Decrease in Peptide Methionine Sulfoxide Reductase in Alzheimer's Disease Brain


  • S. Prasad Gabbita,

  • Michael Y. Aksenov,

  • Mark A. Lovell,

  • William R. Markesbery

  • Abbreviations used : AD, Alzheimer's disease ; CK, creatine kinase ; GS, glutamine synthetase ; IPL, inferior parietal lobule ; MetSO, methionine sulfoxide ; MsrA, methionine sulfoxide reductase ; PBS-T, phosphate-buffered saline containing 0.05% Tween 20 ; SMT, superior and middle temporal gyri.

Address correspondence and reprint requests to Dr. W. R. Markesbery at 101 Sanders-Brown Building, University of Kentucky, Lexington, KY 40536-0230, U.S.A.


Abstract : Previous studies have shown that the pathophysiology of Alzheimer's disease (AD) is linked to oxidative stress. Oxidative damage to different biomolecular components of the brain is a characteristic feature of AD. Recent evidence suggests that methionine may act as an antioxidant defense molecule in proteins by its ability to scavenge oxidants and, in the process, undergo oxidation to form methionine sulfoxide. The enzyme peptide, methionine sulfoxide reductase (MsrA), reverses methionine sulfoxide back to methionine, which once again is able to scavenge oxidants. The purpose of this study was to measure the activity of MsrA in the brain of AD patients compared with control subjects. Our results showed that there was a decline in MsrA activity in all brain regions studied in AD and this decline reached statistical significance in the superior and middle temporal gyri (p < 0.001), inferior parietal lobule (p < 0.05), and the hippocampus (p < 0.05) in AD. An elevation of protein carbonyl content was found in all brain regions except the cerebellum in AD and reached statistical significance in the superior and middle temporal gyri and hippocampus. Messenger RNA analysis suggests that the loss in enzyme activity may be the result of a posttranslational modification of MsrA or a defect of translation resulting in inferior processing of the MsrA mRNA. Our results suggest that a decline in MsrA activity could reduce the antioxidant defenses and increase the oxidation of critical proteins in neurons in the brain in AD.

Experimental data show that oxidative stress is involved in the pathogenesis of Alzheimer's disease (AD) (Beal, 1994 ; Markesbery, 1997 ; Markesbery and Carney, 1999). Oxidative damage to biomolecular components of the brain is a characteristic feature of AD. Studies show an increase in oxidative damage to lipids (Subbarao et al., 1990 ; Lovell et al., 1995), proteins (C. D. Smith et al., 1991 ; Hensley et al., 1995 ; M. A. Smith et al., 1996), and DNA (Mecocci et al., 1994 ; Gabbita et al., 1998) in different brain regions in AD.

Oxidation of amino acid residues in proteins can have deleterious effects on functional properties of cell homeostatic balance. Excessive oxidative damage to proteins can result in neuron death and accelerated aging of the brain (Stadtman, 1992). The levels of oxidative damage to proteins in aging are pertinent, because aging is the most consistent risk factor for AD. Typical oxidative modifications of protein residues in neurons are mediated by a variety of physiological systems, including the following : oxidases ; formation of superoxide, hydrogen peroxide, and hydroxyl radicals ; trace metal-catalyzed oxidation ; oxyradical “leakage” from the mitochondrial transport chain ; and autooxidation of flavins or xenobiotics.

Although susceptibility to oxidation of different amino acids varies, methionine, cysteine, and tryptophan are among the most vulnerable to oxidation. The principal product of methionine oxidation is methionine sulfoxide (MetSO). Oxidation to form MetSO deprives the methionine residue of the ability to function as a methyl donor (Paik and Kim, 1975). Multiple studies demonstrate that proteins lose their biological activity when specific methionine residues are oxidized to MetSO (Carp et al., 1982 ; Brot and Weissbach, 1983 ; Swaim and Pizzo, 1988 ; Vogt, 1995).

Recent studies suggest that methionine residues act as a last line of antioxidant defense in proteins (Levine et al., 1996). High concentrations of methionine in proteins allow for effective scavenging of oxidants to form MetSO before an attack on residues critical to structure or function. The enzyme responsible for a reversal of MetSO back to methionine by reduction is the peptide methionine sulfoxide reductase (MsrA). Most biological systems contain disulfide reductases and MetSO reductases that can convert the oxidized forms of cysteine and methionine residues, respectively, to their unmodified forms. Methionine and cysteine oxidative modification products are the only residues that can be repaired within proteins (Berlett and Stadtman, 1997). Moskovitz et al. (1996 a, b) studied high-level expression, purification, and development of assays to measure MsrA, and reported its expression in a variety of organisms and animal tissues. A recent study demonstrated high levels of MetSO in yeast MsrA mutants when exposed to oxidative stress compared with wild-type strain, indicating that MsrA possesses an antioxidant function (Moskovitz et al., 1998). The ubiquitous presence of MsrA and its specific function of repairing oxidative damage in vivo indicate that MsrA could have an important role in providing cells with a defense system against oxidative stress.

The brain has a high metabolic rate and therefore is under constant stress by reactive oxygen species generated during cellular and mitochondrial processes. To counter the deleterious effects of these, the brain has a complex antioxidant defense system to attempt to minimize oxidative damage to proteins and other macromolecules. Most of these enzyme systems, such as superoxide dismutase, glutathione peroxidase, and catalase, quell reactive oxygen species by a direct interaction. Because the brain is under increased oxidative stress in AD and evidence indicates that MsrA has the ability to repair oxidative damage to methionine, which serves as an antioxidant within proteins, we compared the activities of MsrA in different brain regions of AD patients and normal control subjects.



L-[methyl-3H]Methionine (88 Ci/mmol, 1 mCi/ml) was purchased from Amersham. Protease inhibitors, dithiothreitol, methionine, and hydrogen peroxide were purchased from Sigma (St. Louis, MO, U.S.A.). TriZOL for RNA was purchased from GIBCO Life Technology. The Access RT-PCR kit was purchased from Promega, and the oxyblot kit (S7150) was purchased from Intergen.

Brain specimens

Brain specimens used in this study were removed at autopsy from 11 AD patients (three females and eight males) and 10 control subjects (seven females and three males). The mean ± SEM age of AD patients was 80.4 ± 1.9 years, and that of control subjects was 81.9 ± 2.1 years. The mean ± SEM postmortem interval was 2.8 ± 0.33 h for AD patients and 2.55 ± 0.14 h for control subjects. All AD patients showed progressive cognitive decline and met standard clinical diagnostic criteria for probable AD (McKhann et al., 1984). For histopathologic diagnosis, multiple sections of neocortex, hippocampus, amygdala, entorhinal cortex, basal ganglia, brainstem, and cerebellum were stained with hematoxylin and eosin, the modified Bielschowsky stain, and 10D-5 (for β-amyloid), ubiquitin, and α-synuclein immunohistochemistry. All cases met standard histopathologic criteria for the diagnosis of AD (National Institute on Aging-Reagan Institute, 1997). Control subjects were individuals without a history of dementia, other neurologic disorders, or systemic diseases affecting the brain. All control subjects underwent annual prospective neuropsychological testing and had test scores within the normal range. Neuropathologic examination, using the same methods as for AD patients, confirmed the absence of significant gross or microscopic alterations in control subjects.

For this study, 200-300-mg specimens from the left middle frontal gyrus, superior and middle temporal gyri (SMT), inferior parietal lobule (IPL), amygdala, hippocampus, and cerebellar hemispheres were removed at the time of autopsy, snapfrozen in liquid nitrogen, and subsequently stored at -70°C until used for analyses.

Preparation of tissue extract

The brain samples for enzyme activity measurement were thawed, suspended, and homogenized in 5 ml of 10 mM HEPES buffer (pH 7.4) containing 137 mM NaCl, 4.6 mM KCl, 1.1 mM KH2PO4, 0.6 mM MgSO4, and proteinase inhibitors [leupeptin (0.5 mg/ml), pepstatin (0.7 mg/ml), aprotinin (0.5 mg/ml), and phenylmethylsulfonyl fluoride (40 mg/ml)] using a motor-driven, Teflon-coated Dounce homogenizer with 20 strokes. Three-milliliter aliquots of the homogenate were centrifuged at 100,000 g for 1 h at 4°C. The resultant supernatant was kept on ice and used immediately for the MsrA assay. The remainder of the supernatant was frozen and later used for protein carbonyl measurement.

Synthesis of N-acetyl[3H]MetSO

The synthesis of N-acetyl[3H]MetSO was carried out in two stages (Brot et al., 1982). Methionine was first oxidized to MetSO, which was then acetylated to form N-acetylMetSO. L-[methyl-3H]Methionine (1 mCi) was added to 2 μl of 0.1 M L-methionine, 100 μl of 0.5 M HCl, and 5 μl of 30% H2O2 and incubated at 23°C for 1 h. The mixture was lyophilized, and the residue was dissolved in 15 μl of water and applied as a narrow band on a silica gel TLC plate. Standards containing methionine and MetSO were spotted on the plate remote from the narrow band of the reaction mixture. The plate was developed with n-butanol/acetic acid/water (60:15:25), and only the part of the plate where the standards were spotted was sprayed with ninhydrin. From the migration of the standard, the radioactive MetSO area was determined, and this region was scraped from the plate and extracted twice with 1 ml of glacial acetic acid. The [3H]MetSO was then acetylated by adding 2 ml of acetic anhydride and incubating at 23°C for 2 h. After the addition of 2 ml of water, the mixture was lyophilized. The resulting residue was dissolved in 1 ml of water, relyophilized, and the N-acetyl[3H]MetSO dissolved in 100 μl of water.

Assay for the reduction of N-acetyl[3H]MetSO

The assay conditions for MsrA activity in different brain regions were adapted from the method described by Brot et al. (1982). Each reaction mixture contained, in a final volume of 300 μl, 50 mM Tris-HCl (pH 7.4), 20 mM MgCl2, 30 mM dithiothreitol, and 3 μl of N-acetyl[3H]MetSO. The reaction mixture was incubated for 30 min at 37°C with 200 μl of the tissue extract from each brain region. For the corresponding blank, 200 μl of tissue extract was heat-denatured by boiling for 15 min to ensure the absence of any MsrA activity. The reaction was terminated by adding 1 ml of 0.5 M HCl. The acidified mixture was extracted with 3 ml of ethyl acetate and centrifuged for 3 min in a tabletop centrifuge. A 2-ml aliquot was removed from the organic layer and placed in a scintillation vial, followed by addition of 5 ml of Scintiverse scintillation liquid. Radioactivity was assayed using a liquid scintillation counter and the activity expressed in picomoles per milligram per minute (~2,000 cpm corresponds to 1 pmol).

Messenger RNA study of MsrA

The expression of genes encoding MsrA in brain was analyzed in the cerebellum and SMT in AD patients and control subjects using semiquantitative RT-PCR. Total RNA was extracted from 200 mg of tissue using TriZOL reagent following the method of Chomczynski and Sacchi (1987). The RNA concentration was determined by 260/280 absorbance measurement. The integrity of the extracted RNA was monitored using 1% agarose formaldehyde-gel electrophoresis (Maniatis et al., 1982). RT-PCR was performed using the Access RT-PCR kit. Each reaction mixture (25 μl) included AMW/Tfl reaction buffer, 1 mM MgSO4, 0.2 mM of dNTP, 5 U of AMV reverse transcriptase, 5 U of Tfl DNA polymerase, 50 pmol of each primer, and a known amount of total RNA. The sequence of primers was adapted from Moskovitz et al. (1996a) : sense primer : 5′-GTGGTGTTCCA GCCCGAGCACAT-3′ ; antisense primer : 5′-ATGTCGGTGGTGAT CAGGCCGAA-3′ ; product size : 228 bp.

Preliminary experiments were performed to confirm the specificity of the RT-PCR product and to ensure that the yield of specific RT-PCR product was proportional to the total RNA load. For preliminary experiments, RT-PCR reaction mixtures were loaded with 100 μg of total RNA. The conditions for RT-PCR included the cDNA first-strand synthesis step (45 min at 48°C, 2 min at 94°C) and multiple (22-32) cycles of amplification (90°C, 30 s, melting ; 55°C, 1 min, annealing ; 68°C, 2 min, elongation). RT-PCR products were visualized by staining of 12% polyacrylamide gels with SYBR Green I (Molecular Probes).

As a result of RT-PCR amplification of total human brain RNA preparation, a cDNA product of the expected size was obtained (Fig. 1A). The identity of RT-PCR product was confirmed by sequencing.

Figure 1.

MsrA-specific RT-PCR product was separated by polyacrylamide electrophoresis (A). In RT-PCR reaction mixtures loaded with 100 μg of total human brain RNA, MsrA-specific RT-PCR product was detectable by SYBR Green 1 staining after 22 cycles of amplification, and the yield of the amplified cDNA fragment was proportional to the number of PCR cycles (B). The yield of MsrA cDNA fragment was proportional to total RNA load (C). RT-PCR mixtures containing different amounts of total RNA were subjected to optimized number of PCR cycles after the reverse transcription.

FIG. 1.

Multiple cycles allowed the determination of the number of PCR cycles necessary to detect the PCR product and thereby stay within the linear region of PCR amplification (Fig. 1B).

The yield of MsrA-specific cDNA product was proportional to the RNA load (Fig. 1C) within the range of 0-200 ng of human brain RNA per 25 μl of reaction mixture.

For the analysis of MsrA gene expression, 100 μg of total RNA isolated from AD and control brain samples was added to 25 μl of RT-PCR reaction mixtures. After the completion of first-step reverse transcription, RT-PCR reaction mixtures were subjected to optimized numbers (26-28) of PCR cycles. A negative control (no reverse transcription) was always performed to ensure the absence of DNA. After amplification, cDNA was separated by polyacrylamide electrophoresis. Gels were stained with SYBR Green I digitized by computer-assisted imaging using MCID/M4 software (Imaging Research Inc., St. Catharines, Ontario, Canada). Yields of MsrA-specific product obtained using standard amounts (100 ng/25 μl reaction) of total RNA isolated from AD brain samples were compared with yields of MsrA-specific product obtained using the same amounts of total RNA from age-matched control brain samples.

Protein carbonyl measurement

The same tissue extract used for the MsrA activity measurement was analyzed for protein carbonyl content using the oxyblot kit. Eight AD patients and eight control subjects were studied. The 96-well Biodot apparatus (Bio-Rad) was used for the purpose of dot-blotting the proteins on a nitrocellulose membrane. All AD and control extracts from a particular brain region were spotted on the same nitrocellulose membrane to avoid blot-to-blot variation during subsequent treatment with primary and secondary antibodies and the chemiluminescent detection agent. The typical reaction mixture contained 20 μg of protein and 2,4-dinitrophenyl-hydrazine solution maintaining the required volume-to-volume ratio. The reaction was conducted for 20 min, subsequently neutralized with the neutralization solution, and spotted on the membrane. The same amounts of protein (20 μg) for each sample were spotted in every well of each blot. Each sample was spotted in triplicate. The derivatization-control solution was used as the blank. Following blotting of samples, the membrane was incubated overnight in blocking/dilution buffer composed of 1% bovine serum albumin in phosphate-buffered saline containing 0.05% Tween 20 (PBS-T). The membrane was incubated with rabbit antidinitrophenyl primary antibody for 1 h, washed with PBS-T, and then incubated with goat anti-rabbit, horseradish peroxidase-conjugated, IgG secondary antibody. Following washing with PBS-T, chemiluminescent detection (ECL kit, Amersham) was performed. For the detection of the chemiluminescent signal, membranes were exposed to x-ray film (Hyperfilm-ECL, Amersham) for 3 min. The protein carbonyl levels were expressed as a percentage of the mean of control subject staining intensity. For statistical analysis, a one-way ANOVA and Student's t test were performed.


The highest mean levels of MsrA were present in the hippocampus and amygdala, and the lowest level was present in the cerebellum (Fig. 2). The mean levels of activity were decreased in all brain regions in AD compared with controls. The most pronounced loss in mean MsrA activity was observed in the SMT of the AD brain compared with control subjects (p < 0.001). The hippocampus and IPL also showed a significant loss in MsrA activity in AD (p < 0.05).

Figure 2.

Mean ± SEM MsrA activity in AD and control brain regions. There are decreases in all regions, which reach statistical significance in the SMT (p < 0.001) and the IPL and hippocampus (p < 0.05) in AD. MFG, middle frontal gyrus.

FIG. 2.

Messenger RNA levels were measured in cerebellum and SMT. The results showed that there was a nonsignificant loss of MsrA mRNA levels in both cerebellum and SMT (Figs. 3 and 4). The mean loss in cerebellar mRNA in AD compared with control subjects was 31% (p < 0.27). The mean loss of MsrA activity in AD cerebellum compared with controls was 28%. The mean loss of mRNA in SMT in AD compared with control subjects was 33% (p < 0.1). There was a mean loss of 62% of MsrA activity in AD SMT compared with control subjects.

Figure 3.

Representative gel images illustrating the results of RT-PCR analysis of MsrA mRNA levels. RNA (100 ng) was isolated from the SMT, IPL, and cerebellum (Cer) of AD and control (CON) subjects.

Figure 4.

Mean ± SEM levels of MsrA mRNA in AD and control brain regions. There was a nonsignificant decline in the mean of MsrA mRNA levels in the cerebellum and SMT in AD compared with control subjects.

FIG. 3.

FIG. 4.

Protein carbonyl levels were measured in the same regions of AD and control brains in which MsrA activity was measured. Mean protein carbonyl content was increased in all AD brain regions compared with controls except the cerebellum. A statistically significant increase in mean protein carbonyl content was observed in the SMT (79% ; p < 0.01) and hippocampus (83% ; p < 0.05) in AD compared with control subjects (Fig. 5). Mean protein carbonyl content was also increased in IPL (125%), although this difference did not reach statistical significance due to variance in the individual data. The mean level of protein carbonyls in the AD cerebellum was lower than control, but this difference was not statistically significant.

Figure 5.

Mean ± SEM levels of protein carbonyl in AD and control brain regions. The mean carbonyl levels were increased in all AD regions except the cerebellum and were statistically significant in AD SMT (p < 0.01) and hippocampus (p < 0.05) compared with control subjects. MFG, middle frontal gyrus.

FIG. 5.


This study revealed a decrease in peptide MsrA activity in all brain regions studied in AD compared with control subjects. These decreases reached statistical significance in the SMT, IPL, and hippocampus, regions that show considerable histopathologic changes in AD. Accompanying this was an increase in brain protein carbonyl levels, which reached statistical significance in SMT and hippocampus of the AD brain, but not in the cerebellum. The results of the analysis of MsrA mRNA levels in the AD brain demonstrate that the decrease of the enzyme activity is not the consequence of the specific defect in transcription of the MsrA gene. Levels of MsrA mRNA were 33% lower in vulnerable (SMT) and 31% in nonvulnerable (cerebellum) regions of the AD brain compared with the same regions of the control brain. The decreased level of MsrA mRNA may be the reflection of a general decline of transcription in AD, which was reported in several studies (Sajdel-Sulkowska and Marotta, 1984 ; Guillemette et al., 1986 ; Crapper-McLachlan et al., 1988). The percentages of loss in the message coding for the enzyme (31%) and the enzyme activity (28%) in the cerebellum in AD compared with controls were similar in magnitude. This suggests that the loss in activity is due to a lower level of transcription of MsrA in the AD cerebellum. The percentage of loss in activity was much larger (62%) than that in the mRNA level (33%) in SMT in AD compared with control. Thus, the larger decline in the SMT suggests that the enzyme undergoes an oxidative posttranslational modification that affects its activity. Another possibility is that there may be a translational defect that occurs in cells in the AD SMT and causes an inferior processing of the total message of the enzyme.

Moscovitz et al. (1996a) demonstrated that MsrA immunoreactivity is more pronounced in neurons than in glia and that some populations of neurons express higher levels than others. They also suggested that MsrA may contribute to the resistance of neurons to oxidative damage. Although we do not have specific neuron quantification in our AD cases, there was severe neuron degeneration in the hippocampus, SMT, and IPL, regions where MsrA was significantly diminished. Thus, it is possible that the decrease in MsrA activity in AD is related to neuron degeneration and death in these regions.

Previous evidence suggests that increased free radical formation and oxidative damage to biomolecular components in the brain play a pivotal role in the pathogenesis of AD (Beal, 1994 ; Smith et al., 1995 ; Markesbery and Carney, 1999). Increases in lipid peroxidation (Lovell et al., 1995), a decline in polyunsaturated fatty acids (Prasad et al., 1998), and increases in the lipid peroxidation product, 4-hydroxynonenal, have been found in the brain in AD (Markesbery and Lovell, 1998). Isoprostanes (Montine et al., 1998), neuroprostanes (Roberts et al., 1998), and 4-hydroxynonenal (Lovell et al., 1997) are increased in AD cerebrospinal fluid. Activities of several antioxidant enzymes in different regions of the AD brain were elevated where there was pronounced lipid peroxidative damage (Lovell et al., 1995), although overall multiple studies of antioxidant enzyme activities in the brain in AD have not shown any consistent trends (for review, see Markesbery and Carney, 1999). Nuclear and mitochondrial DNA oxidation is increased in the brain in AD compared with controls (Mecocci et al., 1994 ; Lyras et al., 1997 ; Gabbita et al., 1998), and DNA oxidation products are increased in AD cerebrospinal fluid (Lovell et al., 1999).

Oxidation of the side chains of lysine, proline, arginine, and threonine residues has been shown to yield carbonyl residues. In addition to the present study, several other studies have shown elevated carbonyls in the brain in AD (C. D. Smith et al., 1991 ; Hensley et al., 1995 ; M. A. Smith et al., 1996), and others have demonstrated an increase in nitrotyrosine (Good et al., 1996 ; Hensley et al., 1998). Oxidative damage to glutamine synthetase (GS), a brain enzyme, is associated with an increase in carbonylation of protein residues and a loss of enzymatic activity (Oliver et al., 1990 ; Stadtman et al., 1992). GS activity is markedly diminished in the hippocampus and neocortex in AD (Smith et al., 1991 ; Hensley et al., 1995). Thus, it follows that in the presence of a generalized increase in protein oxidation in the AD brain, it is possible that MsrA could be oxidized leading to a loss of enzymatic activity.

Another enzyme that is sensitive to oxidation is creatine kinase (CK). Two studies have demonstrated a decline in CK in the frontal and temporal lobes in AD (Burbaeva et al., 1992 ; Aksenov et al., 1997). Levine et al. (1996) demonstrated that oxidation by hydrogen peroxide of at least six methionine residues in GS was required before susceptibility to proteolysis occurred in vitro. It is possible that diminished levels of MsrA activity such as we have observed in AD could contribute to failure of reductions of MetSO and lead to oxidative inactivation of GS and CK.

Ciorba et al. (1997) reported that oxidation of methionine in a voltage-dependent potassium channel in Xenopus oocytes modulates its inactivation. Oxidation of methionine to MetSO disrupts inactivation, which is reversed by coexpression with peptide MsrA. This suggests that oxidation and reduction of methionine could play a role in the cellular signal transduction cascade and that the diminished MsrA activity found in our study could possibly interfere with this process in AD.

Overall, the present study suggests that diminished MsrA may make less methionine available to reverse the oxidative damage that occurs to critical proteins in the brain in AD. This potentially could lead to a loss of critical protein functions in neurons and possibly contribute to the cascade of events leading to neuron degeneration. Even if this proves to be a secondary event in AD, it further weakens the defense against free radicals and allows for greater oxidative stress.


This research was supported by NIH grants 1P01-AG05119 and 5P50-AG05144, and a grant from the Abercrombie Foundation. The authors are grateful for the editorial assistance of Paula Thomason and technical assistance of Jane Meara.