. 32
( 45 .)


To carry out the procedure, a 1% Tris-borate agarose gel is made with two
sets of wells, several centimeters apart (upper and lower wells), and 12.5 pL of
each PCR (one quarter of the total) m loading buffer JS loaded mto each of a
pan consistmg of one upper well and one lower well. The aim is to get two
“coptes” of the PCR products on the gel. The DNA 1selectrophoresed into the
gel to a distance of l-2 cm and a Southern transfer (14) is carried out onto a
nylon membrane that will allow reprobing if necessary. The membrane then 1s
cut mto two pieces, each with its “copy” of the transferred PCR amphmers, and
each piece JS hybridized to a different allele-specific (AS) radioacttve probe.
Oligonucleotide hybridization with radioactively labeled AS oligonucleo-
tides (oligos) (16) gives good hybridtzation (detectedusing X-ray film) of an oligo
to its complementary sequencem homozygotesand heterozygotes.The olrgos are
15-mers with the single base mismatch m the middle, i.e., at nucleotide num-
ber 7 or 8 There JS little or no hybridization to a homozygote-mismatched
sequence if the washing temperature and salt concentration (sequence-depen-
dent) are optimized. This means that by making pairs of duplicated filters and
using an ohgo complementary to the Gln ,71sequence to probe one “copy” and
a second olrgo complementary to the Arg17t allele to probe the second “copy,”
an ammal can be genotyped m the followmg way The Gin,,, AS oligo ˜111
hybridize to Gln/Gln,,, DNA and to Arg/Gln,,, DNA but not to Arg/Argr,,
DNA. Therefore, an animal with DNA that hybridizes with only one AS oltgo
is a homozygote and an animal with DNA hybrrdrzmg to both ohgos 1sa het-
erozygote. It is not always possible to see the expected reduction m signal in
heterozygotes compared with homozygotes. Carrying out this procedure with
only one AS ohgo is unsatisfactory because the meaning of a negative result
will not be clear and using one membrane “copy” of the PCR products and
reprobing with the second ohgo is not ideal because of the difficulty of first
removing the hybridization from the first oligo.
Nonradtoactive labelmg methods also give good results but are less flexible
m terms of rewashing if, for some reason, the wash temperature is not exactly
right the first time. The membrane washing temperature depends on the AT
content of the oligethe more As and Ts in the sequence, the lower the wash
temperature. Examples of wash temperatures, e.g., with a 15-mer Gin,,, AS
oligo (S™TGGATWTATAGTA, with 66% AT) washes are m 2X SSC or
SSPE (14) at 37°C whereas the Arg,,, AS oligo (S™TGGATCGGTATAGTA,
with 60% AT) washes are at 4245°C. This method does not work very well
for the codon 136 position because of the sequence composition at this point
In this case, m the 15 bases around the single-base difference, the AT content
218 Hunter

IS <50% and wash temperatures are elevated to 55™C or more. This 1s much
more difficult to control and inconststent results are obtained.
When workmg with pairs of oligos rt is essential to be sure which membrane
“copy” is hybridized with which AS oligo. This step can be controlled by mak-
ing use of known homozygote DNA controls.
2.4. Problems Encountered in Genotyping
Problems encountered when using these techniques for genotypmg are
largely those with which any molecular biologist wt!! be familiar There are
some real mmefields, however, that relate to the mtrmsic problems of PCR and
the highly polymorphtc nature of the sheep PrP gene. The human gene is also
very polymorphic (2 7) although the cattle PrP gene is unusually invariant (I 8).
For PCR to work, tt 1s necessary for ohgo primers to bmd to complementary
sequences m the genomic DNA. If there happens to be an unsuspected sequence
polymorphtsm m the region of the genomic DNA covered by the oligo primer,
the PCR may not work properly and on heterozygote DNA may amplify only
one of the alleles. This will give a mtsleadmg result in subsequent digestion or
sequencing. Scientists all recognize the necessity of repeating results for con-
firmatton, but with the sheep PrP gene, repeat PCRs should be carried out usmg
primmg oligo pan-s from different parts of the sequence m order to mnumize
the risks of hidden mutations resulting m incorrect genotypmg. There are many
anecdotal stories of such problems.
Hidden sequence differences m the middle of the PCR amplrmer can also
lead to spurious results. One of these has already been mentioned. Digestion
with BspHI at codons 136 and 154 wt!! identify correctly the codon present
Failure to digest may result from any sequence change that alters the restrtction
site and may not mean that the expected alternate codon is present. If possrble,
the genotypes of a number of animals should be checked by sequencing.
Differential oligohybridtzatton also has its pitfalls. For example, there are three
codon 171 variants, Gin,,, and Arg,,,, which are described in the preceding
but there is also an allele that encodes His t7r (ref. IO; and Hunter, unpubltshed).
One of the discoveries of this polymorphism came about because of very low
levels of hybridization of the Gin,,, and Arg17i ohgos to a normal amount of
PCR product as seen on ethtdmm-stained gels. Sequencing revealed the unex-
pected finding of a third codon variant at this position. Screenmg by use of a
H˜s,˜, AS oligo (S™TGGATCATTATAGTA, 73% AT content, washes at room
temperature) works well and an example of this oligo used m conJunctton with
on three “copies” of a PCR reaction is shown in Fig. 2.
Gh and Arm
Finally, there 1s the perennial problem of contammation m the PCR reac-
tions. This possibtlrty must always be remembered and controlled. The best
defense against contammation as a source of error is obsessive paranoia and
Sheep PrP Gene and Scrapie

Fig. 2. Genotyping at codon 17 1 by AS oligo hybridization. (1) Eight sheep PrP
gene PCR amplimers, divided into three, on 1% TBE gels. (2) Filters made from (A)
Arg,,,; (C) His,,,. Genotypes are (left to right) Gln/Gln,
probed with (A) Gin,,,;(B)
Arg/His, Gin/His, Gln/Gln, Gin/His, Gin/His, Gln/Gln, Gln/Gln.

suspicion! Control PCR reactions containing all constituents except genomic
DNA should reveal any problems.
3. Use of Genotype Information
Once the genotype of an animal is assigned, what use can be made of it?
There are breed differences in genotypes of animals affected by natural scrapie.
Some breeds, for example Suffolks, do not encode PrP Val,,, alleles and yet
they do contract scrapie (15,19,2(J). In Suffolks, scrapie occurs in animals that
are Gln/Gln,,, . These animals are also Ala/Alals6, a genotype that in Cheviots,
Swaledales, Shetlands, and many other breeds would be expected to be resis-
tant to natural scrapie (20). It is, therefore, important to know the breed of any
tested sheep and whether it is Suffolk-like or Cheviot-like before deciding it is
scrapie resistant.
220 Hunter

Breedmg sheep for resistance to scrapte, by eltmmatton of the Val, ,,,Gln,,,
PrP allele, may now be posstble. However, because the etiology of scrapie is
unknown, it 1s not clear that this 1s a good idea m the long term Scrapte may
be the result stmply of a faulty gene-a genetic disease or tract-m which
case breedmg to reduce the frequency of disease-associated alleles of the PrP
gene should eradicate the dtsease. If scrapie is an infectious agent, however,
then breedmg for reststance to the common natural strains may simply allow
selection of rare scrapie mutant strains that could infect the so-called rests-
tant sheep.
Despite thts uncertainty, the association of sheep PrP genotype wrth scrapte
suscepttbtltty 1s now so well understood that the answer to the genetic disease
vs genetic suscepttbiltty debate 1s within reach at last

1 Parry, H B (1983) Scrapze Academic, London
2 Journals of the House of Commons (1754)
3 Morgan, K L , Nicholas, K , Glover, M J., and Hall, A P. (1990) A questionnaire
survey of the prevalence of scrapie m sheep in Britam Vet, Ret 127, 373-376
4 Wooldridge, M J A, Homville, L J , and Wilesmith, J W (1992) A scrapie
survey by postal questionnaire. alms, problems and results Proc Sac Vet Epzd
Prev Med, l-3 April 1992, pp 78-89
5 Hunter, N., Foster, J D , Goldmann, W., Stear, M , Hope, J , and Bostock, C
(1995) Natural scraple in a closed flock of Cheviot sheep occurs only in specific
PrP genotypes, in preparation.
6 Dickinson, A G and Outram, G. W (1988) Genetic aspects of unconventional
virus mfectlons* the basis of the vmno hypothesis, m Novel Infectlow Agents and
the Central Nervous System (Bock, J and Marsh, J , eds.), Ciba Foundation Sym-
posium vol 135 Wiley-Interscience, London, pp 63-83
7 Goldmann, W , Hunter, N , Benson, G , Foster, J D , and Hope, J (199 1) Dlffer-
ent scrapie-associated fibril proteins (PrP) are encoded by lures of sheep selected
for different alleles of the Sip gene. J Gen Vu-01 72, 2411-2417.
8 Laplanche, J L , Chatelam, J , Westaway, D., Thomas, S , Dussaucy, M , Brugere-
Picoux, J., et al (1993) PrP polymorphisms associated with natural scrapie dls-
covered by denaturing gradient gel electrophoresis. Genomzcs 15,3&37.
9. Goldmann, W , Hunter, N., Foster, J D , Salbaum, J M , Beyreuther, K., and
Hope, J. (1990) Two alleles of a neural protein gene lmked to scrapie in sheep.
Proc Nat1 Acad Scl USA 87,2476-2480
10 Belt, P B. G M , Muileman, 1. H , Schreuder, B , Bos-deRuiJter, J , Gielkens, A.
L J , and Smits, M A (1995) Identification of five allehc variants of the sheep
PrP gene and then association with natural scrapie J Gen Vzrol 76, 509-S 17
11 Goldmann, W , Hunter, N , Smith, G., Foster, J D , and Hope, J (1994) PrP geno-
type and agent effects m scrapie: change in allehc interaction with different ISO-
lates of agent m sheep, a natural host of scrapie. J Gen Vwol 75,989-995
SheepPrPGeneandScrapie 221

12 Collinge, J , Palmer, M S., and Dryden, A J (1991) Genetic predtsposmon to
ratrogenic Creutzfeldt-Jakob disease Lancet 337, 1441-1442.
13 McPherson, M J , Hames, B D., and Taylor, G R (1995) PCR2 A Practuxl
Approach IRL Press at Oxford Universtty Press, Oxford.
14 Sambrook, J , Frrtsch, E F., and Mantatts, T (1989) Molecular Clotung A Labo-
ratory Manual Cold Sprmg Harbor Laboratory, Cold Spring Harbor, NY
15 Westaway, D , Zuham, V , Mrrenda-Cooper, C , Da Costa, M , Neuman, S , Jenny,
A L , et al (1994) Homozygosrty for prton protem alleles encodmg glutamme-
17 1 renders sheep susceptrble to natural scrapte Gen. Dev. 8,959-969
16 Deslys, J. P., Marce, D , and Dormont, D. (1994) Similar genettc suscepttbtltty m
tatrogentc and sporadrc Creutzfeldt-Jakob drsease .I Gen. Vzrol 75,23-27
17 GaJdusek, D C (1994) Spontaneous generation of infectious nucleating amylords
m the transmrssible and nontransmtssrble cerebral amylordoses A4of Neurobiol
8, l-13
18 Hunter, N , Goldmann, W , Smith, G., and Hope, J (1994) Frequencies of PrP
gene vartants in BSE affected and healthy cattle rn Scotland Yet. Ret 135,40&403
19 Laplanche, J-L., Chatelam, J, Beaudry, P , Dussaucy, M., Bounneau, C , and
Launay, J -M (1993) French autochthonous scrapred sheep without the 136Val
PrP polymorphrsm Mum Gen 4,463-464
20 Hunter, N , Goldmann, W , Smrth, G , and Hope, J (1994) The assoctatton of a
codon 136 PrP gene varrant with the occurrence of natural scrapte Arch Vwol
Strain Typing Studies of Scrapie and BSE
Moira E. Bruce

1. Introduction
The basis of stram vartation m scrapie and other transmissible spongrform
encephalopathies is a crucial issue in the ongoing debate about the nature of
the mfectious agent. The clear evtdence for the existence of multiple strains
leads us to conclude that these agents carry some form of stram-specific infor-
mation that determmes the characteristics of the disease Any valid molecular
model must, therefore, include an informational component that can be repli-
cated in the infected host. Further, the behavior of strains when serially pas-
saged m different host species or genotypes hmtts the range of possrble models
for the informational component of the agent In thts chapter the methods used
for agent strain discrimmatton are described. The strategies used for the tsola-
tion and passagmg of dtfferent strains are presented and the tmplicattons of the
results of these simple experiments are discussed.
2. Agent Strain Discrimination
2.7. Disease Characteristics in Mice
Most research into strain variation in the spongtform encephalopathtes has
been cart-ted out in mouse models of these diseases.From long-term studies of
mice infected with many different isolates of scrapie, started by Alan Dtckmson
tn the late 196Os, tt has become clear that there are numerous strains of agent
that produce distinct patterns of disease m the infected host (1,2). Because
there are no serological tests and the molecular basts of strain variation IS not
yet known, strain discrimmation relies on simple measurements of disease
charactertsttcs. The major crtteria are the incubatton periods produced in mice
of defined genotypes and the severity and distribution of pathologtcal

From Methods m Molecular Medmne Pnon &eases
Edtted by H Baker and R M Rldley Humana Press Inc , Totowa, NJ

224 Bruce

changes seen in the brains of these mice. Using these criteria, about 20 pheno-
typically distinct strains of scrapie and BSE have been isolated by serial pas-
sage in mice
2.2. Incubation Period Measurement
The mcubation period is the interval between mmal mfection and a standard
clmical endpoint (3). For strain-typing studies of mouse-passaged isolates at
the Neuropathogenesis Unit m Edinburgh, mice are injected mtracerebrally
with 20 uL of a 1% homogenate of brain from a clmically affected donor.
Injected mice are coded and scored weekly for neurological signs, when they
are classified as “unaffected,” “ possibly affected” or “defimtely affected ” The
endpoint IS the date on which the animal receives an unambiguous score of
“defimtely affected.” Slightly different methods of chmcal assessment are used
m other laboratories.
For people outside the field, it is often difficult to appreciate the precision of
the mcubation period measurement. Inoculum contammg a high dose of a
single scrapie strain, injected intracerebrally mto genetically uniform mice,
will generally give a very tight grouping of incubation periods, with standard
errors of <2% of the mean For a single scrapie strain, the mcubation period also
is remarkably repeatable using inocula prepared from different brains and
remains constant over many serial passages in a single mouse genotype.

2.3. Effect of the Sine Gene on Incubation Period
The host Sznc gene (short for scrapie incubation) exerts a major influence on
the mcubation period of experimental scrapie and related diseases m mice The
action of the Smc gene was first recognized m mice infected with the ME7
strain of scrapte (3). Two alleles of this gene were identified, designated s7 and
p7, which gave, respectively, short orprolonged incubation periods with ME7.
Later it was shown that the Sznc gene controls the mcubation period of all other
strains of agent (1,4). It IS now clear that the Sine gene encodes PrP, Suzcs7 and
SzncP™ mice differ consistently m the sequence of the protein, by two ammo
acids (5,6).
For routme agent strain typing, an isolate is injected mtracerebrally mto
groups of C57BL (Szncs7), VM (Sznd”), and C57BL x VM F, (SzncS7P7) mice
Each scrapie or BSE strain has a characteristic and highly reproducible pattern
of mcubation periods m these three mouse genotypes (1,2) (see Fig. 1). Agent
strains differ:
1. In their incubation periods within a single Sznc genotype;
2. In their relative incubation periods in Szncs7and SzncP7homozygotes; and
3 In the apparent dominance characteristics shown by the two alleles in the F,
Strain Typing Studies of Scrapie and BSE 225

&K” passaged
22L l *A

139A l A l

79A 0 WA

ME7 l l A

22c * l A

301c m
l A

22F A* l
87A l l A


. 32
( 45 .)