P

T

Aa

b

ARRAA

KPFIIPP

1

umrhe(vasmp

ut1nPds

p

T

0d

Journal of Virological Methods 169 (2010) 253–258

Contents lists available at ScienceDirect

Journal of Virological Methods

journa l homepage: www.e lsev ier .com/ locate / jv i romet

rotocols

he depletion of � and � PrP from complex mixtures

ndrew McMahona,∗, Sen Hanb, Ian Walkera

The University of Melbourne, Department of Veterinary Science, Parkville, Victoria, AustraliaDepartment of Biochemistry and Molecular Biology, Bio21 Molecular Science Institute, University of Melbourne, Parkville, Victoria, Australia

rticle history:eceived 5 November 2009eceived in revised form 21 June 2010ccepted 28 June 2010vailable online 13 July 2010

a b s t r a c t

Prion disorders occur when endogenous prion protein (PrPC) undergoes a conformational change from apredominantly �-helix-rich structure to an insoluble �-sheet-rich structure (PrPSc). The resulting PrPSc

then in some way facilitates the progressive transformation of nearby PrPC to PrPSc. In time this resultsin the deposition of insoluble PrPSc aggregates in the brain; aggregate deposition is irreversible and isultimately fatal. Prion diseases are transmissible orally or through transplantation (including blood trans-fusion). Current diagnostic methods are limited in that they lack the ability to distinguish qualitatively

eywords:rionserric compoundsronsolation and purification

between PrPC and PrPSc. PrP has been shown to bind divalent cations including copper and zinc, thesecations are toxic and thus of limited use in the removal of PrP from solutions destined for administrationto subjects. We have immobilised Fe3+ to an inert Sepharose resin; this resin was capable of quantitativelyremoving endogenous and recombinant PrPC and recombinant � PrP from complex solutions. The low

hat thdestin

eptide libraryhage display

toxicity of Fe3+ suggests tPrP from blood products

. Introduction

Prion protein is an endogenous cell surface glycoprotein ofndetermined function which is expressed almost ubiquitously inammals (Hu et al., 2007) and is most highly expressed in neu-

ons and lymphocytes (Cashman et al., 1990; Li et al., 2001). PrPas been implicated in mediating neural cell adhesion (Manget al., 2002), inhibiting Bax-induced apoptosis in human neuronsRoucou and LeBlanc, 2005), promoting T cell proliferation and acti-ation (Ballerini et al., 2006), inhibition of macrophage phagocyticctivity (de Almeida et al., 2005) and in binding divalent cationsuch as zinc, copper and manganese (Choi et al., 2006). Despite itsany implied functions prion protein knockout mice show no gross

henotypic abnormalities (Bueler et al., 1992).Prion disorders are triggered when native prion protein (PrPC)

ndergoes a conformational change from an �-helix rich struc-ure (PrPC) to an insoluble �-sheet rich structure (PrPSc) (Prusiner,982). This event is thought commonly to be triggered by an exoge-

ous source of PrPSc in acquired cases, by germline inheritancerP gene mutation(s) in inherited cases, and (possibly) as a ran-om spontaneous genetic event in sporadic cases. Subsequent touch a quantal event, PrPC is believed to be progressively trans-

Abbreviations: IMAC, immobilised metal ion affinity chromatography; PrP, prionrotein; PrPC, endogenous prion protein; PrPSc, infectious prion protein.∗ Corresponding author at: 2/26 Railway Rd., New Lambton, NSW 2305, Australia.

el.: +61 428471084.E-mail address: Andrew.mcmahon@gmail.com (A. McMahon).

166-0934/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.jviromet.2010.06.016

e resin described in this report may be of practical use in the depletion ofed for human use.

© 2010 Elsevier B.V. All rights reserved.

formed to the PrPSc form, eventually resulting in the deposition ofinsoluble PrPSc aggregates in the brain parenchyma. Such depositsare thought to be the principal cause of neuronal toxicity in priondisorders (Prusiner, 1982).

In human prion diseases, the deposition of PrPSc is not lim-ited to neural tissue. Although PrPSc accumulates primarily in thebrain, PrPSc has also been found in peripheral tissues. These tissuesinclude spleen, tonsils, lymph nodes, the Peyer’s patch (Head et al.,2004; Wadsworth et al., 2001), spleen, lung, liver, kidney (Bruce etal., 2001), skeletal muscle (Glatzel et al., 2003; Peden et al., 2006)and blood (Llewelyn et al., 2004; Peden et al., 2004). As the presenceof PrPSc in a tissue correlates strongly with the infectivity of priondisease the detection of PrPSc in these tissues is essential, particu-larly with respect to iatrogenesis. With respect to blood productsa molecular filter capable of removing PrPSc from solutions couldfind extensive application at the production level.

It is known that PrPC has a capacity for binding various met-als. It has been shown to bind divalent cations such as copper,zinc and manganese (Choi et al., 2006). This binding takes placein 4 octapeptide repeat sequences (HGGGW) of the N-terminus ofPrPC (Hornshaw et al., 1995) and a non-octapeptide copper bind-ing region involving histidine-96 (Jackson et al., 2001; Burns et al.,2003). This region also reduces copper ions in vitro (Ruiz et al.,2000). PrPC has been associated with neuroprotection against high

levels of copper in vivo and its gene, Prnp also shows upregulationin the presence of copper (Varela-Nallar et al., 2006).

In a recent unrelated study conducted by our laboratory it wasfound serendipitously that magnetic porous glass (MPG) particlesrich in Fe3+ depleted recombinant PrP from solution quantitatively

2 ologic

(bssd�rPrtril1

qs

2

2

dHsIwIamf6SP

l�a2∼ltaStors8E

2

(Nt(obsloP

54 A. McMahon et al. / Journal of Vir

data not shown). In light of the aforementioned ability of PrP toind divalent cations the use of an Fe3+ rich matrix to deplete theo-called recombinant � and � forms of human PrP from complexolutions was explored. The � and � terms are used in this report toistinguish differing conformational isoforms of recombinant PrP.PrP refers to the ‘native’ conformation of recombinant PrP, �

eferring specifically to the �-helices present in both PrPC and �rP. The � PrP isoform is a soluble monomeric recombinant PrPich in �-pleated sheets. Apart from the obvious similarly of � PrPo PrPSc in regard to �-pleated sheets, � PrP also exhibits partialesistance to proteinase K digestion and its propensity to assemblento fibrillar structures. There is no evidence to date of � PrP stimu-ating the conversion of PrPC to PrPSc or being infectious (Collinge,999).

It was found that Fe3+ conjugated Sepharose beads removeduantitatively recombinant � and � PrP from a range of complexolutions including blood products and homogenised neural tissue.

. Materials and methods

.1. Immunochemicals and PrP

Recombinant murine � PrP and � PrP was synthesised asescribed in appendices and was provided kindly by Dr Andrewill (Department of Biochemistry and Molecular Biology, Univer-

ity of Melbourne). Monoclonal antibody 7D9 (Signet) is a mousegG1 Kappa monoclonal antibody raised against recombinant � PrP

hich is reactive to bovine, sheep, mule, deer, elk and mouse PrPC.t recognises a non-linear epitope on both PrPC and PrPSc (Adler etl., 2003). Monoclonal antibody SAF32 (Cayman) is a mouse IgG2aonoclonal raised against a preparation of scrapie associated fibrils

rom infected hamster brain; it binds to a region between residues3 and 94 of hamster PrPC in the N-terminal octarepeat region.AF32 is reactive to human, hamster, bovine, ovine and murinerPC and PrPSc isoforms.

PrP.1 refers to a phage-displayed 7 mer cyclised peptide iso-ated by our laboratory with the ability to specifically bind the

PrP and � forms of murine PrP. The enrichment of PrP.1 fromcombinatorial phage library will be briefly described in Section.4. The Ph.D. C7C phage display library contains approximately1.28 × 109 unique clones. Each discrete clone within the Ph.D. C7C

ibrary displays at most 5 copies of a unique peptide upon the N-erminus of its pIII coat protein with the first cysteine preceded byn alanine and the second followed by a short spacer (Gly-Gly-Gly-er) and then the wild type pIII sequence. Extensive sequencing ofhe naïve library has revealed a wide diversity of sequences with nobvious positional biases (New England Biolabs). Anti-M13 horseadish peroxidase (HRP) conjugated monoclonal antibody (Amer-ham Biosciences) reacts specifically with the major coat proteinof bacteriophage M13 and facilitates visualisation of phage in

LISAs.

.2. PrP detection assay

1 �g/ml of the SAF32 mAb was coated to 96 well MaxisorpNUNC) microtitre plates in 100 �l of coating buffer (15 mMa2CO3, 34 mM NaHCO3, pH 9.6 [NaOH]). After overnight incuba-

ion (4 ◦C) wells were washed 3 times with 300 �l of PBS-TweenPBS, 50 mM Na2HPO4, 140 mM NaCl pH 7.4 [HCl], 0.05% (v/v) poly-xyethylenesorbitan monolaurate [Tween 20]). Wells were then

locked for 2 h at room temperature with 350 �l of 10% (w/v)kimmed milk or 5% (w/v) casein dissolved in coating buffer. Fol-owing 3 PBS-Tween washes, 100 �l of appropriate concentrationsf depleted/control PrP solutions were added to the relevant wells.lates were then incubated for 1 h at room temperature and then

al Methods 169 (2010) 253–258

washed 3 times as above. 100 �l of the appropriate PrP bindingreagent (either 7D9 or PrP.1 as indicated in the text) at an appropri-ate concentration was then added and following a 1 h incubationand 3 further PBS-Tween washes; 100 �l of the appropriate sec-ondary reagent then followed, as did a further 1 h incubation atroom temperature. After five final washes, colour development wasperformed by the addition of 100 �l of tetramethyl benzidine per-oxidase substrate (KPL). This reaction was then stopped with 50 �lof 0.1 M H2SO4 and the colour intensity measured spectrophoto-metrically at 450 nm (Labsystems Multiscan MS).

2.3. Metal conjugated Sepharose bead matrix preparation

1.5 ml of resuspended stock chelating fast flow Sepharose resin(Amersham Biosciences) was washed twice with 600 �l ddH2O ina 5 ml column. 4 ml of a solution of the sulphate salt of the selectedmetal ion (5 mg/ml) was then applied to the column. After 1 h,the resin was washed twice in 4 ml ddH2O. The estimated con-centration of metal ion attached to beads following this processis 30–37 �mol ion/ml resin (Amersham Biosciences).

2.4. PrP depletion

PrP depletion experiments were conducted by incubating PrPwith a ligand matrix and assaying the unbound fraction for the pres-ence of PrP. PrP depletion matrices prepared in Section 2.2 wereadded in appropriate concentrations to samples of PrP (or alter-nately an IgG control) and incubated at room temperature for 3 h.Samples were then centrifuged and the supernatant removed andassayed for PrP as in Section 2.2.

2.5. Isolation and analysis of phage displayed 7mer cyclised PrPbinding peptides

To obtain PrP-specific peptide binders a Ph.D. C7C phage dis-play peptide library was enriched over three rounds for bindersto recombinant � PrP in accordance with the manufacturer’sinstructions (New England Biolabs). Briefly, the enrichment matrixconsisted of a nunc maxisorb plate coated with 2 �g/ml of recom-binant PrP (Prionics) with 2 �g/ml of free recombinant PrP in PBSserving to elute any bound phage. Following 3 rounds of enrich-ment the resulting phage clones were assayed for their ability tobind � and � PrP as described in Section 2.2.

3. Results

3.1. The attachment of PrP to Fe3+

In a series of preliminary experiments (designed for anotherpurpose) it was noted that a sample of MPG beads coated withStreptavidin bound PrP (data not shown). It seemed most likely thatthe agent responsible for PrP binding within the bead matrix wastrivalent Fe, one of its prominent constituents. It seemed essen-tial to test this notion directly. To accomplish this an alternativematrix was needed and to this end an immobilised metal ion affin-ity chromatography (IMAC) matrix was constructed. Chelating fastflow Sepharose (Amersham biosciences) was used to create sucha medium. The iminodiacetic acid groups of chelating Sepharosehave the capacity to form complexes with transition metal ions(30–37 �mol Cu2+/ml). Sepharose resin was washed in ddH2Oand incubated with ferric sulphate (5 mg/ml) for an hour; beads

acquired a distinct orange colour following this incubation. Fol-lowing further ddH2O washes the resulting iron chelated resinwas incubated with PrP. Following this incubation, centrifugationallowed the removal of resin free supernatant which was thenassayed for the presence of PrP (Fig. 1).

A. McMahon et al. / Journal of Virological Methods 169 (2010) 253–258 255

Fig. 1. Depletion of recombinant bovine PrP using Fe3+ conjugated chelatingSepharose beads. Aliquots (200 �l) of bovine recombinant PrP (1 �g/ml) wereincubated with Fe3+ conjugated chelating Sepharose beads at the indicated concen-trations in 5% skimmed milk. After 1 h, beads were removed and the supernatantassayed for the presence of PrP. Duplicate depleted samples were analysed for bind-ing to plates coated with SAF32 (1 �g/ml) monoclonal antibody and blocked with5% Casein. 7D9 (1 �g/ml) monoclonal antibody bound any SAF32 captured PrP. TherIta

S2bPpP

epcaiaP

3s

ap

FScrmPT

Fig. 3. Depletion of recombinant human PrP and human IgG using Fe3+ conju-gated chelating Sepharose beads. Aliquots (200 �l) of human recombinant PrP(1 �g/ml) and human IgG (10 �g/ml) were incubated with Fe3+ conjugated chelatingSepharose beads at the indicated concentrations in 5% skimmed milk. After 1 h, beadswere removed and the supernatant assayed for the presence of PrP. Human IgG wasanalysed for binding to plates coated with sheep anti-human Ig monoclonal anti-

esulting ‘sandwich’ was visualised with the addition of 0.5 �g/ml biotinylated antigG1 monoclonal (binding the IgG1 constant region of 7D9) and SA-HRP. Visualisa-ion and general ELISA procedures were performed as in Section 2.1. No significantbsorbance was produced in the absence of PrP, of SAF32 or of 7D9.

It is clear that the attachment of Fe3+ ions to the chelatingepharose matrix enabled the depletion of PrP from a solution.�l of packed Fe3+ resin beads depleted 30% of PrP and 20 �l ofeads depleted about 95%. Sepharose beads alone did not depleterP from solution (data not shown). The ability of PrP to bind ironrompted an investigation of whether other metal ions would bindrP (Fig. 2).

Nickel sulphate and copper sulphate conjugated beads producedssentially quantitative depletion of 0.5 �g/ml of PrP. Ferric sul-hate conjugated beads exhibited a 91% depletion of murine PrP,omparable to previous data. Ferrous sulphate showed depletorybility markedly lower, removing 52% of PrP from solution. Of notes the efficiency of depletion at lower PrP concentration levels, withll ion types completely absorbing detectable levels of 0.25 �g/mlrP.

.2. The specific quantifiable removal of PrP from complexolution

The specificity of the Fe3+ IMAC bead interaction with PrP hadlready been demonstrated in Figs. 1 and 2. These depletions tooklace in 5% skimmed milk PBS; the presence of high concentrations

ig. 2. Depletion of recombinant murine PrP using Fe3+ conjugated chelatingepharose beads. Aliquots (200 �l) of murine recombinant PrP (Prionics) at theoncentrations indicated were incubated with IMAC beads conjugated to 5% fer-ic sulphate, ferrous sulphate, copper sulphate and nickel sulphate in 5% skimmedilk. After 1 h, beads were removed and the supernatant assayed for the presence of

rP. All data points represent the average of two duplicate depletion experiments.he assay used was identical to that of Fig. 1.

body (1 �g/ml) (Gentaur, Kampenhout, Belgium) and bound human Ig was detectedusing HRP-conjugated anti-human Ig antibodies. Visualisation and general ELISAprocedure was performed as in Fig. 1. All data points represent the average of twoduplicate depletion experiments visualised in duplicate using the above ELISAs.

of milk protein did not affect PrP depletion. In order to demonstratefurther that the above interaction was PrP-specific, the ability ofFe3+ IMAC beads to deplete HRP conjugated human anti-mouse IgG(Thermo Scientific, Rockford, IL, USA) from solution was assessed.Depletions were completed as detailed above with the addition of10 �g/ml of human IgG (Jena Bioscience, Jena, Germany). Super-natants were then analysed for the presence of PrP and IgG.

Two points were apparent from the data of Fig. 3. Firstly it isclear that Fe3+ IMAC beads did not deplete human IgG from solu-tion to any significant extent. Secondly, beads appear to reach theirPrP binding capacity at concentrations between 12.5% and 6.25%packed beads. This data concords with Fig. 1; thus approximately20 �l of Fe3+ IMAC beads had the capacity to bind 200ng of PrP(10 �g PrP/ml Fe3+ IMAC).

3.3. The removal of endogenous PrP from blood-derived fractionsand brain lysate

It seemed important to know whether a relevant complexprotein mixture would interfere with Fe3+ IMAC bead mediateddepletion of PrP from solution. Accordingly, several protein richsamples, intermediates in a commercial human blood fractiona-tion process were obtained (supplied kindly by Dr. Hung Pham

of CSL Ltd). Each sample contained a discrete enriched ‘cut’ ofwhole human plasma; exact protein concentrations and compo-sitions were not defined. In brief, 1 �g/ml of PrP was spiked into allsamples shown in Fig. 4 and all samples were depleted with 10%Fe3+ IMAC beads.

Fig. 4. Depletion of recombinant human PrP using unconjugated paramagneticbeads in blood fractionation products. Aliquots (200 �l) of human recombinant PrP(1 �g/ml) were incubated with 10% (v/v) Fe3+ chelating Sepharose beads. After 1 h,beads were removed and the supernatant assayed for the presence of PrP. Visuali-sation and the general ELISA procedure was performed as in Fig. 1. All data pointsrepresent the average of two duplicate depletion experiments visualised in duplicateby ELISA.

2 ological Methods 169 (2010) 253–258

pdCdvwwpd

dueclttof

bb(n(t(wbcersssb

lT8waq

Fbictpd

Fig. 6. PrP.1 binds recombinant � and � PrP. Serial dilutions of PrP.1 were analysed

56 A. McMahon et al. / Journal of Vir

Levels of PrP depletion were comparable with those achievedreviously; thus the nature of the blocking protein used in theepletion assay did not affect markedly the experimental outcome.SL blood fractionation products did however affect the PrP stan-ard curves used in our ELISA system. Specifically the absorbancealues obtained from undepleted, PrP-spiked samples were some-hat lower than equivalent amounts of PrP made up in PBS: thisas thought to be due to the inordinately high protein levelsresent in the CSL samples. The net result of this effect is thatepletion values indicated in Fig. 4 may be underestimates.

Whilst it was clear that Fe3+ IMAC beads had the capacity toeplete recombinant PrP from complex solutions it was as yetndetermined if Fe3+ IMAC beads had the capacity to depletendogenous PrP from complex solutions. Recombinant PrPC, whilstlosely matching endogenous PrPC in conformational structure,acks glycosylation. This along with other possible unknown struc-ural differences may interfere with the capacity of Fe3+ IMAC beadso deplete endogenous PrPC from solution. As the fundamental aimf this study was to develop a means of depleting endogenous PrProm blood products a source of endogenous PrP was required.

10% Mouse brain lysate (Tg20 strain) was provided generouslyy Dr James Thyer of CSL Bioplasma and consisted of 10% mouserain lysate in PBS. Tg20 mice are known to overexpress PrPFischer et al., 1996) and were thus an attractive source of endoge-ous PrPC. 10% mouse brain lysate was diluted 1/10 in CSL fraction 1Fig. 4), effectively a 100-fold dilution. Samples were then subjectedo depletion with ferric sulphate, in the presence or absence of 0.2%v/v) sarcosyl. Sarcosyl was included in the depletion mixture as itas a concern that without it PrP depletion may be inhibited as

inding regions may be occluded in membrane bound PrP. Thisoncern appears to have been unwarranted. No detectable differ-nce was seen between the depletory ability of the Fe3+ conjugatedesin in the presence or absence of sarcosyl. Following depletion,arcosyl was added to all samples as it was found to give a maximalignal to noise ratio in mouse brain homogenate assays (data nothown). Fig. 5 illustrates the concentration of PrP present in mouserain samples with and without depletion by Fe3+ IMAC beads.

The results of Fig. 5 show that the ‘natural’ PrP in mouse brainysate could be quantitatively removed with Fe3+ IMAC beads.his assay indicates that neat mouse brain contains approximately

�g/g of PrP, ∼100-fold more than that stated in the literature inild type mice (Legleiter et al., 2007). It should be noted that the

ssay utilised in the detection of PrP is likely to have been of littleuantitative use. Our capture assay used to quantify PrP in solu-

ig. 5. Depletion of endogenous PrP using Fe3+ conjugated chelating Sepharoseeads. A clarified mouse brain lysate (0.1 g/ml) was diluted 10-fold in PBS contain-

ng no sarcosyl or containing 0.2% (w/v) sarcosyl, 1% casein and Fe3+ conjugatedhelating Sepharose (10%) in 5% skimmed milk. After 1 h, beads were removed andhe supernatant assayed for the presence of PrP. Visualisation and the general ELISArocedure was performed as in Fig. 1. All data points represent the average of twouplicate depletion experiments visualised in duplicate by ELISA.

for binding to plates containing immobilised recombinant murine PrP (coated at2 �g/ml) and then blocked with 10% (w/v) skimmed milk protein, or blocked withouta PrP coat as in the legend to Fig. 3. Visualisation and the general ELISA procedurewas performed as in Fig. 1. All data points represent the average of two duplicates.

tion was able to detect consistently concentrations of PrP greaterthan ∼100 ng/ml in solutions of PBS (Data not shown). The PrP con-centration stated by Legleiter et al. (2007) was determined with asandwich ELISA similar to our own; this assay appears to be at least1000 times more sensitive than our own.

Of greatest significance is the quantitative removal of endoge-nous PrP from mouse brain, further supporting the potential use offerric sulphate as a molecular filtering agent to remove PrP fromcomplex solutions.

3.4. The quantifiable depletion of ˇ PrP from complex solution

Thus far it has been shown that Fe3+ IMAC beads have the capac-ity to deplete both endogenous murine and recombinant humanPrPC from complex solutions. It was considered that whilst Fe3+

IMAC beads were capable of binding PrPC they may not likewisebind PrPSc. If Fe3+ IMAC bead mediated PrP depletion was to be ofuse in the depletion of PrPSc from blood products then its capac-ity to deplete PrPSc needed to be assessed. PrPSc was not available;however, � PrP, a �-pleated sheet rich isoform of PrP which closelymimics the structure of PrPSc was discussed in Section 4.

Depletions of � and � PrP were carried out as in Fig. 1, visual-isation however required an alternative assay as the7D9 anti-PrPmonoclonal antibody (the capture reagent used in Figs. 1–5) wasshown experimentally not to bind � PrP (data not shown). Analternative assay was used which utilised the SAF32 monoclonalantibody as a capture reagent and PrP.1 phage ‘antibody’ as the pri-mary antibody. This ‘sandwich’ was visualised with an anti M13HRP monoclonal antibody. It should also be noted that in this caseboth � and � PrP were of a murine origin, both SAF32 monoclonalantibody and PrP.1 phage have demonstrated the ability to bind �and � recombinant murine PrP (Fig. 6).

Fig. 7 illustrates the percentage of a 1 �g/ml sample of � and� PrP removed from solution by 10% Fe3+ IMAC beads. Efficien-cies of depletion were equal to those observed in Figs. 1–5. � and� PrP were both removed from solution with similar efficiency.The results of Fig. 5 show that � and � PrP could be quantitativelyremoved from complex solutions with Fe3+ IMAC beads.

4. Discussion

Previous experimental data prompted the coupling of Fe3+ tofast flow chelating Sepharose, an alternate matrix bearing no chem-

ical similarities to streptavidin-MPG beads which has the capacityto bind 30–37 �mol Fe3+/ml. The resulting IMAC beads were capa-ble of binding PrP, unchelated beads were not. The two matricescannot be compared directly in terms of surface area, Fe3+ con-centration and PrP capacity per bead. However, a 1000-fold higher

A. McMahon et al. / Journal of Virologic

Fig. 7. Depletion of recombinant murine � and � PrP using Fe3+ conjugated chelatingSepharose beads. Aliquots (200 �l) of murine � and � PrP (1 �g/ml) were incu-bated with Fe3+ conjugated 10% (v/v) chelating Sepharose beads at the indicatedconcentrations in 5% skimmed milk. After 1 h, beads were removed and the super-npe

ca

oT(F

sanninci

osdib

rrba

abtobmtF�lP

wmaJ

atant assayed for the presence of PrP (Fig. 1 assay, SAF32 capture reagent, PrP.1rimary antibody). All data points represent the average of two duplicate depletionxperiments visualised in duplicate by ELISA.

oncentration of Fe3+ IMAC beads is necessary to bind equivalentmounts of PrP to streptavidin MPG beads.

In this study regardless of species (human, bovine and murine)r conformational isoform (� and �) PrP has an affinity for Fe3+.he binding of Cu2+ to these forms of PrP is well documentedMillhauser, 2007). To the authors’ knowledge this is the first timee3+ has been observed to bind � PrP and recombinant bovine PrP.

Fig. 2 indicates that copper, nickel and ferric sulphate have aimilar affinity for PrP; by contrast ferrous sulphate binds withn apparent lower affinity. A more thorough examination of PrPickel and copper binding may have revealed affinity differencesot apparent in Fig. 2 which may not represent threshold PrP bind-

ng capacity in copper and nickel IMAC beads. Whilst it appears thatickel ions may possibly be more effective as depletion reagentsoncerns over toxicity issues during later commercial use of theseons prompted the focus on ferric sulphate.

The interaction between PrP and Fe3+ IMAC beads is a specificne. Throughout this study the presence of saturating amounts ofkimmed milk, casein, mouse brain and other protein rich solutionsid not inhibit the ability of IMAC beads to bind PrP. Furthermore,

t has been shown that human IgG was not bound by Fe3+ IMACeads at concentrations as high as 25% (v/v) (Fig. 3).

The affinity of Fe3+ IMAC beads for PrP is not an artefact ofecombinant PrP constructs. As is evident in Fig. 5 Fe3+ IMAC beadsemoves detectable levels of endogenous murine PrP from mouserain lysate in both the presence and absence of sarcosyl. Thus itppears that membrane associated PrP is also bound by the beads.

All the above data suggest that Fe3+ IMAC beads bind PrP fromvariety of species in complex solutions. This binding appears toe specific. To be useful clinically Fe3+ IMAC beads must facilitatehe removal of PrPSc from blood products. Due to the unavailabilityf infectious PrPSc within this department, this property has yet toe examined. Readily available samples of � and � recombinanturine PrP isoforms provided the ‘next best’ means of assessing

he ability of IMAC beads to bind a distinctly different PrP isoform.ig. 7 illustrates the capacity of Fe3+ IMAC beads to deplete both theand � conformations of PrP from complex solution. It thus seems

ikely that the conformationally similar infectious PrPSc isoform ofrP would also be bound by Fe3+ IMAC beads.

It should be noted that the present study did not investigate

hether PrPSc per se binds Fe3+ ions; however, the known confor-ational relationship between the � isoform of PrP used herein and

uthentic PrPSc suggest that the latter material should bind Fe3+.ackson et al. (2001) have shown (consistent with our results) that �

al Methods 169 (2010) 253–258 257

PrP binds several heavy metal ions. However, there is contradictoryevidence in the literature concerning whether PrPSc binds metalions. On the one hand, Shaked et al. (2001) were unable to demon-strate the attachment of PrPSc to Cu2+ conjugated IMAC beads. Onthe other hand, Wadsworth et al. (1999) claimed the contrary. It isof obvious importance to resolve whether or not PrPSc binds metalions in the same way as its PrPC counterpart.

The ability of PrP to bind Fe3+ is not apparent in the literaturealthough observations have been made that PrPSc infected murinecells show a deficiency in iron levels. This deficiency is linked to adecreased level of cellular ferritin (Fernaeus et al., 2005). A lack offerritin appears to be linked to the iron related oxidative stress seenin PrPSc infected cells which is believed to contribute to cell death(Fernaeus and Land, 2005). The likely site of PrP attachment to Fe3+

IMAC beads is its N-terminal octapeptide repeat as this is the site ofcopper binding (Choi et al., 2006; Hornshaw et al., 1995). Anothermetal binding region located around histidines 96 and 111 (Jacksonet al., 2001; Burns et al., 2003) is also a potential site of attachment.Mutation and/or removal of one or both of these metal binding sitesmay help elucidate the Fe3+ binding region.

A major aim of this study was to construct an affinity matrixcapable removing PrPC and PrPSc from solution. It is clear in regardto PrPC that this aim was met. Fig. 7 illustrates the Fe3+ IMAC beadmediated removal of � PrP from solution, providing tentative evi-dence that the conformationally similar PrPSc may also be removed.No PrPSc was available at the time of writing. As opposed to otherPrP binding cations the low toxicity of iron makes it an ideal reagentfor use as a PrP depletory agent in blood products destined forhuman use. Affinity column based filtration is likely to be an effec-tive and cheap method of guaranteeing that a solution has beendepleted of PrP by at least an order of magnitude. Given the low con-centrations of PrP in blood such depletion would effectively be total.

Acknowledgements

Thanks are given to Dr. Andrew Hill (Department of Biochem-istry and Molecular Biology, University of Melbourne) and his teamfor providing our laboratory with high quality recombinant PrP ina variety of conformations; these materials were invaluable.

Appendix A.

A.1. Generation of full-length murine PrP23-231

Full-length murine PrP23-231 was PCR amplified frommouse genomic DNA cloned into the pTrcHis expression vec-tor (Invitrogen) and was sequence verified by Bigdye (ABI) DNAsequencing. The expression construct was then transformed intoBL21(DE3) competent E.coli cells and expression of mid-log phasegrowth cells was induced for 4 h with 1 mM isopropyl-�-D-thiogalactopyranoside. The cell pellet collected was resuspended in50 ml of lysis buffer consisting of 25U benzonase (Novagen), 500 �lprotease inhibitors (Sigma P8465) and 25 mg of lysozyme (resus-pended in 10 mM Tris, pH 8) before incubating at room temperatureon a spinning wheel 20 min. 1% (v/v) Nonidet-P40 (Fluka) wasadded before incubating for a further 10 min on spinning wheel.Inclusion bodies were isolated as a pellet after centrifugation for15 min at 6000 × g.

A.2. Purification of recombinant PrP

Inclusion body pellets were solubilized in IMAC binding buffer(8 M urea, 0.1 M sodium phosphate (Na2HPO4), 10 mM Tris-HCland 10 mM reduced glutathione pH 8.0) before loading onto an Ni-NTA column (GE Healthcare). After washing off unbound protein

2 ologic

(I(ti5tEtoa1b1iu

A

paPpd(aatstR(oSw

A

(w1riert

R

A

B

B

B

58 A. McMahon et al. / Journal of Vir

flow-through) with the binding buffer, recPrP was eluted withMAC (immobilized metal affinity chromatography) elution buffer8 M urea, 0.1 M Na2HPO4, 10 mM Tris-HCl, 10 mM reduced glu-athione, pH 4.5) in a single step (100% elution buffer) and collectedn fractions supplemented with EGTA at a final concentration ofmM. Collected fractions were analysed by SDS-PAGE and PrP con-

aining fractions were pooled and oxidised overnight with 5 mMGTA and 0.2 mM oxidized glutathione at a final protein concentra-ion of 0.4 mg/ml. HPLC purification of oxidised protein was carriedut with HPLC buffer A: 0.01% trifluoroacetic acid (Supelco) in waternd buffer B: 0.01% trifluoroacetic acid in acetonitrile using a Jupiter0u C4 520 mm × 10 mm column (Phenomenex). A gradient withuffer B was applied in the following manner;0–25% in the first5 min followed by 25–35% for the next 65 min and finally 35–100%

n 15 min. The purified protein was lyophilised and stored at −80 ◦Cntil use.

.3. Refolding into ˛ and ˇ-PrP

�-PrP and was obtained by reconstituting post-HPLC purifiedrotein into buffer of choice (for example water, 10 mM sodiumcetate or 10 mM 3-(N-morpholino) propanesulfonic acid (MOPS)).reparation and refolding of �-PrP was carried out based on therotocol in Jackson et al. (1999). Briefly, HPLC purified �-PrP wasiluted to 0.04 mg/ml and unfolded overnight in �-unfolding buffer100 mM dithiothreitol (DTT), 10 mM sodium acetate, 10 mM Triscetate, 6 M guanidine hydrochloride, pH 8) on a rotating wheelt room temperature. The unfolded protein was then refoldedhrough dialysis with �-PrP refolding buffer (1 mM DTT, 10 mModium acetate, 10 mM Tris acetate, pH 4). Typically, 50 ml of pro-ein was dialysed with four changes of 4 L of refolding buffer.efolded protein was concentrated with Vivaspin20 spin columnsGE Healthcare) to 2 ml and centrifuged at 13,000–15,000 g at 4 ◦Cvernight to spin down precipitated material from soluble �-PrP.oluble protein was collected in the supernatant and quantitatedith A280 nm readings.

.4. CD analysis of recombinant PrP

To confirm the conformation of PrP, far-UV circular dichroismCD) spectra were recorded on a JASCO J-815 CD spectrometerith a protein concentration of 0.2 mg/ml in quartz cuvettes withmm pathlength (Hellma), using a scan speed of 50 nm/min and a

esponse time of two seconds. Multiple scans were averaged (typ-cally n = 5). Spectra scanned over 185–260 nm was collected andxpressed as mean residue ellipticity (MRE, [�] deg cm2 dmol-1esidue-1). The results were analysed with the online CD analysisool Dichroweb for secondary structure content analysis.

eferences

dler, V., Zeiler, B., Kryukov, V., Kascsak, R., Rubenstein, R., Grossman, A., 2003. Small,highly structured RNAs participate in the conversion of human recombinantPrP(Sen) to PrP(Res) in vitro. J. Mol. Biol. 332, 47–57.

allerini, C., Gourdain, P., Bachy, V., Blanchard, N., Levavasseur, E., Gregoire, S.,Fontes, P., Aucouturier, P., Hivroz, C., Carnaud, C., 2006. Functional implica-tion of cellular prion protein in antigen-driven interactions between T cells anddendritic cells. J. Immunol. 176, 7254–7262.

ruce, M.E., McConnell, I., Will, R.G., Ironside, J.W., 2001. Detection of vari-ant Creutzfeldt-Jakob disease infectivity in extraneural tissues. Lancet 358,208–209.

urns, C.S., Aronoff-Spencer, E., Legname, G., Prusiner, S.B., Antholine, W.E., Gerfen,G.J., Peisach, J., Millhauser, G.L., 2003. Copper coordination in the full-length,recombinant prion protein. Biochemistry 42, 6794–6803.

al Methods 169 (2010) 253–258

Bueler, H., Fischer, M., Lang, Y., Bluethmann, H., Lipp, H.P., DeArmond, S.J., Prusiner,S.B., Aguet, M., Weissmann, C., 1992. Normal development and behaviour ofmice lacking the neuronal cell-surface PrP protein. Nature 356, 577–582.

Cashman, N.R., Loertscher, R., Nalbantoglu, J., Shaw, I., Kascsak, R.J., Bolton, D.C.,Bendheim, P.E., 1990. Cellular isoform of the scrapie agent protein participatesin lymphocyte activation. Cell 61, 185–192.

Choi, C.J., Kanthasamy, A., Anantharam, V., Kanthasamy, A.G., 2006. Interaction ofmetals with prion protein: possible role of divalent cations in the pathogenesisof prion diseases. Neurotoxicology 27, 777–787.

Collinge, J., 1999. Variant Creutzfeldt-Jakob disease. Lancet 354, 317–323.de Almeida, C.J., Chiarini, L.B., da Silva, J.P., PM, E.S., Martins, M.A., Linden, R., 2005.

The cellular prion protein modulates phagocytosis and inflammatory response.J. Leukoc. Biol. 77, 238–246.

Fernaeus, S., Halldin, J., Bedecs, K., Land, T., 2005. Changed iron regulation in scrapie-infected neuroblastoma cells. Brain Res. Mol. Brain Res. 133, 266–273.

Fernaeus, S., Land, T., 2005. Increased iron-induced oxidative stress and toxicity inscrapie-infected neuroblastoma cells. Neurosci. Lett. 382, 217–220.

Fischer, M., Rulicke, T., Raeber, A., Sailer, A., Moser, M., Oesch, B., Brandner, S.,Aguzzi, A., Weissmann, C., 1996. Prion protein (PrP) with amino-proximal dele-tions restoring susceptibility of PrP knockout mice to scrapie. Embo. J. 15,1255–1264.

Glatzel, M., Abela, E., Maissen, M., Aguzzi, A., 2003. Extraneural pathologic prionprotein in sporadic Creutzfeldt-Jakob disease. N. Engl. J. Med. 349, 1812–1820.

Head, M.W., Ritchie, D., Smith, N., McLoughlin, V., Nailon, W., Samad, S., Masson,S., Bishop, M., McCardle, L., Ironside, J.W., 2004. Peripheral tissue involvementin sporadic, iatrogenic, and variant Creutzfeldt-Jakob disease: an immunohisto-chemical, quantitative, and biochemical study. Am. J. Pathol. 164, 143–153.

Hornshaw, M.P., McDermott, J.R., Candy, J.M., Lakey, J.H., 1995. Copper binding tothe N-terminal tandem repeat region of mammalian and avian prion protein:structural studies using synthetic peptides. Biochem. Biophys. Res. Commun.214, 993–999.

Hu, W., Rosenberg, R.N., Stuve, O., 2007. Prion proteins: a biological role beyondprion diseases. Acta Neurol. Scand. 116, 75–82.

Jackson, G.S., Hosszu, L.L., Power, A., Hill, A.F., Kenney, J., Saibil, H., Craven, C.J.,Waltho, J.P., Clarke, A.R., Colling, J., 1999. Reversible conversion of monomerichuman prion protein between native and fibrilogenic conformations. Science283 (5409), 1935–1937.

Jackson, G.S., Murray, I., Hosszu, L.L., Gibbs, N., Waltho, J.P., Clarke, A.R., Collinge, J.,2001. Location and properties of metal-binding sites on the human prion protein.Proc. Natl. Acad. Sci. U.S.A. 98, 8531–8535.

Legleiter, L.R., Ahola, J.K., Engle, T.E., Spears, J.W., 2007. Decreased brain copper dueto copper deficiency has no effect on bovine prion proteins. Biochem. Biophys.Res. Commun. 352, 884–888.

Li, R., Liu, D., Zanusso, G., Liu, T., Fayen, J.D., Huang, J.H., Petersen, R.B., Gambetti, P.,Sy, M.S., 2001. The expression and potential function of cellular prion protein inhuman lymphocytes. Cell Immunol. 207, 49–58.

Llewelyn, C.A., Hewitt, P.E., Knight, R.S., Amar, K., Cousens, S., Mackenzie, J., Will,R.G., 2004. Possible transmission of variant Creutzfeldt-Jakob disease by bloodtransfusion. Lancet 363, 417–421.

Mange, A., Milhavet, O., Umlauf, D., Harris, D., Lehmann, S., 2002. PrP-dependent celladhesion in N2a neuroblastoma cells. FEBS Lett. 514, 159–162.

Millhauser, G.L., 2007. Copper and the prion protein: methods, structures, function,and disease. Annu. Rev. Phys. Chem. 58, 299–320.

Peden, A.H., Head, M.W., Ritchie, D.L., Bell, J.E., Ironside, J.W., 2004. Preclinical vCJDafter blood transfusion in a PRNP codon 129 heterozygous patient. Lancet 364,527–529.

Peden, A.H., Ritchie, D.L., Head, M.W., Ironside, J.W., 2006. Detection and local-ization of PrPSc in the skeletal muscle of patients with variant, iatrogenic,and sporadic forms of Creutzfeldt-Jakob disease. Am. J. Pathol. 168, 927–935.

Prusiner, S.B., 1982. Novel proteinaceous infectious particles cause scrapie. Science216, 136–144.

Roucou, X., LeBlanc, A.C., 2005. Cellular prion protein neuroprotective function:implications in prion diseases. J. Mol. Med. 83, 3–11.

Ruiz, F.H., Silva, E., Inestrosa, N.C., 2000. The N-terminal tandem repeat region ofhuman prion protein reduces copper: role of tryptophan residues. Biochem.Biophys. Res. Commun. 269, 491–495.

Shaked, Y., Rosenmann, H., Hijazi, N., Halimi, M., Gabizon, R., 2001. Copper bindingto the PrP isoforms: a putative marker of their conformation and function. J.Virol. 75, 7872–7874.

Varela-Nallar, L., Toledo, E.M., Larrondo, L.F., Cabral, A.L., Martins, V.R., Inestrosa, N.C.,2006. Induction of cellular prion protein gene expression by copper in neurons.Am. J. Physiol. Cell Physiol. 290, C271–281.

Wadsworth, J.D., Hill, A.F., Joiner, S., Jackson, G.S., Clarke, A.R., Collinge, J., 1999.

Strain-specific prion-protein conformation determined by metal ions. Nat. CellBiol. 1, 55–59.

Wadsworth, J.D., Joiner, S., Hill, A.F., Campbell, T.A., Desbruslais, M., Luthert, P.J.,Collinge, J., 2001. Tissue distribution of protease resistant prion protein in variantCreutzfeldt-Jakob disease using a highly sensitive immunoblotting assay. Lancet358, 171–180.