Biochem Biophys Res Commun 232(2), 508-512 (1997) It has been proposed that the most highly conserved sequence segment within the prion protein (PrP) may be involved in dimer formation within both the normal (PrPC) and misfolded (PrPSc) forms. This hypothesis is now examined in the context of amino acids known to be involved in species barriers or in disease modifying polymorphisms, and the structure of a mouse PrP fragment. These locations can be plausibly explained on the basis of the specific dimer model, so that a potential role for a conserved dimerisation element in prion disease progression cannot be excluded.
A 60-kDa prion protein with properties of both the normal and scrapie forms Priola SA, Caughey B, Wehrly K, Chesebro B J Biol Chem 270(7), 3299-3305 (1995) .... Several hypotheses predict that oligomeric forms of either the normal or abnormal PrP may act as intermediates in the conversion process. We have now identified a 60-kDa PrP derived from hamster PrP expressed in murine neuroblastoma cells. Peptide mapping studies provided evidence that the 60-kDa PrP was composed solely of PrP and, based on its molecular mass, appeared to be a PrP dimer. The 60-kDa PrP was not dissociated under several harsh denaturing conditions, which indicated that it was covalently linked. It was similar to the disease-associated form of PrP in that it formed large aggregates. However, it resembled the normal form of PrP in that it was sensitive to proteinase K and had a short metabolic half-life....
Purified scrapie prions resist inactivation by UV irradiation. Bellinger-Kawahara C, Cleaver JE, Diener TO, Prusiner SB J Virol 61(1), 159-166 (1987) Species specificity and a model for the scrapie species barrier Kocisko DA, Priola SA, Raymond GJ, Chesebro B, Lansbury PT Jr, Caughey B Proc Natl Acad Sci U S A 92 (9): 3923-3927 (Apr 1995) Scrapie is a transmissible neurodegenerative disease that appears to result from an accumulation in the brain of an abnormal protease-resistant isoform of prion protein (PrP) called PrPsc. Conversion of the normal, protease-sensitive form of PrP (PrPc) to protease-resistant forms like PrPsc has been demonstrated in a cell-free reaction composed largely of hamster PrPc and PrPsc. We now report studies of the species specificity of this cell-free reaction using mouse, hamster, and chimeric PrP molecules. Combinations of hamster PrPc with hamster PrPsc and mouse PrPc with mouse PrPsc resulted in the conversion of PrPc to protease-resistant forms. Protease-resistant PrP species were also generated in the nonhomologous reaction of hamster PrPc with mouse PrPsc, but little conversion was observed in the reciprocal reaction.
Glycosylation of the PrPc precursors was not required for species specificity in the conversion reaction. The relative conversion efficiencies correlated with the relative transmissibilities of these strains of scrapie between mice and hamsters. Conversion experiments performed with chimeric mouse/hamster PrPc precursors indicated that differences between PrPc and PrPsc at residues 139, 155, and 170 affected the conversion efficiency and the size of the resultant protease-resistant PrP species. We conclude that there is species specificity in the cell-free interactions that lead to the conversion of PrPc to protease-resistant forms. This specificity may be the molecular basis for the barriers to interspecies transmission of scrapie and other transmissible spongiform encephalopathies in vivo.