1
Chapter 5: The effects of clusterin on proIAPP amyloid formation and cytotoxicity
5.3 RESULTS
5.3.1 The effect of clusterin on amyloid formation measured by ThT assays
The effect of clusterin on proIAPP amyloid formation was investigated using ThT assays. 250 µg/mL proIAPP was incubated in the absence and presence of 25, 50 and 100 µg/mL clusterin at 37 oC. These concentrations of clusterin are physiologically relevant as clusterin is present in human serum at 35-105 µg/mL (Murphy et al., 1988). The amyloid formation of proIAPP in the absence of clusterin showed a sigmoidal pattern as previously described in Section 4.2.3. The presence of clusterin had no effect on the lag phase, which was approximately 120 h, in the presence and absence of clusterin. In the absence of clusterin, there was a rapid growth phase of approximately 50 h and by 170 h amyloid formation had gone to completion. In the presence of clusterin, the growth phase was slowed in a dose-dependent manner (Figure 5.2). In the presence of 25 µg/mL clusterin, amyloid formation had not gone to completion by 340 h and the ThT fluorescence was approximately half that of proIAPP in the absence of clusterin. Clusterin at higher concentrations could not prevent the initiation of the growth phase but the growth phase appeared so slow as to have plateaued with a low ThT fluorescence, indicating that very little amyloid had formed.
The inhibition of amyloid formation corresponded to molar ratios of 1 clusterin to 25-100 proIAPP. These molar ratios are similar those seen over which clusterin inhibits amyloid formation of apolipoprotein C II (1:90) (Hatters et al., 2002) and A42 (1:50) (Oda et al., 1995). The molar ratio of clusterin to proIAPP also compares favourably with the minimal molar ratios of clusterin to proteins, stressed by heat or reduction, that are required to almost completely abolish their precipitation (Humphreys et al., 1999). For example the minimal molar ratio of clusterin to glutathione-S-transferase required for clusterin to prevent the reduction-induced precipitation of glutathione-S-transferase was approximately 1:3.2. Similarly the minimal molar ratios to inhibit the heat-induced precipitation for catalase, BSA and -lactalbumin were 1:1.3, 1: 2.3 and 1:11 respectively (Humphreys et al., 1999).
Figure 5.2 Clusterin inhibits proIAPP amyloid formation. 250 µg/mL proIAPP at 37 oC was incubated in the absence of clusterin (red) and in the presence of 25 µg/mL (blue), 50 µg/mL (green) and 100 µg/mL clusterin (black). Amyloid formation was measured using a ThT assays as described in Section 2.2.16a.
The substoichiometric ratios over which clusterin suppresses proIAPP amyloid formation suggest that clusterin is either interacting transiently with proIAPP or is interacting with a small proportion of proIAPP species, such as the nuclei of amyloid formation. In this context Hatters et al. have suggested that clusterin interacts with apolipoprotein C II nuclei leading to the dissociation of the monomeric subunits (Hatters et al., 2002). It is not likely to be the case for clusterin and proIAPP. If clusterin dissociated proIAPP nuclei, this would have been as a longer lag phase as the time required for nucleation is prolonged. This was not observed as the lag phase in the presence and absence of clusterin was the same, suggesting that clusterin was unable to prevent the assembly of proIAPP monomers into a stable nucleus (Figure 5.2). Clusterin slows the growth rate of amyloid formation in a dose-dependent manner, suggesting that clusterin may inhibit elongation. Hatters et al. have suggested that clusterin inhibits fibril growth because clusterin reduced the average sedimentation coefficient of apolipoprotein C II fibrils (Hatters et al., 2002). Clusterin inhibits A amyloid formation leading to the formation of soluble low molecular weight amyloidogenic oligomers that are soluble and highly cytotoxic (Lambert et al., 1998). As low molecular weight oligomers do not enhance ThT fluorescence (LeVine 1993), based on the ThT assay we cannot exclude the possibility that the inhibition of proIAPP amyloid formation results in an increase in proIAPP oligomers rather than fibrils.
5.3.2 Analysis of proIAPP and clusterin mixtures by transmission electron microscopy
TEM was used to characterise the morphology of amyloid fibrils formed in the presence of clusterin. 500 µg/mL proIAPP in the presence of various concentrations of clusterin up to 100 µg/mL was incubated for 10 days at 37 oC. At the end of 10 days, aliquots were adsorbed to Formvar-coated copper grids, negatively stained with 1% phosphotungstic acid and analysed by TEM. ProIAPP in the absence of clusterin showed typical fibril morphology compared to that seen previously (Figure 5.3A and B, compare Figure 4.8). The fibrils were approximately 10 nm wide and could be up to hundreds of nm long. The fibrils also formed a net-like mesh (Figure 5.3A, E, G and H) as was previously described for proIAPP (Figure 4.8) (Krampert et al., 2000). In the presence of 12.5 µg/mL clusterin, proIAPP was fibrillar (Figure 5.3G) and looked very similar to the fibrils formed in the absence of clusterin (Figure 5.3A). There were far fewer fibrils in the presence of 50 and 100 µg/mL clusterin, and very little evidence of amorphous aggregate (Figure 5.3E and F and Figure 5.3C and D). This also indicated that the reduced ThT fluorescences of proIAPP seen in the presence of clusterin were unlikely to be the result of the formation of amorphous aggregate rather than fibrillar aggregate and were likely to be due to a suppression of amyloid formation.
Small spherical structures were seen in proIAPP mixtures with 50 and 100 µg/mL clusterin (Figure 5.3D and F, arrows). These spherical structures were seen in proIAPP incubations in the absence of clusterin (Figure 4.10D) and may represent proIAPP oligomers. Electron micrographs of clusterin alone also showed spherical structures of similar size (Figure 5.18G), indicating these may be oligomeric clusterin but these were less numerous than in the proIAPP and clusterin mixtures.
ProIAPP was also incubated with 100 µg/mL ovalbumin (Figure 5.3H) as a control for a non-specific effect of a protein. ProIAPP has a molecular weight of 7.5 kDa and an isoelectric point of 9.5. Clusterin is a glycoprotein with a molecular weight of around 75 kDa and an isoelectric point of 4.5-5.2 (Murphy et al., 1988). At pH 7.4, an electrostatic interaction may partially contribute to the suppression of amyloid formation therefore ovalbumin, as a protein unrelated to amyloidosis, was chosen for its similar properties to clusterin: a relatively large glycoprotein with a molecular weight of 45 kDa and an isoelectric point of 4.5-4.9. Similar amounts of amyloid were formed in the presence of ovalbumin (Figure 5.3H) compared to proIAPP alone (Figure 5.3A) indicating that the effect of clusterin on proIAPP was unlikely to be non-specific.
(Figure on previous page)
Figure 5.3 Electron microscope images of proIAPP fibrils formed in the presence and absence of clusterin. 500 µg/mL proIAPP was incubated for 10 days at 37 oC in PBSaz in the absence of clusterin (A and B) and in the presence of 100 µg/mL (C and D), 50 µg/mL (E and F), and 12.5 µg/mL clusterin (G). 500 µg/mL proIAPP was incubated with 100 µg/mL ovalbumin (H) as a non-specific protein. Specimen were negatively stained with 1% (w/v) phosphotungstic acid and examined as described in Section 2.2.19.
5.3.3 The binding of clusterin to freshly redissolved proIAPP measured by direct ELISA
The data presented above show that clusterin prolongs proIAPP in a soluble state, preventing amyloid formation (Section 5.3.1 and 5.3.2). One way in which clusterin could inhibit amyloid formation would be to interact with soluble amyloid-prone proIAPP species and inhibit their conversion to amyloid. Therefore the binding of clusterin to freshly redissolved proIAPP was measured to establish whether clusterin could bind to natively unfolded proIAPP. The binding was measured by direct ELISA, in which one binding partner, in this case proIAPP, was adsorbed to the ELISA plate and the other partner, in this case clusterin was in solution.
The binding of clusterin to a variety of native and denatured proteins has been shown to be enhanced at pH 5.5 compared to 7.4 (Hochgrebe et al., 2000). Clusterin was added at various concentrations up to 100 µg/mL at pH 7.4 and 5.5. The unbound clusterin was then washed away. Clusterin that was bound to the proIAPP was detected using a monoclonal antibody against clusterin and a secondary antibody conjugated to horseradish peroxidase. The amount of clusterin bound was proportional to the coloured product produced by horseradish peroxidase and was quantified by its absorbance at 490 nm. Clusterin bound to proIAPP at both pH 7.4 and 5.5 in a dose-dependent manner. The binding of clusterin to proIAPP did not reflect non-specific binding as non-specifically bound clusterin was removed by washing with 0.1% (v/v) Triton X-100 in PBSaz prior to the detection of clusterin. Moreover, there was no binding of clusterin to blocker only.
The binding of clusterin to proIAPP was enhanced at pH 5.5 compared to 7.4, as has been seen for other proteins. The increased binding of clusterin to proIAPP is likely to be due to conformational changes in clusterin rather than proIAPP, which is adsorbed to the ELISA tray. Clusterin is predominantly present in an oligomeric form at pH 7.4 but at pH 5.5 a greater proportion of clusterin is present in a monomeric form (Hochgrebe et al., 2000). The increase in monomer is associated with an increase in exposed hydrophobic surface (Hochgrebe et al., 2000). This increase in exposed hydrophobic surface at pH 5.5 is likely to be linked to the increase in activity as chaperones like clusterin mainly interact with their client proteins through hydrophobic interactions.
Figure 5.4 ELISA measurements of clusterin binding to proIAPP at pH 7.4 and 5.5. ProIAPP (100 µg/mL) was absorbed to an ELISA plate. The binding of clusterin in 1% BSA (w/v) in MES/PBSaz at pH 7.4 (blue) or 5.5 (black) was measured as in Section 2.2.21. The data shown are means ± standard deviations of triplicate measurements.
5.3.4 The binding of clusterin to freshly redissolved proIAPP by competition ELISA
ProIAPP is likely to undergo conformational changes prior to amyloid formation. Soluble proIAPP is likely to comprise monomers and an increasing concentration of a heterogenous assortment of oligomers, once going down the amyloid pathway. A time course experiment was therefore devised to measure the binding of clusterin to proIAPP at different times along the amyloid formation pathway, as this would enable us to correlate the binding of clusterin to proIAPP with the effect of clusterin on amyloid formation from it. A direct ELISA protocol as described in Section 5.3.3, in which one binding partner, in this case proIAPP, is adsorbed to the ELISA plate and the other binding partner in this case clusterin is in solution, would not be suitable for a time course experiment. The assortment of proIAPP species are likely to change with time of progression down the amyloidogenic pathway and these different proIAPP species are unlikely to adsorb to an ELISA plate equally well. A change in the amount of proIAPP adsorbed to the ELISA plate could be falsely interpreted as a change in the extent of clusterin binding to proIAPP.
Detecting proIAPP bound to clusterin adsorbed to an ELISA tray would also pose difficulties, as proIAPP-specific antibodies may not recognise different proIAPP species in the amyloid formation pathway equally well. This has been shown in other amyloid systems: a polyclonal antibody that has been raised against oligomeric A does not bind to monomeric A or oligomers of A smaller than an octamer (Kayed et al., 2003). These changes in the effectiveness of detecting bound proIAPP would also be falsely interpreted as a change in clusterin binding to proIAPP. To avoid these possibilities, the binding of clusterin to proIAPP was measured indirectly in a competition ELISA format.
In a competition ELISA, a protein (in this case clusterin) can either bind to a solution phase ligand (in this case proIAPP) or solid phase ligand (adsorbed to the ELISA plate). The binding of the protein (clusterin) to the solution phase ligand (proIAPP) is seen as a change in the binding of clusterin to the solid phase ligand. The exact region(s) of the “chaperone active site” of clusterin are not known, however ELISA studies suggest that the binding site for native ligands is distinct from misfolded proteins (Lakins et al., 2002). Pre-incubation of clusterin with native GST in solution inhibited the binding of clusterin to solid-phase GST in a dose-dependent manner but did not inhibit the binding of clusterin to heat-stressed GST. In addition, pre-incubation of clusterin with native A1-40 or heparin had no effect on the binding of clusterin to lysozyme or insulin, stressed by reduction. As clusterin was considered likely to bind to proIAPP at the same site as denatured proteins, denatured proteins were chosen as competitors for clusterin binding to proIAPP.
Denatured lysozyme was initially chosen as the denatured target because it had been previously characterised as a ligand for clusterin (Lakins et al., 2002) and its high isoelectric point ~11 would reduce the likelihood of an electrostatic interaction with proIAPP, which has a similarly high isoelectric point, ~9.5. Lysozyme was adsorbed to the plate and subsequently denatured by reduction with 15 mM DTT. Clusterin was pre-incubated with freshly dissolved proIAPP for 15 min prior to being applied to the denatured lysozyme in the ELISA plate and incubated for a further 1 h. Clusterin was detected as previously described in Section 5.3.3.
Figure 5.5 Clusterin binding to denatured lysozyme in the presence of proIAPP. Lysozyme was adsorbed to an ELISA plate and denatured by reduction with 15 mM DTT. Clusterin at 25 µg/mL (red bars), 50 µg/mL (blue bars) or 100 µg/mL (green bars) was added to the wells in the presence of proIAPP and clusterin in PBSaz, pH 7.4 that was bound to the denatured lysozyme was detected by ELISA as described in Section 2.2.22. The data shown are means standard deviations of triplicate measurements. The significance of the amount of clusterin bound to denatured lysozyme in the absence of proIAPP, compared to the amount bound in the presence of proIAPP was assessed using Student’s t-test (^ p < 0.05, * p < 0.005, # p < 0.0005).
In the absence of proIAPP, clusterin bound to denatured lysozyme in a dose-dependent manner (Figure 5.5), consistent with previous findings (Lakins et al., 2002). At all fixed concentrations of clusterin, proIAPP enhanced the binding of clusterin to denatured lysozyme in a dose-dependent manner (Figure 5.5). At all concentrations of clusterin with 50, 100 and 200 µg/mL proIAPP, the amount of clusterin bound to denatured lysozyme was approximately 1.5-fold, 1.75-fold and 2.2-fold respectively compared to clusterin bound in the absence of proIAPP (Figure 5.5). This indicated that the enhancement of clusterin binding to denatured lysozyme was dependent on the concentration of proIAPP and independent of the concentration of clusterin.
Figure 5.6 Clusterin binding to denatured lysozyme in 1% (w/v) BSA in the presence of proIAPP. Lysozyme was adsorbed to an ELISA plate and denatured by reduction with 15 mM DTT. Clusterin at 25 µg/mL (red bars), 50 µg/mL (blue bars) or 100 µg/mL (green bars) was added to the wells in the presence of proIAPP and clusterin in 1% (w/v) BSA in PBSaz, pH 7.4 that was bound to the denatured lysozyme was detected by ELISA as described in Section 2.2.22. The data shown are means standard deviations of triplicate measurements. The significance of the amount of clusterin bound to denatured lysozyme in the absence of proIAPP, compared to the amount bound in the presence of proIAPP was assessed using Student’s t-test (^ p < 0.05, * p < 0.005, # p < 0.0005).
One potentially artefactual reason for the proIAPP enhancement of clusterin binding to denatured lysozyme may be that during the 15 min pre-incubation step, protein adsorbs non-specifically to the tube walls. Therefore when clusterin is pre-incubated in the absence of proIAPP more clusterin is adsorbed to tube walls and less is applied to the ELISA plate than when clusterin and proIAPP are pre-incubated together. To exclude this possibility 1% (w/v) BSA in PBSaz was used a diluent for clusterin and proIAPP mixtures. The enhancement of clusterin binding to the denatured lysozyme in the presence of proIAPP was seen regardless of the presence of 1% (w/v) BSA (Figure 5.6). This indicated that enhanced binding of clusterin to denatured lysozyme in the presence of proIAPP was not an artefact of non-specific adsorption losses of clusterin during the pre-incubation step.
5.3.5 The binding of clusterin to different denatured proteins in the presence of proIAPP at pH 7.4
Another potentially artefactual reason for the enhanced binding of clusterin to denatured lysozyme in the presence of proIAPP is that proIAPP could be binding to the denatured lysozyme and making another surface for clusterin to attach to. Therefore, the enhancement of clusterin binding to a denatured lysozyme in the presence of proIAPP was investigated further to see if this effect was specific to lysozyme or to the mode of denaturation. Lysozyme has a molecular weight of 14.3 kDa, a pI of around 11 and was denatured in the ELISA experiments by reduction with 15 mM DTT. Denatured alcohol dehydrogenase and insulin were chosen as solid-phase ligands for competition ELISA experiments. Insulin has a molecular weight of 5.8 kDa, a pI of around 5.3 and was stressed by reduction with 15 mM DTT. Alcohol dehydrogenase (ADH) has a molecular weight of 147 kDa, a pI of around 5.4-5.8 and was denatured by heating at 80 oC. Clusterin in 1% (w/v) BSA was pre-incubated in the presence and absence of proIAPP at various concentrations and its binding to denatured insulin or ADH monitored by competition ELISA as described above.
In the absence of proIAPP, clusterin bound to denatured insulin and ADH in a dose-dependent manner (Figure 5.7) consistent with previous findings (Lakins et al., 2002). ProIAPP enhanced the binding of clusterin to both denatured ADH and insulin at pH 7.4 in a dose dependent manner. The binding of clusterin to denatured insulin was enhanced by approximately 1.5-fold, 3-fold and 5-fold in the presence of 50 µg/mL, 100 µg/mL and 200 µg/mL respectively at all fixed concentrations of clusterin (Figure 5.7A). The binding of clusterin to denatured ADH was enhanced by 2-fold, 3-fold and 4-fold in the presence of 50 µg/mL, 100 µg/mL and 200 µg/mL proIAPP respectively regardless of clusterin concentration (Figure 5.7B). Thus, the observation of proIAPP-mediated enhancement of clusterin binding a variety of denatured proteins that do not share similar size or charge properties or mode of denaturation, suggests that proIAPP has a general effect on enhancing the binding of clusterin to denatured proteins rather than proIAPP specifically binding to these different denatured proteins and creating another surface for clusterin to attach to.
Figure 5.7 Clusterin binding to denatured insulin and ADH in 1% (w/v) BSA in the presence of proIAPP. Insulin (A) and ADH (B) were adsorbed to an ELISA plate and denatured by reduction with 15 mM DTT or heat respectively as described in Section 2.2.22 . Clusterin at 25 µg/mL (red bars), 50 µg/mL (blue bars) or 100 µg/mL (green bars) was added to the wells in the presence absence of various concentrations of proIAPP in 1% BSA in PBSaz, pH 7.4. Clusterin that was bound to the denatured protein was detected by ELISA as described in Section 2.2.22. The data shown are means standard deviations of triplicate measurements. The significance of the amount of clusterin bound to denatured protein target in the presence of proIAPP, compared amount bound in the absence of proIAPP was assessed using Student’s t-test (^ p < 0.05, * p < 0.005, # p < 0.0005).