Increased levels of oxidative stress markers detected in the brains of mice devoid of prion protein

Authors


Address correspondence and reprint requests to Man-Sun Sy, Institute of Pathology, BRB 933, Case Western Reserve University School of Medicine, 10900 Euclid Ave, Cleveland, OH 44106, USA. E-mail: mxs92@po.cwru.edu

Abstract

Although minor abnormalities have been reported in prion protein (PrP) knock-out (Prnp−/–) mice, the normal physiological function of PrP, the causative agent implicated in transmissible spongiform encephalopathies (TSE), remains unresolved. Since there are increasing correlations between oxidative stress and amyloidoses, we decided to investigate whether PrP plays a role in oxidative modulation. We found higher levels of oxidative damage to proteins and lipids in the brain lysates of Prnp−/– as compared to wild-type (WT) mice of the same genetic background. These two indicators, protein oxidation and lipid peroxidation, are hallmarks of cellular oxidative damage. Elevated levels of ubiquitin-protein conjugates were also observed in Prnp−/– mice, a probable consequence of cellular attempts to remove the damaged proteins as indicated by increased proteasome activity. Taken together, these findings are indicative of a role for PrP in oxidative homeostasis in vivo.

Abbreviations used
4-HNE

4-hydroxyalkenals

DNPH

2,4-dinitrophenylhydrazine

MDA

malondialdehyde

Prnp – / –

prion protein knock-out mice

PrP

prion protein

ROS

reactive oxygen species

SOD

superoxide dismutase

WT

wild-type

Transmissible spongiform encephalopathies (TSE) belong to a growing family of diseases linked to abnormal protein conformations (Dobson 1999). These protein misfolding disorders occur when normally soluble autologous proteins aggregate as abnormal insoluble amyloid fibrils and accumulate in tissues, causing structural and functional damage leading to disease (Dobson 1999). Unlike others, TSE can be infectious, and the implicated aetiological agent has been suggested to be a post-translationally derived abnormal isoform (PrPSc) of a normal cellular copper-binding prion protein (PrPC) (Prusiner 1998).

Although minor abnormalities have been reported (Collinge et al. 1994; Sakaguchi et al. 1996; Tobler et al. 1996) in transgenic mice with disrupted prion protein gene (Prnp–/–), the normal physiological function of prion protein is still under considerable debate. As there is an increasing correlation between oxidative stress and amyloidosis-associated proteins in vivo (Sayre et al. 1999), we decided to investigate whether there was a similar link with PrP by using prion protein knock-out (Prnp–/–) mice. In this report, we demonstrated increased oxidative damage to proteins and lipids in the brain lysates from Prnp–/– as compared to wild-type mice of the same genetic background. Protein oxidation and lipid peroxidation are established markers of cellular oxidative damage (Halliwell and Gutteridge 1999). Elevated ubiquitination and proteasome activity were also observed in Prnp–/– mice, probable consequences of increased protein damage. All these data imply a role for PrP in modulating the oxidative environment in vivo.

Materials and methods

Unless described, all chemicals and enzymes were purchased from Sigma (MO, USA) and Roche Diagnostic (IN, USA), respectively.

Mice

The generation of the prion protein knock-out mice (Prnp–/–) used in this study has been described (Manson et al. 1994). The wild-type (WT) mice used were of 129/Ola background (Harlan, Bicester, UK), the same genetic background as Prnp–/– mice. The numbers of animal used were kept to a minimum, but all experiments were performed in triplicate using three age-matched animals (n = 3).

Antibodies

Both monoclonal antibodies to ubiquitin and actin were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). Rabbit polyclonal antibody to macropain (20S proteasome) was kindly provided by Dr George DeMartino (University of Texas, Dallas, TX, USA) (McGuire et al. 1988).

Immunoblot analysis

The procedures for total brain protein extraction and subsequent immunoblotting have been described (Wong et al. 2000). Briefly, brain proteins were extracted from whole mouse brain by gentle homogenization on ice in 9 vol (w/v) of an extraction buffer (phosphate-buffered saline (PBS), 0.5% Nonidet P-40, 0.5% sodium deoxycholate) supplemented with Complete Mini EDTA-free protease inhibitor tablet (Roche Diagnostic, IN, USA). The homogenates were spun at 5000 g for 10 min at 4 °C and the cleared supernatant removed for protein concentrations measurement at 280 nm. The supernatant was then aliquoted and stored at − 80 °C until used. Identical amount of total proteins were boiled in loading buffer before electrophoretic separation, transferred to nitrocellulose, probed with primary and secondary antibodies, and finally analysed for immunoreactivity using the POD chemiluminescence kit (Roche Diagnostic, IN, USA). Pre-stained protein standards (Sigma, MO, USA) were used to calculate the apparent molecular weight of the protein bands. The relative optical densities were quantified using the Scion Image (Beta version 4.0.2) software.

Immunodetection of oxidized proteins (reactive carbonyl assay)

Protein oxidation in total brain homogenates were measured by assaying the amount of carbonyl groups on proteins using the OxyBlot™ Protein Oxidation Detection Kit (Intergen, NY, USA). The procedure was performed according to manufacturer's recommendation. Briefly, 20 µg of total brain lysates were derivatized with or without 2,4-dinitrophenylhydrazine (DNPH) and samples loaded onto a 4–20% linear gradient SDS-PAGE gel (Novex, CA, USA). After separation, proteins were electrotransferred to a nitrocellulose membrane and incubated with an antibody against the derivatized carbonyl groups. The oxidized proteins were visualized by using the POD chemiluminescence kit (Roche Diagnostic, IN, USA) and BioMax ML film (Kodak Eastman, NY, USA). The relative optical densities were quantified using the Scion Image (Beta version 4.0.2) software.

Proteasome activity analysis

Brain extracts were diluted with the proteasome assay buffer (PBS, 5 mm MgCl2, 1 mm dithiothreitol, 5 mm ATP) to a concentration of 10 mg/mL prior to analysis. The proteasome activity, performed in triplicates, was measured by adding 170 µL of the proteasome assay buffer and 30 µL of either the chymotrypsin fluorogenic substrate succinyl-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (Calbiochem, CA, USA) (2 m stock in DMSO) or the postglutamyl peptidase fluorogenic substrate succinyl-Leu-Leu-Glu-7-amino-4-methylcoumarin (Calbiochem, CA, USA) (2 m stock in DMSO), to 100 µL of the diluted brain extracts. The mixture was incubated for 30 min at 37 °C and the activity was determined as the increase in fluorescence of the reaction products. The fluorescence for chymotrypsin activity was monitored at 380 nm excitation and 410 nm emission, whereas the postglutamyl peptidase activity was monitored at 335 nm excitation and 410 nm emission, as previously described (Keller et al. 2000a). Background fluorescence was determined by incubating the brain lysates with the proteasome inhibitor lactacystin (50 µm) (kindly provided by Dr Clifford Harding) for 30 min before adding the proteasome substrate.

Lipid peroxidation assay

Lipid peroxidation was determined in total brain lysates using the N-methyl-2-phenylindole based LPO-586™ lipid peroxidation kit (Oxis International, OR, USA) (White et al. 1999b). This assay measures the level of malondialdehyde (MDA) and 4-hydroxyalkenals (4-HNE). Standard curves of both MDA and 4-HNE were established using 1,1,3,3 tetramethocypropane and 4-HNE, respectively. To minimize non-specific oxidation during sample preparation, 5 mm butylated hydroxytoulene (BHT dissolved in acetonitrile) was added to the extraction buffer. The assay was performed in triplicates according to manufacturer's recommendation using 4 mg of total brain lysates per reaction, and the results were calculated as picomoles of MDA or 4-HNE per mg of protein. Briefly, 200 µL of brain lysates (4 mg) was added with 10 µL of 0.5 m BHT and 650 µL of 0.1 mmN-methyl-2-phenylindole, followed by gently mixing. For MDA measurement, 150 µL of reagent grade HCl (∼36%) was added, whereas for the 4-HNE measurement 150 µL of 15.4 m methanesulphonic acid was added. The sample was mixed and incubated for 1 h at 45 °C. After incubating, the sample was centrifuged, and the supernatant was extracted and measured at 586 nm.

Results

Increased level of ubiquitin-protein conjugates in the brains of Prnp−/− mice

The ability to degrade oxygen radical-damaged protein is essential to all cells. These proteins are usually marked for degradation by the covalent attachment of multiubiquitin conjugates. Thus, an increase in the level of protein ubiquitination is indicated on immunoblots by increased ubiquitin immunoreactivity and a shift in the reaction product to higher position on sodium dodecyl sulfate polyacrylamide gel electophoresis (SDS-PAGE) (Ramanathan et al. 1999). Immunodetection on the above products were performed using an antiubiquitin monoclonal antibody. Our results in Fig. 1(a) clearly indicate a significant increase in ubiquitin immunoreactivity and shift in the molecular weight of the ubiquitin–protein conjugates in the brains of 8-week-old prion protein knock-out mice (Prnp–/–) in comparison with aged-matched WT mice of the same genetic background. Densitometric analysis indicates a more than threefold increase in the levels of ubiquitin–protein conjugates in Prnp–/– (p < 0.05, n = 3) as compared to the WT mice (Fig. 1b). The ∼7 kDa band is indicative of free ubiquitin as described previously (Ramanathan et al. 1999). The actin immunoblot indicates equal amounts of total brain lysates were loaded in the gel.

Figure 1.

Elevating levels of ubiquitin–protein conjugates in total brain homogenates of prion protein knock-out (Prnp–/–) mice. (a) Immunodetection with a monoclonal antibody against ubiquitin after separation on a reducing 4–20% gradient SDS-PAGE using 20 µg of total brain proteins as starting material. Immunoblotting with an antiactin antibody was used as a control to ensure loading of identical amount of starting materials. (b) The relative optical density on the levels of ubiquitin–protein conjugates in total brain lysates (white bar for WT and black bar for Prnp–/–). Data are mean ± SD of three independent experiments using three age-matched mice (n = 3; 8 weeks old). *p < 0.05 compared with wild-type (WT) mice. Quantification was performed by plot profiling the whole series of bands in a prefixed area for each lane using Scion Image (Beta version 4.0.2) software.

Total protein reactive carbonyls

Biological processes involving the generation of free radicals usually lead to oxidative post-translational modifications of proteins. An example is the introduction of site-specific reactive carbonyl groups, and the level of formation reflects the intensity of the oxidative stress experienced (Smith et al. 1996). To detect the presence of carbonyl groups, total brain lysates were derivatized with 2,4-dinitrophenylhydrazine (DNPH). The DNP-derivatized protein samples (with carbonyl groups attached) were separated by polyacrylamide gel electrophoresis followed by immunoblotting using a primary antibody against the DNP moiety attached to the proteins. As shown in Fig. 2(a), an observable increase of 49% (p < 0.05, n = 3,Fig. 2b) in the level of carbonyl groups was detected in the derivatized (DNPH +) brain lysates from 8-week-old Prnp–/– in comparison to WT mice of similar age. No immunoreactivity was detected in the non-derivatized brain lysates (DNPH –). These results are indications of higher degree of oxidative damage in the brains of Prnp–/– mice.

Figure 2.

Increased protein oxidation (reactive carbonyl groups) in the brain lysates of prion protein knock-out (Prnp–/–) mice. (a) Carbonyl assay using 20 µg total brain lysates derivatized without (–) or with (+) 2,4-dinitrophenylhydrazine (DNPH), followed by electrophoretic separation on a 4–20% gradient SDS-PAGE and immunoblotted against the derivatized carbonyl groups. (b) The relative optical density on the level of reactive carbonyl groups detected in brain lysates (white bar for wild-type (WT) and black bar for Prnp–/–). Data are mean ± SD of three independent experiments using three age-matched mice (n = 3; 8 weeks old). *p < 0.05 compared with WT mice. Quantification was performed by plot profiling whole series of bands for each lane in a prefixed area using Scion Image (Beta version 4.0.2) software.

End products of lipid peroxidation (malonaldehyde and 4-hydroxyalkenals)

Another well-established indicator of oxidative stress is the measurement of the reactive products generated as a result of lipid peroxidation. Lipid peroxides are unstable and decompose to form a complex series of reactive aldehydes. Two of these are analysed in this report: malonaldehyde (MDA) and 4-hydroxyalkenals (4-HNE). In Prnp–/– mice, the levels of MDA and 4-HNE in the 8-week-old brains were 59% (p < 0.05, n = 3) (Fig. 3a) and 24% (p < 0.05, n = 3) (Fig. 3b) higher when compared with age-matched WT mice. No background interference by BHT was observed. Control reactions without brain lysates were measured and subtracted from the actual measurements.

Figure 3.

Increased level of lipid peroxidation end products in the brains of prion protein knock-out (Prnp–/–) (black bar) as compared to wild-type (WT) mice (white bar). Detection of (a) malonaldehyde (MDA) generation in hydrochloric acid (HCl), and (b) 4-hydroxyalkenals (4-HNE) formation in methanesulphonic acid using 4 mg of total brain homogenates as starting material. Data are mean ± SD of three independent experiments using three age-matched mice (n = 3; 8 weeks old). *p < 0.05 compared with WT mice.

To determine what effect different ages has on lipid peroxidation, we prepared total brain homogenates from mice of various ages. Since birth, the brains of Prnp–/– mice started to experience a greater level of lipid peroxidation than age-matched WT mice (Fig. 4). The level of MDA different between the two mice seen to be greater as Prnp–/– mice progresses beyond 6-month-old (> 40%, p < 0.05, n = 3; Fig. 4a). In contrast, 4-HNE different between the two mice remains relatively constant (∼20–35%, p < 0.05, n = 3; Fig. 4b). The higher levels of lipid peroxidation experienced by both mice at birth is probably a consequence of trauma (Brown et al. 1996).

Figure 4.

The effect of ages on lipid peroxidation in the total brain homogenates from wild-type and prion protein knock-out mice. The detection of (a) malonaldehyde (MDA) and (b) 4-hydroxyalkenals (4-HNE) formation in hydrochloric acid and methanesulphonic acid, respectively, using 4 mg of total brain homogenates from different ages of prion protein knock-out (Prnp–/–) (broken line) and wild-type mice (WT) (straight line). Data are mean ± SD of three independent experiments using three age-matched mice (n = 3). *p < 0.05 compared with WT mice.

Analysis of the levels of proteasome subunit and activity

To determine whether the increased level of ubiquitin–protein conjugates was due to proteasome dysfunction, we determined the levels of a major subunit of proteasome and two enzymatic activities associated with this cellular complex. Immunoblotting using an antibody against macropain (20S proteasome subunit) indicates similar levels in the brains of both 8-week-old Prnp–/– and WT mice (Fig. 5a). The apparent molecular weight of ∼ 29 kDa is similar to those previously reported after separation by SDS-PAGE (McGuire et al. 1988). To determine whether the proteasome was functioning, we measured two associated enzymatic activities (Keller et al. 2000a): chymotrypsin and postglutamyl peptidase protease activities. Both activities were only modestly higher in the brains of 8-week-old Prnp–/– mice as compared with WT mice of similar age. The increase was 26% for the chymotrypsin activity (p < 0.05; n = 3; Fig. 5b) and 14% (p < 0.05, n = 3; Fig. 5c) for postglutamyl peptidase activities in Prnp–/– mice. These results clearly indicate that the proteasome is not only functioning but has increased its activity to cope with the elevated levels of ubiquitin-tagged protein conjugates.

Figure 5.

Analysis of proteasome subunit expression and associated enzymatic activities. (a) Immunodetection of ∼ 29 kDa macropain (20S proteasome) expression using a rabbit polyclonal antibody described previously (McGuire et al. 1988) in the brain lysates from both wild-type (WT) and prion protein knock-out (Prnp–/–) mice. Detection of two proteasome-associated enzymatic activities (b) chymotrypsin and (c) postglutamyl peptidase, in prion protein knock-out mice (Prnp–/–) (black bar) and WT mice (white bar). Data for (b) and (c) are mean ± SD of three independent experiments using three age-matched mice (n = 3; 8 weeks old). *p < 0.05 compared with WT mice.

Discussion

The results presented in this report implicate a role for PrP in oxidative homeostasis in vivo. We have detected elevated levels of various oxidative markers, e.g. ubiquitination (Fig. 1), protein oxidation (Fig. 2) and lipid peroxidation (Figs 3 and 4) in the brains of PrP knock-out mice (Prnp–/–) as compared to age-matched WT mice of the same genetic background. To handle the increased ubiquitin–protein conjugates, higher levels of proteasome activity were also detected (Fig. 5).

Oxidative stress is a condition in which the generation of reactive oxygen species (ROS), a ubiquitous by-product of aerobic metabolism, overwhelms the cellular antioxidant defence mechanism (Halliwell and Gutteridge 1999). Lipid peroxidation has been taken as an indicator of cellular oxidative damage. It is an autocatalytic process that leads to the generation of reactive aldehydes as a result of oxidative attack to polyunsaturated fatty acids, a component of both cellular and subcellular membrane (Halliwell and Gutteridge 1999). The extent of lipid peroxidation is associated with the level of breakdown products generated. In this report, we measured two of these aldehydes, MDA and 4-HNE. In most instances, MDA is the most abundant arizing from lipid peroxidation, while 4-HNE is the most neurotoxic (Halliwell and Gutteridge 1999). To minimize non-specific amplification of peroxidation during sample preparation, we have incorporated the chain-breaking antioxidant BHT to the lysates. As shown in Fig. 3, increased levels of MDA and 4-HNE were observed in the brains of Prnp–/– mice. Although a high level of lipid peroxidation can be detected in Prnp–/– mice from birth (Fig. 4), the changes in the MDA level between the two mice seem to be greater (> 40%) as they aged beyond 6-month-old (Fig. 4a), while 4-HNE differences remain relatively constant (Fig. 4b). Coincidently, altered lipid composition has been documented in experimental prion disease (Guan et al. 1996).

Another consequence of oxidative damage is the post-translational oxidative modifications of proteins, commonly detected by increased reactive carbonyl groups (Halliwell and Gutteridge 1999). As shown in Fig. 2, the level of reactive carbonyl groups in proteins derived from the brains of Prnp–/– is ∼ 49% higher than in WT mice. Oxidatively damaged proteins are markers for ubiquitin conjugation and subsequent proteolysis, a process where they are more susceptible to than their normal counterparts (Alves-Rodrigues et al. 1998). The ability to eliminate oxidatively damaged-ubiquitinated proteins is important to maintain cellular homeostasis (Alves-Rodrigues et al. 1998), and failure to do so can lead to undesirable accumulation of damaged cellular proteins. Therefore, it is not surprising that more than threefold increases in the levels of ubiquitinated proteins were observed in the brains of Prnp–/– mice (Fig. 1). Interestingly, correlations between ubiquitination and prion diseases have been documented (Lowe et al. 1990; Ironside et al. 1993). Immunocytochemistry studies of prion diseases have found increased ubiquitin–protein conjugates along the periphery of PrP amyloid plaques (Ironside et al. 1993) and within intraneuronal inclusion bodies (Lowe et al. 1990; Ironside et al. 1993). It is intriguing to consider that accumulation of ubiquitin–protein conjugates, a feature of several neurodegenerative diseases (Perry et al. 1987; Lowe et al. 1988; Manetto et al. 1988), that are marked by oxidative damage are also a response to an increase in oxidatively damage proteins.

The proteasome is one of the main proteolytic systems within eukaryotic cells and is responsible for the degradation of ubiquitinated proteins (Alves-Rodrigues et al. 1998). It is a 26S ATP-dependent multicomponent enzymatic complex including the 20S catalytic core. To determine whether the increased level of ubiquitin–protein conjugates was due to the malfunctioning of this degradation system, we determined the level of two enzymatic activities associated with the catalytic core (Keller et al. 2000a). Our results demonstrated that the proteasome was not only functioning but has modestly increased its activity in order to cope with the elevated levels of ubiquitin-tagged protein conjugates (Figs 5b and c) in the brains of Prnp–/– mice. However, this increased activity was not associated with any detectable increased in the levels of one of the proteasome subunits (macropain) (Fig. 5a). Nevertheless, the observation of significant level of non-degraded ubiquitin-tagged proteins indicate either (a) the proteasome was not sufficiently activated to handle the task well, (b) was blocked with oxidative damaged products, or (c) could indicate dysfunction in other ubiquitin-linked degradation pathways, e.g. lysosome system (Alves-Rodrigues et al. 1998). Recent reports have indicated that elevated levels of 4-HNE could affect proteasome function (Okada et al. 1999; Keller et al. 2000b). Although our study did not observe any reduction in the proteasome associated-enzymatic activities, but the modest increase as compared to the significant elevation of ubiqutin–protein conjugates could be signs of possible inhibition. The lack of actual reduction in the activities could be due to the modest level of 4-HNE detected in this report as compared to previous studies (Okada et al. 1999; Keller et al. 2000b). Taken together, these results suggest that the proteasome is able to tolerate slight increase of 4-HNE, but as 4-HNE accumulates probably during inflammation (Keller et al. 2000b), accumulating 4-HNE will bind to the proteasome and observable functional reduction is inevitable (Okada et al. 1999).

In conclusion, we have presented a correlation between the loss of PrP and an increased level of oxidative stress markers in the brains of Prnp–/– mice. However, we may be underestimating the extent of oxidative stress resulting from Prnp–/– mice due to compensation in cellular repair mechanisms (e.g. increased proteasome activity), which explained the lack of pivotal phenotypic traits in these mice. This is consistent with the view that oxidative balance is modulated by an overlapping webs of interdependence antioxidant mechanisms, and is unlikely to be affected by the removal of a single factor, no matter how seemingly critical (Perry et al. 2000). Our findings also imply that Prnp–/–-derived neurones might be oxidatively impaired and, as such, displayed increased susceptibility to oxidative stress in vitro (Brown et al. 1997; White et al. 1999a). Although it is probable that the increased oxidative stress is due to the loss of oxidative stress amelioration by PrP (Brown et al. 1999; Wong et al. 2000), we cannot rule out the possible involvement of PrP interacting with other protein(s) that are directly modulating the oxidative environment. Further studies are needed to clarify these relationships.

Acknowledgements

We thank Dr George DeMartino (University of Texas, Dallas, USA) for providing the antimacropain polyclonal antibody, Dr Clifford Harding for providing the proteasome inhibitor lactacystin, and Dr Wieslaw Swietnicki for assistance with the fluorescence measurements. DRB was supported by a fellowship from BBSRC (UK). This work was supported in part by grants from NIH (#AG14359 to PG and MS), the Britton Fund (to PG), BBSRC (8/BS410551 to DRB) and EU Biomed Commission (976051 to DRB).

Ancillary