Methionine oxidation of clusterin in Alzheimer’s disease and its effect on clusterin’s binding to beta-amyloid

Abstract

Clusterin is a secreted glycoprotein that participates in multiple physiological processes through its chaperon function. In Alzheimer’s disease, the brain functions under an increased oxidative stress condition that causes an elevation of protein oxidation, resulting in enhanced pathology. Accordingly, it is important to determine the type of human brain cells that are mostly prone to methionine oxidation in Alzheimer’s disease and specifically monitoring the methionine-oxidation levels of clusterin in human and mice brains and its effect on clusterin’s function. We analyzed the level of methionine sulfoxide (MetO)-clusterin in these brains, using a combination of immunoprecipitation and Western-blott analyses. Also, we determine the effect of methionine oxidation on clusterin ability to bind beta-amyloid, in vitro, using calorimetric assay. Our results show that human neurons and astrocytes of Alzheimer’s disease brains are mostly affected by methionine oxidation. Moreover, MetO-clusterin levels are elevated in postmortem Alzheimer’s disease human and mouse brains in comparison to controls. Finally, oxidation of methionine residues of purified clusterin reduced its binding efficiency to beta-amyloid. In conclusion, we suggest that methionine oxidation of brain-clusterin is enhanced in Alzheimer’s disease and that this oxidation compromises its chaperon function, leading to exacerbation of beta-amyloid’s toxicity in Alzheimer’s disease.

Introduction

The brain during Alzheimer’s disease (AD) is under severe oxidative attack by reactive oxygen species that lead to methionine oxidation of proteins. Besides the oxidation of the sole methionine (Met35) of beta-amyloid (Aβ), oxidation of methionine residues of other extracellular proteins may be one of the earliest events contributing to the toxicity of Aβ and other proteins. In our recent published studies [1], we performed an active immunization of transgenic AD mice using a recombinant methionine sulfoxide (MetO)-rich protein as the antigen. This treatment induced the production of auto anti-MetO antibody and caused alleviation of key AD-related phenotypes [1]. One of these phenotypes was a reduction of total Aβ levels (presumably through clearance of the MetO-Aβ forms) in both blood-plasma and brains [1]. This data prompted us to investigate the possible clearance of other extracellular proteins by this immunization, in which their methionine residue/s oxidation may hamper their function, leading to an enhanced development of AD. One such extracellular protein is the heat-shock protein, clusterin (apolipoprotein J: ApoJ), that also functions as a chaperon and participates in restoring proteins folding, including Aβ. Clusterin is a 70-kDa component of high-density lipoproteins in human blood-plasma, comprises of two disulfide-linked subunits: alpha (34–36 kDa), and beta (36–39 kDa). It is highly expressed in the brain especially in the hippocampal region) and blood-plasma. Clusterin’s extracellular form is the most common type that exhibits a major chaperon function through binding to several target proteins. For example, clusterin has been linked to modulation of lipid metabolism and immune response as well as to oxidative stress related diseases [2]. Although the role of the intracellular form of clusterin is not clear, it is consistent with the view that has evaded the secretion pathway or is a result of re-entering of the secreted form into the cell [2]. In AD, clusterin seems to have an important role in modifying the aggregation of Aβ and fostering its clearance. This function of clusterin suggests a neuroprotective role of the protein against AD [3–11]. However, other researchers show that clusterin may reduce the clearance of Aβ [3, 12–15] and may promote Aβ-induced neurotoxicity [16,17]. Thus, it has been suggested that outcome of the interaction between Aβ and clusterin depends on the clusterin:Aβ ratio [18,] and the factor in excess might determine whether clusterin exhibits neuroprotective or neurotoxic effects. Genetically, polymorphism variants of the ApoJ gene are the third most significant genetic risk factors for late onset of AD, depending on the origin of the tested population [19]. The evident complexity of clusterin’s physiological and pathological functions contributes to the difficult task of understanding the nature of the factors that determine the role of clusterin in AD pathophysiology. MetO-Clusterin and its relation to Aβ and oxidative stress affecting the development of AD. The oxidation of clusterin protein in general and specifically of its methionine residues have not been studied, if at all, according to the published literature. Oxidation of the hydrophobic amino acid residue, methionine to MetO can cause a conformational change that may hamper the targeted protein function. The protein sequence of clusterin contains a 3.6% as methionine, a level that is significantly higher than the common average of 2.0% in a typical protein. Moreover, most of the methionine residues of clusterin are located towards the C-terminus of the protein, a feature that increases the chances of having MetO patches upon methionine oxidation. Accordingly, oxidation of specific methionine residue/s of clusterin may interfere with its chaperon function, leading to compromised biological performances of its binding partners (like aggregated Aβ). In the current studies we examined the possibility of having a correlation between the levels of MetO-Clusterin and AD as judged by relevant analyses of postmortem human and mouse brains. We further hypothesized that methionine-oxidation of clusterin negatively affects its ability to interact with Aβ, as judged by relevant in vitro assays.

Materials and Methods

Antibodies

Anti-Neun antibody was purchased from BioLegend (San Diego, CA; USA). Anti-Clusterin antibody was purchased from GeneTex (Irvine, CA, USA). Anti-mouse/rabbit IgG HRP-conjugated antibody was purchased from Santa-Cruz Biotechnology (Santa Cruz, CA, USA). Anti-rabbit/mouse fluorescent-labeled antibody was purchased from Thermo-Fisher Scientific (Wathham, MA, USA). The rabbit anti-MetO antibody was produced in our lab, as previously described [1]. Mouse GFAP Monoclonal Antibody (ASTRO6) was purchased from Thermo-Fisher Scientific.

Postmortem human and mouse brains

We have obtained postmortem human brains from the NIMH Neurobiobank that has provided us with five postmortem human AD brains (hippocampus region at late stages of AD) and comparable five control non-AD postmortem patients (all brain specimens were from age-matched subjects). AD-model mice (5xFAD (B6SJL-Tg(APPSw-FlLon,PSEN1*M146L*L286V)6799Vas/Mmjax) and their corresponding controls were purchased from Jackson’s labs. All animals were housed in the animal care facility of the University of Kansas under pathogen-free conditions. No more than five animals were present in one cage, where they were housed under a 12-h light and dark cycle with free access to food and water. The animals were euthanized as needed in accordance with the American National Institute of Health’s animal care (all procedures were approved by the University of Kansas’ animal care unit-approved protocol). Accordingly, postmortem mouse brain samples of early stages of AD and their controls (4 months of age; n=5 per strain) and of late stages of AD (8 months of age; n=5 per strain) were processed.

Immunohistochemistry analyses for MetO-proteins and neurons in postmortem AD-human brains

Each postmortem human brain was fixed in 4% paraformaldehyde in PBS at 4° C and then flash-frozen by exposure to dry ice. Coronal sections (30 μm) were cut using a cryotome and transferred to gelatin-coated glass slides. For immunohistochemistry staining, sections were blocked with 3% (w/v) gelatin in PBS (1 h at 37 °C), treated with 0.1% Triton X-100 in PBS (15 min at 23 °C), and reacted (overnight at 4 °C plus 1 h at 23°C) with the selected primary antibody. After rinsing in PBS, the sections were incubated (2 hr at room temp) with either primary antibody-matched fluorescent dye-labeled secondary antibodies (Alexa 568 goat anti-rabbit or Alexa 488 goat anti-mouse; Thermo Fisher Scientific) or HRP labeled secondary goat anti-mouse/rabbit antibody (Santa-Cruz Biotechnology). Then, the sections were rinsed with PBS followed by Tris-EDTA buffer, mounted, viewed, and analyzed by using a light microscope.

Immunoprecipitation and Western blot analyses for the detection of clusterin and MetO-Clusterin levels in postmortem human and mouse brains

To determine the levels of MetO-Clusterin and total clusterin in the extracellular fraction of postmortem human brains, isolation of extracellular MetO-proteins the protein extraction was performed, as previously described [20]. Briefly, whole brains were washed five times with icecold saline containing 0.32M sucrose (buffer A), 1mM calcium acetate (pH, 7.4), and protease inhibitors cocktail (Sigma-Aldrich, St. Louis, MO). Thereafter, the brain tissues were incubated in buffer A for 4 hr at 0°C with gentle agitation to facilitate the extraction process. At the end of the incubation period, the brains were separated from the medium by filtration through a fine-meshed gauze. Following high-speed centrifugation (10,000 rpm for 10 min), the soluble fraction was concentrated by 3000 Da cut-off centricon (Thermo-Fisher Scientific, Waltham, MA) and protein was measured using the Bio-Rad protein assay (Bio-Rad, Hercules, CA). Equal amounts of the resulting extracellular protein moiety from each brain were subjected to immunoprecipitation (IP), using the rabbit anti-MetO antibody [1]; followed by western-blot (WB) analysis, using a primary mouse anti-human clusterin antibody (Proteintech, Rosemont, IL). The acquired signal level from the WB that follows the IP procedure was divided by the relative total expression level of clusterin, as determined by the WB analysis using the same anti-human clusterin antibody.

To investigate whether methionine oxidation of clusterin is also prevalent in mouse models of AD, we have performed the same analyses for determining the levels of MetO-Clusterin in whole brains of the 5xFAD (B6SJL-Tg(APPSwFlLon,PSEN1*M146L*L286V)6799Vas/Mmjax) model mice and their corresponding controls. Accordingly, postmortem mouse brain samples of early stages of AD and their controls (4 months of age; n=5 per strain) and of late stages of AD (8 months of age; n=5 per strain) were processed as follows. Whole brains were extracted in PBS in the presence protease inhibitors (Sigma-Aldrich, St. Louis, MO). The resulting extracted protein moiety contains both cellular and extracellular proteins in each brain. Equal amounts of protein extracts per brain were subjected to immunoprecipitation (IP), using the rabbit anti-MetO antibody [1]; followed by WB analysis, using a primary goat anti-mouse clusterin antibody (Proteintech, Rosemont, IL).

Competitive Binding capability of active recombinant clusterin (ApoJ) and MetO-Clusterin to aggregated Aβ

Aβ₄₂ was commercially obtained from Echelon Biosciences (Salt Lake City, UT). An active pure clusterin was commercially obtained from BioLegend (San Diego, CA). Methionine oxidation of the recombinant ApoJ protein was performed by incubating the protein with 50mM H₂O₂ for 3 hr at room temperature. The residual H₂O₂ was quenched by multiple washes in PBS and concentration using an Amicon Ultra-0.5 Centrifugal Filter Unit (3000 Da cut-off). The oxidation of the methionine moiety of clusterin was confirmed by both dot-blot and western blot analyses using the anti-MetO antibody. Mass spectrometry analysis confirmed that only the methionine residues were oxidized (data not shown). Aggregation kinetics of Aβ₄₂ were followed by using an in situ Thioflavin-T (ThT) fluorescence assay (with minor modifications) [2], based on the increase of the fluorescence signal of ThT when bound to sheet-rich structures. We used a Aβ₄₂ at concentration of 5.0μM and recombinant native and MetO-Clusterin concentrations of 0.3 μM using this assay. Accordingly, freshly prepared Aβ₄₂ was diluted at the indicated concentration in 20 mM phosphate buffer (pH 8.0) containing 0.02% NaN₃ and 200 μM EDTA.

Then the Aβ₄₂ was incubated, in the presence and absence of active recombinant human clusterin or MetO-Clusterin, under quiescent conditions at 37 °C in microplate wells (Corning Inc. Life Sciences, Acton, MA) in the presence of 20 μM ThT (100 μl solution/well). ThT fluorescence was measured every 4.0 min for 18 hr, using a Biotek Synergy HT Microplate Reader (Marshal Scientific, Hampton, NH). A parallel analysis was performed using BSA instead clusterin, at the indicated concentration, serving as a non-chaperone control protein. The ThT dye was excited at 440 nm, and the emission was measured at 495 nm. Fluorescence data are reported after normalization on the corresponding maximal ThT value after subtracting ThT basal background signal value.

In-silico Clusterin sequences alignment and hydrophobicity

We obtained and aligned the Clusterin sequences for the indicated species in UniProt. The aligned sequences were imported in Jalviewer to color code for hydrophobicity.

Results

Protein-MetO in cells of postmortem AD- human brains

Post-mortem human brains were acquired from the NIMH Neurobiobank and the relevant information regarding age, sex, and AD-related pathology is described in Table 1. The main goal of the detection of MetO-containing proteins in these brains is to determine the type of cells that are enriched with MetO-proteins in AD. The selected brains were from subjects that were clinically diagnosed with AD (while they were alive) and this diagnostic was verified by the presence molecular markers of AD in their postmortem brains (as described in Table 1). We applied the anti-MetO antibody to detect MetO-proteins in immunohistochemistry analyses of these brains, showing that neurons and astrocytes in the hippocampal region mainly exhibit MetO-proteins compared to other types of cells. As shown in Figure 1A, neurons of the hippocampus contain MetO-proteins mainly in the nucleus region, as detected with the primary anti-MetO antibody (emitting a red fluorescence). Validation of this observation is provided by the immunostaining of the brain region with the primary antibody against Neun antibody (emitting a green fluorescence, a marker for neuronal detection) and merging it with the yellow signal emitted by the anti-MetO antibody staining created a orange or red signal (Figure 1B). In addition, staining with the anti-MetO antibody of the hippocampus shows that astrocytes contain relatively high levels of MetO-proteins, as judged by the common structure that is typically observed for astrocytes (Figure 1C) that was confirmed by the staining with anti-GFAP antibody (a protein marker for astrocytes) (Figure 1D). The main objection of this these analyses is to show that in human AD there are brain cells that are enriched with MetO-proteins and these proteins may also be present in a non-AD brain as well. However, to follow with accuracy reliability on the correlation between MetO-proteins to an AD disease stage, monitoring the methionine oxidation of a single protein target at a time is preferred (e.g., clusterin as shown in Figures 2 & 3).

Table 1. Postmortem human brains description.

The Table contains information about the postmortem brains with respect to the subjects age, sex, and neurological assessment.

Age (years)	Sex	Neuropathology changes that are associated with Alzheimer’s disease (AD)
89	Female	Cerebral atherosclerosis (sever); Multiple cerebral and cerebellar infarcts; Braak’s stage: IV-V.
81	Female	AD neurophatological changes; Thal stage V for beta-amyloid deposition; Braak’s Baark’s tangle stage: IV-V; CERAD age-related plaque score: C; ABC score: A3B3C3 (NIA-AA criteria)- high probability of dementia.
91	Male	Thal stage V for beta-amyloid deposition; Braak’s Baark’s tangle stage: IV-V; CERAD age related plaque score: B; ABC score: A3B3C2 (NIA-AA criteria)-high probability of dementia.
74	Female	AD neurophatological changes (severe); Thal stage IV for beta-amyloid deposition; Braak’s Baark’s tangle stage: IV-V; ABC score: A3B3C3 (NIA-AA criteria)- high probability of dementia.
74	Male	AD neurophatological changes (severe); Thal stage IV for beta-amyloid deposition; Braak’s Baark’s tangle stage: VI; CERAD age-related plaque score: Frequent; ABC score: A3B3C3 (NIA-AA criteria)-high probability of dementia.
		Note: The control postmortem subjects were non-AD subjects with the following specification: two males and three females with ages that were within a close match to the age of the AD-subjects. The cause of death was not related to AD and thus was not included.

Figure 1. Accumulation of methionine-sulfoxide containing proteins in postmortem AD human brains.

The Hippocampal region of postmortem brains was processed and analyzed using immunohistochemistry methods, as described under “Materials and Methods”. A. Detection of MetO-proteins using the anti-MetO antibody in neurons (red-color signal). B. Detection of MetO-proteins in neurons as described in A (yellow-color) along with the detection of neurons using anti-Neun antibody (green-color signal) and superimposed of the two signals (resulting in orange or red colors signals). C. Detection of MetO-proteins in astrocytes (green-color signal). D. Detection of GFAP protein (a protein marker of astrocytes) (red-color signal). Triangle arrows are pointed towards a single body cell of each selected astrocyte.

Figure 2. MetO-Clusterin levels are predominantly expressed in postmortem brains of late-stage AD patients.

A. Postmortem human brain samples of late stages of AD (hippocampus region, n=5) and at late stages of non-AD brains (hippocampus region, n=5) were incubated in PBS in the presence of 0.32M sucrose and protease inhibitors (Sigma-Aldrich) for 4 hours at 4°C. Equal amounts of the resulting extracellular protein moiety from each brain were subjected to immunoprecipitation (IP), using the rabbit anti-MetO antibody; followed by western-blot (WB) analysis, using a primary mouse anti-human clusterin antibody (Proteintech). lanes 1–3, loads of one processed extracted brain per each lane (representatives out of the total brains’ number).

B. Western blot analysis of extracellular brain protein extracts obtained for the brains described in panel A (using the same mouse anti-human clusterin antibody). Equal amounts of protein extracts were loaded per each lane.

C. Quantification of the clusterin and MetO-Clusterin bands shown in the WB in Figure 2 (using the NIH-Image-J program) showed that the AD brains had about 3.5-fold higher MetO-Clusterin levels relative to the non-AD brains (*, t-test, P<0.001) after correction for the total expressed levels of clusterin in panel B.

Figure 3. MetO-Clusterin levels are predominantly expressed in postmortem AD-model mouse brains as a function of age.

A. Postmortem mouse brain samples at 4 months of age (Lanes 1 & 3) and at 8 months of age; Lanes 2 & 4). Lanes 3 & 4 represent protein loads of 5xFAD mouse model of AD and Lanes 1 & 2 represent protein loads of the corresponding controls. Whole brains (n=5 per strains per age) were extracted in PBS in the presence of protease inhibitors (Sigma-Aldrich). The resulting extracted protein moiety contains both cellular and extracellular proteins in each brain. Equal amounts of protein extracts per brain were subjected to immunoprecipitation (IP), using the rabbit anti-MetO antibody; followed by western-blot (WB) analysis, using a primary goat anti-mouse clusterin antibody (Proteintech). The WB depicts representative data for one brain per strain per age. kDa, molecular mass markers.

B. Western blot analysis of whole brain protein extracts obtained for the brains described in panel A (using the same goat anti-mouse clusterin primary antibody). (Lanes 1 & 3) and at 8 months of age; Lanes 2 & 4). Lanes 3 & 4 represent protein loads of 5xFAD mouse model of AD and Lanes 1 & 2 represent protein loads of the corresponding controls. The WB depicts representative data for one brain per strain per age (total n=5 per strainper age). The β-actin expression levels serve as loading controls (using anti-β-actin primary antibody, Thermo-Fisher Scientific). kDa, molecular mass markers.

C. Quantification of the clusterin and MetO-Clusterin bands shown in the WB in Figure 3 A&B (using the NIH-Image-J program). The data showed that the 5xFAD brains had higher MetO-Clusterin levels relative to the non-AD brains (*, t-test, P<0.001) after correction for the total expressed levels of clusterin in panel B.

MetO-Clusterin levels in postmortem human and mouse AD-brains compared to controls Postmortem human brains

To determine the presence and relative levels of MetO-Clusterin in comparison to total clusterin in postmortem human brains (at late stage of AD compared to non-AD brains), we have performed the following experiments. We have obtained five postmortem human AD brains (hippocampus region at late stages of AD) and comparable five control non-AD postmortem patients (all brain specimens were from age-matched subjects). We first determine the levels of MetO-Clusterin and total clusterin in the extracellular fraction of the brains. Accordingly, the isolation of extracellular MetO-proteins the protein extraction was performed, as described under Materials and Methods section. Equal amounts of the resulting extracellular protein moiety from each brain were subjected to immunoprecipitation (IP), using the rabbit anti-MetO antibody [1]; followed by western-blot (WB) analysis, using a primary mouse anti-human clusterin antibody (Proteintech, Rosemont, IL). The acquired signal level from the WB that follows the IP procedure was divided by the relative total expression level of clusterin, as determined by the WB analysis using the same anti-human clusterin antibody. As shown in Figure 2A, the resulting data show that the levels of MetO-Clusterin are significantly higher in the human AD brains compared to controls. To determine the total expressed levels of clusterin in these brains, WB analysis was performed on the brain extracts obtained for the brains described in Figure 2A, and the results are depicted in Figure 2B. Accordingly, the data show no significant difference in the expression of clusterin between the AD and the Non-AD brains (Figure 1B). Quantification of the clusterin and MetO-Clusterin bands shown in the WB of Figure 2A (using the NIH-Image-J program) showed that the AD brains had ~3.5-fold higher MetO-Clusterin levels relative to the non-AD brains (t-test P<0.001, after adjustment for the total expressed levels of clusterin (Figure 2C)). Thus, we have concluded that oxidation of the methionine moiety of clusterin occurs dominantly in the late stage of AD in human brain, when compared to non-AD control brains. It is yet to be determined how this phenomenon corelates with earlier stages of AD.

Postmortem mouse brains

To investigate whether methionine oxidation of clusterin is also prevalent in mouse models of AD, we have performed the same analyses for determining the levels of MetO-Clusterin in whole brains of the 5xFAD model mice and their corresponding controls. Accordingly, postmortem mouse brain samples of early stages of AD and their controls (4 months of age; n=5 per strain) and of late stages of AD (8 months of age; n=5 per strain) were processed as described under “Material and Methods” section. The resulting extracted protein moiety contains both cellular and extracellular proteins in each brain. Equal amounts of protein extracts per brain were subjected to immunoprecipitation (IP), using the rabbit anti-MetO antibody [1]; followed by WB analysis (Figure 3A), using a primary goat anti-mouse clusterin antibody (Proteintech, Rosemont, IL). As shown in Figure 3A, the resulting data show that the levels of MetO-Clusterin are significantly higher in the 5xFAD brains compared to controls at 4 & 8 months of age (Figure 3A). To determine the total expressed levels of clusterin in these brains, WB analysis was performed on the brain extracts obtained for the brains described in Figure 3A and the results are depicted in Figure 3B. Accordingly, the data showed no significant difference in the expression of clusterin between the AD and the control mouse brains (Figure 3B). Quantification of the clusterin and MetO-Clusterin bands displayed in the WB of Figure 3 (using the NIH-Image-J program) showed that indeed the 5xFAD brains had 2.75 and 3.25-fold higher MetO-Clusterin levels in the 4-months and 8-months old brains, respectively, relative to the control brains (t-test P<0.001; these data were adjusted to the total expressed levels of clusterin (Figure 3C)). Thus, we have concluded that oxidation of the methionine moiety of clusterin occurs dominantly in the AD-model mouse brains, when compared to controls, and that these accumulation of MetO-Clusterin in the AD-model mice are seemingly age-dependent changes.

Effect of methionine oxidation of clusterin on its binding efficiency to Aβ₄₂ in vitro

Recombinant clusterin and its methionine-oxidized form were used to determine their ability to bind to Aβ42, following the relevant procedures described under the “Materials and Methods” section. The oxidation of the methionine moiety of clusterin was confirmed by both dot-blot and western blot analyses using the anti-MetO antibody, as the primary antibody (Figure 4A). Additionally, methionine residues were identified in the sequence of clusterin as being oxidized following oxidation with hydrogen peroxide, using mass-spectrometry analyses of tryptic digested clusterin resulting peptides (non-oxidized clusterin had no oxidized methionine, respectively). It is important to note that this in vitro methionine oxidation determination was limited by the number of produced tryptic peptides and it was not verified in vivo. Aggregation kinetics of Aβ₄₂ were followed by using an in situ ThT fluorescence assay, based on the increase of the fluorescence signal of ThT when bound to sheet-rich structures. Figure 4B shows that Aβ₄₂ has aggregated over-time and that addition of BSA to its reaction mixture has not reduced the rate of aggregation nor its maximum value. Moreover, it increased the maximum value of the aggregation by an average of 17% (a yet to be explained phenomenon). However, in the presence of active recombinant clusterin, the slope-rate of Aβ₄₂ fibrillation was significantly lower compared with Aβ₄₂ or BSA alone, (represents a 100% fibrillation). Finally, the inhibition effect of MetO-Clusterin on Aβ₄₂ fibrillation was compromised compared to the non-oxidized clusterin, as it reduced the maximum level of Aβ₄₂ fibrillation only to 76%, relative to the non-oxidized clusterin (i.e., which reduced the Aβ₄₂ fibrillation to 49%) (Figure 4B). In summary, methionine oxidation of clusterin seems to reduce its ability to protect against Aβ₄₂ fibrillation in vitro. Moreover, the presented preliminary data in Figure 4 provides the foundation for further detailed investigations on the effect of methionine oxidation on the ability of clusterin to bind and affect the fibrillation of Aβ42.

Figure 4: Oxidation of active recombinant ApoJ and the binding kinetics of recombinant ApoJ (clusterin) and its oxidized form (MetO-ApoJ) to Aβ₄₂, in the presence of ThT.

A. Methionine oxidation of an active ApoJ protein (50μg) was performed by incubating the protein with 50mM H₂O₂ for 3 h at room temperature. The residual H₂O₂ was quenched by multiple washes in PBS and concentration using an Amicon Ultra-0.5 Centrifugal Filter Unit (3000Da cut-off). The oxidation of the methionine moiety of ApoJ was confirmed by both dot-blot (right panel) and western blot (left panel) analyses using the anti-MetO antibody (as the primary antibody) and 1μg of non-oxidized and oxidized ApoJ. N, non-oxidized ApoJ; Ox, oxidized ApoJ; kDa, molecular mass markers in kilo-Daltons. B. Effects of ApoJ and MetO-ApoJ on the kinetics of Aβ₄₂ fibril formation. Freshly prepared Aβ₄₂ was diluted at the indicated concentration in 20 mM phosphate buffer (pH 8.0) containing 0.02% NaN₃ and 200 μM EDTA. Then the Aβ₄₂ was incubated, in the presence and absence of active recombinant human ApoJ or MetO-ApoJ, under quiescent conditions at 37 °C in microplate 96-wells Microplate. in the presence of 20 μM ThT. ThT fluorescence was measured every 4.0 min for 18h, using an M200 Infinity plate reader. A parallel analysis was performed using BSA instead ApoJ, at the indicated concentration, serving as a non-chaperone control protein. The ThT dye was excited at 440 nm, and the emission was measured at 495 nm. Fluorescence data are reported after normalization on the corresponding maximal ThT value after subtracting ThT basal background signal value. The figure shows a representative data of three repetitive independent experiments. Abeta, Aβ₄₂.

In-Silico a cross-species sequence analyses for conserved methionine residues of clusterin and their possible effect on its structure

Since the location of the Met residues that may be oxidized in vivo is not known, we predict that oxidation of Met residues in the hydrophobic clusters identified in silico. Oxidation of methionine residues in proteins may affect their three-dimensional structure, leading to change of function. This phenomenon is driven by the idea that oxidation of the hydrophobic amino acid methionine will cause it to change its physical characteristics to a hydrophilic amino acid [21]. Given the negative effect of methionine oxidation on clusterin’s function (Figure 4), we used in-Silico methods to determine the location of the most conserved methionine residues among several species, while selecting the ones that are located within hydrophobic regions. Oxidation of these selected methionines is predicted to cause conformational changes, in which their severity is dependent on the hydrophobicity level within their location. Accordingly, the acquired data in Figure 5 show that there are five prominent conserved methionines within the sequence of clusterin, Met96, Met204, Met 205, Met455, and Met538, and that these methionines are located within hydrophobic clusters. Based on these data we hypothesize that the strongest effect of methionine oxidation on clusterin’s structure-function will occur when Met 204, or Met 205, or both are oxidized. The other three conserved methionines seem to have the same level of hydrophobicity clusters. Thus, their oxidation is predicted to have a similar effect on the structure-function of clusterin.

Figure 5. In-Silico analyses for the location of clusterin’s conserved methionine residues and their associated hydrophobicity regions.

The number indicates the position of a methionine residue in the sequence alignment. The intensity of the red color indicates the level of hydrophobicity, and the intensity of the blue color indicates the level of hydrophilicity (a denser color-coded amino acids corelates with either stronger hydrophobic or hydrophilic cluster, respectively). The Uniprot program and the align tool were used to acquire the clusterin sequences of the indicated organisms. The amino acid numbers are based on the longest Clusterin sequence in the indicated species.

Discussion

Methionine oxidation of proteins in post-mortem AD-human brains was previously described in the literature [22]. However, the published data lacked information about the type of cells harboring enhanced level of MetO-proteins within the hippocampus region. Here we showed that most MetO-protein accumulations in the human brain occur in neurons and astrocytes (Table 1 and Figure 1). Indeed, elevated levels of MetO-proteins were also detected in astrocytes of AD-model mouse brains [1], suggesting that this is a common feature across mammalian species. Interestingly, clusterin is one of the proteins that is produced and secreted from astrocytes upon a need to protect the brain from faulty formation of Aβ, through its chaperon function. Thus, it was expected that clusterin would serve as a target for methionine oxidation both in these cells and in the extracellular hippocampal region. The presented data in Figure 2 shows that in humans, the methionine oxidation of clusterin is dramatically enhanced in postmortem AD-brains compared to controls. This phenomenon strongly suggests that during the clusterin is prone to oxidation through the processes leading to AD. Corroboration for this interpretation is demonstrated by the data showing that in AD-model mice the methionine oxidation of clusterin is enhanced in an age-dependent manner, compared to controls (Figure 3). In vitro methionine oxidation of clusterin compromised its chaperon function on Aβ₄₂ {Figure 4). As far as we know, the effect of methionine oxidation on clusterin function has never been studied. Accordingly, this observation suggests that the oxidation of specific methionine residue/s of clusterin causes conformational alterations leading to a compromised chaperon activity. Indeed, In-silico analyses on the clusterin protein-sequence and hydrophobicity clusters showed that there are four methionine residues in which their oxidation could cause these suspected structural changes (Figure 5). It is proposed that replacement of the selected methionine residues with valines may protect clusterin from the observed oxidation effect. The most suitable approach to follow on this idea is to accordingly mutate these methionines to valines, using a clusterin-expression plasmid that also maintains the glycosylation state of clusterin. One such clusterin-expression plasmid was created and the clusterin characterization and functions were validated accordingly [23]. We plan to acquire this type of plasmid for future related studies. The methionine sulfoxide reductase (Msr) system comprises of MsrB and MsrA that reduce R-MetO or S-MetO enantiomers, respectively [24]. Both MsrA and the selenoprotein MsrB1 (SelR) were shown to protect against oxidative damage to brain cells, and thus play an important role in neurodegenerative diseases [24]. Oxidation of Met³⁵ in Aβ can produce reactive oxygen species (ROS), leading to the accumulation of toxic forms of oxidized and aggregated Aβ. Previously, it was demonstrated that SelR interacts with the central region of clusterin (AA 315–381), and that this interaction increases SelR activity and reduces intracellular ROS in N2aSW cells [25]. We hypothesize that an enhanced cellular interaction between SelR with may have a positive effect on clusterin activity, through a reduction of, yet to be determined, clusterine’s MetO residue/s. This situation is predicted to promote the secretion of relatively more natively folded clusterin from brain astrocytes. In turn, this MetO-reduced type of clusterin will exert its chaperon function in a better manner towards Aβ aggregates. More future studies are warranted to elucidate the role of methionine oxidation of clusterin and its relation to Aβ structure and neurotoxicity.

Conclusions

The data presented in this manuscript show that the chaperon protein, clusterin, is highly oxidized in its methionine moiety in postmortem human and mouse brains of AD subjects. This methionine oxidation of clusterin compromises its ability to protect Aβ from getting aggregated in vitro. Thus, it is suggested that preventing the methionine oxidation of clusterin will provide a better protection against the development or progression of AD.