1
Chapter 5: The effects of clusterin on proIAPP amyloid formation and cytotoxicity
5.3.8 The effect of clusterin on proIAPP amyloid formation monitored by CD spectropolarimetry
Far-UV CD had been used to characterise the secondary structure of proIAPP as it proceeds down the amyloidogenic pathway (Section 4.2.2). ProIAPP remained stable as a random coil prior to a decrease in the CD signal likely to be due to an oligomerisation event or fibril formation. One way in which clusterin may inhibit amyloid formation would be to stabilise the natively unfolded conformation of proIAPP. To investigate this possibility a far-UV CD time course of proIAPP in the presence and absence of clusterin was carried out. Since clusterin exerts its effects on proIAPP at substoichiometric ratios of clusterin to proIAPP, concentrations of clusterin were used that would not alter the CD signal due to proIAPP. 100 µg/mL proIAPP in 10 mM phosphate, pH 7.4 was incubated in the presence and absence of 1, 10 and 20 µg/mL clusterin. These concentrations correspond to molar ratios of clusterin:proIAPP of 1:100 to 1:50, which are within the range over which clusterin exerts its effects on proIAPP amyloid formation (Figure 5.2).
Figure 5.12 The aging of proIAPP in the presence and absence of clusterin followed by far-UV CD spectroscopy. Spectra are 100 µg/mL proIAPP in 10 mM phosphate, pH 7.4 in the absence of clusterin (red trace) and in the presence of clusterin at 1 µg/mL (blue trace), 10 µg/mL (green trace) and 20 µg/mL (black trace) taken at 0 h (A) and day 3 (B). Spectra were taken as described in Section 2.2.15.
At 0 h, proIAPP in the absence of clusterin, was predominantly random coil, seen by the strong minima at 200 nm (Figure 5.12A) as previously discussed in Section 4.2.1. The spectra of proIAPP in the presence of clusterin at 0 h were indistinguishable from those of proIAPP alone (Figure 5.12A). By day 3 the spectra of proIAPP alone and proIAPP with 1 µg/mL clusterin were still indicative of random coil but the intensity of the signal had diminished (Figure 5.12B). The samples containing proIAPP with 10 µg/mL and 20 µg/mL clusterin had lost CD signal completely. This indicates that clusterin did not stabilise the natively unfolded conformation of proIAPP. Complete loss of CD signal had been previously observed when proIAPP alone was aged in the absence of clusterin after 12 days (Figure 4.4). The loss of CD signal for proIAPP in the presence of 10 µg/mL and 20 µg/mL clusterin is likely to be due to oligomerisation or fibril formation of proIAPP as discussed in Section 4.2.2. This suggests that clusterin does not inhibit the conversion of proIAPP to an amyloidogenic intermediate and may even promote this conversion. The inability of clusterin to prevent the conversion of proIAPP to an amyloidogenic intermediate is consistent with the ThT assay results, showing that although clusterin suppresses amyloid formation it is unable to prevent the initiation of the growth phase or extend the lag phase of amyloid formation (Figure 5.2).
5.3.9The formation of complexes between clusterin and proIAPP
As described above, the interaction between proIAPP and clusterin was studied using direct ELISA, competition ELISA and CD spectropolarimetry. The direct ELISA showed that clusterin binds to proIAPP at pH 7.4 (Figure 5.4), the pH at which inhibition of amyloid formation was observed by ThT assay. However the direct ELISA experiments do not give an indication of stoichiometry nor whether clusterin binds to all proIAPP species or specifically to certain species. The competition ELISA experiments shows that the interaction between clusterin and soluble proIAPP changes during the lag phase of amyloid formation (Figure 5.11). The enhancement of clusterin binding to denatured protein targets mediated by proIAPP at pH 7.4 masks any possible reduction in clusterin binding to the denatured proteins that could be used to deduce the formation of proIAPP-clusterin complexes. The CD time course in the presence and absence of clusterin show that clusterin does not stabilise the natively unfolded structure of proIAPP (Figure 5.12) however it does not provide any direct evidence for a clusterin proIAPP complex. As the previous experiments showed that there was an interaction between clusterin and proIAPP (Figure 5.4) and that this interaction changed depending on the proIAPP species (dependant on the stage of amyloid formation) (Figure 5.11), we decided to attempt to obtain more direct evidence for the existence of proIAPP-clusterin complexes.
Zone electrophoresis was used in attempts to detect putative soluble proIAPP-clusterin complexes. Zone electrophoresis separates on the basis of native charge and size. Clusterin, (pI ~5) and proIAPP, (pI ~9.5) are negatively and positively charged respectively at pH 8.0. Any complex of the two would have a reduced net charge and the migration of which would be retarded relative to clusterin or proIAPP alone. Immunoblotting for clusterin in soluble and insoluble fractions of proIAPP and clusterin mixtures was used to assess whether clusterin was forming complexes with fibrillar proIAPP (Section 5.3.10).
Mixtures of 500 µg/mL proIAPP containing clusterin at concentrations of up to 100 µg/mL were incubated at 37 oC for 144 h to allow amyloid formation to occur. At various times during this incubation aliquots were removed. Amyloid formation was measured with ThT assays (Figure 5.13). Samples were also taken for analysis by zone electrophoresis or immunoblotting for clusterin in soluble and insoluble fractions.
The ThT assay time course (Figure 5.13) shows a similar pattern to that seen previously (Figure 5.2). Amyloid formation was preceded by a lag phase, which in this case lasted 24 h, after which amyloid became detectable by ThT. Consistent with Figure 5.2, clusterin inhibited amyloid formation in a dose-dependent manner.
Figure 5.13 The effect of clusterin and ovalbumin on proIAPP amyloid formation. 500 µg/mL proIAPP was incubated in the absence of clusterin (red) and in the presence of 12.5 µg/mL (blue), 25 µg/mL (green), 50 µg/mL (black) and 100 µg/mL clusterin (pink). 500 µg/mL proIAPP was incubated with 100 µg/mL ovalbumin (yellow), as a control for a non-specific effect of protein on amyloid formation. Samples were taken through out the time course for analysis by zone electrophoresis (Figure 5.14) and dot-blotting (Figure 5.15). Amyloid formation was monitored by the increase in ThT, as described in Section 2.2.16b.
Zone electrophoresis was used to search for putative clusterin-proIAPP complexes as high molecular weight complexes between clusterin and stressed proteins have been detected using this method (Humphreys et al., 1999). Samples were mixed with non-denaturing loading dye and separated on a 1% (w/v) agarose gel in TAE buffer at 60 V for 1 h. Capillary transfer was used to transfer the separated proteins to a PVDF membrane. The proteins were stained with amido black, a general protein stain, and the clusterin detected by immunoblotting. No complex between clusterin and proIAPP could be detected even after amyloid started accumulating in proIAPP and clusterin mixtures (Figure 5.14). This suggested that clusterin does not form complexes with bulk proIAPP.
The migration of proIAPP was the same regardless of whether it had or had not been incubated with clusterin up to 48 h. By 72 h the majority of proIAPP in the absence of clusterin, had converted to amyloid and was not detectable by amido black staining (Figure 5.13). However proIAPP in the presence of clusterin was detectable at 72 h consistent with the ThT assay showing that in the presence of clusterin, amyloid formation was suppressed. The migration of proIAPP changed with incubation time. The distance that proIAPP migrated was compared to the distance that clusterin migrated in the opposite direction. At 0 and 28 h, proIAPP migrated as a discrete band approximately 75% the distance of clusterin (Figure 5.14A and B). This is likely to be the distance proIAPP monomer migrates as CD analysis of freshly dissolved proIAPP is of a random coil. At 48 h proIAPP both in the presence and absence of clusterin, migrated as a more diffuse band indicating that the size distribution of proIAPP was becoming more heterogeneous (Figure 5.14C). At 48 h the outer boundary of proIAPP migrated approximately 75% of the distance of clusterin but the inner boundary migrated approximately 60% of the distance (Figure 5.14C). This was not due to proIAPP degrading as SDS-PAGE analysis of proIAPP showed that proIAPP was the full-length size, 7.5 kDa (Figure 4.8). The change in mass to charge ratio indicates that proIAPP oligomerised. By 72 h the majority of proIAPP in the absence of clusterin had converted to amyloid and soluble proIAPP could not be detected by zone electrophoresis.
Figure 5.14 Zone electrophoresis of clusterin and proIAPP mixtures. Samples of clusterin and proIAPP mixtures were taken at 0 (A), 24 (B), 48 (C) and 72 h (D) and analysed by zone electrophoresis on a 1% (w/v) agarose in TAE gel. Lanes contained: 500 µg/mL proIAPP in the absence of clusterin (lane 1) or 500 µg/mL proIAPP in the presence of clusterin at 12.5 µg/mL (lanes 2 and 7), 25 µg/mL (lanes 3 and 8), 50 µg/mL (lanes 4 and 9) and 100 µg/mL (lanes 5 and 10). Lanes 6 and 11 contain 100 µg/mL clusterin in the absence of proIAPP. After zone electrophoresis, capillary transfer was used to transfer the proteins to a PVDF membrane. Lanes 1-6 were stained with amido black. Lanes 7-11 were blotted for clusterin as described in Section 2.2.24.
At 72 h proIAPP in the presence of clusterin migrated as oligomers as observed at 48 h. The inability of clusterin to inhibit proIAPP oligomerisation is consistent with the far-UV CD showing that clusterin did not stabilise the natively unfolded structure of proIAPP (Figure 5.12) and the ThT assay showing that clusterin could not prevent nucleation of amyloid formation (Section 5.3.1). However this data is consistent with clusterin maintaining proIAPP in a soluble form, consistent with the ThT assay showing a dose-dependent inhibition of amyloid formation in the presence of clusterin.
The absence of detectable clusterin-proIAPP complexes does not exclude the possibility that a small proportion of clusterin may be forming complexes with a small proportion of proIAPP that is below the limit of detection by this method. The limit of detection of clusterin using this method is below 12.5 µg/mL. This idea would be consistent with clusterin forming complexes with the nuclei, which is likely to be a small proportion of the oligomers, or short fibrils and inhibiting elongation.
5.3.10 Dot-blotting for clusterin in soluble and insoluble fractions of proIAPP and clusterin mixtures
Although the failure to detect putative clusterin complexes with nuclei or short fibrils may be due to a lack of sensitivity of the amido black and clusterin immunoblotting techniques used, another reason may be that these proIAPP species are insoluble and unable to enter the zone electrophoresis gel. To investigate this possibility, possible interactions between clusterin and insoluble (fibrillar) proIAPP was investigated by dot-blotting for clusterin in soluble and insoluble fractions of proIAPP and clusterin mixtures. The samples were then boiled in SDS-PAGE loading buffer and dotted on to a nitrocellulose membrane and immunoblotted for as described in Section 2.2.25. The intensity of the dots was obtained by quantitiative densitometry as described in Section 2.2.25. Dot-blotting was used instead of Western blotting because, as shown previously, the fibrils do not completely dissociate to monomer in SDS-loading resulting in aggregates that do not enter the resolving gel (Figure 4.8).
Clusterin was present in the insoluble fraction at 48 h when amyloid first became detectable by ThT although very little amyloid was present at 48 h in samples containing clusterin (Figure 5.13). In samples containing 12.5 µg/mL and 25 µg/mL clusterin, this was approximately half of the total clusterin being present in the insoluble fraction at 48 h and by 96 h all the clusterin has been removed from the soluble phase (Figure 5.15A and B). In proIAPP mixtures containing 50 and 100 µg/mL clusterin, approximately a quarter of the clusterin was in the insoluble fraction at 48 h and this proportion increased slightly but at 144 h clusterin was still present in soluble fraction (Figure 5.15A and 5.16C and D). The inhibition of amyloid formation by clusterin was correlated with the amount of clusterin present in the soluble phase.
(Figure previous page)
Figure 5.15 Dot-blotting for the presence of clusterin and ovalbumin in soluble and fibrillar fractions during proIAPP amyloid formation. ProIAPP was incubated in the presence of 12.5, 25, 50 and 100 µg/mL clusterin (A) and 100 µg/mL ovalbumin (B) for 144 h. Amyloid was detectable by 48 h by ThT (figure 4.4.10) and mixtures were separated by centrifugation into soluble and insoluble (fibrillar) fractions. Samples were mixed with SDS-PAGE loading dye and boiled for 10 min. Samples were then allowed to adsorb onto a nitrocellulose membrane and clusterin or ovalbumin were detected as described in Section 2.2.24. 5 µL samples were blotted onto nitrocellulose membrane with the exception of 500 µg/mL proIAPP and 12.5 µg/mL clusterin (15 µL), and 500 µg/mL proIAPP and 25 µg/mL clusterin (10µL). Dot blots for 12.5 µL clusterin in the absence of proIAPP (15 µL) were carried out and indicated that this level of clusterin was detectable by this method (data not shown).
Figure 5.16 Densitometric analysis of dot blot for clusterin and ovalbumin. The intensity of the dots (Figure 5.15) corresponding to total (red), soluble (blue) and insoluble (green) were obtained by quantitative densitometry using Quantity One Software (BioRad) as described in Section 2.2.25. The mixtures were 12.5 µg/mL clusterin and 500 µg/mL proIAPP (A), 25 µg/mL clusterin and 500 µg/mL proIAPP (B), 50 µg/mL clusterin and 500 µg/mL proIAPP (C), 100 µg/mL clusterin and 500 µg/mL proIAPP (D) and 100 µg/mL ovalbumin and 500 µg/mL proIAPP (E).
The presence of clusterin in the insoluble fraction could be due to clusterin being non-specifically trapped in proIAPP precipitating as amyloid or related to the inhibitory role of clusterin. Ovalbumin was used as a negative control for non-specific trapping of clusterin in insoluble material as it is not known to have a role in amyloid formation. TEM analysis of proIAPP and ovalbumin incubated for 10 days had shown similar amounts of amyloid were formed in the presence of ovalbumin (Figure 5.2H) compared to proIAPP alone (Figure 5.2A), indicating that ovalbumin was unlikely to suppress amyloid formation. 500 µg/mL proIAPP was also incubated with 100 µg/mL ovalbumin and the amyloid formation measured by ThT assay (Figure 5.13). In the presence of 100 µg/mL ovalbumin, amyloid formation went to completion at 120 h (Figure 5.13) compared to 96 h for proIAPP alone (Figure 5.13). However the effect of 100 µg/mL ovalbumin was insignificant compared to the suppression of amyloid formation by 100 µg/mL clusterin. Aliquots of proIAPP and ovalbumin were processed and immunoblotted in the same way as for clusterin.
Blotting for ovalbumin in soluble and insoluble fractions of ovalbumin and proIAPP mixtures show that minimal ovalbumin is detectable in the insoluble fraction and that the majority of ovalbumin is in solution at the end of 144 h (Figure 5.15 B and Figure 5.16E). This suggests that the presence of clusterin in the insoluble fraction is not a by-product of proIAPP precipitating out of solution but is related to the inhibitory effect of clusterin.
The intensity for total sample in proIAPP mixtures containing clusterin decreases with incubation time but not for mixtures containing ovalbumin. Clusterin may be more prone than ovalbumin to non-specific adsorption to the tube, in which the solutions are incubated, due to its chaperone activity.
The proIAPP species that clusterin binds are likely to be large insoluble oligomers or short fibrils, with patches of exposed hydrophobic surface, a major way in which molecular chaperones recognise non-native conformations. This is consistent with the increased exposed hydrophobic surface of proIAPP, measured by bisANS fluorescence that occurs soon after amyloid becomes detectable (Figure 4.9).
5.3.11 Binding of clusterin to mature proIAPP fibrils
If clusterin is able to recognise and bind exposed hydrophobic patches on forming fibrils such as fibril ends, clusterin should be able to bind to mature fibrils. Mature fibrils were formed by incubating 1 mg/mL proIAPP for a week at 37 oC. Clusterin was incubated with 125 µg/mL, 250 µg/mL and 500 µg/mL proIAPP fibrils formed in the absence of clusterin for 15 minutes and then separated into soluble and insoluble fractions by centrifugation. The presence of clusterin in total, soluble and insoluble fractions was determined by blotting for clusterin (Figure 5.16A). The binding of clusterin to preformed proIAPP fibrils is seen as an increase in clusterin co-localised with fibrillar proIAPP (Figure 5.16A and 5.17A). This is also seen by the concurrent decrease in clusterin remaining in the soluble fraction.
The binding of ovalbumin to fibrillar proIAPP was also measured as a control. Approximately 80% ovalbumin remained in the soluble fraction regardless of the concentration of fibrillar proIAPP (Figure 5.17B). The amount of ovalbumin present in the insoluble fraction did not increase with fibrillar proIAPP concentration, indicating that ovalbumin did not bind to fibrillar proIAPP and was not non-specifically trapped in the insoluble material (Figure 5.15 B and Figure 5.16B).
Figure 5.16Clusterin binds pre-formed fibrillar proIAPP. Fibrillar proIAPP was formed by the incubation of proIAPP at 37 oC for a week. 125, 250 and 500 µg/mL fibrillar proIAPP was incubated with 25 µg/mL clusterin (A) or ovalbumin (B) for 15 min at 37 oC. The mixtures were separated into soluble and insoluble fractions by centrifugation at 13000 g for 15 min. Total (uncentrifuged), soluble and insoluble fractions were boiled in SDS-PAGE loading dye and blotted on to a nitrocellulose membrane as described in Section 2.2.25. Clusterin and ovalbumin were detected as described in Section 2.2.25.
Figure 5.17 Amount of clusterin and ovalbumin bound to pre-formed fibrillar proIAPP. The intensity of the dots (Figure 5.16) was obtained by quantitative densitometry using Quantity One software (BioRad). The clusterin (A) or ovalbumin (B) present in the soluble (red) and insoluble (blue) fractions were compared to the intensity of the total dots
5.3.12 Clusterin bound to proIAPP fibrils detected by immunoelectron microscopy
Immunoelectron microscopy was used to see where clusterin was bound on the proIAPP fibril. Mixtures of clusterin and proIAPP were adsorbed on to Formvar-coated copper grids. Clusterin was detected using a monoclonal antibody against clusterin and a secondary antibody conjugated to 10 nm gold particles, which could be seen by as black dots under TEM. Clusterin could be seen at the ends of the fibrils (Figure 5.18A, B and C). Clusterin could also be seen surrounding small spherical structures, which may represent amyloidogenic oligomers. Similar structures were seen in samples containing clusterin alone but these were less numerous (Figure 5.18G). The localization of clusterin is consistent with the predicted hydrophobicity of fibril ends (Figure 4.11). This may be functionally relevant as fibrils grow through elongation at fibril ends and represents one way that clusterin could inhibit the growth phase of amyloid formation. Clusterin could also be seen binding along the length of the fibril (Figure 5.18D and E).
Figure 5.18 ImmunoEM piccies