Báo cáo khoa học: Ligand binding promotes prion protein aggregation – role of the octapeptide repeats potx

Here, we report our findings that GAG and Cu2+ promote the aggregation of recombinant human PrP rPrP.. There-fore, the octapeptide-repeat region is critical in the aggregation of rPrP, ir

of the octapeptide repeats

Shuiliang Yu1, Shaoman Yin1, Nancy Pham2, Poki Wong1, Shin-Chung Kang1, Robert B Petersen1, Chaoyang Li1and Man-Sun Sy1

1 Department of Pathology, Case Western Reserve University, Cleveland, OH, USA

2 Cleveland Clinic Lerner College of Medicine of Case Western Reserve University, Cleveland, OH, USA

Prion diseases are a group of fatal neurodegenerative

disease in humans and animals It is believed that all

prion diseases are caused by the conversion of a

nor-mal cellular prion protein (PrPC) to a pathogenic and

infectious isoform, commonly referred to as scrapie

prion (PrPSc) or proteinase-resistant prion (PrPRES) [1]

The majority of human prion diseases are sporadic,

and the cause of the disease is not known A small

number of prion diseases, such as Kuru, iatrogenic

Creutzfeldt–Jacob disease and variant Creutzfeldt–

Jacob disease are contracted through an infectious

mechanism By contrast, familial or inherited human

prion disease, which accounts for 10–15% of human prion diseases, is the result of mutations in the germ-line prion protein gene, PRNP More than 30 different pathogenic mutations in human PRNP have been iden-tified [2,3] These mutations are either insertional or point mutations The insertion mutations occur solely

in the octapeptide-repeat region; wild-type human PrP has five octapeptide repeats The number of pathogenic insertions ranges from two to nine However, point mutations occur along the entire PrP molecule, but tend to cluster in the C-terminal globular domain It is thought that the mutant prion protein is inherently

Keywords

aggregation; copper; glycosaminoglycan;

octapeptide repeat; prion

Correspondence

M.-S Sy, Room 5131 Wolstein Research

Bldg, School of Medicine, Case Western

Reserve University, 2103 Cornell Road,

Cleveland, OH 44106-7288, USA

Fax: +1 216 368 1357

Tel: +1 216 368 1268

E-mail: man-sun.sy@case.edu

(Received 7 July 2008, revised 18 August

2008, accepted 10 Sepember 2008)

doi:10.1111/j.1742-4658.2008.06680.x

Aggregation of the normal cellular prion protein, PrP, is important in the pathogenesis of prion disease PrP binds glycosaminoglycan (GAG) and divalent cations, such as Cu2+and Zn2+ Here, we report our findings that GAG and Cu2+ promote the aggregation of recombinant human PrP (rPrP) The normal cellular prion protein has five octapeptide repeats In the presence of either GAG or Cu2+, mutant rPrPs with eight or ten octa-peptide repeats are more aggregation prone, exhibit faster kinetics and form larger aggregates than wild-type PrP When the GAG-binding motif, KKRPK, is deleted the effect of GAG but not that of Cu2+ is abolished

By contrast, when the Cu2+-binding motif, the octapeptide-repeat region,

is deleted, neither GAG nor Cu2+ is able to promote aggregation There-fore, the octapeptide-repeat region is critical in the aggregation of rPrP, irrespective of the promoting ligand Furthermore, aggregation of rPrP in the presence of GAG is blocked with anti-PrP mAbs, whereas none of the tested anti-PrP mAbs block Cu2+-promoted aggregation However, a mAb that is specific for an epitope at the N-terminus enhances aggregation in the presence of either GAG or Cu2+ Therefore, although binding of either GAG or Cu2+ promotes the aggregation of rPrP, their aggregation pro-cesses are different, suggesting multiple pathways of rPrP aggregation

Abbreviations

GAG, glycosaminoglycan; PBST, NaCl ⁄ P i ⁄ 0.05% Tween; PrP, prion protein; PrP C , normal cellular form of PrP; PrP Sc , the infectious and pathogenic scrapie PrP; rPrP, recombinant wild-type PrP; rPrP D51-90 , recombinant PrP with deletion of octapeptide-repeat region; rPrP DKKRPK , recombinant PrP with deletion of GAG binding motif, KKRPK at the beginning of the N-terminal; rPrP10OR, recombinant PrP with 10 octapeptide-repeats; rPrP 8OR , recombinant PrP with 8 octapeptide-repeats.

unstable, prone to misfold and aggregate, forming a

structure which acts as a ‘seed’ to recruit additional

mutant proteins, eventually leading to the formation

of pathogenic and infectious PrPSc[4]

Recombinant bacterial-produced wild-type PrP, rPrP

and rPrP with pathogenic mutations have been used

extensively as model systems for studying the

conver-sion processes [5] Some mutant rPrPs have been

shown to acquire certain physical characteristics

simi-lar to PrPSc, such as the content of b-sheet structure,

partial resistance to proteinase K and a propensity to

aggregate [6–8] However, the mechanisms leading to

these changes are not completely understood

Biophy-sical studies suggest that thermo-instability is not the

major contributing factor in the conversion process [9]

Accumulated in vivo and in vitro evidence suggest that

the conversion process may require the participation of

other proteins, such as ‘protein X’ or non-protein

mac-romolecules, such as nucleic acids, glycosaminoglycans,

lipids or divalent cations [1,10]

Recently, we found that rPrP with a pathogenic

mutation of three additional insertions, rPrP8OR, has a

more exposed N-terminus, binds better to

glycosami-noglycans (GAGs) and is more susceptible to oxidative

attack than wild-type rPrP The aberrant properties

associated with rPrP8OR are also observed in another

insertion mutant prion protein with five extra repeats,

rPrP10OR; the aberrations are even more profound in

rPrP10OR [11] In addition, we also found that under

denaturing conditions and low pH, the insertion

mutant proteins are more prone to aggregate, and the

degree and kinetics of aggregation are proportional to

the number of inserts [12]

Here we report further studies on the consequences

of binding of GAG and Cu2+to rPrPs We found that

both GAG and Cu2+promote the aggregation of rPrP

in proportion to the number of inserts Furthermore,

we found that the octapeptide-repeat region is critical

for rPrP aggregation irrespective of whether

aggrega-tion is promoted by GAG or Cu2+ Blocking with

anti-PrP mAb revealed that GAG and Cu2+ promote

the aggregation of rPrP differently Because

aggrega-tion is an essential step in PrPC to PrPSc conversion,

the significance of these findings with respect to the

pathogenesis of inherited human prion disease is

dis-cussed

Results

Enhancement of rPrP aggregation with GAG

We previously reported that insertion mutant rPrPs

such as rPrP8OR and rPrP10OR bind much better to

GAG than rPrP [11] Furthermore, the level of GAG binding is proportional to the number of inserts [11]

We also showed that at low pH, for example pH 4.0, rPrPs aggregate spontaneously, again proportional to the number of inserts [12] We therefore investigated whether heparin, a GAG, promotes the aggregation of rPrPs, and whether the degree of enhancement is pro-portional to the number of inserts These experiments were carried out in NaCl⁄ Pi at pH 7.4, with low con-centrations of rPrPs and GAG; conditions that are more physiological At pH 7.4, heparin enhances the aggregation of all three rPrPs, and the enhancement is greatest for rPrP10OR followed by rPrP8OR and then rPrP (Fig 1) Heparin does not promote the aggre-gation of rPrPDKKRPK, which lacks the GAG-binding motif, KKRPK, the first five amino acids at the

Fig 1 Aggregation of rPrP is enhanced by heparin (A) Comparison

of the heparin-enhanced aggregations of rPrP, rPrP 8OR , rPrP 10OR and rPrP DKKRPK rPrPs (1 l M ) were mixed with various concentra-tions of heparin in NaCl ⁄ P i (pH 7.4) at 25 C, and A 405 was recorded 300 s after mixing The results are means ± SEM for three experiments (B) Kinetics of the heparin-enhanced aggrega-tions of rPrP, rPrP8OR, rPrP10ORand rPrP DKKRPK rPrPs (1 l M ) were mixed with 1 lgÆmL)1heparin in NaCl ⁄ P i at 25 C, and A 405 was monitored as described in the Experimental Procedures The enhanced aggregation is given here as an increased percentage of starting turbidities [P = (T ⁄ T 0 )1) · 100; P, percentage increase; T, turbidity; T0, starting turbidity] All experiments were carried out at least three times with different batches of rPrPs.

N-terminus These findings provide the first evidence

that enhanced GAG binding has biological

conse-quences on insertion mutant proteins, allowing them

to bind GAG better, which then facilitates aggregate

formation

Because commercially purchased heparin is

heteroge-neous in its molecular mass, we next investigated

whether heparin with a defined molecular mass of

3 kDa, which contains nine sugar residues, also

pro-motes rPrPs aggregation We obtained similar results

with this low molecular mass heparin However,

hepa-rin with only two sugar residues did not promote

aggregation, indicating that a minimal size is required

for aggregate promotion (Fig 2A,B)

We next used an ELISA to determine whether the

aggregates contain GAG A biotinylated GAG was

used to promote the aggregation of rPrP or

rPrPDKKRPK After aggregation, the rPrP aggregates

were collected by repeated centrifugation and washing

Aggregates were then resuspended, diluted in various

amounts of NaCl⁄ Pi and added to individual ELISA

wells, which had been pre-coated with an anti-PrP

mAb, 11G5, to capture the rPrP An

avidin-conju-gated enzyme was then added to the wells to detect

bound biotinylated GAG Much stronger

immunore-activity is detected in samples containing rPrP than

rPrPDKKRPK, which cannot bind GAG (Fig 2C)

These results suggest that rPrP aggregates indeed

contain GAG

Sucrose-gradient centrifugation of rPrP–GAG

aggregates

We used sucrose-gradient centrifugation to compare

the relative sizes of rPrP–GAG and PrP10OR–GAG

aggregates rPrPDKKRPK was used as a control rPrP–

GAG aggregates and controls (without GAG) were

centrifuged on 5–50% sucrose gradients Ten fractions

from each gradient were collected, run on 12%

SDS⁄ PAGE and immunoblotted with mAb 8H4

Without GAG, rPrP, rPrPDKKRPK and rPrP10OR were

detected in the upper fractions (Fig 3) By contrast,

when mixed with 3 kDa GAG, rPrP immunoreactivity

is detected in all fractions, with the bottom fractions

containing most immunoreactivity These results

sug-gest that rPrP–GAG aggregates exist in different sizes

By contrast, when rPrP10OR is mixed with 3 kDa

GAG, all the immunoreactivity is detected in the

bot-tom fraction Therefore, rPrP10OR forms much larger

aggregates than wild-type rPrP In rPrPDKKRPK, which

does not bind GAG, when mixed with GAG and

centrifuged under identical conditions, all the

immuno-reactivity remained on the top of the gradient

Fig 2 Characterization of the heparin enhanced aggregation of rPrP (A) Comparison of the aggregation of rPrP enhanced by hepa-rin, low molecular mass heparin (LMW hepahepa-rin, 3 kDa) and heparin disaccharide rPrP (5 l M ) was mixed with various concentrations of heparin, LMW heparin or heparin disaccharide, respectively in NaCl ⁄ P i at 25 C, and the A 405 was recorded 300 s after mixing The results are means ± SEM for three experiments (B) Kinetics

of the aggregation of rPrP enhanced by heparin, LMW heparin and heparin disaccharide rPrP (5 l M ) was mixed with 10 lgÆmL)1 of heparin, LMW heparin or heparin disaccharide respectively in NaCl ⁄ P i at 25 C, and the A 405 was monitored as described in the text (C) Detection of biotinylated heparin in the aggregates of rPrP rPrP (5 l M) was mixed with 10 lgÆmL)1 biotinylated heparin in NaCl ⁄ P i and the aggregates were harvested by centrifugation at

13 000 g for 10 min The pellet was washed with NaCl ⁄ P i three times and dissolved in NaCl ⁄ P i containing 0.1% Triton X-100 as described in the text Various dilutions of the resolved aggregate solution were incubated with mAb 11G5 pre-coated plates and the biotinylated heparin, which bound in the aggregates was detected using horseradish peroxidase–streptavidin.

Enhancement of rPrP aggregation by Cu2+or

Zn2+but not Mg2+or Mn2+

rPrP binds divalent cations such as Cu2+ and Zn2+

[13,14] We next determined whether Cu2+ or Zn2+

influences the aggregation of rPrP, rPrP8OR and

rPrP10OR At low pH, neither Cu2+nor Zn2+has any effect on the aggregation of rPrP (not shown) The failure of these cations to modulate rPrP aggregation

is most likely due to the effects of pH on the octapep-tide repeat, rendering it unable to bind divalent cations [15] However, when the aggregation assay was carried out at pH 7.4, Cu2+ and Zn2+, but not Mg2+ or

Mn2+, promote rPrP aggregation in a concentration-dependent manner (Fig 4A–D) Again, the levels of enhancement are proportional to the number of inserts These results are in good accord with earlier findings that PrP binds Cu2+and Zn2+but not Mg2+

or Mn2+[13,16]

Furthermore, although the KKRPK deletion mutant was totally unable to form aggregates in the presence

of heparin, in the presence of Cu2+, the KKRPK dele-tion mutant behaved identically to wild-type rPrP (Fig 5A,B) In addition to the octapeptide-repeat region, two additional Cu2+-binding sites have been identified in the rPrP C-terminal globular domain [16,17] To investigate whether the octapeptide-repeat region is important in Cu2+-induced aggregation, we deleted the octapeptide-repeat region and created rPrPD51-90 In contrast to wild-type rPrP, Cu2+ does not promote the aggregation of rPrPD51-90(Fig 5C,D) Therefore, the octapeptide-repeat region is the critical motif that mediates Cu2+-induced rPrP aggregation Unexpectedly, GAG also failed to promote the aggregation of rPrPD51-90 (Fig 6A) This deficit is not because rPrPD51-90 does not bind GAG rPrPD51-90 does bind GAG albeit with lower avidity (Fig 6B)

Fig 3 Sucrose-gradient centrifugation of rPrP–GAG aggregates.

rPrP (1 l M ) was mixed with 5 lgÆmL)1low molecular mass heparin

(3 kDa) in NaCl ⁄ P i and incubated at 25 C for 30 min The mixture

was loaded on to a 5–50% sucrose gradient and centrifuged at

4 C, 100 000 g for 2 h Ten fractions were drawn from top to

bot-tom An equal volume of each fraction was loaded onto a 12%

SDS ⁄ PAGE and PrPs were detected by immunoblotting with mAb

8H4.

Fig 4 Aggregation of rPrPs is enhanced

by metal ions One micromole rPrP, rPrP8OR

or rPrP 10OR was mixed with various

concen-trations of CuCl2(A), ZnCl2(B), MnCl2(C)

and MgCl 2 (D) respectively in NaCl ⁄ P i , and

A405was recorded 300 s after mixing The

results are means ± SEM of at least three

experiments All the enhanced aggregation

are given here as an increased percentage

of starting turbidities [P = (T ⁄ T 0 )1) · 100;

P, percentage increase; T, turbidity; T 0 ,

starting turbidity].

At higher protein concentrations, rPrPD51-90 and rPrP

have comparable GAG-binding activity (Fig 6B)

These results suggest that the octapeptide-repeat

region is the nucleation center of rPrP aggregation,

irrespective of whether aggregation is initiated with

GAG or Cu2+ Furthermore, these results also

pro-vide strong epro-vidence that although the KKRPK motif

is the GAG-binding site, the octapeptide-repeat region

also contributes to the total affinity between PrP and

GAG Although Cu2+ and GAG bind to different

sites on PrP, we did not observe a synergistic effect

when both Cu2+ and GAG were added to the rPrPs

(results not shown)

Sucrose-gradient centrifugation of rPrP–Cu2+

aggregates

We also used sucrose-gradient centrifugation to

com-pare the relative sizes of rPrP–Cu2+ and PrP10OR–

Cu2+ aggregates rPrPD51-90was used as a control As

expected, without Cu2+, rPrP, rPrPD51-90 and

rPrP10OR were detected in the top fractions (Fig 7)

By contrast, when rPrP is mixed with Cu2+, most of

the PrP immunoreactivity is detected in the bottom

fractions However, upon longer exposure, PrP

immu-noreactivity is also present in the intermediate

frac-tions (not shown) By contrast, when rPrP10OR is

mixed with Cu2+, all the immunoreactivity is detected

in the bottom fraction Therefore, rPrP10OR also

forms much larger aggregates than wild-type rPrP

rPrPD51-90, which does not bind Cu2+, remained on

the top of the gradient

Modulation of GAG- or Cu2+-promoted aggregation of rPrP10ORwith anti-PrP mAbs

We next investigated whether GAG- or Cu2+ -pro-moted aggregation of rPrP10OR can be amended with anti-PrP mAbs The epitopes of these mAbs are dia-grammatically presented in Fig 8A Of all the anti-PrP mAbs tested, one, 8B4, consistently enhanced the aggregation of rPrP10ORin the presence of either GAG

or Cu2+(Fig 8B,C) mAb 8B4 alone does not induce the aggregation of rPrP10ORwithout the PrP ligands Four mAbs, SAF32, 11G5, 7A12 and 8H4 consis-tently blocked the aggregation of rPrP10ORin the pres-ence of GAG (Fig 8B) However, none of the tested mAb was able to block the effects of Cu2+ (Fig 8C) The inability of these mAbs to block Cu2+ induced aggregation is not because Cu2+ prevents the binding

of these mAbs as shown by ELISA; Cu2+ does not inhibit the binding of these mAbs to rPrP (results not shown)

Discussion

Aggregation of PrP is an essential step in the conver-sion of PrP to PrPSc [1] Here we describe four new findings on the aggregation of rPrPs: (a) in the pres-ence of PrP ligands, such as GAG or the divalent cation Cu2+, rPrPs aggregate in proportion to the number of octapeptide inserts, thus rPrPs with inser-tional mutations, such as rPrP8OR and rPrP10OR form more and larger aggregates with faster kinetics than wild-type rPrP; (b) whereas GAG-induced aggregation

Fig 5 Copper enhances aggregation of rPrP and rPrP DKKRPK , but not rPrP D51-90 Various dilutions of CuCl2were mixed with rPrP (A), rPrP DKKRPK (B) and rPrP D51-90 (C), respectively in NaCl ⁄ P i and A 405 was recorded 300 s after mixing (D) A compari-son of the aggregation of rPrP, rPrP DKKRPK and rPrP D51-90 enhanced by 50 l M CuCl2in NaCl ⁄ P i All the enhanced aggregation are given here as an increased percentage of starting turbidities [P = (T ⁄ T 0 )1) · 100; P, percentage increase; T, turbidity; T 0 , starting turbidity] And the results are means ± SEM for at least three experiments.

requires the GAG-binding motif, Cu2+-induced

aggre-gation requires the octapeptide repeat; (c) the

octapep-tide-repeat region is essential for both GAG- and

Cu2+-promoted rPrP aggregation; (d) aggregation

induced by GAG and Cu2+ share common features,

yet each one has its own unique features, suggesting

multiple pathways leading to rPrP aggregation

Bacterial produced rPrP has been used extensively as

a model system for studying the aggregation process

[5] In previous studies, aggregation of rPrP required

denaturation, low pH and relatively high

concentra-tions of rPrP [18–22] In this study, aggregation of

rPrP was carried out at pH 7.4 and with relatively low

concentrations of full-length rPrP; these conditions are

physiologically more relevant Accumulated evidence

suggests that binding of GAG may be important in

the pathogenesis of prion diseases [23–27] PrPSc

parti-cles formed in vivo contain GAG [28] In vitro, GAG facilitates the conversion of PrP to PrPSc [24], and greatly increases the infectivity of non-aggregated PrPres [25] Reduction of cellular GAG significantly decreases the biogenesis of PrPSc in scrapie-infected cells [29] Cell-surface GAG has also been reported to

be the receptor for PrPSc [23,27] However, exogenous GAG and GAG analogs, such as low molecular mass heparin, suramin, pentosan polysulfate and dextran sulfate can inhibit PrPSc formation in cells, and pro-long the incubation time of experimental prion diseases [10] It has been postulated that exogenous GAG and GAG analogs block PrPSc formation by competing with the endogenous GAG which is critical for PrPSc generation [10]

GAG may function as a scaffold for concentrating PrP, creating a reservoir of PrP for conversion We reported earlier that rPrP8ORand rPrP10ORbind GAG better than rPrP, and the level of binding is propor-tional to the number of inserts [11] Our current find-ings that GAG also promotes the aggregation of rPrP8OR and rPrP10OR proportional to the number of inserts are in good accord with our earlier results Enhancement of rPrP aggregation is most apparent when the concentration of rPrP is low, such as 1 lm

At this concentration, rPrP by itself does not aggre-gate A small GAG, with nine sugar residues is as

Fig 6 Heparin enhances aggregation of rPrP, but not rPrPDKKRPK

and rPrP D51-90 (A) rPrPs (1, 3, 5 l M ) were mixed with 5 lgÆmL)1

heparin in NaCl ⁄ P i and A 405 was measured 300 s after mixing All

the enhanced aggregation are given here as an increased

percent-age of starting turbidities [P = (T ⁄ T 0 )1) · 100; P, percentage

increase; T, turbidity; T0, starting turbidity] The results herein are

means ± SEM for three experiments (B) Detection of rPrP D51-90

binding to heparin by ELISA Heparin (10 lgÆmL)1) was coated onto

plates at 4 C overnight and blocked with 3% BSA BSA was

coated as a control Different concentrations of rPrP, rPrP DKKRPK or

rPrP D51-90 were incubated with the plates for 2 h at 25 C After

three washes with PBST, appropriate dilution of mAb 8H4 was

used to detect the bound rPrP The results are means ± SEM for

three wells and this experiment was repeated at least three times.

Fig 7 Sucrose-gradient centrifugation of rPrP–Cu 2+ aggregates rPrPs (1 l M ) was mixed with 20 l M CuCl 2 in NaCl ⁄ P i and incubated

at 25 C for 30 min The mixture was loaded on top of a 5–50% sucrose gradient and centrifuged at 4 C, 100 000 g for 2 h Ten fractions were drawn from top to bottom An equal volume of each fraction was loaded onto 12% SDS ⁄ PAGE and the PrPs were detected by immunoblotting with mAb 8H4.

effective as larger GAG in promoting rPrP

aggrega-tion However, a disaccharide of GAG is unable to

cause aggregation, suggesting that the minimum unit

of GAG required for rPrP aggregation is between

three and nine sugar residues The promotion of

aggre-gation by GAG is not only limited to rPrP with

inser-tion mutainser-tions GAG also promotes the aggregainser-tion

of rPrPs with pathogenic point mutations, albeit at

lower levels [30] We hypothesize that enhanced

bind-ing to GAG, leadbind-ing to aggregation is a common

feature in inherited human prion disease

The precise mechanism by which GAG promotes

rPrP aggregation is not known GAG may promote

aggregation by serving as a scaffold If this is the case,

the rPrP aggregates should contain GAG

Alterna-tively, GAG may simply serve as a platform for rPrPs

to be physically close to each other, resulting in

aggre-gation between rPrPs, without including GAG Our

ELISA results suggest that some rPrP–GAG

aggre-gates contain GAG However, our sucrose-gradient

centrifugation experiments revealed that rPrP–GAG

aggregates exist in many different sizes Because the GAG used in these experiments has a molecular mass

of 3 kDa, it is probable that some of the larger rPrP– GAG aggregates are composed mainly of rPrP Thus, GAG serves as a scaffold as well as a platform in facil-itating rPrP aggregation In contrast to rPrP, when mixed with 3 kDa GAG, all the rPrP10OR is detected

in the bottom fraction of the sucrose gradient This is

in good accordance with our earlier finding that under denaturing and low pH condition; rPrP10OR has the propensity to spontaneously aggregate, in a protein concentration-dependent manner When incubated with GAG, rPrP10OR is concentrated, thus able to form much larger aggregates It is interesting to note that in PrPSc infected mouse brain homogenate centri-fuged under identical conditions, most of the PrP immunoreactivity is present in the bottom fractions of the sucrose gradient [31] However, in contrast to

in vivo-derived PrPSc aggregates, the rPrP aggregates formed in the presence of GAG are PK sensitive (results not shown)

Fig 8 Blocking of rPrP10ORaggregation enhanced by heparin or copper using anti-PrP mAbs (A) The location of mAb-binding epitopes along the length of PrP LS, leader sequence; GPI, glycosylphosphatidylinositol anchor (B) Blocking of the aggregation of rPrP 10OR enhanced by heparin rPrP 10OR (1 l M ) was mixed with 1 lgÆmL)1heparin and 0.125 l M mAbs in NaCl ⁄ P i and A405 was monitored as described in text NS mAb, non-specific mAb (C) Blocking of the aggregation of rPrP 10OR enhanced by cop-per rPrP 10OR (1 l M ) was mixed with 20 l M

CuCl 2 and 0.125 l M mAbs in NaCl ⁄ P i and

A405was recorded The aggregation is given

as an increased percentage of starting tur-bidities [P = (T ⁄ T 0 )1) · 100; P, percentage increase; T, turbidity; T0, starting turbidity] The two experiments were repeated at least three times.

rPrP binds divalent cations, such as Cu2+and Zn2+

but not Mg2+or Mn2+[32] A metal imbalance in the

central nervous system has been speculated to play a

role in neurodegenerative diseases, including prion

dis-ease [32] However, the physiological significance of

the interaction between PrP and Cu2+ remains poorly

understood Some studies found that Cu2+ causes

aggregation of rPrP [33–35] Others reported that

Cu2+ inhibits rPrP conversion to amyloid [36,37]

Copper chelators also inhibit PrPSc replication in vitro

[38] Some studies suggest that treatment with Cu2+

causes PrP to acquire PK resistance [34–36,39,40]

However, this interpretation is complicated by the

recent finding that Cu2+ inhibits proteinase K activity

[41]

We found that at neutral pH and low concentrations

of rPrPs, Cu2+ and Zn2+ but not Mg2+ and Mn2+

promote aggregation of rPrP, rPrP8ORand rPrP10ORin

a concentration-dependent manner For rPrP8OR and

rPrP10OR, the enhancement can be observed in as low

as 1 lm of Cu2+ or Zn2+, a concentration that is

physiologically relevant [42] Again the degree of

enhancement is proportional to the number of

octa-peptide repeats, and Cu2+is consistently more efficient

in promoting aggregation than Zn2+ Cu2+ does not

promote the aggregation of rPrPD51-90, which lacks the

repeat region Therefore, the

octapeptide-repeat region is important in rPrP aggregation This

finding is consistent with an earlier report suggesting

that the octapeptide-repeat region constitutes a

pH-dependent folding and aggregation site of PrP [22]

Our result is also consistent with another study

show-ing that when Cu2+ binds to the octapeptide-repeat

region, it serves as a ‘copper switch’, which is

impor-tant in PrP aggregation [43] However, we were

sur-prised to find that GAG was also unable to promote

the aggregation of rPrPD51-90, because the

octapeptide-repeat region is not required for the binding of GAG

Furthermore, under high rPrPD51-90concentration, low

pH and denaturing conditions, rPrPD51-90also failed to

aggregate spontaneously (results not shown)

There-fore, the octapeptide-repeat region is critical for rPrP

aggregation irrespective of whether aggregation is

ligand initiated or spontaneous It should be noted

that others have identified additional Cu2+-binding

sites at the C-terminus of PrP [16] It is possible that

these binding sites may not be essential for PrP

aggre-gation

The precise mechanisms by which the divalent

cations promote aggregation are not known Cu2+and

Zn2+can bind PrP intramolecularly as well as

molecularly [44] We speculate that it is the

inter-molecular binding of Cu2+ that enhances aggregation

Presumably, by having more octapeptide repeats, rPrP10OR is more readily to interact with Cu2+ and

Zn2+ The failure of either Mg2+or Mn2+to enhance aggregation provides the most appropriate control for the specificity of the interactions This interpretation is also supported by results from the sucrose gradient centrifugation experiments

It has been reported that GAG promotes the aggre-gation of rPrP and that the aggregate is stabilized by the binding of Cu2+[26] However, we did not observe

a synergistic effect between GAG and Cu2+ in our aggregation assay (results not shown) It should be noted that in our assay the concentrations of rPrPs (1 versus 4 lm), GAG (0.1 versus 2 lm) as well as Cu2+ (1–20 versus 500 lm) were much lower than is typically used in this type of experiments Furthermore, our assay only detects the amount of aggregate that is gen-erated rather than the stability of the aggregate Hence, it is possible that the aggregate formed with GAG alone is different from the aggregate formed in the presence of high concentrations of GAG and

Cu2+

We reported earlier that under low pH and denatur-ing conditions, only mAbs which react with an epitope

in the octapeptide-repeat region and the helix 1 region respectively, block the spontaneous aggregation of rPrPs [12] In the current study, we found that mAb 8B4, which reacts with an epitope at the N-terminus further promotes GAG- and Cu2+-induced rPrP aggregation We suggest that mAb 8B4 is able to align the rPrP in the same orientation, in parallel, pairing the N-terminus of two PrPs, which then facilitates the binding of either GAG or Cu2+ It should be noted that mAb 8B4 does not cause the aggregation of rPrP without the participation of either GAG or Cu2+

In addition to mAbs that are specific for the octa-peptide-repeat region, such as SAF32, other mAbs, such as 7A12, 11G5 and 8H4 also blocked GAG-induced aggregation These results suggest that the entire C-globular domain including the helix 1, b2 and helix 2 regions are all important in the aggregation process We could not evaluate whether mAb 8H4 inhibits spontaneous aggregation of rPrP because mAb 8H4 does not bind PrP at pH 4.0 This observation is

in good accordance with a recent study suggesting that the opening of the helix 1 region, followed by confor-mational changes in helix 2 of rPrP, is critical in rPrP aggregation [45] Finally, we showed that mAb 8F9, which reacts with an epitope at the C-terminal end, does not block GAG-induced aggregation These results suggest that in the presence of GAG, aggrega-tion of rPrP starts at the end of N-terminus, proceed-ing into the octapeptide-repeat region, the b1-sheet

region, helix 1 region and then the helix 2, in a

‘zip-per’-like manner This interpretation is also in good

agreement with another recent finding showing that

PrP fibril formation proceeds by aligning PrP

mole-cules in parallel, face to back, like a ‘zipper’ [46]

The underlying reason that none of the anti-PrP

mAbs is able to block Cu2+-induced rPrP aggregation

is not known Accumulated evidence suggests that

there are multiple pathways in the PrP aggregation

process [10,47] Our results suggest that GAG-induced

and Cu2+-induced aggregation proceed via different

pathways

All the studies described here were based on findings

using rPrPs Normal PrP has two highly conserved

N-linked glycosylation sites and is present on the cell

membrane with a glycosylphosphatidylinositol anchor

Therefore, it is possible that the presence of N-linked

glycans as well as the placement of the cell membrane

can further modulate the interactions between PrP and

its ligands Based on our findings, we hypothesize that

an increase in the number of octapeptide repeats

causes conformational changes at the N-terminus,

resulting in an enhancement in the binding of PrP

ligands, such as GAG, eventually leading to PrP

aggre-gation Because all these aberrant features are

propor-tional to the number of insertions, our earlier and

current findings provide a biochemical explanation for

the observation that patients with more

octapeptide-repeat insertions have earlier disease onset and shorter

disease duration [3,48]

Experimental procedures

Plasmid construction and recombinant protein

preparation

Cloning, generation and purification of human rPrP,

rPrP8OR, rPrP10OR and rPrPDKKRPK were performed as

described previously with slight modification [11,30] After

refolding and purification, these rPrPs were dialyzed against

20 mm NaAc, pH 5.5 and filtered through a 0.2 lm

mem-brane For human rPrPD51-90, codons 51–90 were removed

from the prion protein coding sequence by annealing the

primer 5¢-GGCAACCGCTACCCA ⁄ CAAGGAGGTGG

CACC-3¢ ( ⁄ marks the site between codon 50 and codon 91)

to a phagemid containing the PrP-coding sequence

Muta-genesis was performed using the BioRad Muta-Gene

phage-mid in vitro mutagenesis kit The PrP mature fragment

(codons 23–231 with deletion of residues 51–90) was cloned

to the vector of pET42a(+) (Novagen, Gibbstown, NJ,

USA) [11], termed pET–rPrPD51-90 The insertion sequence

was verified by using the Applied Biosystems 3730

sequen-cer (Foster City, CA, USA)

Freshly transformed BL21 (DE3) star Escherichia coli (Invitrogen, Carlsbad, CA, USA) containing plasmid pET– rPrPD51-90 was transferred to 1 L Luria–Bertani media with

50 lgÆmL)1kanamycin at 37C until A600reached 0.6 and induced for 4 h with 1 mm isopropyl thio-b-d-galactoside Bacteria were harvested by centrifugation at 4000 g for

15 min at 4C, resuspended in 20 mm Tris ⁄ HCl, pH 7.4,

150 mm NaCl, 1 mm phenylmethanesulfonyl fluoride, 0.1 mgÆmL)1 lysozyme, 1 mm EDTA, 0.1% Triton X-100 and incubated at 25C for 30 min before further lysis by sonication Samples were centrifuged at 13 000 g for

15 min, and the protein pellets were extensively washed using 20 mm Tris⁄ HCl, pH 7.4 with 0.5% Triton X-100 twice, then washed with the same buffer containing 2 m NaCl and 2 m urea respectively The pellets were then resuspended in 20 mm Tris⁄ HCl, pH 8.0, 8 m urea, 10 mm b-mercaptoethanol The protein was refolded by dialysis against 20 mm Tris⁄ HCl, pH 8.0 buffer with decreasing urea and b-mercaptoethanol gradient concentrations All refolded rPrPs were further dialyzed against 20 mm NaAc,

pH 5.5, filtered through 0.2 lm membrane, stored at )80 C and used for experiments within one week after refolding SDS⁄ PAGE and Coomassie Brilliant Blue stain-ing showed that the purity of the recombinant protein is consistently > 95% (not shown) Protein concentration was determined with a Bio-Rad Protein Assay Kit All of the recombinant prion proteins were freshly purified before use

Antibodies

The generation, purification and characterization of all the anti-PrP murine mAbs have been described in detail previ-ously [49,50] mAb 8B4 recognizes an epitope at residues 35–45; SAF32 reacts with residues 63–94 covering the octa-peptide-repeat sequences [51]; 7A12 interacts with helix 1 between residues 143 and 155; 11G5 reacts with residues 115–130 covering b-sheet 1; 8H4 recognizes residues 175–

185 of helix 2; 8F9 reacts with residues 220–231 mAbs 8B4, SAF32, 7A12, 8H4 and 8F9 are IgG1, whereas mAb 11G5 is IgG2b All mAbs were affinity purified using Protein G chromatography The concentration of mAbs was determined with a BCA protein assay Kit (Pierce, Rockford, IL, USA)

Turbidity measurement

The assays were performed at 25C in flat-bottomed 96-well plates Heparin (from porcine intestinal mucosa; Sigma, St Louis, MO, USA) or CuCl2was added into the wells before addition of 200 lL NaCl⁄ Pi(pH 7.4) contain-ing 1 lm rPrPs After mixcontain-ing as quickly as possible, turbidi-ties were monitored within 15 s by reading the absorbance

at 405 nm in a Beckman Coulter AD340 micro-ELISA

plate reader, using a kinetic photometric model (interval

time 30 s, 30 cycles with 1 s shaking before every cycle)

Similar processes were performed with ZnCl2, MgCl2 and

MnCl2

To investigate whether anti-PrP mAbs can block the

hep-arin enhanced aggregation of rPrP, 10 lL hephep-arin (final

con-centration 1 lgÆmL)1) was mixed with 2.5 lL mAbs (final

concentration 0.125 lm) Then 200 lL NaCl⁄ Picontaining

1 lm rPrP10OR was added and mixed quickly Turbidities

were recorded as described in above A similar procedure

was carried out to investigate the effect of anti-PrP mAbs on

copper enhanced aggregation of rPrP An irrelevant mAb

9C1, anti-(brain-derived neurotrophic factor), was used as a

negative control All experiments were carried out at least

three times with different batches of rPrPs

Detection of rPrP binding to heparin

Flat-bottomed, 96-well Costar plates (Corning, Corning,

NY, USA) were coated with 10 lgÆmL)1 heparin at 4C

overnight and blocked with 3% BSA in NaCl⁄ Pi at 25C

for 3 h BSA was coated onto the plates as a control

Appropriate dilutions of rPrPD51-90or rPrP were added into

the plates in triplicate and incubated at 25C for 2 h After

three washes with phosphate-buffered saline⁄ 0.05% Tween

(PBST), bound rPrP was detected with mAb 8H4

Horse-radish peroxidase-conjugated goat anti-mouse IgG

(Chem-icon, Billerica, MA, USA) was used as the secondary

antibody and A405was measured for

2,2¢-azinobis-(3-ethyl-benzthiazoline-6-sulfonic acid) (Roche Diagnostics,

India-napolis, IN, USA) All experiments were carried out at

least three times with different batches of rPrPs

Detection of biotinylated heparin in the

aggregates of rPrPs

mAb 11G5 was previously shown to be able to react with

PrP aggregates [52] mAb 11G5 was coated onto the

flat-bottomed, 96-well Costar plates at 5 lgÆmL)1at 4C

over-night and blocked with 3% BSA in NaCl⁄ Pi at 25C for

3 h BSA was coated onto the plates as a control Five

micromoles of either rPrP or rPrPDKKRPK was mixed with

10 lgÆmL)1 biotinylated heparin (from porcine intestinal

mucosa, Sigma) in 400 lL NaCl⁄ Pi respectively and

incu-bated at 25C for 30 min The aggregates were collected by

centrifugation at 16 000 g for 10 min at 25C Supernatants

were removed and the pellets were washed three times with

NaCl⁄ Piby vortexing followed by centrifugation at 16 000 g

for 5 min The aggregates were dissolved with 50 lL

NaCl⁄ Pi containing 0.1% Triton X-100 by incubation at

42C for 10 min NaCl ⁄ Pi (450 lL) was then added into

the Eppendorf tubes to a final volume of 500 lL Various

dilutions of this original aggregate solution in NaCl⁄ Piwere

then incubated with mAb 11G5-coated plates at 4C

over-night After three washes with PBST, the bound biotinylated

heparin was detected by adding horseradish peroxidase-con-jugated streptavidin (Chemicon) at 1 : 10 000 dilutions 2,2¢-Azinobis-(3-ethylbenzthiazoline-6-sulfonic acid) was added and A405was recorded All experiments were carried out at least three times with different batches of rPrPs

Sucrose-gradient fractionation

To form a 5–50% step sucrose gradient, 5, 10, 15, 20, 30,

40, 50% sucrose solution prepared in NaCl⁄ Piwere loaded into ultraclear centrifuge tubes (13· 51 mm) rPrPs (1 lm) were mixed with 5 lgÆmL)1 low molecular mass heparin (3 kDa) or 20 lm CuCl2in NaCl⁄ Pi After incubation for

30 min at 25C, 0.5 mL of the mixture was loaded on top

of the sucrose gradient Ultracentrifugation was carried out

in SW55 rotor (Beckman, Fullerton, CA, USA) at

100 000 g, 4C for 2 h Fractions of 0.5 mL were collected from the top of the tubes rPrP present in different sucrose-gradient fractions was detected by immunoblotting 10 lL

of each fraction was mixed with 2· SDS loading buffer and heated at 95C for 10 min before separation on 12% SDS⁄ PAGE The gel was transferred to a nitrocellulose membrane and probed with mAb 8H4 Blue dextran (Sigma) with a molecular mass of 2000 kDa was used as a marker in the gradient

Statistical analysis

A two-way ANOVA program was used to determine the P-value between various groups P > 0.05 is considered to

be not significant (ns)

Acknowledgements

We would like to thank Dr Jacques Grassi (Atomic Energy Commission, Saclay, France) for his gift of mAb SAF32 This work was supported in part by NIH (National Institutes of Health) grant

NS-045981-01 and an award⁄ contract from the US Department of the Army, DAMD17-03-1- 286 (to MSS)

References

1 Prusiner SB (1998) Prions Proc Natl Acad Sci USA 95, 13363–13383

2 Mead S (2006) Prion disease genetics Eur J Hum Genet

14, 273–281

3 Kovacs GG, Trabattoni G, Hainfellner JA, Ironside

JW, Knight RS & Budka H (2002) Mutations of the prion protein gene phenotypic spectrum J Neurol 249, 1567–1582

4 Cohen FE, Pan KM, Huang Z, Baldwin M, Fletterick

RJ & Prusiner SB (1994) Structural clues to prion repli-cation Science 264, 530–531