Amyloid precursor protein carboxy-terminal fragments modulate G-proteins and adenylate cyclase activity in Alzheimer’s disease brain

Molecular Brain Research 117 (2003) 73–82 www.elsevier.com / locate / molbrainres

Research report

Amyloid precursor protein carboxy-terminal fragments modulate Gproteins and adenylate cyclase activity in Alzheimer’s disease brain ¨ Riina Mahlapuu b , Kaido Viht a,c , Lajos Balaspiri a,d , Nenad Bogdanovic e , Kulliki Saar a , ¨ Langel a , * Ursel Soomets a,b , Tiit Land a , Mihkel Zilmer b , Ello Karelson a,b , Ulo a

¨ 21 A, S-10691 Stockholm, Sweden Department of Neurochemistry and Neurotoxicology, Stockholm University, Svante Arrhenius vag b Department of Biochemistry, Tartu University, 50411 Tartu, Estonia c Department of Chemical Physics, Tartu University, 50411 Tartu, Estonia d Chemical Research Centre, Hungarian Academy of Sciences, Institute of Chemistry, Department of Bioorganic Chemistry, H-1025 Budapest, Hungary e Geriatric Department, NEUROTEC, Karolinska Institute, S-14186 Huddinge, Sweden Accepted 2 July 2003

Abstract The influence of three C-terminal sequences and of transmembrane domain from amyloid precursor protein (APP) on the activity of G-proteins and of the coupled cAMP-signalling system in the postmortem Alzheimer’s disease (AD) and age-matched control brains was compared. 10 mM APP(639–648)–APP(657–676) (PEP1) causes a fivefold stimulation in the [ 35 S]GTPgS-binding to control hippocampal G-proteins. APP(657–676) (PEP2) and APP(639–648) (PEP4) showed less pronounced stimulation whereas cytosolic APP(649–669) (PEP3) showed no regulatory activity in the [ 35 S]GTPgS-binding. PEP1 also showed 1.4-fold stimulatory effect of on the high-affinity GTPase and adenylate cyclase activity in control membranes, whereas in AD hippocampal membranes the stimulatory effect of PEP1 was substantially weaker. The PEP1 stimulation of the [ 35 S]GTPgS-binding to the control membranes was significantly reduced by 1.5 mM glutathione, 0.5 mM antioxidant N-acetylcysteine and, in the greatest extent, by 0.01 mM of desferrioxamine. In AD hippocampus these antioxidants revealed no remarkable reducing effect on PEP1-induced stimulation. Our results suggest that C-terminal and transmembrane APP sequences possess receptor-like G-protein activating function in human hippocampus and that abnormalities of this function contribute to AD progression. The stimulatory action of these sequences on G-protein mediated signalling suggests the region-specific formation of reactive species.  2003 Elsevier B.V. All rights reserved. Theme: Disorders of the nervous systems Topic: Degenerative disease: Alzheimer’s—beta amyloid Keywords: APP C-terminus; G-protein; Hippocampus; Alzheimer’s disease; Oxidative stress

1. Introduction Amyloid precursor protein (APP) represents an ubiquitous transmembrane glycoprotein with a large N-terminal ectodomain, a single membrane-spanning domain, and a short cytoplasmatic carboxy-terminus (CT). APP can function both as a membrane-anchored receptor-like molecule and as a secreted derivative that acts upon other cells [11,32]. *Corresponding author. Tel.: 146-8-161-793; fax: 146-8-161-371. E-mail address: [email protected] (.U. ¨ Langel). 0169-328X / 03 / $ – see front matter  2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0169-328X(03)00292-4

The cytoplasmatic CT of APP is suggested to regulate APP metabolism and functions in normal and Alzheimer’s disease (AD) brain. The specific CT sequences of APP are known to act as recognition sites for cytoplasmatic signals directing APP to subcellular compartments [31,42]. The completely conserved cytoplasmatic APP sequence, His657–Lys676, is reported to form a complex and to activate G o , a major GTP-binding protein in the brain [3,24]. When the His657–Lys676 sequence is connected to a APP transmembrane sequence Thr639–Leu648, its potency to stimulate G o is enhanced 20 times compared to the transmembrane or the cytoplasmic sequence alone [24].

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

74

Fig. 1. Location of the peptides PEP1, PEP2, PEP3 and PEP4 in amyloid precursor protein APP695 .

There are some suggestions that abnormalities in APP cytoplasmic domain can weaken / disrupt G-protein stimulating function of APP [25]. It has been shown that G-protein associated signalling pathways in the AD post-mortem brains are altered, particularly, the G-protein mediated phosphoinositide hydrolysis and adenylate cyclase pathways. Disruption in signalling can be caused by an uncoupling of G-proteins from receptors or by decreased levels of G-proteins in different regions of AD brain, including frontal cortex, hippocampus, etc. (for a review, cf. Ref. [7]). In addition, a decrease in G-protein GTP hydrolysis activity in early stages of AD in cortex and hippocampus has been reported [10]. Excessive production of APP and / or reduced activity of lysosomal / endosomal enzymes may induce the accumulation of amyloidogenic APP fragments in AD and Down’s syndrome brain [22,30]. In accordance with recent studies, not only amyloid-b peptides, but also the other APP reactive products, e.g., CT sequences, might be involved in the amyloidogenesis and neurodegeneration through freeradical generated oxidative stress [16,37,38]. Experiments with various cell-lines have revealed that APP CT-fragments can interact with plasma membranes to form nonion-selective channels or pores resulting in cell death and neuronal degeneration [8,12,15]. However, the mechanism behind functional activity of APP CT-fragments in the

human brain membranes is not known. In addition, several unanswered questions still remain about alterations in the activity of CT-fragments in AD brain where diseaseassociated disturbances exist in the neuronal membrane architecture and turnover [5,29]. Our aim was to study the effects of APP CT peptides on G-proteins and adenylate cyclase activity in the membranes from AD and age-matched control hippocampus, a region that shows high level of amyloidogenesis and profound neuronal degeneration in AD [1,33,34]. Here we report the effects of four APP CT peptides: PEP1 [APP(639–648)–APP(657–676) amide], PEP2 [APP(657– 676) amide], PEP3 [APP(649–669) amide] and PEP4 [APP(639–648) amide] (Fig. 1, Table 1). Since the PEP1 was found to stimulate hippocampal G-proteins as well as purified G o [24] in a higher degree than the PEP2, PEP3 or PEP4, an attempt was made to elucidate a role of free radicals in the PEP1 stimulation of G-proteins in the hippocampal region in AD brain.

2. Materials and methods

2.1. Materials Tritiated cyclic AMP (59 Ci / mmol) was purchased from Amersham (Buckinghamshire, UK). [ 35 S]GTPgS (1250

Table 1 The peptide sequences from the C-terminus of amyloid precursor protein (APP) used in this study Name

Peptide

Sequence

PEP1 PEP2 PEP3 PEP4

APP(639–648)–APP(657–676) amide APP(657–676) amide APP(649–669) amide APP(639–648) amide

TVIVITLVMLHHGVVEVDAAVTPEEHLSK amide HHGVVEVDAAVTPEERHLSK amide KKKQYTSIHHGVVEVDAAVTP amide TVIVITLVML amide

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

Ci / mmol) was obtained from NEN Life Science Products (Boston, MA, USA). GDP, dithiothreitol, cyclic AMP, phosphoenol pyruvate, bacitracin and bovine serum albumin were from Fluka or Sigma–Aldrich (Deisenhofen, Germany). All reagents used were of highest analytical grade.

2.2. Peptide synthesis Peptides were synthesized in a stepwise manner in a 0.1 mmol scale manually or on the Applied Biosystem Model 431 A peptide synthesizer (Applied Biosystem, Foster City, CA, USA) on solid support using dicyclohexylcarbodiimide / hydroxybenzotriazole activation strategy. tert.Butyloxycarbonyl amino acids were coupled as hydroxybenzotriazole esters to a p-methyl-benzhydrylamine resin (1.1 mmol / g; Bachem, Switzerland) to obtain C-terminally amidated peptides. The dinitrophenyl protecting groups on histidine were removed by treatment for 1 h at room temperature with 20% (v / v) thiophenol / dimethylformamide. The peptides were cleaved from the resin with liquid hydrofluoric acid (HF) at 0 8C for 1 h in the presence of scavenger p-cresol or thiocresol /p-cresol mixture when Met was present in the sequence. The peptides were purified using high-performance liquid chromatography (HPLC; Gynkotek, Munich, Germany) on a Nucleosil C 18 reversed-phase column (2532.4 cm), and the purity of peptides was .99% as demonstrated with HPLC on an analytical column (1030.4 cm). The molecular masses of the peptides were determined with a plasma desorption mass spectrometer (Bioion 20, Uppsala, Sweden), and the calculated values were obtained in each case.

2.3. Brain tissue sampling and preparation of membranes Hippocampal and frontal cortical tissues of postmortal human brain were obtained from Huddinge Brain Bank, Sweden. The present study included the regions from eight female AD patients and eight female control subjects (mean age6S.D.: 7968 and 84612 years, respectively). The AD patients met clinical and histological criteria for AD [2,20]. The control group consisted of patients without history of neurological or psychiatric disorders. The brain regions were rapidly dissected and kept at 270 8C prior to experiment. The postmortem time for all brains was under 24 h. The hippocampal (and frontocortical) membranes for the [ 35 S]GTPgS binding measurement and for the GTPase and adenylate cyclase assay were mainly prepared according to the protocol [13]. For the final resuspension of the membrane pellet either 10 mM Tris–HCl, pH 7.4 ([ 35 S]GTPgS binding, GTPase activity) or 30 mM Tris– HCl, containing 1.5 mM theophylline, 8.25 mM MgCl 2 , 0.75 EGTA, 7.5 mM KCl and 100 mM NaCl (adenylate cyclase activity) was used.

75

2.4. [ 35 S] GTPg S-binding assay The brain membranes with the final protein concentration of 0.04 mg / ml were incubated in a reaction cocktail containing standard TE buffer (10 mM Tris–HCl10.1 mM EDTA, pH 7.4), GDP (1 mM), dithiothreitol (1 mM), MgCl 2 (5 mM), NaCl (150 mM) and [ 35 S]GTPgS (50,000–70,000 cpm in an aliquot of the reaction cocktail). Incubation was carried out for 2 min at 26 8C in a total volume of 0.1 ml and in the absence (basal value) and presence of various concentrations of peptides. Bound and free [ 35 S]GTPgS were separated by vacuum filtration through GF / B filters (Whatman International, Mainstone, UK), which were washed three times with 5 ml of ice-cold TE buffer. Radioactivity was quantified by liquid scintillation counting. In data analysis, the basal [ 35 S]GTPgS binding was defined as 100%.

2.5. Measurement of GTPase activity The activity of GTPase was assayed radiometrically according to the protocol by McKenzie [21], modified by Zorko et al. [44]. Diluted membranes (final protein concentration 0.2 mg / ml) were added to the ice-cold reaction cocktail, containing in standard TE buffer 1 mM ATP, 1 mM 59-adenylylimido-disphosphate, 1 mM ouabain, 10 mM phosphocreatine, 2.5 units / ml creatine phosphokinase, 4 mM dithiothreitol, 5 mM MgCl 2 , 100 mM NaCl and trace amounts of [g- 32 P]GTP to give 50,0002100,000 cpm in an aliquot of the reaction cocktail. To achieve the required total concentration of 0.5 mM of GTP, an additional non-labeled GTP was added to the cocktail. Background low-affinity hydrolysis of [g- 32 P]GTP was assessed in the presence of 100 mM GTP. Reaction in standard TE buffer was carried out for 12 min at 25 8C and terminated by adding of 5% charcoal suspension containing 20 mM phosphoric acid. After centrifugation at 1200 g for 12 min the radioactivity in the supernatant was determined by a Packard 3255 liquid scintillation counter. In control hippocampus and frontal cortex, the basal (unaffected) GTPase activity showed values of 1.060.1 and 1.360.1 pmol Pi / min / mg protein, respectively. In AD brain regions, the corresponding values were lower: 0.660.1 and 1.060.1 pmol Pi / min / mg protein, respectively.

2.6. Measurement of adenylate cyclase activity The basal adenylate cyclase activity was assayed by determining the formed cAMP in the reaction media (final volume of 150 ml) containing a reaction buffer (30 mM Tris–HCl buffer, 1.5 mM theophylline, 8.25 mM MgCl 2 , 0.75 mM EGTA, 7.5 mM KCl and 0.1 M NaCl, 100 mg / ml bacitracin, 0.03% bovine serum albumin, 10 mM phosphoenol-pyruvate, 30 mg / ml pyruvate kinase) and hippocampal membranes with the final protein concen-

76

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

tration of 0.04 mg / ml [41]. The peptides, dissolved in the reaction buffer, were added 2 min before the reaction was initiated by 10 mM ATP/ 10 mM GTP. The reaction at 30 8C was terminated after 15 min by the addition of 100 mM EDTA and boiling the samples for 3 min. The cyclic AMP content in the tubes was measured by a competitive protein saturation assay [4]. Mean basal values of the adenylate cyclase activities in the membranes from control and AD hippocampus were 4863 and 3761 pmol cyclic AMP/ min / mg protein, respectively.

2.7. Treatment of hippocampal membranes with H2 O2 To induce oxidative stress, TE-buffered control or AD hippocampal membranes (protein 0.4 mg / ml) were treated with 10 mM of H 2 O 2 at 30 8C for 5 min. The reaction was terminated by 10-fold dilution of samples with ice-cold [ 35 S]GTPgS-binding media (see above). The effect of H 2 O 2 on the stimulation by PEP1 of [ 35 S]GTPgS-binding was estimated as a difference in the stimulation of binding found in the absence or presence of H 2 O 2 . The effect of H 2 O 2 -treatment on the basal [ 35 S]GTPgS-binding was also elucidated. The effects were expressed as percentages of basal (unaffected) radionucleotide binding (5100%).

2.8. Incubation of hippocampal membranes with antioxidants To elucidate the effects of antioxidants on the stimulation by PEP1 of [ 35 S]GTPgS-binding in control or AD hippocampal membranes, 0.01 mM of ferrous ion chelator desferrioxamine (DFA) as well as 1.5 mM of reduced

glutathione (GSH) or 0.5 mM of N-acetyl-L-cysteine (NAC) were added to the medium before the peptide. The effects of the antioxidants were estimated as a difference in the stimulation of binding in the absence or presence of antioxidants. In parallel, the effect of antioxidants on the basal [ 35 S]GTPgS-binding was studied.

2.9. Statistical analysis Data were expressed as the mean6S.E.M. for each statistical group (n53–6 experiments in duplicate). Student’s t-test for independent samples was used to determine significant differences between the groups. Values for P#0.05 were taken to indicate significant difference.

3. Results The effects of four peptides, derived from APP Cterminus (PEP1, PEP2, PEP3) or from APP transmembrane domain (PEP4), on [ 35 S]GTPgS-binding to G-proteins in the human control and AD hippocampal membranes are shown in Fig. 2A and B. In the control membranes, the PEP1 produced a dose-dependent stimulation of [ 35 S]GTPgS-binding with the maximal value of 440.1% at 10 mM (EC 50 54.6 mM). In AD hippocampal membranes, the maximal stimulation of [ 35 S]GTPgS-binding by the PEP1 was about three times lower (133.3%; EC 50 51.3 mM). The PEP2 stimulated [ 35 S]GTPgS-binding in both control and AD membranes by 100% and the PEP4, the transmembrane sequence, by 153 and 120% in Co and AD, respectively. At the same time, the PEP3 had no significant effect on [ 35 S]GTPgS-binding (Figs. 2A and

Fig. 2. Effect of the peptides PEP1 (j), PEP2 (h) and PEP3 (♦) and PEP4 (s) on [ 35 S]GTPgS binding in the hippocampal membranes (protein 0.04 mg / ml) from the age-matched control (A) and AD (B) brain. 100% corresponds to the basal [ 35 S]GTPgS binding (mean6S.E.M.; for the PEP1 n56; for the other APP derived peptides n53–4).

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

77

Fig. 3. Effect of the PEP1 on [ 35 S]GTPgS binding in the hippocampal (A) and frontocortical (B) membranes (protein 0.04 mg / ml) from the age-matched control (h) and AD (j) brain. For hippocampus, n56; for frontal cortex, n54.

1B). In control brain tissues, the maximal stimulation of G-proteins by PEP1 was 1.5-fold higher in the hippocampal as compared to frontocortical membranes (Fig. 3), whereas in AD patients, the stimulatory effect of the peptide showed no remarkable difference between these regions. The effects of PEP1 on GTPase activity in the hippocampal and frontocortical membranes of control and AD brains are presented in Fig. 4A and B. The GTPase activity in control hippocampal and control frontal cortex membranes was stimulated to a maximal effect of 42.8 and 39.7%, respectively, with EC 50 values at 5.1 and 4.8 mM. However, in AD hippocampus, the maximal stimulation of

GTPase activity was declined to 14.9% (EC 50 at 3 mM) and in AD frontal cortex to 25.2% (EC 50 at 3.9 mM). The effect of the PEP1 on the adenylate cyclase activity in the control and AD hippocampal membranes was also studied (Fig. 5). In control brain membranes, the peptide produced the stimulation of the enzyme activity with maximal value of 43% (EC 50 57.6 mM), the effect being similar to that found for the hippocampal GTPase. In AD hippocampal membranes, the stimulatory effect of the PEP1 on adenylate cyclase activity was lower (maximal value of 25.8%; EC 50 511 mM). To elucidate the mechanism of stimulatory effect of PEP1 on [ 35 S]GTPgS-binding in the hippocampal mem-

Fig. 4. Effect of the PEP1 on the GTPase activity in the hippocampal (A) and frontocortical (B) membranes (in both, protein 0.2 mg / ml) from the age-matched control (h) and AD (j) brain. 100% corresponds to the basal GTPase activity (n53).

78

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

3–10 mM of PEP1, the binding of radio-nucleotide increased approximately 1.5-fold (P,0.1) over the H 2 O 2 induced level. The effect of antioxidants on G-proteins stimulation by PEP1 in the control and AD hippocampal membranes was studied. We found that classical antioxidants, GSH and NAC, as well as ferrous iron chelator, DFA, enhanced basal [ 35 S]GTPgS binding (Fig. 7). In both, control and AD membranes, the stimulation of [ 35 S]GTPgS binding by 0.5 mM of NAC (90.9 and 72.5%, respectively) and 0.01 mM of DFA (86 and 59%, respectively) was higher than by 1.5 mM of GSH (48.7 and 26.7%, respectively). The simultaneous presence of 5–10 mM of PEP1 and antioxidants in the incubation media significantly decreased stimulation of [ 35 S]GTPgS binding by the PEP1. In control membranes, DFA caused a 2.5-fold reduction in stimulatory effect of the peptide at 10 mM; the protective effect of DFA tended to be more potent than that of GSH or NAC. In AD membranes, all used antioxidants showed weaker protection against PEP1-induced stimulation of G-proteins as compared to control, whereas the protective effect of NAC did not reach a significant value (Fig. 6). Fig. 5. Effect of the PEP1 on the adenylate cyclase activity in the hippocampal membranes (protein 0.04 mg / ml) from the age-matched control (h) and AD (j) brain. 100% corresponds to the basal adenylate cyclase activity (in Co, n53; in AD, n54).

branes, we investigated whether this effect could be modified by H 2 O 2 as a reactive oxidant. Pre-treatment of control membranes with 10 mM H 2 O 2 induced the stimulation of basal [ 35 S]GTPgS-binding by 258%, the effect being 1.5-times higher than in AD membranes (Fig. 6). In both cases, co-treatment of membranes with PEP1 and H 2 O 2 enhanced the stimulation induced by H 2 O 2 alone. At

4. Discussion Recent studies on cell transplantation models and transgenic mice have shown that APP CT fragments exert a variety of effects on the cellular membrane functions [6,9,19,27]. Induction of non-selective ion currents, stimulation of G o -proteins and other membrane processes induced by CT fragments appear to be involved not only in APP physiological activity, but also in the neurodegeneration and AD development [31,38,39]. However, the se-

Fig. 6. Effect of the PEP1 on [ 35 S]GTPgS binding in the aged-matched control (A) and AD (B) hippocampal membranes (protein 0.04 mg / ml) before (h) and after (j) their treatment with 10 mM H 2 O 2 (5 min 30 8C); n54 (after treatment of Co membranes) or n53 (after treatment of AD membranes).

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

79

Fig. 7. Effect of the PEP1 on [ 35 S]GTPgS binding in the hippocampal membranes (protein 0.04 mg / ml) from the aged matched control (A) and AD (B) brain in the presence of 1.5 mM glutathione (GSH), 0.5 mM of N-acetyl-L-cysteine (NAC) or 0.01 mM of desferrioxamine (DFA); mean6S.E.M.; n53–4.

quences of APP C-terminus, which might affect signal transduction in human brain membranes, have not been established yet. In the present study, we compared the effect of three CT sequences and transmembrane domain of APP on the [ 35 S]GTPgS binding to G-proteins in the hippocampal and frontocortical membranes obtained from both, control and AD postmortem brains. Furthermore, the effect of the most active CT sequence, truncated peptide PEP1, on the high-affinity GTPase and adenylate cyclase activity was examined. In addition, the involvement of plasma-membrane oxidative processes in the stimulation of [ 35 S]GTPgS binding by the PEP1 was elucidated. To our knowledge, this is the first report concerning the functional activity of APP CT-peptides at the level of human brain signalling systems in the case of AD. Our results showed a structure–activity relationship among the tested APP sequences with respect to the stimulation of [ 35 S]GTPgS binding to the control and AD hippocampal membranes (see Fig. 2). In the control membranes, the PEP1, consisting of transmembrane and cytosolic sequences of APP, elicited unexpectedly high, fivefold at 10 mM concentration, stimulation of the radionucleotide binding. PEP2, the CT fragment without the transmembrane domain, induced only twofold and the transmembrane PEP4 2.5-fold stimulation. Nishimoto et al. demonstrated strong potentiating effect of transmembrane APP(639–648) on the cytosolic APP(657–676) stimulatory function in regard to the vesicle-incorporated G o protein [24]. Our data confirm more than twofold augmentative role of the same transmembrane domain for the PEP2 stimulation of G-proteins in the human control hippocampus. Decrease at the level of G o - and minor types of Gproteins [7,17,33], decline in the G-protein GTP hydrolysis [10,28] and impairment of the coupled signal transduction pathways have been described for the more injured AD

brain regions. This study revealed that PEP1 stimulation of [ 35 S]GTPgS binding to AD hippocampal membranes is significantly lower than in the corresponding control region. Similarly, PEP1-induced stimulation of [ 35 S]GTPgS binding was markedly lower in AD frontal cortex compared to control (see Figs. 2 and 3). Moreover, PEP1 stimulation of high-affinity GTPase in AD hippocampus and frontal cortex revealed significant decline from the values detected for the control regions (see Fig. 3). These data suggest that AD leads to considerable dysfunction / down-regulation of the G-protein (sub)types preferentially stimulated by the PEP1 sequence in the control regions. Certain differences in the decline of PEP1 stimulatory effect on G-proteins in AD hippocampus and frontal cortex (Figs. 2 and 3) might be related to the region-specific alterations in the membrane composition and in the G-protein garniture, induced by AD [7,17]. Investigations of G-protein levels and activities in postmortem AD and age-matched control brains have shown that G-protein subtypes and related signal transduction pathways might be differently affected. While the functional activity of G o -protein, major contributor to high-affinity GTPase, appears to be reduced, the ability of G i to inhibit cAMP signalling system is reported to be unaltered in the affected regions [26,28]. Furthermore, a decrease in the activity of G s -proteins and of the coupled cAMP system has been demonstrated in the more injured AD brain regions compared to controls [26]. Our studies have revealed a significant 1.4-fold stimulation of adenylate cyclase by the PEP1 sequence in the control hippocampal membranes whereas in the corresponding AD brain region this effect was significantly lower (see Fig. 5). The latter might be explained by dysfunction of the G s proteins, which at normal conditions mediate a marked stimulatory signal from the truncated peptide to the enzyme. The 1.5-fold stimulation of [ 35 S]GTPgS binding

80

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

by PEP1 in the membranes from Sf9 G s overexpressing cells (data not shown) corroborates the assumption that G s proteins are involved in the stimulation of [ 35 S]GTPgS binding by this peptide. Reactive oxygen species have been repeatedly shown to act as stimulators of signal transduction pathways [40]. More recently, H 2 O 2 gained attention as a stimulator of [ 35 S]GTPgS binding to cardiac plasma membranes and as a direct activator of the G i and G o protein a-subunits [23]. Our results are consistent with these findings and show H 2 O 2 -stimulation of the [ 35 S]GTPgS binding to the control hippocampal membranes, the effect being remarkably higher than in the corresponding AD brain region. The more injured AD brain regions (including hippocampus) revealed significantly higher levels of peroxidation intermediates and end products than the regions from control [14,18]. These oxidative stress associated alterations could lead to the lower content of readily peroxidable substrates and thereby to the decline in H 2 O 2 stimulation of [ 35 S]GTPgS binding to AD hippocampal membranes. Enhancement by 1.5-fold of the H 2 O 2 -induced stimulation of [ 35 S]GTPgS binding by PEP1 in control and AD hippocampal membranes (Fig. 6) suggests that stimulatory effect of the peptide on the hippocampal G-proteins realizes via a complex mechanism possibly implying an oxidative modification of the specific target(s) in the regional G-protein garniture. Such suggestion is predicted by the recent studies, which have revealed that Ab(25–35) and the other functionally active APP sequences increase cellular levels of reactive (including oxygen) species [37,43]. In addition, classical free radical scavangers, GSH and NAC, were shown to protect neuronal cells and rat brain adenylate cyclase from Ab(25–35) toxic effects [35]. Consistently, in the present study, GSH and NAC, as well as ferrous iron chelator desferrioxamine, caused the significant decrease in the potent stimulatory effect of the truncated APP sequence on G-proteins in the hippocampal membranes. Desferrioxamine tended to be more potent protector than NAC or GSH, suggesting that, in the control as well in AD hippocampus, the preferential stimulatory mechanism of the PEP1 implies iron-derived formation of reactive species. Decreased protective ability of antioxidants in AD hippocampal membranes compared to the same area of control (Fig. 7) suggests that G-proteins and coupled effectors of AD hippocampus have much lower prerequisite to be protected against the PEP1-induced oxidative stimulation. Significant decline in the antioxidant capacity in the more injured regions of AD brain as hippocampus and associative cortex has previously been demonstrated [14,18]. The accumulation of redox active iron in AD hippocampus, an important source of highly reactive free radicals and contributor of oxidative damage [36], might be an additional causative factor, lowering the protective effect of antioxidants against PEP1-stimulation of G-proteins. In conclusion, the studied APP fragments sequence-

dependently stimulate the activity of G-proteins in the human control and AD hippocampal membranes. In the control membranes, the PEP1, consisting of transmembrane and short cytosolic sequence of APP, seems to function as a receptor, capable for the potent stimulation of [ 35 S]GTPgS binding, activation of high-affinity GTPase and for transducing the stimulatory signals to adenylate cyclase. The cytosolic PEP2 and transmembrane PEP4 reveal a relatively weak stimulation of G-proteins whereas the cytosolic sequence PEP3 is not capable to function as a signalling structure. The stimulatory effect of the PEP1 on the signal transduction appears to have a dependence on the brain region G-proteins’ garniture. In the membranes from the more affected AD brain regions (hippocampus and frontal cortex) the stimulatory activity of the PEP1 was substantially declined, probably due to the (region) specific dysfunctions in G-proteins and coupled signalling systems at which severe oxidative stress might serve as a main reason. The potent stimulatory effect of PEP1 on the control hippocampal G-proteins appears to be mediated by free radical induced mechanism prevented by GSH, NAC and, in the greatest extent, by desferrioxamine. The preferential pro-oxidant mechanisms of the PEP1 might imply the iron-derived formation of reactive species. Therefore, antioxidants could form the basis for the development of drugs modulating the actions of APP CT fragments on signal transduction in the normal brain and in AD progression.

Acknowledgements This work was financially supported by the Estonian Scientific Foundation (grant Nos. 3260, 4913), by the Swedish Institute, Stockholm, Sweden (scholarships to E.K. in 1998 and 2000), by The New Visby Project, Swedish Institute (U.S.) by the L&H Osterman Foundation and by the Karolinska Institute Foundation for Elderly Research. The authors sincerely thank Mrs. Inga Volkman from the Geriatric Department, Neurotec, Karolinska Institute, Sweden for her excellent technical assistance.

References [1] M.J. Ball, M. Fisman, V. Hachinski, V. Blume, A. Fox, V.A. Kral, A.J. Kirchen, H. Fox, H. Merskey, A new definition of Alzheimer’s disease: a hippocampal dementia, Lancet 1 (1985) 14–16. [2] N. Bogdanovic, J.C. Morris, Diagnostic criteria for Alzheimer’s disease in multicentre brain banking, in: F.F. Cruz-Sanchez, R. Ravid, M.L. Cuziner (Eds.), Neuropathological Diagnostic Criteria For Brain Banking, Biomedical and Health Research Series, Vol. 10, IOS Press, Amsterdam, 1995, pp. 20–29. [3] E. Brouillet, A. Trembleau, D. Galanaud, M. Volovitch, C. Bouillot, C. Valenza, A. Prochiantz, B. Allinquant, The amyloid precursor protein interacts with Go heterotrimeric protein within a cell compartment specialized in signal transduction, J. Neurosci. 19 (1999) 1717–1727.

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82 [4] B.L. Brown, R.P. Ekins, J.D. Albano, Saturation assay for cyclic AMP using endogenous binding protein, Adv. Cycl. Nucl. Res. 2 (1972) 25–40. [5] R. Buchet, S. Pikula, Alzheimer’s disease: its origin at the membrane, evidence and questions, Acta Biochim. Pol. 47 (2000) 725– 733. [6] Y.H. Chong, J.H. Sung, S.A. Shin, J.-H. Chung, Y.H. Suh, Effects of the b-amyloid and carboxyl-terminal fragment of Alzheimer’s amyloid precursor protein on the production of the tumor necrosis factor-a and matrix metalloproteinase-9 by human monocytic THP1, J. Biol. Chem. 276 (2001) 23511–23517. [7] R.F. Cowburn, C. O’Neill, W.L. Bonkale, T.G. Ohm, J. Fastbom, Receptor-G-protein signalling in Alzheimer’s disease, Biochem. Soc. Symp. 67 (2001) 163–175. [8] S.P. Fraser, Y.H. Suh, M.B. Djamgoz, Ionic effects of the Alzheimer’s disease beta-amyloid precursor protein and its metabolic fragments, Trends Neurosci. 20 (1997) 67–72. [9] K.I. Fukuchi, D.D. Kunkel, P.A. Schwartzkroin, K. Kamino, C.E. Ogbrun, C.E. Furlong, G.M. Martin, Overexpression of a C-terminal portion of the beta-amyloid precursor protein in mouse brains by transplantation of transformed neuronal cells, Exp. Neurol. 127 (1994) 253–264. [10] A. Garcia-Jimenez, R.F. Cowburn, T.G. Ohm, H. Lasn, B. Winblad, N. Bogdanovic, J. Fastbom, Loss of stimulatory effect of gunasine triphosphate on [(35)S]GTPgammaS binding correlates with Alzheimer’s disease neurofibrillary pathology in entorhinal cortex and CA1 hippocampal subfield, J. Neurosci. Res. 67 (2002) 388–398. [11] J. Kang, H.-G. Lemaire, A. Unterback, J.M. Salbaum, C.L. Masters, ¨ K.H. Grezeschik, G. Multhaup, K. Beyereuther, B. Muller-Hill, The precursor of Alzheimer’s disease amyloid A4 protein resembles a cell-surface receptor, Nature 325 (1987) 733–736. [12] M. Kawahara, Y. Kuroda, Molecular mechanism of neurodegeneration induced by Alzheimer’s b-amyloid protein: channel formation and disruption of calcium homeostasis, Brain Res. Bull. 53 (2000) 389–397. [13] E. Karelson, J. Laasik, R. Sillard, Regulation of adenylate cyclase by galanin, neuropeptide Y, secretin and vasoactive intestinal polypeptide in rat frontal cortex, hippocampus and hypothalamus, Neuropeptides 28 (1995) 21–28. [14] E. Karelson, N. Bogdanovic, A. Garlind, B. Winblad, K. Zilmer, T. Kullisaar, T. Vihalemm, C. Kairane, M. Zilmer, The cerebrocortical areas in normal brain aging and in Alzheimer’s disease: noticeable differences in the lipid peroxidation level and antioxidant defence, Neurochem. Res. 26 (2001) 353–361. [15] S.-H. Kim, C.H. Park, S.H. Cha, J.-H. Lee, S. Lee, Y. Kim, J.-C. Rah, S.-J. Jeong, Y.-H. Suh, Carboxyl-terminal fragment of Alzheimer’s APP destabilizes calcium homeostasis and renders neuronal cells vulnerable to excitotoxicity, FASEB J. 14 (2000) 1508– 1517. [16] S.-H. Kim, Y.-H. Suh, Neurotoxicity of a carboxyl-terminal fragment of the Alzheimer’s amyloid precursor protein, J. Neurochem. 67 (1996) 1172–1182. [17] K. Kolasa, L.E. Harrell, D.S. Parsons, R. Powers, Densitometric analysis of Galphao protein subunit levels from postmortem Alzheimer’s disease hippocampal and prefrontal cortical membranes, Alzheimer Dis. Assoc. Disord. 14 (2000) 53–57. [18] M.A. Lovell, W. Ehmann, S.M. Butler, W.R. Markesbery, Elevated thiobarbituric acid-reactive substances and antioxidant enzyme activity in the brain in Alzheimer’s disease, Neurology 45 (1995) 1594–1601. [19] Y. Matsumoto, S. Watanabe, Y.H. Suh, T. Yamamoto, Effects of intrahippocampal CT105, a carboxyl terminal fragment of betaamyloid precursor protein, alone / with inflammatory cytokines on working memory in rats, J. Neurochem. 82 (2002) 234–239. [20] G. McKhann, D. Drachman, M. Folstein, D. Katzman, D. Price, E.M. Stadlan, Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of Depart-

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

81

ment of Health and Human Services Task Forces on Alzheimer’s disease, Neurology 34 (1994) 939–994. F.R. McKenzie, Basic techniques to study G-protein function, in: G. Milligan (Ed.), Signal Transduction. A Practical Approach, Oxford University Press, Oxford, 1992, pp. 31–56. R.L. Neve, A. Kammesheidt, C.F. Hohman, Brain transplants of cells expressing the carboxy-terminal fragment of the Alzheimer amyloid precursor protein cause specific neuropathology in vivo, Proc. Natl. Acad. Sci. USA 89 (1992) 3448–3452. M. Nishida, Y. Maruyama, R. Tanaka, K. Kontani, T. Nagao, H. Kurose, G alpha (i) and G alpha(o) are target proteins of reactive oxygen species, Nature 408 (2000) 492–495. I. Nishimoto, T. Okamoto, Y. Matsuura, S. Takahashi, T. Okamoto, Y. Murayama, E. Ogata, Alzheimer amyloid protein precursor complexes with brain GTP-binding protein G o , Nature 362 (1993) 75–79. T. Okamoto, S. Takeda, Y. Murayama, E. Ogata, I. Nishimoto, Ligand-dependent G-protein function of amyloid transmembrane precursor, J. Biol. Chem. 270 (1995) 4205–4208. C. O’Neill, B. Wiehager, C.J. Fowler, R. Ravid, B. Winblad, R.F. Cowburn, Regionally selective alterations in G-protein subunit levels in the Alzheimer’s disease brain, Brain Res. 636 (1994) 193–201. M.L. Oster-Granite, D.L. McPhie, J. Greenan, R.L. Neve, Agedependent neuronal and synaptic degeneration in mice transgenic for C terminus of the amyloid precursor protein, J. Neurosci. 16 (1996) 6732–6741. B.M. Ross, M. McLaughlin, M. Roberts, G. Milligan, J. McCulloch, J.T. Knowler, Alterations in the activity of adenylate cyclase and high affinity GTPase in Alzheimer’s disease, Brain Res. 622 (1993) 35–42. G.S. Roth, J.A. Joseph, R.P. Mason, Membrane alterations as causes of impaired signal transduction in Alzheimer’s disease and aging, Trends Neurosci. 18 (1995) 203–206. C. Russo, S. Salis, V. Dolcini, V. Venezia, X.-H. Song, J.K. Teller, G. Schettini, Identification of amino-terminally and phosphoptyrosinemodified carboxy-terminal fragments of the amyloid precursor protein in Alzheimer’s disease and Down’s syndrome brain, Neurobiol. Dis. 8 (2001) 173–180. D.J. Selkoe, Cell biology of the amyloid beta-protein precursor and the mechanism of Alzheimer’s disease, Annu. Rev. Cell Biol. 10 (1994) 373–403. D.J. Selkoe, The cell biology of beta-amyloid precursor protein and presenilin in Alzheimer’s disease, Trends Cell Biol. 8 (1998) 447– 453. S. Shimohama, S. Kamiya, T. Taniguchi, Y. Sumida, S. Fujimoto, Differential involvement of small G proteins in Alzheimer’s disease, Int. J. Mol. Med. 3 (1999) 597–600. G. Simic, I. Kostovic, B. Winblad, N. Bogdanovic, Volume and number of neurons of the human hippocampal formation in normal aging and Alzheimer’s disease, J. Comp. Neurol. 379 (1997) 482– 494. ¨ U. Soomets, R. Mahlapuu, R. Tehranian, J. Jarvet, E. Karelson, M. ¨ Langel, Regulation ¨ Zilmer, K. Iverfeldt, M. Zorko, A. Graslund, U. of GTPase and adenylate cyclase activity by amyloid b-peptide and its fragments in rat brain tissue, Brain Res. 850 (1999) 179–188. M.A. Smith, P.L.R. Harris, L.M. Sayre, G. Perry, Iron accumulation in Alzheimer disease is a source of redox-generated free radicals, Proc. Natl. Acad. Sci. USA 94 (1997) 9866–9868. M.A. Smith, C.A. Rottkamp, A. Nunomura, A.K. Raina, G. Perry, Oxidative stress in Alzheimer’s disease, Biochim. Biophys. Acta 1502 (2000) 139–144. Y.-H. Suh, An etiological role of amyloidogenic carboxyl-terminal fragments of the b-amyloid precursor protein in Alzheimer’s disease, J. Neurochem. 68 (1997) 1781–1791. Y.-H. Suh, H.S. Kim, J.P. Lee, C.H. Park, S.I. Jeong, S.S. Kim, J.C. Rah, J.H. Seo, S.S. Kim, Roles of A beta and carboxyl terminal

82

R. Mahlapuu et al. / Molecular Brain Research 117 (2003) 73–82

peptide fragments of amyloid precursor protein in Alzheimer disease, J. Neural Transm. Suppl. 58 (2001) 65–82. [40] Y.J. Suzuki, H.J. Forman, A. Sevanian Oxidants as stimulators of signal transduction, Free Radic. Biol. Med. 22 (1997) 269–285. ¨ Langel, ´ [41] A. Valkna, A. Jureus, E. Karelson, M. Zilmer, T. Bartfai, U. Differential regulation of adenylate cyclase activity in rat ventral and dorsal hippocampus by rat galanin, Neurosci. Lett. 187 (1995) 75–78. [42] T. Yamazaki, E.H. Koo, D.J. Selkoe, Trafficking of cell-surface amyloid beta-protein precursor. II. Endocytosis, recycling and

lysosomal targeting detected by immuno localization, J. Cell Sci. 109 (1996) 999–1008. [43] S.M. Yatin, M. Aksenova, M. Aksenov, W.R. Markesbery, T. Aulick, D.A. Butterfield, Temporal relations among amyloid beta-peptideinduced free-radical oxidative stress, neuronal toxicity and neuronal defensive responses, J. Mol. Neurosci. 11 (1998) 183–197. ¨ Langel, Differential [44] M. Zorko, M. Pooga, K. Saar, K. Rezaei, U. regulation of GTPase activity by mastoparan and galparan, Arch. Biochem. Biophys. 349 (1998) 321–328.