The FASEB Journal express article 10.1096/fj.01-0696fje. Published online March 12, 2002.
Effects of reactive γ-ketoaldehydes formed by the isoprostane pathway (isoketals) and cyclooxygenase pathway (levuglandins) on proteasome function Sean S. Davies*, Ventkataraman Amarnath‡, Kathleen S. Montine‡, Nathalie Bernoud-Hubac*, Olivier Boutaud*, Thomas J. Montine*,‡, and L. Jackson Roberts II*,† Departments of Pharmacology*, Medicine†, and Pathology‡, Vanderbilt University Medical Center, Nashville, Tennessee Corresponding author: L. Jackson Roberts II, Department of Pharmacology, 522 RRB, Vanderbilt University, Nashville, TN 27232-6602. E-mail:
[email protected] ABSTRACT Oxidative stress can impair proteasome function, both of which are features of neurodegenerative diseases. Inhibition of proteasome function leads to protein accumulation and cell death. We discovered recently the formation of highly reactive γ-ketoaldehydes, isoketals (IsoKs), and neuroketals (NeuroKs) as products of the isoprostane and neuroprostane pathways of free radical-induced lipid peroxidation that are analogous to cyclooxygenase-derived levuglandins (LGs). Because aldehydes that are much less reactive than IsoKs have been shown to inhibit proteasome function, we explored the ability of the proteasome to degrade IsoKadducted proteins/peptides and the effect of IsoK and IsoK-adducted proteins/peptides on proteasome function. Adduction of IsoK to model proteasome substrates significantly reduced their rate of degradation by the 20S proteasome. The ability of IsoK to inhibit proteasome function directly was observed only at very high concentrations. However, at much lower concentrations, an IsoK-adducted protein (ovalbumin) and peptide (Aβ1-40) significantly inhibited chymotrypsin-like activity of the 20S proteasome. Moreover, incubation of IsoK with P19 neuroglial cultures dose-dependently inhibited proteasome activity (IC50 = 330 nM) and induced cell death (LC50 = 670 nM). These findings suggest that IsoKs/NeuroKs/LGs can inhibit proteasome activity and, if overproduced, may have relevance to the pathogenesis of neurodegenerative diseases. Key words: isoketal • free radical • neurodegenerative
A
ccumulation of aggregated proteins is the hallmark of many neurodegenerative diseases, including Alzheimer’s Disease (AD), Parkinson’s Disease, and Huntington’s Disease, which suggests an impaired ability to degrade proteins. One of the major pathways for protein degradation is via the multicatalytic proteinase complex or proteasome. The central core of the proteasome, often referred to as the 20S proteasome, is a 700-kD macromolecule composed of 28 subunits (1, 2) with chymotrypsin-like, trypsin-like, and peptidyl-glutamyl-peptide hydrolyzing activities.
Inhibition of the proteasome with selective proteasome inhibitors has a profound effect on many cell types. Adding these inhibitors to PC12 cells, an NGF-responsive cell line, induces accumulation of ubiquinated proteins (3) and phosphorylated neurofilaments (4), as well as apoptosis (5, 6). Injection of these inhibitors into the lateral ventricle of the brain in rats induces apoptosis in the hippocampus (7). These results suggest the intriguing possibility that the accumulation of polyubiquinated, hyperphosphorylated tau in the neurofibillary tangles of AD (8) and eventual neuronal death may be related to impaired proteasome activity during the progression of the disease. Indeed, short postmortem interval autopsied brains from patients with AD demonstrate reduced proteasomal chymotrypsin-like and postglutamyl peptidase activity versus age and sex-matched controls (50). This decreased proteasome activity appears to be due to functional deficiency rather than decreased proteasomal subunit expression. The mechanism(s) of decreased proteasome activity is unknown, but significant reduction of activity occurs only in the same brain regions in which increased oxidative damage occurs, as assessed by a variety of parameters [reviewed in (13)]. The association between oxidative stress and reduced proteasome activity may be important, because oxidative stress has been shown to modulate proteasome activity in vitro (14–23). Lipid peroxidation generates a number of lipid aldehydes, including 4-hydroxy-2-nonenal (HNE), acrolein, malondialdehyde, isoketals (IsoKs) (24, 25), and neuroketals (NeuroKs) (26), which can adduct to proteins and can induce protein cross-linking and aggregation under certain conditions. HNE is the most studied of these aldehydes, and addition of HNE to P19 neuroglial cells has been found to induce an accumulation of high-molecular-weight ubiquinated proteins (27). HNE may induce protein accumulation and proteasome inhibition by three mechanisms. First, adduction of oxidized protein with sufficient amounts of HNE to crosslink the protein reduces its rate of degradation by the 20S proteasome (28). Second, HNE-adducted protein inhibits 20S proteasome degradation of non-adducted proteasome substrates (28, 29). Finally, mM concentrations of HNE added directly to the proteasome modestly inhibit proteasome activity (30). Although HNE is the most studied of the aldehydes formed by lipid peroxidation, levuglandins (LGs) and recently characterized isoketals (IsoKs) and neuroketals (NeuroKs) are more than an order-of-magnitude more reactive than HNE (24, 31, 32). IsoKs and LGs are γ-ketoaldehydes formed by the isoprostane (IsoP) pathway of free radical catalyzed oxidation of arachidonic acid and the cyclooxygenase pathway, respectively (24, 33, 34) (Fig. 1). NeuroKs are formed in an analogous manner to IsoKs, but they derive from the neuroprostane (NeuroP) pathway of free radical-induced oxidation of docosahexanoic acid, which is highly enriched in the brain (32). The cyclooxygenase pathway forms LGE2 and LGD2, whereas the isoprostane pathway forms four analogous E2-IsoK and four D2-IsoK regioisomers, each of which consists of four racemic diastereomers. One of the E2- and D2-IsoK regioisomers is the same as LGE2 and LGD2, but can differ in regard to constituent stereochemistry (Fig. 1). During in vitro oxidation of arachidonic acid, IsoKs are formed to a somewhat greater extent than F2-IsoPs (24) and IsoK adducts dramatically increase during oxidative stress in vivo, paralleling other products of the isoprostane pathway (unpublished data).
IsoKs/NeuroKs/LGs react with the ε-amine of lysine residues to form an initial, reversible Schiff Base adduct (35) and then a pyrrole, which rapidly undergoes autooxidation to form stable lactams and hydroxylactams (24, 36). Although the half-life for HNE adduction to proteins is about 1 h, the half-life for IsoK/NeuroK/LG adduction is less than 5 min (24). In addition, IsoKs/NeuroKs/LGs exhibit a remarkable proclivity to crosslink proteins that is not shared by HNE (31, 37, 38). Thus, IsoKs/NeuroKs/LGs may have more profound biological effects than HNE. These findings led us to explore the effects of E2-IsoK and E2-IsoK-adducted proteins on proteasome activity, both in cell lysates and in intact cells. MATERIALS AND METHODS Materials The mouse macrophage cell line RAW 264.7 and the mouse embryonal carcinoma cell line P19 were purchased from American Type Culture Collection (Manassas, VA). Carboxylbenzoylvaline-lysine-methionine-amourilide methyl courain (VKM-AMC), suc-leucine-leucine-valinetyrosine-AMC (LLVY-AMC), synthetic lactacystin, MG-115, and recombinant Methanosarcina thermophila 20S proteasome were purchased from Calbiochem (La Jolla, CA). Chicken egg ovalbumin (OVA), the monoclonal anti-chicken egg albumin clone OVA-14, dimethylformamide (DMF) and 3-4,5-dimethylthiazol-2-yl-2,5-diphenyl tetrazolium bromide (MTT), and trichloroacetic acid (TCA) were purchased from Sigma (St. Louis, MO). Phosphate buffered saline (PBS), Hanks balanced salt solution (HBSS), and Alpha Minimum Essential Medium (MEM) were purchased from GibcoBRL (Gaithersburg, MD). Amyloid β1–40 (Aβ1–40) was purchased from Calbiochem. Preparation of synthetic E2-IsoK [8(R)-acetyl-9(R)-formyl-12(S)-hydroxy-5(Z),10(E)heptadecadienoic acid] The O-tert-butyl-dimethylsilyl ether, isopropylidine precursor of E2-IsoK, 8-acetyl-9-(3,3dimethyl-2,4-dioxolanyl)-12-(t-butyldimethylsilyloxy) heptadeca-5(Z),10(E)-dienoic acid, was synthesized by methods previously published (25, 39, 40). The precursor was then hydrolyzed in 2:1 acetic acid–water, oxidized with NaIO4, quenched with ethylene glycol, and purified, and the identity and concentration were determined by nuclear magnetic resonance as reported (41). Preparation of proteasome Lysates containing proteasome activity were isolated from RAW 264.7 cells by similar methods as used previously for cultured liver epithelial cells (42). Cells were grown to a density of about 2 × 106 cells/ml, scraped gently, and centrifuged to remove media, and cells were lysed with 1 mM dithiothreitol (DTT) for 1 h while being stirred vigorously. Lysates were centrifuged at 14,000 x g for 30 min at 4oC to remove membrane debris, nuclei, and unlysed cells. Aliquots of the supernatant were stored at –70oC until use. We measured protein content by using the Pierce (Rockford, IL) BCA protein assay, which was modified as suggested by the manufacturer to minimize DTT background in the assay. To test the proteasomal contribution to the chymotrypsin-like activity of the lysate, the selective proteasome inhibitors lactacystin or MG-
115 (0 to 50 µM final concentrations) were preincubated with 3.75 µg protein of RAW lysate for 30 minutes. Hydrolysis of 50 µM final concentration LLVY-AMC was measured by fluorescence at 380 nm excitation and 440 nm emission by using free AMC as standards. Lactacystin and MG-115 inhibited greater than 75% and 95% of LLVY-AMC hydrolysis, respectively. We used recombinant Methanosarcina thermophila 20S proteasome without further purification (reported purity 95% by sodium dodecylsulfate polyacrylamide gel electrophoresis [SDS-PAGE]). Adduction of peptides and proteins with synthetic E2-IsoK We resuspended VKM-AMC at 2 mM in dimethyl sulfoxide (DMSO) and adducted VKM-AMC by incubating 100 µM VKM-AMC with 100 µM E2-IsoK in PBS for 4 h at 37oC. For experiments in which E2-IsoK was reduced before incubation with VKM-AMC, we incubated E2-IsoK with 10 mM sodium borohydride in DMF for 30 min at room temperature, which was then neutralized with PBS before the addition to VKM-AMC. In experiments in which E2-IsoKadducted VKM-AMC was post-treated with sodium borohydride, an aliquot of the adducted peptide was treated with 10 mM sodium borohydride for 30 min before neutralization with PBS. OVA (100 µM) was incubated with either 100 µM or 1 mM E2-IsoK for 2 h, and then a portion was treated with sodium borohydride as above. We oxidized OVA by using FeSO4, as described previously for oxidation of glucose-6-phosphate dehydrogenase (28), and adducted oxidized OVA with E2-IsoK as described above. We incubated Aβ1–40 (50 µM) with 100 µM E2-IsoK for 24 h, and then treated a portion with sodium borohydride as before. Measurement of isoketal-adducted fluorogenic peptide hydrolysis Final concentration of VKM-AMC (10 µM) treated with the various conditions described above was added to RAW lysate proteasome preparation (30 µg protein) in 1 ml final volume of 50 mM Tris buffer, pH 7.8, with 1 mM DTT, 20 mM potassium chloride, and 500 µM magnesium acetate. Aliquots (180 µl) of each triplicate reaction were removed at the stated times, an equal volume of ice-cold ethanol was added, and then 800 µl of 125 uM sodium borate buffer, pH 9.0, was added. We measured released AMC as before. Measurement of protein degradation OVA (50 µg) or E2-IsoK-adducted OVA were added to 1 µg of recombinant proteasome in 250 µl total of 50 mM HEPES buffer, pH 7.5, containing 10 mM magnesium chloride, 100 mM potassium chloride, 0.1 mM CaCl2, and 1 mM DTT (43). SDS (0.03%) was added to aid protein unfolding and stimulate proteasome activity (44, 47). We removed aliquots (50 µl) from the reaction at the appropriate times, added an equal volume of ice-cold TCA (20%), and then centrifuged at 14,000 x g for 15 min. The supernatant was then neutralized with potassium hydroxide and HEPES buffer and was derivatized with fluorescamine (0.3 mg/ml) (17). The fluorescence was quantified at 390 nm excitation and 470 nm emission by using glycine as standard.
Visualization of protein cross-links Sample buffer containing reducing agent was added to adducted and unadducted proteins, and samples were run on 10% SDS-PAGE and visualized by immunoblotting by using monoclonal anti-OVA antibody or Coomassie blue staining for Aβ1-40. Measurement of proteasome chymotrypsin-like activity Unadducted or E2-IsoK-adducted Aβ1-40 (10 nM to 5 µM), 20 µM MG-115, or 100 µM lactacystin was preincubated with RAW 264.7 lysate proteasome preparation (3.75 µg protein) for 2 h in 100 µl total HEPES buffer as above, but without SDS. LLVY-AMC (50 µl of 150 µM) was then added, and hydrolysis of LLVY-AMC was measured after 1 h as before. One-hundred percent proteasome activity was set equal to the amount of AMC released in the control reaction. We used GraphPad Prism version 3.00 for Windows from GraphPad Software, (San Diego CA) to calculate the IC50 for E2-IsoK-adducted Aβ1-40 and to estimate the IC50 for unadducted Aβ1-40. For adducted and unadducted OVA, we incubated 200 nM to 1 µM of each OVA with RAW lysate (15 µg protein) for 4 h at 37oC in 495 µl total of 50 mM Tris buffer as above. We then added LLVY-AMC (5 µl of 5 mM) in DMSO for 1 h, and quenched the reaction and released AMC measured as above. For inhibition of proteasomes by synthetic E2-IsoK, RAW lysate proteasome preparation (3.75 µg protein) or 10 nM recombinant proteasome was incubated at 37oC for 2 h with stated concentrations of E2-IsoK, lactacystin, or MG-115 in 100 µl total HEPES buffer. LLVY-AMC (50 µl of 150 µM) was then added and fluorescence was measured as before. For Hanes-Woolf analysis, 50 µl of LLVY-AMC stock solutions (60 µM to 210 µM) were added. Inhibition of P19 neuroglial culture proteasomal activity P19 embryonal carcinoma cells were induced to differentiate into neuroglial cultures with retinoic acid as described (45) and were plated at 1 × 106 cells per well in 24-well plates. Cultures were used 7 days after plating. Media were removed, and the stated concentrations of E2-IsoK, lactacystin, or MG-115 were added to six replicate wells in 500 µl total of HBSS with 0.5% DMSO for 1 h at 37oC, followed by the addition of 500 µl of MEM with 10% fetal bovine serum for another 23 h. To determine viability, we added MTT to duplicate wells after 22 h of total incubation for each treatment. At 24 h of total treatment, the media were removed and 500 µl of acidic isoproponal was added to the wells with MTT; absorbance at 570 nm was measured on duplicate aliquots. We set the average value for untreated wells as 100% viability. To measure proteasome chymotrypsin-like activity, we removed the treatment media after 24 h of total treatment and added 500 µl of 1 mM DTT to the remaining four wells. The plates were then shaken vigorously at 4oC for 1 h. The lysates were transferred to microfuge tubes and were centrifuged at 14,000 x g for 30 min. The supernatant (150 µl) was transferred to a 96-well plate; 50 µl of 200 µM LLVY-AMC in HEPES buffer was added and AMC was released measured as before. Protein measurement of supernatants showed no significant differences in concentration between treatment groups (data not shown). We calculated proteasome activity by using maximal chymotrypsin-like activity.
RESULTS Effect of adduction of proteasomal substrates with E2-IsoK on their degradation by the 20S proteasome To test both the effects of both IsoKs, which arise from isoprostane pathway, and LGs, which derive from the cyclooxygenase pathway, on proteasome function, we synthesized an E2-IsoK isomer identical to LGE2. Although numerous IsoK isomers are formed by the isoprostane pathway, the reactivity and hydrophobicity of all of the isomers—and thus their biological effects—would be expected to be essentially identical. To test the effect of IsoK adduction of proteasomal substrates on their degradation, we used the lysine containing fluorogenic proteasomal substrate VKM-AMC (46), which releases the AMC fluorophore on hydrolysis by the proteasome. VKM-AMC was reacted with an equimolar concentration of E2-IsoK, and a portion of this reaction was reduced by treatment with sodium borohydride to convert any remaining aldehyde to its nonreactive alcohol. In an additional reaction, E2-IsoK was reduced with sodium borohydride for 30 min before incubation with VKM-AMC. Proteasome activity was measured in lysates from RAW 264.7 cells. Incubation of vehicle-treated VKM-AMC with RAW cell lysates resulted in a rapid and linear increase in free AMC over 150 min (Fig. 2A). In contrast, incubation of E2-IsoK-adducted VKM-AMC with RAW cell lysates resulted in negligible release of AMC. The lack of proteolysis was dependent on adduction of the peptide substrate, as reduction of E2-IsoK before incubation with VKMAMC abolished the observed effect. Reduction of the incubation containing E2-IsoK-adducted VKM-AMC did not increase proteolysis significantly compared with unreduced incubations, which suggests that the lack of proteolysis was not due to direct inhibition of the proteasome by any unreacted E2-IsoK that may still be present in the preparation. We then determined the effect of IsoK adduction on degradation of a model protein, OVA. OVA, which contains 20 lysine residues, was adducted with E2-IsoK and its rate of proteolysis compared with that of vehicle-treated OVA. We tested the effect of E2-IsoK adduction on OVA degradation by using a commercially available preparation of recombinant Methanosarcina thermophila 20S proteasome, stimulated with 0.03% SDS to aid protein unfolding and increase proteasome activity (44, 47). Proteolysis of OVA was measured by TCA precipitation and derivatization of free amines with fluorescamine. We used purified recombinant Methanosarcina thermophila proteasome, rather than RAW cell lysates in these experiments, to minimize interference of low-molecular-weight peptide amines present in cell lysates with the fluorescamine assay. Degradation of OVA treated with one-molar equivalent E2-IsoK was reduced significantly compared with vehicle-treated OVA after 5 and 20 h of incubation (P
