THE PRION DISEASES

The prion diseases: a molecular and genetic perspective

1. Aims of the study
2. What are the animal and human prion diseases?
3. Phylogeny
4. Why do we study them?

 

Section II: Epidemiology and Clinical features of the different diseases Jump

5. Creutzfeldt-Jacob disease (sporadic)
6. Clinical features of CJD
7. Iatrogenic CJD
8. New Variant CJD
9. Different (atypical) forms of CJD
10. Kuru

 

Section III: Inherited prion diseases Jump

11. Gerstmann-Sträussler-Scheinker disease
12. Fatal Familial Insomnia
13. Familial CJD

 

Section IV: Introduction to the ‘prion’ Jump

14. The discovery of the prion protein
15. What is the evidence that prions are infectious?

Section V: The prion concept Jump

16. The prion structure
17. The same amino acid sequence but two different structures
18. The three dimensional structure of PrP
19. Thermodynamics
20. The number of a -helices in PrP^c
21. The prion gene and inherited prion diseases
22. The challenge for familial prion diseases

 

Section VI: The function of PrP^c Jump

23. Studies on transgenic mice to deduce the function of PrP^c
24. The overexpression of PrP^c in mice
25. Demonstration of the requirement of PrP^c for PrP^sc infectivity and the proposed mechanisms for neuronal damage
26. Alzheimer's Twist?
27. Other studies directed at determining the function of PrP^c

 

Section VII: The Replication of PrP^Sc Jump

28. Evidence for the conversion of PrP^c into PrP^sc
29. Evidence for Protein X
30. Models of Propagation

 

Section VIII: The Species Barrier Jump

31. What is the species barrier?
32. The 'different strains' idea about incubation and how we should now look at 'the species barrier' 68
33. The link between BSE and CJD
34. Other Possible PrP^sc interactions
35. B Lymphocytes
36. Copper
37. Ancillary factors

 

Section IX: Nature of the agent: Prion, Virus or even Bacteria? Jump

38. Prion or Virus?
39. The coup de grace?
40. Is there a radical bacterial link?

 

Section X: Diagnosis Yeast and Drugs – Recent Developmental Work Jump

41. New Diagnostic Techniques
42. Yeast Link
43. Development of chemotherapeutic agents

 

Section XI: End discussion Jump

44. A summary
45. Into the future, and the unknown incubation factor

This study is aimed at looking at the prion diseases from a molecular point of view, taking into account the genetic as well as the transmissible and sporadic aspect of the diseases. It attempts to first fully identify these novel diseases to give an exact idea of what the research is about and why it is taking place. The molecular side starts by looking at the molecular and genetic evidence from the beginning of the discovery of the 'prion' hypothesis. Then, by following a fascinating path of logical deductions and discovery with exciting consequences that were revolutionary for molecular biology and resulted in uncertainties and fierce scientific debates, the properties of the prion diseases are investigated. The report intends to not only critically analyse the 'prion' hypothesis, but to compare it to the viral and bacterial hypotheses as well, ending in identifying the progress that needs to be made in the future, and above all, the major uncertainty of variant CJD.

The animal and human prion diseases are characterised by their ability to cause an invariably fatal degeneration of the nervous system. When the brain tissue is viewed under a microscope, the sections have a sponge-like appearance, shown in Figure 1. The classic neuropathological features of human prion disease include spongiform degeneration, gliosis, and neuronal loss in the absence of an inflammatory response. Essentially, the brain cells fall apart, leading to symptoms such as a progressive sub-acute or chronic decline in cognitive or motor function.

Figure 1 (Adapted from Owen, 1998).

Figure 1 (A) shows normal brain cells (x400) (B) Shows damaged brain cells (x1300)

Due to their characteristic appearance, the diseases are known as the transmissible spongiform encephalopathies (TSEs) as they can be transmitted both experimentally and naturally. The TSEs consist of several closely related prion diseases, of which four similar diseases affect humans. These are shown in Table 1. These diseases can be sporadic (no known antecedent events), acquired (by contamination with infectious agent) or familial (genetically inherited).

Sporadic diseases include CJD, which occurs worldwide with incidence of about 1 in a million people. An example of an acquired disease is kuru, meaning shivering or trembling, which was discovered in Eastern Highlands of Papa New Guinea amongst members of the Foré tribe. It is also known as the laughing disease due to the facial grimaces it causes and is thought to be acquired through the ritual cannibalism of dead relatives who had caught the disease. Kuru-like symptoms include the loss of co-ordination, shaking, and often dementia. There is also iatrogenic CJD and new variant CJD. New variant CJD is commonly thought to result from eating food contaminated with the BSE agent.

Familial diseases are genetically inherited and include familial CJD, which accounts for about 15% of all CJD cases. Also inherited are GSS, with a global incidence of 1 in 10 million, and FFI, where insomnia is a major symptom, in addition to other symptoms including dementia. GSS diseases comprise dementing, ataxic and atypical GSS with neurofibrillary tangles. Included is the atypical prion disease, which does not easily fit the diagnostic criteria for prion disease.

There are also several types of animal prion diseases affecting animals like those shown in Table 2, including animals uncommon to the UK such as Nyala and Kudu, (related to the antelope and found in Mozambique and South Africa respectively). This is explained by the fact that many are zoo animals possibly exposed to the agent in their feed.

Scrapie is a rare endemic brain disease of sheep and goats, so called because an animal suffering from the condition scrapes itself up against posts to relieve an itch, damaging its coat and itself. Transmissible mink encephalopathy (TME) is a disease of farmed mink, probably caused by feeding on scrapie-contaminated meat. Chronic wasting disease (CWD) is a scrapie-like disease of obscure origin found in wild and captive mule deer and Rocky Mountain elk.

Bovine spongiform encephalopathy (BSE) is an epidemic disease in dairy cows, observed mainly in the UK, caused by cows fed with foodstuffs containing rendered remains of scrapie affected sheep and BSE affected cattle. Feline spongiform encephalopathy (FSE) is a disease seen in domestic cats, and in a few other members of the cat family shown above. The spongiform encephalopathies (feline and non feline) are assumed to be caused by feeding BSE-infected material to the animals in question, often in zoos. Spongiform encephalopathies appear in other species listed in Table 2 (Ridley and Baker, 1998). Mice, hamsters and monkeys have proved experimentally to be the most useful models of prion diseases.

Figure 2. Phylogeny of animals relevant to TSE

Figure 2 shows the phylogeny of animals relevant to TSE (adapted from the website www.mad-cow.org). The tree is based on a review of about 300 recent scientific publications. Data from short interspersed elements (SINEs) and satellite DNA were used to refine and resolve tree topology deduced from classical morphological analysis [reflected inGenBank taxonomy] and sequence alignment of many genes by standard methods.

The coloured bars represent independent genetic markers that definitively partition artiodactyls. Everything 'downstream' of a bar forms a monophyletic clade (uniquely sharing a last common ancestor) which is a crucial property of repeat elements. Grey boxes show unresolved nodes where conflicting trees are still prevalent in the literature. For example, all three possible topological trees for Tylopoda (camels), Suidae (pigs), and Ruminantia (cattle) appeared in 1997 in single-gene studies with various degrees of support. The species shown in red are suggested strategic candidates for prion gene sequencing. Species such as white-tailed deer and cheetah have experienced TSE; reindeer are at risk for scrapie in Norway; squirrels have been associated with CJD in Kentucky and American bison are potentially at risk to CWD. The other species are selected for balance (baleen minke whale), for their maximally informative phylogenetic position (chevrotain, nilgai, cape buffalo, rhino, hippo), or to eliminate long branches (peccary, okapi, nutria, guinea pig). These sequences, if completed, would vastly improve our ability to reconstruct ancient phylogenetic nodes, our understanding of what constitutes the wild type allele in sheep and cattle, and our understanding of fixed mutational events during the evolution of this protein. In turn, this helps understand which protein domains are highly conserved and which are random-connecting loops of little normal function. For example, the complete absence of change in the pre-repeat region over 100 million years conflicts with a recent study asserting no fixed structural role for this domain. It then might be asked what the significance of this is to prions? Suppose that a sequence of a whale prion had just been completed. As whales shared a common ancestor with sheep and cattle that pigs and camels did not, its sequence provides information about the structure and function of the prion gene at the time of appearance of the ancestral ruminant. The species barrier to BSE may well be lower to whale that to pig than to camel. The dolphin node helps us understand when and where certain changes took place in the development of the current cattle or sheep prion and whether certain polymorphisms are new or have persisted for tens of millions of years or have occurred in multiple lineages. Prion sequences can be clamped to the above tree, using both its topology and dated divergences. Instead of trying to deduce the tree and dates from prion sequences, they used a consensus tree derived from many genes and fossils, and asked where, how fast, and when mutations in the prion gene become accepted as dominant alleles. This is wholly interesting and relevant as inherited prion diseases are usually due to a dominant allele like with Huntingdon's chorea. It also raises the question of the highly conserved nature of the gene, and perhaps its relation to its structure, which is discussed later.

The prion diseases can affect both humans and animals, and it is invariably fatal due to the neurological damage it causes. Presently, there are no vaccines or medicines that will prevent the disease and, apart from nursing care, there are no treatments to halt or slow its inevitable course. The course of the disease is alarmingly rapid compared with other neurodegenerative disorders such as Alzheimer disease, Parkinson's disease and motor neurone disease. However, they are rare neurodegenerative disorders affecting only 1 in a million people, which is only about 50 or so new cases in the United Kingdom each year. Despite this, remarkable attention has recently been focused on these diseases. This is due to their unique biology (described later) and also because of the fears that an epidemic of a newly recognised bovine prion disease (BSE (recognised in 1986)) among UK cattle could pose a threat to public health through dietary exposure to infected tissues (Collinge & Palmer, 1997, Ridley & Baker, 1998). The exposure of the infected tissues may not have been considered as a problem due to modern cooking techniques and chemical treatments such as UV light and radiation. Unfortunately, this agent happens to be particularly resistant to both, resulting with little or no loss of infectivity. It could then be said that a significant part of the UK population has been exposed to BSE infected food, with no known consequences. The effects could be minor with only a handful of people succumbing to the disease or the effects could be disastrous with many thousands of people falling foul of the prion disease. Interest has therefore been directed at the infectivity of the prion agent between individuals and across the species barrier in relation to the incubation time, and possible treatments, cures or vaccinations.

Creutzfeldt-Jacob disease (sporadic)

In a number of countries, there have been extensive epidemiological studies on the numbers of CJD cases. All of these countries obtained broadly the same results, approximating about 0.5-1 case per million. There is no significant case clustering other than in familial clusters. Cases are distributed apparently at random with a frequency related only to local population density. There is no evidence for case to case spread (other than with respect to iatrogenic CJD discussed below). There is now evidence of an association with local scrapie prevalence, since CJD is as common in scrapie free areas such as Australia and New Zealand, as it is in the UK, where scrapie is endemic. Despite this, the BSE epidemic has led to a renewed interest in this area and since 1990 an attempt has been made to monitor all suspected cases of CJD in the UK by a National CJD Surveillance Unit in Edinburgh. Parallel studies have also taken place in a number of other European countries, some of which have started to report BSE cases in their cattle population. Despite the attention focussed on links between BSE and CJD, the overall incidence figures do not appear to have risen more in the UK than in countries, which do not have BSE. Although an overall increase has been observed, it is likely to be due to increased accurate identification of cases due to increased attention, which is now focussed on these diseases. This is especially relevant to the increased incidence noted in the elderly, where other dementing illnesses are common and could be confused with CJD. Occasional weak statistical associations with particular foodstuffs have been reported, although these have tended to disappear in subsequent years of the study and no consistent new risk factors have emerged. Despite this, concern has been expressed about a number of occurrences of CJD in UK dairy farmers, (Collinge & Palmer, 1997).

According to Collinge & Palmer (1997) the core clinical syndrome of CJD is of "rapidly progressive multifocal dementia usually with myoclonus. The onset is usually in the 45-75 year age group with peak onset between 60-65. The clinical progression is typically over weeks progressing to akinetic mutism and death, often within two to three months. Around 70% of cases die in under 6 months. Prodromal features include extrapyramidal signs, cerebellar ataxia, pyramidal signs and cortical blindness. About 10% of the cases present initially with cerebellar ataxia. Routine haematological and biochemical investigations are normal although occasional cases have been noted to have raised serum transaminases or alkaline phosphatase. There are no immunological markers and acute phase proteins are not elevated." They go on to mention that "Prospective epidemiological studies have demonstrated that cases with a progressive dementia and two or more of myoclonus, cortical blindness, pyramidal, cerebellar or extrapyramidal signs, or akinetic mutism with a typical EEG nearly always turn out to be confirmed as histologically definite CJD if neuropathological examination is performed". This depends on neuropathological confirmation being "by demonstration of spongiform change, neuronal loss, and astrocytosis". PrP amyloid plaques are usually not seen with CJD, although the presence of protease resistant PrP can be demonstrated by immunoblotting of brain homogenates (Collinge & Palmer, 1997).

Iatrogenic CJD is usually referred to as CJD despite cases arising from peripheral inoculation (not intracerebral) usually having a prominent cerebellar syndrome more reminiscent of kuru (discussed later). Transmission may occur by any one of a number of routes involving accidental inoculation with human prions as a result of medical procedures; the routes involving the use of inadequately sterilised neurosurgical instruments due to the high heat resistance of the prion, dura mater and corneal grafting, as well as the use of gonadotrophin derived from the pituitary gland from cadavers. Iatrogenic CJD actively demonstrates the transmissible properties of the prion (described later). The first case was with a 55 year old woman who received a corneal transplant from a donor later found to have died from CJD. In such cases of neurosurgical transmission, the equipment was found to have been sterilised with 70% ethanol and formaldehyde vapour. This is normally enough for conventional agents, but is ineffective against prions. In view of this, the new sterilisation protocol involves immersion in 1M NaOH for one hour to inactivate the prions. (Collinge & Palmer, 1997)

Teenagers rarely get sporadic CJD, and when in 1995 two cases were reported in the UK, both were unusual in having kuru-type plaques, which is only found in about 5% of CJD cases. Shortly after, a third young case emerged and a link was suggested with BSE. In the following months further cases emerged showing similar histological patterns which were unique from those observed with other prion diseases. These cases were named new variant CJD, despite being atypical in their clinical presentation with most cases not meeting the accepted clinical diagnostic criteria for probable CJD. These cases were extensively compared with other prion diseases and the lack of correlation seemed to confirm the arrival of a new strain of prion disease in the UK due to the incredibly small statistical probability of these similar cases occurring by chance. The symptoms of the new variant CJD start with behavioural and psychological disturbances, which include anxiety, depression, persistent pain in the limbs/face, withdrawal and behavioural changes which progress. Overt neurological features later become apparent often after most patients have gone to see a psychiatrist due to the preliminary symptoms. Several weeks later, progressive cerebellar syndrome develops with gait and limb ataxia. Later dementia develops progressing to akinetic mutism; myoclonus was also seen in most patients and sometimes was preceded by choreoathetosis. The neuropathological appearances were very consistent with widespread spongiform change, gliosis and neuronal loss, most severe in the basal ganglia and thalamus, the most remarkable feature being the abundant PrP amyloid plaques in the cerebral and cerebellar cortex. The plaques consisted of kuru-like 'florid' plaques (surrounded by spongiform vacuoles) and multicentric plaque types. The 'florid' plaques were very unusual and seen previously only in scrapie, but were very consistent. There was also an extensive amount of PrP in the cerebral cortex, cerebellar cortex, and in the molecular layer of the cerebellum. The unusual young age of these patients and the link with BSE may point to a risk involving dietary exposure. The risk involved however is as yet undetermined, and will only be determined as time goes by and more or fewer cases are identified. This is due to the potentially long incubation time related to the species barrier, (discussed later) (Collinge & Palmer, 1997).

These forms of the disease are well recognised, and around 10% of cases have a longer clinical course of over two years. These cases may be represented by people with heterozygous PrP polymorphisms. 10% of cases involve cerebellar ataxia rather than cognitive impairment, and are called ataxic CJD. Other types of CJD can include Heidenhain's variant of CJD, which refers to cases where cortical blindness predominates with severe involvement of the occipital lobes. There is also the panencephalopathic type of CJD where there is extensive degeneration of the grey matter; this was mainly reported in Japan. Amyotrophic variants of CJD have been described with prominent early muscle wasting and are usually seen in late disease when other features are established. However most cases of dementia with amyotrophy are not experimentally transmissible, and their relationship with CJD is not clear. It is possible that these are variants of motor neurone disease with associated dementia. (Collinge & Palmer, 1997)

The study of kuru is immensely important for the understanding of these diseases, providing the most extensive clinical experience of acquired prion disease in humans, the epidemiology of kuru clearly indicates the infectious, but non-contagious, properties of prions and provides extensive evidence that vertical transmission from pregnancies does not occur.

Kuru reached epidemic levels in the Foré people in the Okapa district of the Eastern Highlands of Papua New Guinea other areas affected included Keiagana, Kanite, Yate, Usurfa and Gima, which are in the close vicinity. The highlanders were usually of Aboriginal Australian origin. Papa New Guinea (South East Asia) is thought of by many as the "new cosmos" due to the immense diversity within its population of 4.2 million, there being around 850 different languages. Work done by Gajdusek in an epidemiological survey that started in 1957, uncovered the link between the ritualistic cannibalism and the disease occurrence. This included the differential age and sex incidence which was due to the women and children preparing the food, and becoming exposed to the agent, as well as eating the less desirable brains and other internal organs, even the bones were pounded up to be eaten with greens later. The whole of the deceased sufferer was recycled by way of endo-cannibalism. This is a type of cannibalism. It could be termed 'transumption' as it is seen as disposing of the dead relative, recycling the remains back into the community, and most importantly as a spiritual and virtuous gift to the family members. The ratio of occurrence was about 5 females to 1 male, with about 1% of the population affected. In the years after these practices were stopped, the only cases were from the older and not the younger people (who won't have been cannibals) indicating no evidence for vertical transmission, and due to the mortality rate much later, an incubation time of as much as 40 years or more.

The clinical duration had a mean of 12 months with a range of 3 months to three years, the course tending to be shorter in children. Cerebellar ataxia is the central feature with dementia usually absent (contrasting to CJD). Gross dementia can occur but is usually rare. There is a prodrome stage followed by three clinical stages which are:

1. Prodromal stage Kuru typically begins with prodromal symptoms consisting of headache, aching of limbs, and also joint pains which can last for several months.
2. Ambulatory stage At this stage there may be no objective signs of disease, but it can be self diagnosed as patients experience an onset of unsteadiness in standing or walking, or of dysarthria or diplopia. Gait ataxia worsens and patients develop a broad-based gait, truncal instability, and titubation. There is a postural tremor present which is coarse and accentuated by movement; patients characteristically hold their hands together in the midline to suppress this. Standing with feet together reveals clawing of toes to maintain posture. This clawing response is regarded as pathogonomic of kuru. Therefore signs at this stage are of midline truncal ataxia, presumably due to paleocerebellar disease, which is reflected by the only clear macroscopic brain abnormality seen in kuru, which is atrophy of the cerebellar vermis. As the gait ataxia and astasia worsens the patients require a stick, as ataxia progresses the patient passes to the next stage of the disease (Sedentary). The mean clinical duration of the first stage is around 8 months and correlates closely with total duration.
3. Sedentary stage At this stage patients cannot walk, but are able to sit. Walking attempts lead to a high steppage, wide based gait with reeling instability, and flinging arm movements in an attempt to maintain posture. Hyper-reflexia is seen, although usually plantar responses remain flexor and abdominal reflexes intact. Clonus is characteristically short lived. Athetoid and choreiform movements and parkinsonian tremors may occur. There is reduced muscle power but no paralysis, and obesity may be common at this stage but possibly linked to bulimia. Characteristically there is emotional lability and bizarre uncontrollable laughter, which has lead to the disease being referred to as 'laughing death'; one sufferer may start and this spreads throughout the group, although this is only a small aspect of the disease. Myoclonic jerking is rarely seen (unlike in CJD) in this stage which lasts about two to three months.
4. Tertiary stage Hypotonia and hyporeflexia develop and the terminal stage is marked by flaccid muscle weakness. Plantar responses remain flexor and abdominal reflexes intact. The patient is unable to sit unsupported, dysphagia occurs and the patients become incontinent. Inanition and emaciation develop and transient conjugate eye signs and dementia may occur with primitive reflexes developing in some cases. Brainstem involvement and both bulbar and pseudobulbar signs occur. Respiratory failure and bronchopneumonia eventually lead to death. This final stage lasts one or two months. Often death may occur earlier due to the nature of the illness, for example the sufferers may role into a fire, and burn themselves when in the later stages of the disease. The sufferer is always very well cared for by their family.

In the last five years, there have only been 18 deaths in the south Foré, with only one male death in 1998 and none yet in 1999, indicating the 'beginning of the end' of the kuru epidemic (Collinge & Palmer, 1997, Prusiner, 1982, Alpers, 1999).

This was first described by Gerstmann in the 1920s-30s as a 'peculiar heredo-familial disease of the central nervous system'. The classical presentation is with a chronic cerebeller ataxia accompanied by pyramidal features, with dementia occurring later in a much more prolonged clinical course than that seen in CJD. The mean duration being around five years, with onset usually in the third or fourth decades of life. The histological hallmark is the presence of the multicentric amyloid plaques. Spongiform change, neuronal loss, astrocytosis and white matter loss are also usually present. This is an autosomal dominant disorder, which can now be classified within the spectrum of inherited prion disease.

This disorder was first described in 1986 and is an autosomally dominant inherited sleep disorder. This presents itself as a progressive untreatable insomnia and autonomic dysfunction. Thalamic atrophy is the main histological finding, with the anterior-ventral and medial dorsal thalamic nuclei being consistently affected, while the cortex shows moderate asterocytosis. (Collinge & Palmer, 1997)

This inherited disease occurs later on in life at a similar time to GSS with similar symptoms except for a more rapidly advancing dementia and an abnormal EEG.

(Prusiner et al., 1992)

The discovery of the prion protein

In 1982 S B Prusiner published an article in Science journal claiming to identify the causative agent of scrapie to be a 'novel proteinaceous infectious particle'. He also proposed a then new term 'prion' (pronounced pree on), to describe this agent, representative of its proteinaceous and infectious nature. The idea of a protein only infectious agent was first proposed by Griffiths in 1967. However, it was not until the co-purification of the prion protein with hamster scrapie infectivity that Prusiner was able to distinguish it from a virus. It was the unanswered question of the chemical structure of the scrapie agent that Prusiner endeavoured to solve at the time, as the agent had reportedly been transmitted to sheep when 18,000 were vaccinated against louping ill virus. The suspensions were treated with formalin to prevent contamination, to which the scrapie-agent showed resistance, and 1500 of the sheep developed scrapie two years later. Prusiner's experiments uncovered six separate and distinct lines of evidence that the scrapie agent contains a protein that is required for infectivity. These lines were:

1. The agent was inactivated as a result of digestion with proteinase K and trypsin,
2. Inactivation was by chemical modification with diethyl pyrocarbonate,
3. Inactivation by SDS (Sodium dodecyl sulphate),
4. Inactivation by chaotropic salts such as guanidinium thiocyanate,
5. Inactivation by phenol,
6. Inactivation by urea.

The agent was also resistant to procedures which attach nucleic acids, such as low pH, ribonucleases, deoxyribonucleases, UV at 254nm, divalent cation hydrolysis, psoralen photoreaction by AMT, HEP, HMT, MMT, TMP, and chemical modification by hydroxylamine.

This compelling evidence can be shown using Tables 3, 4 and 5.

Table. 3

Table. 4

Table. 5

The molecular size of the agent was thought to be between 64,000 to 150,000, which didn't take into account the possible high number of copies of nucleic acid, or efficient nucleic acid repair system which would have an upward trend on the size. However in experiments involving gel filtration, rate-zonal sucrose gradients and gel electrophoresis the scrapie agent was found to be possibly globular and of a molecular weight of less than 50,000 Daltons.

The novel properties of the scrapie agent can be summarised in the table below.

Interestingly, the prion is resistant to psoralen photoreaction; five different psoralens were tested and the scrapie agent was found to be insensitive. The psoralens were used as they can penetrate the protein and lipid coats of conventional viruses, react with their genomes upon photoactivation and cause impressive losses of titre, (Masiarz et al., 1982).

So the molecular properties of the scrapie agent were deduced to be different from those of viruses, viroids, and plasmids. As the evidence gained could not exclude the possibility of a small nucleic acid present, Prusiner came up with two models for the scrapie agent, it was either a (i) a small nucleic acid surrounded by a tightly packed protein coat or (ii) a protein devoid of nucleic acid, which means an infectious protein. Later he came up with the idea that the protein could be associated with a small polynucleotide, (Prusiner, 1991). The postulated mechanisms for the replication of prion proteins ranged from those used by viruses to the synthesis of polypeptides in the absence of nucleic acid templates to posttranslational modifications of cellular proteins, (Prusiner, 1992). Since then subsequent discoveries have narrowed these hypotheses for both the prion structure and the mechanisms of replication for which recent models have been suggested and are discussed later.

Prusiner then also suggested possible mechanisms of the prion replication as shown in Table 7.

Table 7. Possible mechanisms of prion replication. (Prusiner, 1982)

Prions contain undetected nucleic acids

Code for prion protein(s)

Activate transcription of host genes coding for prion protein

Prions are derived from nucleic acids

Activate transcription of host genes coding for prion protein

Code for their own replication be either:

Reverse translation

Protein-directed protein synthesis

Prusiner also mentioned the possibility of prion genes (not discovered until later) and how they must be highly regulated, not readily activated, but yet pose an important function due to their conservation preserved by evolutionary pressure amongst most, if not all mammal species. This was the first major report suggesting the possible lack of nucleic acid and the unusual proteinaceous nature, which was backed by strong evidence, and from this point the molecular investigation into prion biology really began to attract interest, (Prusiner, 1982).

Proteins are not normally termed infectious agents and unlike viruses, bacteria and other infectious organisms, the prions have no obvious nucleic acid. It has been proven that these are present with the disease symptoms, but do these actually cause the disease from step one?

The prion diseases themselves are known to be transmissible, from the early studies of Cuillé and Chelle, when they set out to determine the infectivity of scrapie in 1936, to the likely cannibalistic transmission of kuru. Parallels were drawn to the possible infectivity of similar diseases such as CJD. The infectivity of these diseases were demonstrated by intracerebral inoculation of kuru into chimpanzees by Gajdusek et al. in 1966 and by similar ways with CJD in 1968 by Gibbs et al. With this evidence it can only be said that the disease is transmissible and not strictly due to the prions. The strongest evidence was by Prusiner et al. In 1982 where he further purified the prion by demonstration using 3000-fold enrichment for infectivity, and it was the only macromolecule that could be associated with infectivity. However some investigators think that the agent is merely the pathologic product which coincidentally purifies with the "scrapie virus". Despite these ideas, there is little evidence to support this view as fractions containing <1 prion molecule per ID₅₀ have not been found, if it were to be found, it would indicate that prions are not required for infectivity, (Prusiner, 1991).

There is now a lot of evidence that the discovered 27-30 prion protein is a major and necessary component of the prion disease, this evidence includes:

1. That during the copurification of PrP 27-30 and scrapie infectivity using biochemical methods, the PrP concentration obtained is proportional to the prion titre.
2. The kinetics of proteolytic digestion of PrP 27-30 and infectivity are similar.
3. Copurification of the infectious prion and infectivity by immunoaffinity chromatography is achieved, and a -Prion antisera neutralises infectivity.
4. The novel prion protein is detected only in clones of cultured cells producing infectivity.
5. The prion amyloid plaques are specific for prion diseases of animals and humans, and the deposition of the prion amyloid plaque is controlled at least partially by the prion protein sequence.
6. There is correlation between the infectious prion and disease in humans and animals.
7. Incubation time in mice is affected by base substitutions in the mouse prion gene-Mo-PrP.
8. The existence of the species barrier (discussed later) due to amino acid differences in the protein causes a longer incubation time.
9. There is genetic linkage between mutations in the prion gene (also discussed later) and the onset of disease, with a mutation at codon 102 linked with GSS development and a mutation in codon 200, or insertions of six additional octarepeats linked with familial CJD.
10. Mice expressing prion genes with a point mutation similar to that of GSS develop neurological dysfunction, spongiform brain degeneration and gliosis (Prusiner, 1991).

From the evidence discussed at this point most scientists will believe that there is some involvement of prions in the infectivity and pathology of scrapie, BSE and CJD, but what is the nature of the involvement? Investigators suggest an answer to this with the prion concept.

The isolated prion protein was found to be a product of a cellular gene on the short arm of chromosome 20 in humans and in a syntenic region of chromosome 2 in mice. Studies then on the hamster and mouse models, after uncovering a protein designated PrP 27-30 (27-30Kdaltons) led to the eventual molecular cloning and complete sequence of prion protein cDNA, (Oesch et al., 1985, Locht et al. 1986, Palmer & Collinge, 1997).

The normal product (PrP^c) of the prion gene (called PRNP in humans) is found to be expressed in most cell types, with expression at its highest predominantly in the brain. In 1987 investigations involving PrP^c and PrP 27-30 started where an experiment involving bacterial-derived stearic acid treatment of the prion enabling it to react with antisera indicated that PrP^c is anchored to the cell surface by a glycolipid, components of which include ethanolamine, myo-inositol, phosphate, and stearic acid. Later experiments involving PrP^c indicated that PrP^c is expressed as a glycosylphosphatidyl inositol-anchored glycoprotein found on the outer cell membrane, (Stahl et al., 1987, 1990, 1992). However, this normal form can be distinguished from the infective prion found in diseased brain samples by its non-resistance to proteolytic hydrolysis (Borchelt et al., 1990). The infectious prion protein will now be termed as PrP^sc and the normal prion protein, its function later to be discussed is now termed as PrP^c (the ^"sc" stands for scrapie and the "^c" stands for cellular). Different terminology of the two forms of the protein have also come into light. This is the PrP^sens and PrP^res; these names are given to prions, which are not structurally identical to PrP^c or PrP^sc. They may be truncated or recombinant forms, and are not exactly the same as the scrapie (^sc) and cellular (^c) forms, the 'res' indicates resistance to proteolytic degradation, and the 'sens' is for sensitivity to proteolytic degradation. There are therefore two different isoforms of the prion protein, one is expressed normally and one is present aberrantly.

Initial studies on the different forms of the prion (Borchelt et al., 1990) found there were differences in the proteolytic resistance. Also demonstrated was that PrP^c can be released from normal, and scrapie infected neuroblastoma cells which were treated with phosphatidylinositol-specific phospholipase C (PIPLC), and also be labelled by a sulpho-NHS-biotin. Using pulse chase experiments [³⁵S] methionine was found to incorporate almost immediately into PrP^c. PrP^sc however, was incorporated at a much slower rate. PrP^sc was not released by PIPLC, which suggested structural differences. It also enabled them to monitor the amounts of both PrP^c and PrP^sc in the cell. Although this gave no real clues to the mechanisms involved, it further established the presence of two different forms of the protein.

Mass spectrometry and amino acid sequencing studies were undertaken on the scrapie prion protein, this was done using denatured PrP^sc that was digested with endoproteases and isolated by HPLC. Analysis was done using mass spectrometry and Edman sequencing and the primary structure of PrP^sc was found to be identical to that of the PrP gene sequence, arguing that neither RNA editing nor protein splicing feature in the synthesis of PrP^sc. Mass spectrometry was also used to identify any posttranslational modification-differences between the two isoforms, and none were found contending that PrPsc molecules do now differ from PrPc at the level of an amino acid substitution or a post-translational chemical modification. (Stahl et al., 1993) However, they could not eliminate the possibility of a small fraction of PrP^sc being modified by an yet unidentified post-translational process, or that PrP^c carries a modification that is removed in the formation of PrP^sc. It is probable then, that the differences are in the secondary and tertiary structures.

Work was then done on investigating the structure of the prion comparing PrP^c to PrP^sc. The structure of PrP^c was found to have mainly a -helices and the structure of PrP^sc was found to comprise mainly of b -pleated sheets. This was achieved using a pattern-matching approach (Cohen et al., 1985) and Fourier analysis of the primary sequence hydrophobocities to detect amphipathic regions, interestingly the secondary structure showed some homology to that of the serum amyloid A proteins. The prion was linked with having a post-translational modification to yield amyloid A, which is more evidence for the prion involvement in amyloid plaque formation, (Bazan et al., 1987).

Later in 1993, Pan et al. wrote a paper describing the conversion of a -helices to b -sheets as part of the formation of the scrapie prion proteins. Using Fourier Transform Infra-Red spectroscopy (FTIR) it was demonstrated that PrP^c has a high a -helix content (42%) and no content (3%), and PrP^sc has a high b -sheet content (43%) and a lower a -helix content (30%). An even higher b -sheet content is observed (54%) (a -helix content 21%) when using limited proteolysis to created N-terminally truncated PrP^sc. This was confirmed by circular dichroism measurements. Pan et al. (1993) then remarked on the likeliness of a conformational transition being a fundamental event in the propagation of prions. Baldwin et al. (1994) produced a report which, identified and confirmed the differences between PrP^c and PrP^sc using spectroscopy to be conformational and even possibly transitional. For this FTIR spectroscopic methods were used, and also suggested the possibility of the transition of alpha to beta transition, they found that the N-terminal truncation of PrP^sc by limited proteolysis did not destroy the prion infectivity. Instead, it increased the b -sheet content shifting the FTIR absorption to lower frequencies, which was typical of the cross b -pleated sheets of amyloids. Thus, the formation of PrP^sc from PrP^c involves a conformational transition in which one or more a -helical regions of the protein are converted to b -sheets. This transition was mimicked by synthetic peptides, allowing predictions of domains of PrP involved in prion diseases.

Further evidence for the PrP^sc formation from PrP^c was shown in experiments involving the synthesis of a -helix peptides from regions of the prion protein, and their subsequent conversion to form amyloid peptides composed largely of b -sheets. Although this evidence is quite good, harder evidence was to follow, (Gasset et al., 1992, 1993).

In 1996 Muramoto et al. Performed studies on recombinant forms of PrP^c. It is known that some truncation of the PrP^c N terminus still permits conversion into the PrP^sc isoform and so to assess whether additional segments of the PrP molecule can be deleted, they removed regions of the putative secondary structure. They found that removing a 36-residue loop between the second and third helix did not prevent formation of protease resistant PrP (PrP^sc106) that was also found to be soluble. Also, all the deletions of the 4 putative a -helical structures prevented PrP^sc formation. The results suggested that all of the proposed secondary structures are necessary for the conversion of PrP^c into PrP^sc and that the discovery of PrP^sc106 should facilitate structural studies of the prions and possible development of antibodies against PrP^sc. However, as only three a -helical regions were found in PrP^c in later studies, (Prusiner et al., 1998) the experiment was rendered partially inaccurate as they assumed they were working on a PrP^c model which had 4 a -helical regions. Despite this, it seems reasonable to deduce that the deletion of portions of the PrP^c molecule other than the N terminus in the "helical regions" does confer loss of PrP^sc formation.

A year later in 1997, studies investigating the possible posttranslational modifications involved with the PrP^c conversion to PrP^sc focussed on scrapie infected ScN2a cells, where the metabolism of both PrP isolates involved cholesterol dependant pathways. They demonstrated that both PrP^c and PrP^sc are attached to Trition X-100-insoluble, low-density complexes or "rafts". These "rafts" probably contain cholesterol due to their sensitivity to saponin. However, when extracted from solubility into insolubility using Triton X-100 at 37° c they found two different peaks of different densities, suggesting two populations of PrP-containing "rafts". This may permit the isolation of PrP^c specific rafts. Their findings suggest the possibility of "rafts" hosting the proposed PrP^c to PrP^sc formation, or if not some other aspect in the "life-cycle" of prions. (Naslavsky et al., 1997).

The three dimensional-structure for the cellular prion protein was predicted using circular dichroism and infrared spectroscopy (Huang et al., 1994). A heuristic approach was used, consisting of the prediction of secondary structures and of an evaluation of the packing of secondary structures and of and evaluation of the packing of the secondary structures. By using experimental and theoretical constraints they achieved four theoretical structures consisting of a -helix bundles. They identified a group of amino acids as being important for tertiary structural arrangement.

Although the predicted structures gave a testable model, PrP^c is thought now to have only three a -helix bundles (shown later). Possible inaccuracies could be in the computer programme or rather the experimental, theoretical constraints or data applied to the programme.

Despite this, the structure of PrP^c seems to be far more predictable than that of PrP^sc (Baldwin et al., 1998). This is partially due to the very small amount of the natural protein isoforms isolated from the brain, and the solubility of PrP 27-30 in detergents that reduces infectivity even though the b -sheets can be regenerated. Despite this, the NMR (Nuclear Magnetic Resonance) experiments on recombinant protein can lead to a much better characterisation of the 3D structure of the prion proteins. In the studies involving the use of hamster and mouse models to elucidate the three-dimensional structure, the hamster and mouse prion protein structure was compared, and is shown below in Figure 3. The picture (James et al. (1997)) shows the hamster and mouse prion proteins superimposed, the mouse structure is more refined but shorter in length than the hamster prion, showing a close relation and therefore is a good model for transgenic studies (described later).

Figure 3 (James et al, 1997)

Using a simple, low-resolution lattice model of prion protein folding, Harrison et al. (1999) studied the thermodynamics and problems implicated in the prion protein folding. To define their model proteins and prions they studied 3D 27-mer maximally compact cubic conformations and 2D n-mers for a range of chain lengths n from 11 to 16. The 2D studies allowed examination the effects of variation on chain length and of surface complementarity in the dimer. The 3D studies had a much more limited number of possible dimer packing geometries but showed the dimensionality of real proteins. The studies did require attention regarding the hydrophobic and hydrophilic sequences that may affect the folding structure, as PrP is of intermediate hydrophobicity. Although in their HP lattice model, a minor proportion of the proteins encrypted a native multimeric state and did not require a hydrophobic or hydrophilic sequence to affect this change. This is consistent with real prions as while PrP is of intermediate hydrophobicity, the yeast prions Ure2p and Sup35p are more hydrophilic than the average protein of known structure. They found that if the model were to re-fold upon dimerisation, then a minor proportion of up to about 17% encrypt an alternative native state as a homodimer. The structures in this homodimeric native state re-arrange so that they are very different in conformation from the monomeric native state. They also found that model proteins which are relatively less stable are more susceptible to the formation of alternative native states as homodimers. The results obtained suggest the less-stable proteins have a greater need for a well-designed energy landscape for protein folding to overcome an increased chance of encrypting substantially different native conformations stabilised by multimeric interactions.

Prusiner predicted that there were four a -helices in PrP^c, and later demonstrated that in fact there were only three a -helical regions termed helix A, helix B and helix C. Interestingly the NMR structures of recombinant Syrian hamster PrP^c (rPrP (90-231)) and Mouse PrP^c (MoPrP (121-231)) although similar in many respects do contain substantial differences.

For example: The loop at the NH₂ terminus of helix B being well defined in rPrP (90-230) but is disordered in MoPrP (121-231). Additionally, helix C is composed of residues 200-227 in rPrP (90-231) but in MoPrP (121-231) extends only from 200-217 residues. The undetermined cause of these differences could be due to one of three possibilities:

1. They have different lengths.
2. Species-specific differences in sequences.
3. The conditions used for solving the structures.

Recently NMR studies of the full length SHaPrP (29-231) and MoPrP (23-231) have demonstrated that the NH2 termini are highly flexible and lack identifiable secondary structure under the experimental conditions employed. Also transient interactions between the COOH terminal end of helix B and the highly flexible, NH2-terminal random coil containing the octarepeats, (residues 29-125), (Prusiner et al., 1998).

Figure 4 (Prusiner, 1996).

Figure 4 (Prusiner, 1996) shows the original four a -helices (left-previous page) with the rogue form (right-previous page) and Figure 5 shows the 'corrected' three a -helices (below left) with a relatively unchanged rogue form (below right).

Figure 5 (adapted from www.mad-cow.org).

The main differences between PrP^c and PrP^sc can be summarised in table 6 shown below, adapted from Murray et al., (1997).

Table 8.

Earlier this year, Willie and Prusiner (1999) were able to perform ultrastructural studies on scrapie prion protein crystals produced using reverse micellar solutions. Structural studies on the prions in this crystal form may provide insight into the structural transition which occurs during PrP^sc formation.

The prion gene is called PRNP (in humans) and is found in humans on the short arm of chromosome 20, and in murine in a syntenic region of chromosome 2; these two locations are homologous (Sparkes et al., 1986). The gene has only one protein coding region. In mice the gene is called the Prn-p gene and in other species it is the PrP gene. Only one homologous gene (PrLP) in a non-mammalian organism (chicken) has so far been described. PrP related DNA hybridisation signals have been detected in organisms such as Drosophila, the nematode Caenorhabditis elegans and even yeast (described later). A novel gene from C.elegans was isolated by Iwasaki et al. (1992) using a hamster PrP cDNA probe. They found that the encoded protein was homologous to nuclear ribonucleoproteins but not to mammalian PrP. These results make it more likely that PrP is a late gene in evolutionary terms, (Goldmann, 1993).

Inherited forms of the disease comprise 25% of all cases of prion diseases notably GSS, familial CJD and FFI. In each of the inherited forms, mutations have been found in the ORF (open reading frame) of the PRNP gene. The first half of the PRNP ORF contains about 170bp with a high content (about 80%) of the nucleotides guanidine (G) and cytidine (C), most of this sequence is organised in 24bp (or 27bp) repeats. Few differences are observed between these sequences, and between those in other species suggesting that they are highly conserved through evolution. It has been suggested that control and fidelity of PrP mRNA translation is associated with these high G/C content regions. In situ and northern blot hybridisation has shown that the gene is predominantly expressed in neuronal cells as well as ganglia and nerves of the peripheral nervous system. It also showed that it is not exclusively expressed in the CNS and neurons as the kidney, heart, lung and spleen all contained transcripts except for the liver which was negative. Apart from the possible exception of sheep, inherited forms of the disease have only been found in humans. This raises interesting questions that cannot be readily pursued in animals, (Goldmann, W., 1993, Prusiner, 1996).

These mutations found in the ORF of the PRNP are illustrated in Figure 6, adapted from Prusiner (1996).

Figure 6 (Prusiner et al., 1996)

As Figure 6 shows, there are many mutations that have been identified with the PRNP ORF and are often genetically linked to hereditary prion disease. This includes a proline (P) to leucine (L) mutation at codon 102 that was shown to be linked genetically to development of GSS with a LOD score exceeding three. (The LOD score is a determinant of genetic linkage, a score of 3 or greater is considered as confirmation of linkage, as it would yield a ratio of 1000 to 1 in favour of linkage as opposed to non-linkage, i.e. > 5cM apart, (Muller & Young (1998)). This mutation may be due to the deamination of a methylated deoxycytosine (C) coupled to deoxyguanosine (G) through a phosophodiester bond (CpG) in the germline DNA encoding PrP resulting in the substitution of deoxythymine (T) for deoxycytosine. This P102L mutation has been found in 10 different families in 9 different countries including the original GSS family. Israeli Jews of Libyan origin showed an unusually high incidence of CJD, this was thought for many years to be due to the consumption of lightly cooked sheep brain or eyeballs. However, studies uncovered some Libyan and Tunisian Jews in families having a PrP point mutation at codon 200 resulting in a glutamate (E) to lysine (K) substitution. One of the patients were homozygous for this mutation (E200K) and as her clinical presentation was similar to that of the heterozygote patients, it was suggested that familial prion diseases are true autosomal dominant disorders. The E200K mutation has been found in many areas including Slovaks originating from Orava in North Central Czechoslovakia, in a cluster of families in Chile, in a large German family living in the United States, as well as in British and Japanese families. This diversity suggests a recurrent point mutation, but some investigators have argues that all the E200K mutations originated from a Shepardic Jew whose descendants migrated from Spain and Portugal at the time of the inquisition. Although, others agree that it is more likely that the mutation has arisen independently multiple times by the deamidation of a methylated CpG. This is supported by historical records of Libyan and Tunisian Jews, indicating that they are descended from Jew living on the island of Jerba where they first settled around 500 BC and not from Shepardim.

At codon 178 a mutation involving the substitution of aspartic acid (D) to asparagine (N) has been identified in many families with CJD, and like with the E200K mutation the PrP amyloid plaques are rare, the neuropathologic changes generally consisting of widespread spongiform degeneration. The D178N mutation has been linked with a number of Italian families with cases of insomnia, although the mutation appeared to be incompletely penetrant. The same mutation was also reported in several families affected by a disease phenotypically different from FFI and similar to CJD, except for the longer duration and the lack of sharp-wave electroencephalographic activity in most of the cases. This finding that the same mutation gives two different phenotypes prompted a series of studies to discover the molecular basis of this phenotypic heterogeneity. A detailed analysis of the PRNP genotype in 15 FFI and 15 CJD patients showed that in addition to the D178N mutation, all of the FFI subjects had a methionine at position 129 of the mutant allele while all CJD subjects had valine at this same position. These results have been confirmed in all of the FFI and CJD cases. Therefore this gives two distinct haplotypes, the 129M, D178N haplotype in FFI, and the 129V, D178N haplotype in CJD. As one of the FFI kindreds has an octapeptide repeat deletion in the mutant allele, it is very unlikely that all of the known FFI kindreds originated from a common founder. This finding strongly argues against the possibility that the phenotypic differences are caused by genetic influences other than PRNP codon 129. Interestingly although the methionine or valine at codon 129 on the mutant allele is obligatory in FFI and CJD178 patients respectively, the codon 129 on the normal allele can be either methionine or valine. Therefore, the FFI and CJD phenotypes are likely to be determined by the codon 129 of the mutant allele, which in association with the D178N mutation, results in the expression of two different types of PrP^res. Also, as FFI is usually expressed in the phenotype earlier than CJD, the codon 129 also modulates the duration of the phenotype. Studies on the PrP^res fragments associated with the two proteins differ both in size and in the ratio of the three differently glycosylated PrP^res isoforms. The size variation is the result of the differential N-terminal digestion by proteases and the difference indicates that PrP^res has different conformations, or specific-ligand interactions. The ratio difference however indicates a different post-translational processing of PrP in the two diseases to ultimately give two different phenotypes. Also noted in these cases were the different incubation times in relation to the heterozygosity and homozygosity of the mutant allele. The homozygote duration of the disease was significantly shorter than that of the heterozygotes. The mean age of onset of CJD in homozygotes was 39+/- 8 years and in the heterozygotes it was 49+/-4 years.

A valine (V) to isoleucine (I) substitution at codon 210 produces CJD with classic symptoms and signs, and like the D178N mutation appears to show incomplete penetrance. GSS has been associated with mutations in codons 105 and 114. Other point mutations have been shown at codons 145, 198, 217 and possibly 232 that segregate with inherited prion diseases. Interestingly, synthetic peptides adjacent to and including residues 109 to 122 respectively have readily polymerised into the rod-shaped structures, which have the tinctorial properties of amyloid, (Prusiner, et al., 1996, Parchi et al., 1996).

Other than base substitutions, octapeptide inserts can also cause mutations. An insert of 144bp at codon 53 containing 6 octarepeats was initially described in patients with CJD from four families all residing in southern England. As the human PrP gene only contains 5 octarepeats a single genetic recombination event could not have created this extra insert. Although as the four families were distantly related, a single person born more than 2 centuries ago may be the founder (LOD score being greater than 11). Studies from several laboratories have demonstrated that two, four, five, six, seven, eight or nine octarepeats in addition to the normal five are shown in individuals with inherited CJD. Deletion of one octarepeat has also been identified but without any neurological disease, (Prusiner et al., 1996). Recently, the location of human mutants has been shown on a mouse prion model, this is shown in figure 7 below taken from James et al. (1997).

Figure 7 (James et al., 1997)

Octapeptide repeats are essentially insertional mutations, and the patients usually have a prolonged illness (averaging 7 years) beginning at an early age averaging 38 years, characterised by progressive dementia accompanied by cerebellar and other neurological signs. The genetic mechanism for the generation of extra repeats is probably unequal crossover, as shown in Figure 8 taken from Goldfarb et al., (1991).

Figure 8 (Goldfarb et al., 1991)

The similarity of the observed repeats allows limited frameshifting in the process of duplication; recombination may happen at any site. As in the diagram, the result of the first event could be a 7-repeat allele, which has not so far been observed. The second event then could be the 9-repeat allele, and further crossover events could result in alleles with 10 or more repetitions. The PrP repeat region has been compared with many structural proteins (collagen, silk fibroin, keratin and elastin) with no close resemblance. However, three others have been closer, including the maize ribosomal protein S18 that has a 7-fold heptapeptide repeat at its N terminus that is thought to have some role in RNA binding. Also, the DNA-dependant RNA polymerase II contains many repeats of a heptapeptide in its C terminal domain. It is here where a regular spacing of proline residues appears to be important in maintaining its unusual secondary structure. The variations observed have been thought to act to regulate enzyme activity. Interestingly ice nucleation points also contain large numbers of octapeptide, each contributing equally to the nucleation process.

Essentially, the mechanistic properties of the repeat function are not known, although there are several possibilities, such as the possibility that it may be involved directly in a templating mechanism, whereby the normal protein is converted posttranslationally into an enantiomorphic form of unusual conformation. The presence of extra repeats might be expected to increase the probability of a spontaneous transition to the abnormal enantiomer and thereafter provide an increasing concentration of template for the conversion of further molecules. The cis-trans isomerisation of the proline residues may play an important part, or if due to the extra repeats, pseudoknots are able to form changing the mRNA performance leading to the production of a PrP^res protein. Nevertheless, whatever the mechanisms are by which the extra repeats exert their effect, the presence of 10 or more such elements is associated with a large range of dementing illnesses that may look like CJD but differ in their time of onset, and slow progression. However, it has been shown that they have been experimentally transmitted horizontally, and therefore share the same transmissibility characteristics as other familial and sporadic forms of spongiform encephalopathy, (Goldfarb et al., 1991).

A study by Murmoto et al (1997) was based on the effect of a deletion of one or more of the hypothetical a -helix coding regions in the PRNP gene. They discovered that in mice this deletion predisposed to a heritable disorder resembling neuronal storage disease. This caused CNS dysfunction and numerous neurons with intracellular accumulations of PrP, which are insoluble in non-denaturing detergents, a property which distinguishes it from the PrP^c. Whether this a -helical deletion disease described here is transmissible is unknown. Their findings described strongly argue that the pathogenesis of this disorder differs considerably from that of the normal prion diseases. No human PRNP mutations have been linked to the loss of a -helical regions, but this could be due to the phenotype differing a lot from the classical CJD phenotype. Interestingly there has been little attention given to the molecular genetics of the PRNP gene in childhood CNS diseases, since prion diseases usually present in adults. But, with the findings of children as young as three or four with kuru in Papa New Guinea (acquired syndrome so children can get the disease), there seems no reason why an inherited disease shouldn't strike. Especially as we do not know the mechanisms by which the subjects carrying a PRNP mutation remain asymptomatic until they reach the mature age. It may then be logical to conclude from this that in future, brain sections should be immunostained for PrP in CNS storage diseases of children where the aetiology is unclear and to sequence the ORF of PrP alleles from these patients.

In conclusion it can be said that recent studies of the genetics of familial prion disease have yielded three major results:

1. Significant genetic linkage between ataxic GSS and PRNP indicates that this horizontally transmissible disease is inherited through a locus at or near PRNP.
2. A leucine substitution at PRNP codon 102 is genetically linked to ataxic GSS in families from ethnically and lineally distinct backgrounds, suggesting that this mutation arose independently at least twice and is therefore a well linked genetic marker. The substitution in question may conceivably cause or directly influence the pathogenesis of the disease.
3. This and other mutations in PrP, is found to track with disease in families with a strong, but not absolute correlation of specific mutations in PrP with different clinical forms of the prion disease, (Hsaio and Prusiner, 1990).

The discovery that FFI is a prion disease widens the clinical spectrum of these disorders and raises the possibility of other degenerative diseases being prion related.

Despite all the advances in work done with prions there are still issues which remain to be addressed. Generally, the challenge is to determine the individual steps that lead from the presence of germline mutations in PRNP to the onset of diseases with various phenotypes and duration. Essentially the issues to be resolved are:

(a) The mechanisms by which the subjects carrying a PRNP mutation remain asymptomatic until they reach the mature age.

(b) Whether the conversion of PrP^sens into PrP^res is a requirement for the onset of the disease.

(c) The specific events that predispose the mutant PrP to convert to PrP^res.

(d) The mechanism of selective vulnerability of different neuronal populations to the various mutant PrP.

From all of the evidence put forward above, it seems certain that the production of the aberrant form of the PRNP gene product is responsible in some way for the pathogenesis of the disease. We do know that the PrP^c protein is expressed as a glycosylphosphatidyl inositol-anchored glycoprotein found on the outer cell membrane of neurons and to a lesser extent of lymphocytes and other cells, (Stahl et al., 1987, 1990, 1992). However, the role of the normal cellular form of the prion is unclear and unproven.

In 1992, studies on mice with inactive Prn-p genes (homozygously disrupted) by Büeler et al. showed that the mice developed normally. The genes were disrupted by homologous recombination with a 4.8 kilobase DNA fragments in which codons 4 to 187 of the 254-codon open reading frame were replaced by a neomycin phosphotransferase (neo) gene under the control of the herpes simplex virus thymidine kinase promoter. They then obtained mice heterozygous (Prn-p^0/+) for the disrupted gene, the presence of which was confirmed by Southern analysis. Heterozygous mice were then mated to other heterozygotes to give a new population. Surprisingly, this population contained 24% Prn-p^(0/0)_, 26% Prn-p^(+/+) and 50% Prn-p^(0/+) mice, surprising because they were all phenotypically the same, and Southern analysis confirmed these results. The Prn-p^(0/0) mice were tested for many abnormalities in their behaviour, fitness and fertility and were all confirmed normal for at least 7 months, with no immunological defects. The reason that these results are surprising is that the Prn-p gene is highly conserved through evolution and so it might be thought that the function of this gene is vital to the organism.

In contrast to Büeler et al. (1992), more recent studies have reported that the prion protein is necessary for normal synaptic function (Collinge et al., 1994) and that altered circadian activity rhythms and sleep has been observed in mice also devoid of the prion protein, (Tobler et al., 1996). One might ascertain from PrP knockout mice, that although there appears to be no symptoms, by looking closely at the behaviour of the mice they might be seen to differ phenotypically from the PrP-normal mice. It seems that Büeler et al. in their studies in 1992 may have missed the subtle phenotypic differences observed by Tobler et al. (1996) and Collinge et al. (1994) in later more focused studies.

Collinge et al. (1994) showed that hippocampal slices from PrP null mice have weakened GABA_A (g -aminobutyric acid type A) receptor mediated fast inhibition and long term potentiation. Interestingly, this gives a clue as to why Büeler et al (1994) may have missed any differences, as in their studies the PrP-null mice that were used were reported to perform the Morris swim task as well as the controls. This can be seen as surprising, as this behavioural task depends on hippocampal integrity, this was also reported with Collinge et al (1992) but when the mice were extensively studied, the phenotypic differences were exposed. The experiments by Collinge et al. (1994) involved taking slices from the hippocampus in mice expressing normal PrP^c and comparing the electrophysiological responses to those of PrP-null mice (PrP^0/0). The results were recorded and the extracellular and intracellular responses are shown in Figure 9.

Figure 9 (Collinge et al., 1994)

Part (a) shows the extracellular responses evoked by afferent stimulation. The null mice reveal extra population spikes. In diagram part (b) a pair of action potentials (only one observed in the normal) and a lack of a hyperpolarizing fast i.p.s.p. (conductance) which was shown in the normal, can be observed. Despite this, there was no difference in the resting potential. The conductance abnormalities in the PrP-null mice were consistent with GABA having to diffuse further to a significant proportion of receptors with extrasynaptic properties, which could arise if the receptors were not properly localised, or if the synaptic cleft architecture had changed. The abnormalities of synaptic inhibition observed in these studies are relevant to the epileptiform discharges seen in CJD and scrapie-infected mice. So it might be seen that the loss of function of PrPc could bare some relation to neurodegeneration. A later study by Tobler et al. (1996) showed that the null mice had an alteration in both circadian activity rhythms and sleep patterns. Their results were confirmed by producing null mutants in two different ways to achieve the same results; effectively they show that a loss of PrP affects the circadian activity rhythm and sleep.

They suggest that because of the intriguing similarities to rhythm and sleep alterations in fatal familial insomnia, where there is a profound alteration in sleep and the daily rhythms of many hormones, the loss of function PrP may be related to the normal function of the prion protein. Essentially, it can be seen that the probable function of PrP^c, being a well-conserved protein present in many mammalian species, may be involved in maintaining sleep continuity or regulating sleep intensity. Another study, this time performed by Sakaguchi et al. (1996), looked at the long-term effects of the disrupted gene on mice. At about 70 weeks of age, their PrP-null mice all began to show progressive symptoms of ataxia. A rotorod test evidently showed up a lack of co-ordination in these mice and upon pathological investigation, there was an extensive loss of Purkinje cells in the vast majority of cerebellar folia. This suggests that PrP plays a role in the long-term survival of Purkinje neurones.

A study by Westaway et al. (1994) on the over-expression of PrP in mice led to the discovery that uninoculated older mice harbouring high copy numbers of wild-type (wt) PrP transgenes derived from Syrian hamsters (SHa), sheep (She), and mice developed the typical symptoms of prion disease. These transgenic (Tg) mice also exhibited a profound necrotising myopathy involving skeletal muscle, a demyelinating polyneuropathy, and focal vacuolation of the central nervous system. The development of disease seemed dependent on transgene dosage. For example, half of all Tg SHaPrP^+/+ mice homozygous for the SHaPrP transgene array developed disease by approximately 460 days of age, while no hemizygous Tg SHaPrP^+/0 mice became ill before 650 days. The novel neurologic syndrome found in older Tg (wtPrP) mice implies that overexpression of wtPrP^c is pathogenic and so widens the spectrum of prion diseases. This report further implicates the involvement of the PrP gene product in prion diseases and also suggests that it needs to be carefully regulated as over-expression, and as previously described, under-expression of PrP^c has detrimental effects on the fitness of the mouse.

In 1993 studies by Büeler et al. using transgenic mice with the Prn-p gene inactivated showed that the PrP-null mice were in fact resistant to scrapie. It was thought that if indeed PrP^sc was an essential component of the scrapie agent, then mice devoid of PrP should be resistant to infection, and so developing neither symptoms of scrapie nor allowing the propagation of the infectious agent. However, if the animals not expressing the Prn-p were to succumb to the disease, then the protein only hypothesis would be falsified. After the generation of mice homozygous for the disrupted genes (Prn-p^0/0) (Büeler et al., 1992), it became possible to study the response of the Prn-p^0/0 mice to inoculation with scrapie prions, as well as of animals carrying a single Prn-p allele (Prn-p^+/0) and of Prn-p^0/0 mice reconstituted with Syrian hamster PrP genes. Their results showed that mice devoid of PrP are completely protected against scrapie for at least 13 months after inoculation. Interestingly, even heterozygotes are partially protected, in that 90% of the heterozygote mice showed signs of the disease around 253-322 days after inoculation but were alive after 322 days, with the interval between first symptoms and death being 2 months. The wild type mice died in about 180 days with the interval between first symptoms and death being 13 days. No infectious agents were detected in Prn-P^0/0 animals 25 weeks after inoculation, which was the latest time point for which data was available. So it seems clear that if the infectious agent is propagated in Prn-P^0/0 mice then this would be at a level of about 5 orders of magnitude lower that that in Prn-P^+/+ animals. Büeler et al. (1992) then concluded that the development of scrapie symptoms and pathology is strictly dependent on the presence of PrP and that incubation time and disease progression are inversely related to the level of PrP. There was the possibility that antibodies against PrP might be formed in the Prn-P^0/0 mice, and such antibodies could affect susceptibility to scrapie infection, but Büeler et al. (1993) failed to detect anti-PrP antibodies in inoculated mice. In the same year Prusiner at al. (1993) also demonstrated that the disruption of the Prn-P gene in mice prevented scrapie. However, this time Prusiner et al. also demonstrated the production of anti-PrP-antibodies, they did this using Mo (mouse) prions and SHa (Syrian hamster) prions. The anti-PrP antibodies in sera were detected by Western blotting, and reacted against Mo, SHa and human PrP.

The prolonged incubation times observed with the heterozygotes in the study by Büeler et al. (1993) suggests that resistance to scrapie is due not to the sought after immune response, but to a rate limiting step due to the concentration of PrP^c. This suggests a positive correlation between the amount of PrP^c and the formation of PrP^sc. Ideas from this study have evolved two ideas, the first is that it may be possible to breed sheep or cattle resistant to this disease, either by PrP gene disruption or the introduction of transgenes expressing PrP anti-sense RNA. The second idea involves a moderate reduction of PrP^c synthesis that could be achieved by anti-sense oligonucleotide therapy, which could be used to slow disease expression. This could be used in early cases of CJD, or even better, in inherited CJD where the sufferers know they will get it in their thirties or so, but are not yet expressing any symptoms. Studies by Brandner et al. (1996) support the idea for the requirement of PrP for scrapie infectivity with evidence obtained in an experiment which was slightly different but produced similar conclusions. They grafted brain tissues over-expressing PrP^c 5-8 fold (which would show incubation times of 60 instead of 160 days in mice) into PrP-null mice. After the intracerebral inoculation with scrapie prions, they observed that the grafts accumulated high amounts of PrP^sc and developed severe histological symptoms of the disease. Also, PrP^sc was recorded to spread from the graft into the host brain tissue. Despite this, no pathological changes were seen in the host tissue, even in the immediate vicinity of the grafts. So, interestingly, in addition to being resistant to scrapie infection, brain tissue devoid of PrP^c does not seem to be damaged by exogenous PrP^sc. This opens a new window of discussion as it has been argued that scrapie pathology is due to intrinsic neurotoxicity of PrP^sc rather than to depletion of PrP^c due to studies by Büeler et al. (1992) suggesting that a lack of PrP^c doesn't affect normal development of mice. Conversely Tobler et al. (1996) and Collinge et al. (1994) suggest this lack of PrP^c does affect normal development. This could mean that the intrinsic neurotoxicity may not be the cause of pathology. Deposition of PrP-immunoreactive material co-localises with typical histopathology, and synthetic amyloidogenic PrP fragments as well as liposome-packaged PrP^27-30 have been shown to be neurotoxic in-vitro, (Brandner et al. (1996)). But, in the studies performed by Brandner et al. (1996), the regions adjoining the grafts with high levels of leaked PrP^sc showed no scrapie pathology. This apparent paradox may be reconciled by the idea that PrP^sc is inherently non-toxic and the plaques formed in spongiform encephalopathies are an epiphenomenon rather than a cause of neuronal damage. This is supported by the fact that the extent of PrP deposition in the brains of humans succumbing to prion diseases with similar clinical presentation is extremely variable. Another idea is that PrP^sc is only toxic when it is formed and accumulated within the cell, but not when it is presented from outside. The other idea is that PrP^sc is pathogenic when presented from outside, but only to cells expressing PrP^sc either because it initiates conversion of PrP^c to PrP^sc at the cell surface and/or it is internalised by was of association with PrP^c, which is endocytosed efficiently. The studies by Brandner et al. (1996) suggest this latter idea in that their data implies that it is not the deposition of PrP^sc, but the availability of PrP^c for some intracellular processed elicited by the infectious agent that is directly linked to spongiosis, gliosis and neuronal death. This seems to make sense and fits in with the results of the simplest experiments involving inoculation of the scrapie agent into unaffected sheep, (Cuillé and Chelle (1936)). A review by Goldmann in 1993 identifies what I believe are the key points that need to be resolved to identify the function of the PrP^c transmembrane protein. These are (i) Do other proteins substitute PrP function? and (ii) Does pathology develop only at a certain age or stress?

One might conclude that despite reports to the contrary, (Büeler et al. (1992) there is sufficient evidence that the function of the PRNP gene and PrP^c is important for the health of individuals, if not for the observed changes in synaptic function and health of Purkinje cells (Tobler et al. (1996), Collinge et al. (1994) and Sakaugchi et al. (1996) respectively), but for the fact that it has been highly conserved through evolution. Therefore it is strongly selected for, which indicates an advantageous trait favoured by natural selection over those not possessing the intact gene, and thus vital to the survival and successful reproduction of an individual within a normal population. It might be a good idea to conduct an experiment, where individual mice lacking the gene would be placed into a population of normal mice and see how they compete and integrate into the normal mouse population. Then the gene may or may not be observed to be selected against, as current studies have not looked closely at the behaviour given the conservation of this gene through evolution. In addition, it must be noted that the damage to the Purkinje cells was noticed when the mice were very old and likely to have passed their sexual peak (Sakaguchi et al.,1996). This feature could have less affect on the evolutionary conservation of the gene as natural selection is based on sexual selection, which occurs when mainly when the mice are less than 70 weeks old and possibly asymptomatic. Despite this, there could be lesser but still important effects not observed in the laboratory including mice-family behavioural patterns, which may account for the evolutionary conservation.

It is interesting to find that despite all the recent attention focussed upon the molecular studies of the disease, no one has yet tried to look for similarities between the pathology of the prion diseases and the other neurological disorders, I am specifically referring to Alzheimer's disease. Alzheimer's disease is a dementia affecting 1 in 5 people over 80 years. It is characterised by the deposition of insoluble amyloid plaques (composed principally of b -amyloid) and the formation of neurofibrillary tangles, leading to neurodegeneration in various regions of the brain. In 1997 work done by Fraser et al. investigated the ionic effects of the Alzheimer's disease beta-amyloid precursor protein and its metabolic fragments. Their work is based on substantial evidence that has implicated beta-amyloid (and other amyloidogenic fragments of the precursor protein) with the neurodegeneration observed in Alzheimer's disease. They claim that beta-amyloid precursor protein and its amyloidogenic metabolic fragments have been shown to alter cellular ionic activity, either through interaction with existing channels or by de novo channel formation. This alteration in the ionic homeostasis has also been linked with cellular toxicity and might provide a molecular mechanism underlying the neurodegeneration seen in Alzheimer's disease. I find it remarkable that this work investigating the link with disease so similar to prion diseases has not at least been investigated and comparatively dismissed or subsequently further investigated by scientists. Alzheimer's and prion diseases are similar in that they both are neurodegenerative and amyloid plaque forming and the symptoms are not identical but some are shared. What needs to be addressed is the possibility that like with Alzheimer's disease, the pathology of the disease is related to ion channel incorporation into the phospholipid bilayer of neurones where damage can then occur.

The structure of the amyloid plaque is shown below in Figure 10, and computer-generated models of the fibrils are also shown below (Figures 11, 12 and 13).

Figure 10 An electron micrograph showing PrP amyloid (Smith, 1998)

Figure 11 (www.mad-cow.org)

Figure 11. A model of the generic amyloid fibril structure. A Molecular model of the common core protofilament structure of amyloid fibrils. A number of b -sheets (four illustrated here) make up the protofilament structure. The sheets run parallel to the axis of the perpendicular to the fibril axis. The b -sheets twist around a common helical axis that coincides with the axis of the protofilament, giving a helical repeat of 115.5Å containing 24 b -strands (indicated by the boxed region) (adapted from www.mad-cow.org).

Figures 12 and 13 are shown below. Figure 12 shows 12 fibrils (groupable as 1, 2, 3, 4, or 6) wrapping with a hollow tube of ratio 44:100 and figure 13 shows a non-supercoiled fibre comprised of 4 fibrils.

Figures 12 and 13 (adapted from www.mag-cow.org).

The likeliness that there is a link between the two disorders may be low, but unless it is properly investigated then it cannot be dismissed. There is evidence to help this idea, in the form of studies by Colling et al. (1996) investigating the intrinsic properties of hippocampal CA1 pyramidal cells lacking the prion protein. They compared the resting potentials, time constants, amplitude of the medium after hyperpolarisation (AHP) and spike firing accommodation, all of which, did not differ between the wild type and null mice. What did differ were the lower input resistances observed in the PrP-null mice, as well as a lack of the late AHP and of a charybdotoxin-sensitive summated AHP. They proposed then that Ca²⁺- activated K⁺ currents are disrupted in PrP-null mice. This not only brings the idea of Ca²⁺ channels being related to damage (like with Alzheimer's disease) (and now possibly PrP^sc) into perspective, but also suggests yet another role for the function of PrP^c.

A recent study by Mabbott et al. (1997) has looked at the possibility of the involvement of T-lymphocyte activation as a role of PrP^c, this is due to the fact that PrP^c is a GPI-anchored protein (membrane attached via a glycosylphosphatidy-linositol anchor). Certain GPI anchored proteins on lymphocytes have previously been shown to play an important role in an early step of cell activation, and as the expression of PrP^c has also been identified in lymphoid tissues, they hypothesised that PrP^c like other GPI-anchored proteins may play a role in lymphocyte activation. In their study they demonstrated that the PrP^c is expressed on the surface of murine lymphocytes, and following mitogen stimulation the lymphocyte PrP^c expression was significantly enhanced suggesting a role PrP^c in lymphocyte activation. Also the mitogen responsiveness in PrP^c-deficient lymphocytes was significantly reduced in comparison to wild-type controls. Therefore, the results imply that PrP^c is expressed on the extracellular leaflet of the lymphocyte membrane and involved in lymphocyte activation. Although this report does suggest a role for PrP^c being involved in immune function it must be noted that the PrP^c molecule is not the only GPI-anchored protein expressed on the surface of murine lymphocytes. Other examples include the Thy-1, Ly-6, CD58, CD-59 and Oa-2 proteins. Previous work demonstrates that the non-specific removal of all GPI-anchored proteins with a phospholipase C enzyme abrogates the ability of murine lymphocytes to respond to stimulation by the mitogen Con A, so there is evidence that the GPI-anchored proteins are involved with activation of lymphocytes. However, PrP^c is predominantly produced in the brain and not predominantly in the peripheral organs and those expressing lymphocytes (spleen, thymus gland, lymph nodes, gut wall and bone marrow). But, this does not necessarily mean the lack of lymphocyte involvement of PrP^c. PrP^c may be involved in a multiple role, for example it may have a role in the brain and a role related to lymphocytes, this would explain the highly conserved nature of the gene, as a mutation in the gene would be more likely to give a loss of function. Or there is the possibility that PrP^c can be involved but in a coincidental way, as many other GPI-anchored proteins are known to have the same effect. Work done by DeArmond et al. (1997) has uncovered the possibility that glycosylation can modify the conformation of PrP^c. Glycosylation could also affect the affinity of PrP^c for a particular conformer of PrP^sc, and there by determining the rate of nascent PrP^sc formation and the specific patterns of PrP^sc deposition. They postulated that variations in the complex carbohydrate structure in the CNS could account for selective targeting (i.e. the formation of PrP^sc in particular areas of the brain being due to differences in PrP^c/PrP^sc affinity). The experiments were done using deletions of Asn-linked oligosaccharides of PrP^c in transgenic mice. Recently, Ikeda et al. (1999) have performed experiments on cultured hepatic stellate cells, and by using immunoelectron microscopy, have revealed that PrP^c resides on the plasma membrane of stellate cells. Their results indicate that PrP expression is closely related to stellate cell activation associated with fibrogenic stimuli. This has helped elucidate the pathogenesis of liver fibrosis, and develop possible therapeutic strategies to prevent liver cirrhosis. PrP^c is not expressed in the normal healthy liver, but it was found to increase dramatically in diseased livers associated with stellate cell activation.

Despite the promising work incorporating the PrP-null idea, it must be mentioned that not all of the PrP gene knockout techniques are the same. They all ultimately result in the loss of gene function, but some are due to small deletions and some are due to large deletions. There is also a chance that during embryogenesis that other proteins or mechanisms may work to compensate the effects of the loss of the PrP-null, so this shows that the mouse PrP-null model is not a perfect model for prion diseases. It is clear from the experiments described, that more studies aimed at the function of PrP^c and the behavioural effects of PrP-null mice need to be performed. There have been many questions and possibilities raised, including the possibility of glycosylation differences in PrP^c/PrP^sc affinity, and there seems to be no one-quick-way answer, especially with the differences in the gene knock-out techniques and their corresponding implications and effects. It might then be a good idea to have a kind of standardised way of knocking-out these genes after a full investigation on how the different techniques may affect the mice in different ways.

Evidence for the conversion of PrP^c into PrP^sc

Despite papers suggesting the conversion of PrPc into PrPsc, and the involvement with a -helices changing to b -sheets, (Baldwin et al., 1993; Pan et al., 1993) they did not have any hard evidence to demonstrate this phenomenon. In 1994, Kocisko et al. managed to demonstrate the cell free formation of a protease resistant prion protein. Their strategy was to develop conditions which allowed the reversible denaturation of PrP^sc and then to test if de novo conversion of PrP^c into a proteinase K (PK)-resistant form could be observed under these conditions. They did this by adding S-labelled recombinant hamster PrP^c derived from uninfected tissue culture to a PK-resistant species and incubated for two days. When the proteins were broken into fragments by analysis they produced S-labelled PK-resistant PrP fragments of comparable size to PK-digested PrP^sc from scrapie infected sources. These S-labelled PK-resistant PrP bands were not observed in the absence of PrP^sc (control) and were greatly reduced in the presence of PrP^sc, indicating that some native PrP^sc structure was required for the formation of large resistant species. Thus, their study shows the PrP^c can be converted selectively to PK-resistant forms similar to PrP^sc in a cell-free system. Earlier this year another study on the topic of conversion of PrP^c into PrP^sc (Jackson et al., 1999) supported and further developed this idea. Jackson et al. demonstrates the reversible conversion of monomeric human prion protein between native (PrP^c) and fibrilogenic (b -PrP) conformations. By producing PrPc from amino acids 91 to 231 (PrP^91-231) with the help of Escherichia coli, they were able to conduct structural studies on the predominantly a -helical form of the protein, which was highly soluble, and monomeric with a single disulphide bridge. The reduction of this disulphide bridge and the lowering of the pH to 4.0 in a dilute acetate buffer without additives, generated a highly soluble protein that can be concentrated to at least 12 mg/ml that turned out to be monomeric by using gel filtration. The chemical treatments were used to make a monomeric protein fragments that was still soluble and amenable to structural analysis, but they found that under these conditions, the breakage of the disulphide bond, the structure had changed to one that was dominated by b -sheets. When this protein was treated with a salt concentration similar to that found in brain tissues, the b -PrP precipitated into insoluble aggregates forming fibrils, similar to those found in prion diseases. The b -PrP was also resistant to proteolytic degradation, further conforming to the properties of PrP^sc. Interestingly, this reaction is reversible, and by exposing the b -PrP to a pH of 8, the native a -helical conformation was restored, however, the rate of conversion in either direction was extremely slow requiring days for completion. This slow process can be bypassed by fully denaturing the protein and refolding it at the appropriate pH to generate either isoform. Unusually for a protein with a predominantly helical fold, most of the residues in PrP^91-231 have a preference for b conformation, suggesting that PrP is balanced between radically different folds with a high energy barrier between them, one being the a conformation and the other being the b conformation. The a conformation requires the precise docking of side chains while the b conformation is dictated by the local secondary structural propensity. Jackson et al. (1999) also suggested the link between this unusually balanced structure and amino acid 129 of the PRNP gene. This amino acid mentioned previously (inherited diseases) can be either methionine or valine. Valine is the mutated form and individuals homozygous for valine at the 129 residue are more susceptible to iatrogenic CJD. Valine has also been identified to have a greater b propensity than methionine, suggesting a causative mechanism for valine-129-homozygous individuals. Interestingly the b form can be locked by intermolecular association, thus supplying a plausible mechanism of propagation of a rare conformational state, although the precise sub-cellular localisation of PrP^sc remains controversial evidence does implicate late endosome-like organelles or lysosomes, (Arnold et al., (1995). This pH related conformational change might be relevant to the conditions that PrP^c would encounter within the cell after its internalisation during recycling. Parallels may be drawn to the diphtheria toxin, as reduction and acidification within the endosomal pathway is required for activation of the diphtheria toxin.

The studies on the conversion of PrP from the a to the b conformation (Kocisko et al., 1994 and Jackson et al., 1999) have shown that in the propagation, there is no involvement of the biosynthesis of new macromolecules. Also, it is consistent with PrP^sc being produced with a seeded polymerisation mechanism similar to that of amyloid fibril formation by other proteins, discussed later.

The exact mode of the replication of Prp^sc has not yet been proven. It is thought by some, to involve the conversion of Prp^sc from Prp^c by some kind of protein-protein interaction. The initial studies on mice involving transgenic Syrian hamster mice expressing chimeric forms of the human (Hu) PrP were by Telling et al. (1995). Their work involved the prion propagation using mice that expressed human and chimeric PrP transgenes. They found that Tg(HuPrP) mice were refractory to human prions unless crossed with PrP-null mice, and concluded that it was linked to a third component, namely protein X. Studies two years later by Kaneko et al. (1997) also provides evidence that there is another molecule, presumably a protein, involved in