Cell Line Specific Modulation of Extracellular Aβ42 by Hsp40

Heat shock proteins (Hsps) are a set of molecular chaperones involved in cellular repair. They provide protective mechanisms that allow cells to survive potentially lethal insults, In response to a conditioning stress their expression is increased. Here we examined the connection between Hsps and Aβ42, the amyloid peptide involved in the pathological sequence of Alzheimer’s disease (AD). Extracellular Aβ42 associates with neuronal cells and is a major constituent of senile plaques, one of the hallmarks of AD. Although Hsps are generally thought to prevent accumulation of misfolded proteins, there is a lack of mechanistic evidence that heat shock chaperones directly modulate Aβ42 toxicity. In this study we show that neither extracellular Aβ42 nor Aβ42/PrPC trigger the heat shock response in neurons. To address the influence of the neuroprotective heat shock response on cellular Aβ42, Western analysis of Aβ42 was performed following external Aβ42 application. Five hours after a conditioning heat shock, Aβ42 association with CAD cells was increased compared to control neurons. However, at forty-eight hours following heat shock Aβ42 levels were reduced compared to that found for control cells. Moreover, transient transfection of the stress induced Hsp40, decreased CAD levels of Aβ42. In contrast to CAD cells, hippocampal neurons transfected with Hsp40 retained Aβ42 indicating that Hsp40 modulation of Aβ42 proteostasis is cell specific. Mutation of the conserved HPD motif within Hsp40 significantly reduced the Hsp40-mediated Aβ42 increase in hippocampal cultures indicating the importance of this motif in regulating cellular Aβ42. Our data reveal a biochemical link between Hsp40 expression and Aβ42 proteostasis that is cell specific. Therefore, increasing Hsp40 therapeutically with the intention of interfering with the pathogenic cascade leading to neurodegeneration in AD should be pursued with caution.


Introduction
Alzheimer's disease (AD), an age-dependent neurodegenerative disease that is estimated to affect 35 million people worldwide is characterized by amyloid deposits, neurofibrillary tangles, selective neuronal loss, cognitive decline and memory loss [1,2]. Multiple lines of evidence suggest an imbalance between the production and clearance of Ab 1242 , a 42 residue long b amyloid protein that spontaneously self aggregates into dimers, oligomers, protofibrils, and fibrils and initiates a toxic sequence of events leading to synaptic dysfunction and dementia [3]. Ab 42 as well as Ab 40 are derived from the amyloid precursor protein (APP) by the sequential proteolytic processing of a, b and c secretases (reviewed: [4,5]. Following proteolysis, the peptides can be secreted or transferred to the endosomal/lysosomal system. Intraneuronal Ab 42 is comprised of both uptake of Ab 42 from the extracellular space as well as intracellular cleavage of APP [6]. Synaptic activity increases levels of secreted, extracellular Ab peptides while reducing intracellular levels [7]. Why do the physiological mechanisms that under normal circumstances tightly regulate Ab 42 production, cell association and clearance fail? Deficiencies in cellular chaperone systems are one possibility. Molecular chaperones maintain protein homeostasis by assisting nascent polypeptides to fold, protecting mature proteins from stresses and eliminating misfolded proteins. Protein quality control mechanisms are critical to neural function and defects in proteolytic pathways are widely held to lead to neurodegeneration [8]. The cellular level of chaperones might affect the toxicity of Ab 42. In fact, enhancement of the cellular quality control machinery, has been proposed to prevent or delay the cascade of misfolding in conformational diseases [9,10]. In addition to maintenance of protein homeostasis (proteostasis) by constitutive chaperones, in response to a number of stressful stimuli, there is an induction of stress-induced chaperones (eg Hsp40, Hsp90, Hsp70 and Hsp27). Understanding the biochemical sequence of events that underlies Ab 42 -mediated neurodegeneration requires a clear understanding of the role(s) that chaperones play in the AD pathogenic cascade.
A number of chaperones are implicated in Ab 42 proteostasis [11]. Several chaperones have been found both in association with senile plaques [12][13][14] as well as endogenous Ab 42 [15]. These reports have given rise to the notion that molecular chaperones are suppressors of toxic Ab 42 conformations leading to AD. This idea is consistent with the observations that heat shock genes appear to be induced poorly late in life and that the principal risk factor for AD is age [9]. Further support for this view has come from reports demonstrating that in experimental models, Hsp70 [16,17], Hsp27 [18] and Hsp90 [16] protect against the toxic effects of Ab 42 . Also, Hsp70 is reported to suppress cognitive deficits and pathological phenotypes in AD mice [19]. In vitro Hsp70/40 and Hsp90 suppressed early stages of Ab 42 assembly into aggregates but had no effect on fibrils [16]. Still, many questions remain unanswered regarding the chaperone folding paths for Ab 42 . For example, Mearow and colleagues have shown that heat shock of neonatal rat cortical cultures increases the detrimental effects of Ab 42 on cell survival while overexpressing Hsp27 protects against Ab 42 [18]. In mice models of Alzheimer's disease overall content of the chaperone ab crystallin is reduced [20]. However, in contrast to the concept of therapeutic rescue by chaperones, several molecular chaperones actually support the formation of the toxic Ab 42 oligomeric species [21][22][23]. This promotion of Ab 42 oligomerization by select chaperones has similar features to that observed in response to general anesthesia [24,25].
In addition to cellular chaperones, cellular prion protein (PrP C )/ Ab 42 association could influence Ab 42 quality control. Ab 42 in contrast to Ab 40, associates rapidly with neuronal cells [26]. Two distinct Ab 42, oligomeric conformations accumulate on the surface of living cells [27]. Exposure of Ab 42 to pH = 6 for 24 hours to mimic endosomal conditions increases Ab 42 binding to PC12 cells [28]. The cellular prion protein has been shown to act as a functional high affinity receptor for Ab 42 [29][30][31] Strittmatter and colleagues report that association of PrP C with Ab 42 mediates downstream Ab 42 -impairement of hippocampal long term potentiation and that AD transgenic mice lacking PrP C accumulate Ab 42 but have normal survival and test normal for learning and memory [29,31]. Furthermore, transgenic overexpression of PrP C is shown to enhance amyloid plaque formation in an AD mouse model [32]. Along these lines, pathological levels of Ab 42 have been shown to disrupt PrP C modulation of NMDA (N-Methyl-daspartate) receptor activity [33]. However, in contrast, Balducci et al report that Ab 42 impairs consolidation of long-term recognition memory in mice independent of PrP C , raising questions regarding the role of Ab 42 /PrP C association in AD progression [30]. Additionally, other molecules (eg STI1) are recognized to bind PrP C , but whether these agents compete with Ab 42 for binding is unknown [34]. Following cell association, insoluble Ab 42 aggregates localize to endosome/lysosome compartments [35,36]. Curiously, one report reveals that PrP C inhibits b-secretase cleavage of amyloid precursor protein and reduces Ab 42 levels [37].
That said, which molecular chaperones directly regulate Ab 42 protestasis and/or toxicity remain to be established. In this study we have monitored the association of Ab 42 with cultured neural cells following induction of the heat shock. Our findings demonstrate that heat shock initially increases Ab 42 association with CAD neuroblastoma cells but is followed by a decline in cellular Ab 42 levels at 48 hours. Transient transfection experiments revealed that Hsp40 decreased cellular levels of Ab 42 in CAD cell but increased cellular levels of Ab 42 in hippocampal cultures. We evaluated the influence of exogenously applied soluble PrP C on cellular uptake/processing of Ab 42 to test the hypothesis that soluble PrP C would bind to Ab 42 and reduce association with cell anchored PrP C , thereby blocking an early event in the Ab 42 pathogenic cascade. Here we document that PrP C failed to decrease cellular association of Ab 42 . Our data reveal a biochemical link between cellular levels of Hsp40 and Ab 42 that is cell line specific. This raises the possibility that Hsp40 is involved in the pathogenic cascade leading to dementia and neurodegeneration in AD.

Extracellular Ab 42 does not Trigger the Heat Shock Response
Heat shock chaperones are induced in response to a number of cell stressors such temperature, ischemia and heavy metals. During aging when heat shock genes are thought to be induced poorly, humans are susceptible to AD. To gain insight into the involvement of the heat shock response in Ab 42 pathogenic cascades, we carried out biochemical studies to establish whether the treatment of neuroblastoma cells with Ab 42 triggers the expression of the stress-induced chaperones. Mouse CAD neuroblastoma cells were incubated with a high concentration (25 mM) of Ab 42 for 48 hours, rinsed in PBS and solublized. 30 mg of supernatant (1% TX-100/0.1% SDS soluble proteins) and 10 ml of total pellet (1% TX-100/0.1% SDS insoluble proteins) were subjected to Western analysis. Figure 1 demonstrates that Ab 42 was clearly found in both soluble and insoluble CAD cell fractions. The expression of cellular Hsp70 (heat shock protein of 70 kDa), Hsp25 (Heat shock protein of 25 kDa) and Hsp40 (Heat shock protein of 40 kDa) in CAD cells did not change following treatment with Ab 42 for 48 hours. Hsp70 and Hsp25 were not detectable in either the presence or absence of Ab 42 . Moreover, Hsp40 showed modest expression in control CAD cells as previously described [38], and no change was observed in response to Ab 42 treatment. The expression levels of the constitutive chaperone Hsc70 (Heat shock cognate protein of 70 kDa) also did not change in response to Ab 42 . Actin is shown as a loading control.

Ab 42 does not Block the Heat Shock Response
The heat shock response is a highly conserved cell survival program that enhances cell survival to subsequent insults [9]. Since interference with the heat shock response would be expected to reduce protein surveillance and triage mechanisms and downstream cell survival, we next tested whether Ab 42 altered induction of the heat shock chaperones. CAD cells were incubated with 3 mM Ab 42 then heat shocked at 43uC for 40 min and allowed to recover for 5 hours prior to lysis and Western analysis. Figure 2 clearly shows that Hsp70 is induced by heat shock and that Ab 42 does not alter the induction of this stress inducible chaperone. As expected, Ab 42 treatment of CAD cells triggers apoptotic pathways as shown by activation of caspase 3, a marker of programmed cell death. The Ab 42 activation of caspase 3 was not altered by heat shock. Heat shock treatment increased cellular levels of soluble Ab 42 ,3.5 fold, suggesting that following heat shock neurons are more prone to the cellular toxicity of Ab 42 ( Figure 2). This observation is in line with a previous study demonstrating that Ab 42 decreases cortical neuron cell survival and that heat shock renders neurons more vulnerable to Ab 42 treatment [18]. Taken together, our data demonstrate that, although a large number of stressors activate the neuroprotective heat shock response, Ab 42 failed to increase the expression of the heat shock chaperones. Moreover, removal/disruption of the protective effects of a conditioning heat shock against cell death is not part of the Ab 42 pathogenic cascade.

Soluble PrP C does not Block Cell Association of Ab 42
PrP C has been proposed to act as a functional high affinity receptor for Ab 42 [29][30][31]. Based on these reports, we wanted to determine how PrP C may impact the cellular association of Ab 42 . Therefore, we examined the possibility that soluble recombinant PrP C would competitively displace Ab 42 from CAD cells thereby reducing the association of Ab 42 with the neuronal cell culture. Treatment with 0.5 mM recombinant PrP C did not induce Hsp70 or activation of caspase 3. Also, PrP C failed to alter the heat shock induction of Hsp70 or the Ab 42 induced activation of caspase 3. No difference in cellular Ab 42 association was observed between Ab 42 and PrP C /Ab 42 treated control (no-heat shock) cells. In fact, rather than inhibit cellular association of Ab 42, PrP C /Ab 42 cotreatment followed by heat shock revealed that PrP C increased Ab 42 levels in the soluble CAD cell fractions ( Figure 2). CAD cells express endogenous PrP C , a glycosylphosphatidyl (GPI) anchored plasma membrane protein [39,40] (Figure 3) that is subject to Nlinked glycosylation and non-, mono-and di-glycosylated versions of PrP C simultaneously exist [41]. Recombinant bovine PrP C252232 migrated further on SDS-PAGE then native unglycosylated PrP C therefore rendering cell association of exogenous recombinant PrP C distinguishable from endogenous PrP C . Exogenous recombinant PrP C was observed to associate with cells ( Figure 3). In the absence of CAD cells recombinant PrP C was observed to undergo partial breakdown following heat shock for 40 min at 43uC. Taken together our results show that exogenous PrP C does not block cellular Ab 42 association, in fact, following heat shock PrP C /Ab 42 co-treatment increased cell associated Ab 42 .

Heat Shock Promotes Time Dependent Clearance of Ab 42
Although heat shock facilitated the pathogenicity of Ab 42 as measured by its increased cellular association, we speculated that this may be due to the physical effects of heat shock on membrane permeability rather than the conformational processing of PrP C by stress induced chaperones. To gain further insight into the relationship between cellular uptake/clearance of Ab 42 and the heat shock response, we carried out immunoblot analysis on CAD cells in which the heat shock was given at an earlier time point thereby increasing the time Ab 42 is exposed to the stress-induced chaperones. Figure 4 shows that when a 40 min heat shock was given starting at the time that Ab 42 was applied to CAD cells, the cellular levels of soluble Ab 42 at the 42 hour time point was reduced. Ab 42 does not always resolve as a discrete band by SDS-PAGE depending on abundance and the characteristic wide Ab 42 band is shown in Figure 4. Cellular levels of Hsp70 and Hsp40, which are elevated 3 hours following heat shock [38], remained elevated 48 hours following heat shock and at the 48 time point translocation of Hsp70 (but not Hsp40) to the detergent insoluble fraction was observed. Our observations demonstrate that while soluble Ab 42 is increased 5 hours following heat shock ( Figure 2), soluble Ab 42 is decreased 48 hours following heat shock ( Figure 4) indicating that with time heat shock chaperones increase cellular Ab 42 clearance. Geldanamycintreatment of CAD cells which induces Hsp40 but not Hsp70 [38] was also observed to reduce cellular levels of Ab 42 (data not shown) . Figure 4 (lower panel) clearly shows that CAD cell levels of Ab 42 increase in response to increasing Ab 42 concentrations. Again, PrP C , (either bovine upper panel or mouse lower panel Figure 4) did not reduce cell association of Ab 42 and did not alter the heat shock response. Figure 5 shows that in the absence of cells, heat shock per se does not cause degradation or oligomerization of Ab 42. Also, PrP C does not initiate any shifts in the molecular weight of Ab 42 indicative of proteolysis or SDS-resistant oligomerization. In contrast a 15 kDa breakdown of PrP C was observed following heat shock. Figure 6 shows that neither PrP C nor Ab 42 were found to alter cellular levels of the constitutive chaperones DnaJA1/Hdj2, DnaJA2/Rdj2, DnaJA3/Tid1, DnaJA4, Hsc70 or the stress induced chaperones Hsp70, Hsp40 and Hsp25, indicating that a generalized reduction in these molecular chaperone levels is not an underlying mechanism of Ab 42 induced neurodegeneration.
Taken together, these results indicate that heat shock chaperones have a role in clearance of cellular Ab 42 , and may perhaps be involved in reducing Ab 42 pathogenesis.

Modulation of Ab 42 by Hsp40 is Cell Line Specific
To further investigate the role that specific inducible chaperones play in heat shock induced reduction of Ab 42, CAD cells were transfected with the stress induced J protein Hsp40 and then challenged with the toxic Ab 42 (Figure 7) . Both heat shock and Hsp40 transfection reduced soluble and insoluble Ab 42 (monomer) at 48 hours. Quantitative immunoblotting uncovered a 50% 63 Hsp40-mediated reduction compared to a smaller heat shockmediated decreases 88% 66 in insoluble monomeric Ab 42 . Soluble Ab 42 monomer was found to decrease to 67% 612 following Hsp40 transfection and 48% 615 following heat shock. Hsp40 and heat shock both caused changes in Ab 42 oligomerization, initially increasing the Ab 42 72 kDa oligomer followed by a decrease at 48 hrs (Figure 7). These experiments clearly establish Hsp40 as a chaperone that influences cellular clearance of Ab 42 . Transfection does not induce the stress response ( Figure 7C ). Likewise, residues encoding amino acids 106-126 of PrP C as well as a scrambled control do not induce Hsp70 or increase Hsp40 levels.
We then asked the question whether Hsp40 also regulates Ab 42 levels in primary hippocampal co-cultures of neurons and glia. Rat hippocampal neurons were isolated on postnatal day 0, transfected by electroporation with cDNA for myc-tagged Hsp40 or pCMV vector. Extensive neuritic outgrowth was found in both control and transfected cultures ( Figure 8). Bassoon (presynaptic-red) and neurofilament (green) expressing cells are apparent. DAPI staining is shown in blue. Healthy neurons are visible in both control and Ab 42 -treated cultures. 3 days post-transfection cultures were treated with 10 mM Ab 42 and 4 days later cultures were washed in PBS and cellular (total) Ab 42 was determined by quantitative immunoblotting. Total lysates were prepared by direct cell lysis in TX100 lysis buffer followed by SDS sample buffer to ensure that all Ab 42 present was evaluated by Western blot analysis. Exposure of primary hippocampal cultures to extracellular Ab 42 resulted in rapid Ab 42 clearance. Figure 9 shows that 4 days following 10 mM Ab 42 exposure primary rat hippocampal cultures, which have high endogenous levels of Hsp40, reduce Ab 42 as does vector transfected neurons. To our surprise, in contrast to CAD neural cultures, transfection of hippocampal neurons with Hsp40 increased (59.3 fold) Ab 42 monomer levels over nontransfected cultures. These results reveal stark differences in neuronal processing of Ab 42 following Hsp40 transfection between CAD cell cultures and hippocampal cultures.
To begin to address the mechanism of the Hsp40-mediated increase in Ab 42 hippocampal co-cultures were transfected with a mutated form of Hsp40 predicted to bind to client proteins but not have chaperone activity. Hsp40 interacts with and activates the ATPase Hsp70/Hsc70 via its J domain. An HPD (histidine- proline-aspartic acid) motif located within the J domain of Hsp40 is required for harnessing Hsc70/Hsp70 for conformational work. By mutating the histidine-proline-aspartic acid motif of Hsp40 to alanines (Hsp40 HPD-AAA ) an Hsp40 that binds client protein but does not activate Hsc70/Hsp70 ATPase is generated. Transfection of hippocampal neurons with Hsp40 HPD-AAA increased Ab 42 monomer (15.3 fold) compared to vector control (7.9 fold) ( Figure 9A&B). Although Hsp40 HPD-AAA increased cellular Ab 42 levels over untreated control cells (no transfection, no extracellular Ab 42 ) and vector control cultures (transfection with vector alone followed by application of extracellular Ab 42 ), the Hsp40 HPD-AAA -mediated increase (15.3 fold) was smaller compared to Hsp40 (59.3 fold) indicating that the HPD motif within the J domain of Hsp40 impacts directly on cell processing of Ab 42 . Figure 9C &D show that transfection efficiency is lower in co cultures compared to CAD cells. Like that seen for CAD cells, Ab 42 did not induce the heat shock response in hippocampal neurons. PrP C did not block the association/accumulation of Ab 42 in hippocampal neurons (data not shown). Figure 9D compares total cellular hippocampal cultures exposed to 10, 20, 25 mM Ab 42 at the time of transfection (left panel) and three days following transfection (right panel) to CAD cells treated for 24 hours with 25 mM Ab 42 .
Taken together, these results demonstrate that modulation of Ab 42 by Hsp40 is cell line specific.
In summary we have found that Hsp40 is able to influence cellular levels of Ab 42. Neural processing of extracellular Ab 42 is dynamic. Overproduction and impaired clearance of Ab 42, are implicated in AD [8]. Accumulation of Ab oligomers can lead to synaptic dysfunction and disruption in neural plasticity [42][43][44][45], however, details of the molecular cascade(s) that underlie Ab 42 neuronal toxicity remain unclear. Understanding how cells regulate cellular Ab 42 levels requires identification of the cellular chaperone machinery involved in processing the neural amyloid pool. In this study we provide mechanistic evidence that Hsp40 regulates association/accumulation of extracellular Ab 42 with neurons and that Hsp40-mediated regulation is cell specific.
Hsp40 is an evolutionarily ancient and widely expressed chaperone that almost certainly targets multiple client proteins. In neurons, Hsp40 is found to be concentrated in postsynaptic densities [46], in lipid rafts [47] and in association with presynaptic chaperones [38]. Furthermore, Hsp40 is linked to neurite outgrowth [48]. That said, its precise role in synaptic transmission and neurodegeneration is not yet known. Hsp40 is a member of the J protein family [49]. All J proteins have a tetrahelical Hsp70/ Hsc70-interacting domain called a J domain. Via their J domain, J proteins target a wide array of cellular proteins to the ATPases Hsc70/Hsp70 for conformational work. Although our observation that Hsp40 alters Ab 42 cellular turnover in a cell culture specific manner is in contrast to the notion that heat shock chaperones rid the cell of toxic proteins in all cells, it is possible that in disease conditions Hsp40 may protect a toxic protein (eg Ab 42 ) from triage resulting in acceleration of disease progression. In fact, the J protein family determines which chaperone pathway is pursued by Hsc70/Hsp70 [50]. For example, DnaJB2(HSJ1) stimulates ubiquitination and sorting of substrates to the proteosome [51], while DnaJC6(auxilin) stimulates recycling of clathrin from clathrin coated vesicles [52] and DnaJB6 (Mrj) regulates keratin turnover [53]. Further experimentation is required to understand how chaperones like Hsp40 either aid in folding and maintenance or lead to degradation. Cellular levels of Hsp40 routinely rise and fall as part of the heat shock program. Levels of constitutive Hsp40 vary drastically among neural cell lines [38] and classes of neurons in the adult rat brain [54] but how this correlates with Ab 42 levels remains to be established. Although the heat shock response is a highly conserved cellular program that confers transient cyto-protection via the induction and translocation of stress induced chaperones, differences in the threshold for the induction of the heat shock response, the complement of heat shock proteins induced as well as the constitutive expression of heat shock proteins are observed among neurons. For example, Hsp27 is induced by heat shock in primary hippocampal cultures but not primary cortical cultures [18]. Also, some neurons (eg motor neurons) have a high threshold for inducing the heat shock response [55]. The cellular variations in either constitutive expression or heat shock expression of Hsp40 may well lead to selective vulnerability of neurons. Other heat shock proteins have been shown to protect against Ab 42 neuronal toxicity. Expression of Hsp27 in cortical neuronal cultures prepared from postnatal day 1 rats protects against Ab 42associated toxicity [18]. Also, Hsp70 overexpressing mice have reduced Ab [19]. Our data show that extracellular Ab 42 does not  elicit a heat shock response in neurons nor does it alter induction of the heat shock response. Moreover we demonstrate that while heat shock initially increases neural Ab 42 , with time in culture Ab 42 levels are reduced in neurons following heat shock compared to control cultures.
The basis of the cell specific responses to Hsp40 is likely due to differences in the chaperone networks between the cultured cells. The elaborate chaperone machinery that is present in cells rids the cell of toxic proteins often via assembly of chaperones into active chaperone complexes. Levels of chaperones and chaperone complexes may differ between CAD neuroblastoma cells and primary hippocampal cultures. Further experimentation is required to develop a detailed and comprehensive overview of differences in chaperones among neurons. In conclusion we have shown that (1) Hsp40 modulates processing of extracellular Ab 42.
(2) Hsp40 modulation of Ab 42 is cell specific and dependent on the conserved HPD motif within Hsp40's J domain. (3) Ab 42 does not trigger the heat shock response or alter the threshold for induction of the heat shock response. (4) Soluble PrP C does not block association of extracellular Ab 42. Overall these results elucidate an important link between Hsp40 and cellular levels of Ab 42, which has not been illustrated previously.

CAD (CNS Catecholaminergic Derived) Mouse
Neuroblastoma Cells [38,56,57] CAD (CNS catecholaminergic derived) mouse neuroblastoma cells were seeded into 6 well plates and grown in DMEM/F12 medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin as previously described. Cells were lysed in     PLoS ONE | www.plosone.org 40 mM Tris (pH 7.4), 150 mM NaCl, 2 mM EDTA, 1 mM EGTA, 1 mM Na 3 VO 4 , 0.1% SDS, 1% TX-100, 0.5 mM PMSF and protease inhibitor (Sigma) at 4uC for 1 hour. Lysates were centrifuged at 150006g for 5 minutes at 4uC and the supernatant (soluble fraction) and pellet (insoluble fraction) was collected and stored at 270uC. For transient transfection, CAD cells were washed in PBS and transiently transfected with c-myc tagged rat Hsp40 DNA using Lipofectamine-2000 (Invitrogen) in Opti-MEM. For heat shock experiments cells were incubated at 43uC for 40 minutes and then returned to 37uC. Amyloid-b 1-42 -TFA was dissolved in 1% NH 4 OH as recommended by the supplier (rPeptide) to a concentration of 1 mg/mL. The resulting solution was sonicated for 1 minute, aliquoted and stored at 270uC. With this preparation, Ab 42 will be mainly in the monomeric form (,95%) [58]. Recombinant bovine Prion Protein (Prionics, amino acids PrP C25-242 ) was suspended in H 2 0, aliquoted and stored at 270uC. PrP C106-126 and the scrambled PrP C106-126 (Anaspec) were resuspended in DMSO to a final stock concentration of 1 mM and stored at 220uC [59]. Recombinant mouse His-tagged L42 prion protein was prepared as previously described [60,61]. Aliquots were diluted in culture media immediately prior to treatment of cells. Protein concentration of the soluble CAD cell fraction was determined by Bradford assay (BioRad). The 1% TX-100/0.1% SDS insoluble cell fraction was dissolved directly in sample buffer.
Primary Hippocampal Cell Culture [62,63] Dissociated primary co-cultures of neurons and glia were isolated by dissection from Sprague Dawley rats (Charles River) at postnatal day 0 as previously described [62]. Animals were anesthetized on ice and sacrificed by decapitation. Hippocampi were removed and incubated in 40 ml/ml of papain. After 30 minutes cells were washed three times in fresh Eagle's basal media (GIBCO-Invitrogen) supplemented with B-27, penicillin, streptomycin, L-glutamine and 4% feta bovine serum. Cells were triturated using three decreasing calibers of pipettes and the cell solution was then transfected by electroporation (Bio Rad Gene Pulsor Xcell:settings 150 volts, pulse length 25, 5 pulses, pulse interval 0.1 and cuvete size 4 mm) with cDNA of either myctagged Hsp40 or Hsp40 HPD-AAA or vector control (pCDNA3.1) and then plated. For immunostaining, transfected cells were plated on silicon wafers (Silicon wafers, Silicon Quest, CA). Co-cultures were grown for 7 days in a 5% CO 2 incubator. After transfection, Ab 42 was added to the culture and 3 days later the cells were washed in PBS, lysed in 40 mM Tris (pH 7.4), 150 mM NaCl, 2 mM EDTA, 1 mM EGTA, 1 mM Na 3 VO 4 , 0.1% SDS, 1% TX-100, 0.5 mM PMSF and protease inhibitor (Sigma) at 4uC for 1 hour. Total cell lysates were solubilized directly in sample buffer and evaluated by Western blot analysis. The University of Calgary Conjoint Faculties Research Ethics Board specifically approved this study (protocol number M09008).

Immunoblotting
Proteins were electrotransferred from polyacrylamide gels to 0.2 mm nitrocellulose membrane in 20 mM Tris, 150 mM glycine and 12% methanol. Membranes were blocked with in PBS with 0.1% Tween 20, 4% milk and incubated with primary antibody overnight at 4uC. The membranes were washed and incubated with horseradish peroxidase-coupled secondary antibody. The signal was developed using West Pico reagent (Pierce Biotechnology Inc.) and exposed to Kodak film. Bound antisera were quantitated by Biorad Fluor-S MultiImager Max and Quanti-tyOne 4.2.1 software.