to PrP 106-I 26 presented a typical apoptotic morphology, with condensation
of the chromatin and fragmentation of the nucleus. Agarose gel electrophoresis
of DNA extracted from cultured cells after 7-d treatment with the peptide
showed an apoptotic pattern of DNA fragmentation, resulting from cleavage of
nuclear DNA in internucleosomal regions (20).
272 Tagliavini et al.
Fig. 7. (A,B) Neurotoxicity of a synthetic peptide homologous to residues 106-I 26
of human PrP. Photomicrographs of primary hippocampal neurons exposed for 10 d to
a scrambled peptide (A) or to PrP 106-126 (B) at concentrations of 80 pM. (C,D)
Growth-promoting activity of peptide PrP 106-126 on glial cells. Photomicrographs
of primary astroglial cultures exposed for 14 d to a scrambled peptide (C) or to PrP 106-
126 (D) at concentrations of 50 @I. Following treatment, cultures were immunostained
with an antiserum to GFAP. Bars: (B), 200 pm; (D), 100 pm.
The prolonged exposure of primary astroglial cultures to the peptide PrP
106-l 26 resulted in a remarkable increase in size and density of astroglial pro-
cesses(Fig. 7C,D) (21). Conversely, no effects were observed following treat-
ment of cultures with the other PrP peptides. The hypertrophy of astrocyteswas
associatedwith a striking increase in glial tibrillary acidic protein (GFAP) tran-
scripts as revealed by Northern blot analysis. Densitometric quantitation of GFAP
mRNA normalized for the level of j3-actin messageshowed that this increment
was dependent on peptide concentration, being significant at 10 @4 and result-
ing in three- and fivefold increase above control values at 25 and 50 pM, respec-
tively. The rise of GFAP transcripts was accompanied by a substantial increase
in GFAP, as determined by Western blot analysis (21). The hypertrophy of
Prim Protein Arnylo˜cis 273
astrogltal cells was assoctatedwtth 1.5-fold increase in cell number at a pepttde
concentration of 50 PM. However, the proliferatton rate of astrocyteswas much
higher when cultures were kept m serum-free medium for the duratton of the
experiment. Under these condttions, the uptake of thymtdme after 9-d treatment
was 50 times htgher than the basal values (22). The prohferattve effect of PrP
106-I 26 was aboltshed by cotreatment of astroglial cultures with ntcardtpme, a
blocker of L-type voltage-sensttlve calctum channels. Mtcrofluorimetrtc analy-
sts of tntracellular calcium levels m single astrocytes showed that PrP 106-l 26
induced a rapid Increase in cytosohc calcium concentrattons, followed by a slow
return to basal levels. This effect was not observed with the scrambled pepttde. It
was absent when calcium was removed from the medium and was prevented by
premcubation of cultures with nicardlpine. These data suggest that PrP 106-l 26
stimulates astroglial proliferation via an increase in mtracellular calcium concen-
tratron, through the acttvation of L-type voltage-sensitive calcmm channels (22).
In summary, our studies showed that the amylotd protein in GSS disease is an
N- and C-terminal truncated fragment of PrP that originates from mutant mol-
ecules. This fragment contams a sequence(i.e., residues 106-126) that can adopt
different conformattons m distinct envtronments, although it has a high propen-
sity to form stable P-sheet structures. When synthesized as a peptlde, this
sequenceIS fibrillogemc and partially resistant to protease digestion; tt is toxrc to
neurons but has a growth-promotmg activity on glial cells. This sequence IS an
integral part of PrP pepttdes that accumulate m the CNS of patients with prion
diseases,and might be a major contributor to the molecular characteristtcs and
the pathogenic properties of disease-spectfic PrP isoforms and PrP amylotd.
5.1. Buffers Used in the Methods Described
1 Lysis buffer. 100 mM NaCI, 10 mM EDTA, 0.5% Nomdet P-40, 0 5% sodium
deoxycholate in 10 mM Trrs-HCl, pH 7.4.
2. Buffer A: 10 nnI4 Trrs-HCI, 150 mMNaC1,2 nuI4 EDTA, 2 mM EGTA, 0.4 mM
PMSF, 1% Trlton X-100, pH 7.5
3. Buffer B: 10 mMTris-HCI, 0.6MKI, 2 mMEDTA, 2 WEGTA, 0.4 WPMSF,
0.5% Trlton X-100, pH 7.5.
4. Buffer C: 10 n&I Tns-HCl, 1.5M KU, 2 mA4 EDTA, 2 m&I EGTA, 0.4 mM
PMSF, 0.5% Trlton X-100, pH 7 5.
5. Buffer D* 50 n-&Trts-HCI, 150 mMNaC1, pH 7 5.
6. Buffer E* 50 mM Tns-HC1, 10 mM CaC12, 3 mA4 NaN3, pH 7.5
7. Buffer F. 50 mk! Tris-HCI, pH 7.5
5.2. Antibodies to PrP
Polyclonal antibodies were raised in New Zealand white rabbits to synthetrc
peptrdes homologous to residues 23-40, 58-71, 90-102, 95-108, 127-147,
Taghavinr et al.
15l-1 65, and 220-23 1 of the ammo acid sequence deduced from human PrP
cDNA. Pepttde synthesis and purificatton were carried out as described m Sec-
tion 5.5. I. The peptides PrP 23-40, 90-102, 127-147, and 220-23 1 were
coupled to keyhole hmpet hemocyanm through a cysteme residue added at the
N-terminus (PrP 127-147 and 220-23 1) or at the C-terminus (PrP 23-40 and
90-102). The peptrdes PrP 58-71,95-108, and 15l-l 65 were directly synthe-
sized on a small nonimmunogemc core of branching lysme residues, which
provide a scaffoldmg to support multiple copies of the peptrde antigen (23).
Polyclonal antibodies were generated by multtple mtradermal and/or subcuta-
neous inJections of 100-200 ug pepttde at 2-3-wk intervals. The rabbits were
bled according to the evolutron of the antibody titer against the uncomugated
peptide evaluated by enzyme-linked immunosorbent assay (ELISA). Rabbit
antisera to synthettc peptides homologous to residues 15-40, 90-102, and
220-232 of hamster PrP (provided by S B. Prusmer) (24) as well as the mono-
clonal antibody 3F4 (provided by R. J Kascsak) (25) were also used for the
studies. The latter was obtained against Syrian hamster PrP 27-30 and Its
epttope corresponds to the sequence M-K-H-M, i.e., residues 108-l 11 of ham-
ster PrP and 109-l 12 of human PrP (26).
5.3. Analysis of Crude Brain Extracts
Brain tissues that had been frozen at -80Â°C at the time of autopsy were used
in the followmg protocol.
1 Homogenrzetissuein 9 vol of cold lysis buffer (27) andcentrifuge at 10,OOOg
2. Take ahquots of supernatanteach equivalent to 100 ug protein and digest wrth
protemaseK (20-100 pg/mL) for 1 h at 37Â°C
3 Termmate digestion by adding PMSF to a final concentrationof 3 mM
Selected protemase K-treated samples were subjected to enzymatic deglyco-
sylation with PNGase F (New England Biolabs, Beverly, MA) using reaction
conditions recommended by the manufacturer. Proteins were then fractionated
on 12.5% trrcine-SDS-polyacrylamide mimgels (tricme-SDS-PAGE) (28)
under reducing conditions, electrophoretically transferred to polyvinylidene
difluorrde membranes (Immobilon, Millipore, Bedford, MA) and probed with
anti-PrP antibodies The immunoreacttvity was visualized by enhanced
chemoluminescence (Amersham, Arlington Heights, IL). Specificity of the
reactions was checked by using normal rabbit or mouse sera as primary anti-
bodies and was confirmed by absorption of the PrP synthetic peptide antisera
with the relevant pepttdes. Each antiserum was incubated with 10 mA4 of the
relevant peptide at 37Â°C for 1 h and then at 4Â°C overnight. After centrifugatron
at lO,OOOg, supernatants were used as primary antibodies.
Prion Pro teln Amylolds 275
5.4. Characterization of GSS Amyloid Protein
5.4.1 Protocol for isolation of Amyloid Fibrils from Brain Tissue
Amyloid plaque cores were isolated from 20-40 g of cerebellum, cerebral
cortex, or basal ganglia that had been frozen at -80Â°C at the time of autopsy.
Patients were selected on the basis of the PRNP genotype, whereas the bram
regions were chosen based on a semiquantttative evaluation of the density of
amylold deposits m 7-pm thick paraffin sections lmmunostamed with anti-PrP
antibodies (see Section 5.4.7 j
Remove leptomenmges and large vessels and homogenize tissue with a Brmkmann
homogemzer m buffer A at a sample-to-buffer ratlo of 1 5 (w/v)
Sieve homogenate through a l-mm nylon mesh and centrifuge the filtrate at
10,OOOg for 20 min.
Rehomogemze pellet m buffer B and centrifuge at 10,OOOg for 20 mm, then
rehomogenize pellet m buffer C and centrifuge at 10,OOOg for 20 mm
Wash pellet three times m buffer D and centrifuge at 70,OOOg for 30 mm
Resuspendpellet m buffer E at a sample-to-buffer ratlo of 1 25 (w/v) and digest
with collagenase (Collagenase 188.8.131.52. Type I, Sigma, St LOUIS, MO) for 18 h
at 37â€™C using a 1â€™ 100 (w/v) ratlo of enzyme to pellet.
Centrifuge at 70,OOOg for 60 mm, resuspend pellet, and wash three times m buffer F
Load suspension on to a dlscontmuous sucrose gradlent (1, 1.2, 1 4, 1 7, and 2M
sucrose m 10 n&f Tns, pH 7 5) and centrifuge at 130,OOOg for 120 mm
Each interface was collected, washed, and pelleted three times m buffer F.
Aliquots of each pellet were assessedfor the presence of amylold fibrils by
polarized light microscopy after Congo red staining, fluorescence microscopy
after thloflavme S treatment, and electron microscopy after negative staining
with 2% aqueous phosphotungstlc acid or 5% uranyl acetate. Amyloid plaque
cores were recovered in the 2M sucrose fraction.
5.4.2. Protocol for Gel Filtration Chromatography
1 Suspend amyloid-enriched pellet m 99% formic acid and sonicate four times
for 20 s
2 Add 2 vol of distilled water and centrtfuge at 10,OOOg for 10 mm.
3. Apply supernatant to a cahbrated Sephadex G-100 column (1.2 x 120 cm), equ&
brated with 3M formic acid.
4. Pool protein peaks and concentrate with a speed vacuum concentrator.
The purity and molecular weight of the fractions were determined by trtcine-
SDS-PAGE using 12.5% polyacrylamide minigels under reducmg conditions
and by immunoblot analysis with the antisera to peptides PrP 23-40, 58-71,
90-l 02, 127-147, 15l-165, and 220-23 1 (1: 1,000) and the monoclonal anti-
body 3F4 (1:50,000).
276 Taghavinr et al
5.4 3. High-Performance liquid Chromatography
The fraction obtained by gel filtration that contained the major amylotd sub-
unit (as revealed by SDS-PAGE and tmmunoblot analysis) was further purified
by reverse-phase chromatography on a C4 column (214TP104, Vydac, Alltech,
Deerfield, IL) with a &80% linear gradtent of acetomtrtle contammg 0.1%
(v/v) trifluoroacetrc acid, pH 2.5. The column eluents were monitored at 214
nm and protem peaks were lyophllized. Followmg trrcme-SDS-PAGE and
rmmunoblot analysis, ahquots of the HPLC-purified, PrP-rmmunoreactrve pep-
tides were digested with endoprotemase Lys-C (see Section 5.4.4 ). The pep-
tides generated from enzymatic digests were separated by reverse-phase
chromatography on a Delta-Pak Cl 8 column (0.39 x 30 cm, Waters, Mrlford,
MA) with a O-70% lmear gradrent of acetonitrrle contammg 0.1% (v/v)
trrfluoroacettc acid, pH 2.5. The column eluents were monitored at 2 14 nm and
protein peaks were lyophrhzed.
5.4 4. Protocol for Enzymatic Digestion and Reduction of Peptldes
1 Dissolve HPLC-purified amylold peptides m 25 mMTns, 1 mM EDTA, pH 8 5
2 Incubate for 24 h at 37Â°C with endoprotemases Lys-C (Boehrmger, Mannherm,
Germany) at an enzyme-to-substrate ratlo of 1 30 (w/w)
3 Termmate proteolysls by rapid freezmg
Followmg fracttonatron of enzymattc dtgests by HPLC (see Section 5.4.3.),
ahquots of lyophrhzed peptrdes were dissolved m 5% acetic acid and reduced
with 0.72511/1drthrothreitol at 37Â°C for 30 h (29) The reduced pepttdes were
purified by HPLC on a reverse-phase C4 column.
5.4.5. Amino Acid Sequencing
Sequence analyses of the intact amylotd protein as well as of peptides gener-
ated by enzymatic digestion were carrted out on a 47714microsequencer and the
resulting phenylthiohydantoin ammo acid derrvattves were identified using the
on-lme 120A PTH analyzer and the standard program (Apphed Btosystems,
Foster Crty, CA).
5.4.6. Mass Spectrometry
C-termmal fragments of the amylotd protein generated by endoprotemase
Lys-C digestion were reduced, repurified by HPLC on a reverse-phase C4 col-
umn, and subjected to electrospray mass spectrometry (30). The analysis was
carried out using a VG BtoTech Bra-Q mass spectrometer with quadrupole
analyzer. The mass scale was calibrated with myoglobin.
Prion Protein Amyloids 277
Coronal sections of cerebral hemispheres as well as sections of cerebellum
and brain stem were fixed m 4% formaldehyde or Camoy solution (t.e., etha-
nol:chloroform:acetlc acid 6:3: 1) and embedded in paraplast.
For light microscopy, 7-ym thick serial sections were stained with Congo
red and thioflavme S, or incubated with the antisera to synthetic peptldes cor-
responding to residues 23-40,58-7 1,90-l 02,95-l 08, 127-147, 15l-l 65, and
220-23 1 of human PrP, or residues 15-40, 90-l 02, and 220-232 of hamster
PrP. The antisera were used at a dilution of 1*100/l :200 and were detected by
the peroxldase-antlperoxidase (PAP) method with swine-antn-abblt immune-
globulins, PAP complex (Dako, Santa Barbara, CA) and 3-3â€™-dlaminobenzldine
as chromogen. Before immunostammg, sections from formalm-fixed blocks
were treated with 98% formic acid at room temperature for 30 mm and/or sub-
jected to hydrolytic autoclaving as follows. The sectionswere immersed in 1 mM
hydrochloric acid in deionized water, autoclaved at 121â€œC for 10 mm, and cooled
to room temperature m tap water. In addition, selected sections from Camoy-
fixed blocks were digested with protemase K (25 pg/mL) at 37OC for 10 mm.
For electron microscopy, 5O+m thick paraplast-embedded sectlons were
deparaffmlzed m xylene and rehydrated through graded ethanol solutions.
Immunogold labeling with antl-PrP antisera (1:20) was performed by means of
goat-antirabblt mununoglobulins conjugated with 10 nm colloidal gold par-
tlcles (Blocell [Cardiff, UK], 1:20), followmg a pre-embedding procedure.
Speclficlty of immunoreactlons was checked by using normal rabbit serum as
primary antibody or by peptide absorptton.
5.5. Assembly and Conformation of Synthetic PrP Peptides
5.5.7. Synthesis and Purification of Peptides
The peptldes for investigation were selected on the basis of the predicted
secondary structure and the hydropathlc profile of human PrP. By the analysis
of these structural parameters, the 11 kDa amylold protein purified from GSS
198 was divided mto four peptides, corresponding to restdues 57-64,89-l 06,
106-126, and 127-147 of the ammo acid sequence deduced from human PrP
cDNA. PrP 57-64 (I.e., the octapeptide repeat) and PrP 89-106 (i.e., the
N-terminal region of PrP 27-30) are hydrophilic, whereas PrP 106-126 and
PrP 127-147 contain a sequence of hydrophobic amino acids flanked by
hydrophlhc residues. Scrambled sequences of these peptldes were also used
for the study.
Stepwise solid-phase peptide synthesiswas carried out on a 430A synthesizer
(Applied Blosystems) using 9-fluorenyl-methoxycarbonyl as the protective
group for aminic residues, and N-hydroxybenzotriazole, O-benzotnazol- 1-yl-
278 Taghavim et al.
N,N,Nâ€™,Nâ€™-tetramethyluronium hexafluorophosphate, and N,Nâ€™-dtcyclohexyl-
carbodnmtde as activators of carboxyhc residues. The pepttdes were purified
by reverse-phase HPLC. Identtty and purtty of all pepttdes were determmed by
ammo acid analysts using a Beckman (Fullerton, CA) 6300 analyzer and ammo
actd sequencing by automated Edman degradation.
5 5.2 Morphology and Staining Properties of Peptde Assemblies
To overcome solubthty problems, all pepttdes were dtssolved m 30% fornnc
acid or m the a-hehx stabihzmg solvent 1,1,1,3,3,3-hexafluoro-2-propanol at
a concentratton of 0.5, 1,2.5,5, and 10 mg/mL. The solutions were dialyzed at
room temperature for 24 h against delomzed water or phosphate-buffered
salme, pH 7.4. For light mtcroscopy, 50 pL of each suspension were an dried
on gelatin-coated shdes, stained with 0.2% Congo red m 80% ethanol saturated
with NaCl, and examined under polarized light, or treated with 1% aqueous
thtoflavme S and observed with fluorescent light (excttatton filter 4 1O-490 nm,
barrier filter 520 nm). For electron mtcroscopy, 5 c(L of each suspension were
applied to formvar-coated nickel grids, negatively stained with 5% uranyl ace-
tate, and observed m a Philips EM 410 at 80 kV
5 5.3 X-Ray Diffract/on Analysis
X-ray dtffraction was performed on pepttdes that proved to be fibrtllogemc
at electron mtcroscopy exammation. Peptldes were analyzed as a lyophthzed
powder, allowed to sediment by slow dehydration after suspension m distilled
water, or solubthzed m either 30% formic acid or 1,l 91 3 33 73-hexafluoro-2-
propanol prior to placement mto 0.8-mm diameter sthcomzed Lmdeman glass
capillary tubes. Dtffractton patterns were obtained after exposure to mckel-
filtered CuKa radiation from a sealed X-ray tube generator source operated at
40 kV, 16 mA, with exposure times of l-2 h. Spacings were measured directly
on CEA Reflex 25 X-ray films, using a specimen-to-film distance of 7.5 cm