Biochimie 106 (2014) 91e100

Contents lists available at ScienceDirect

Biochimie journal homepage: www.elsevier.com/locate/biochi

Research paper

Anti-amyloidogenic property of human gastrokine 1 Filomena Altieri a, Chiara Stella Di Stadio a, Valeria Severino b, c, Annamaria Sandomenico c, Giuseppina Minopoli a, g, Giuseppina Miselli a, Antimo Di Maro b, Menotti Ruvo c, Angela Chambery b, d, Vincenzo Quagliariello e, Mariorosario Masullo f, Emilia Rippa a, **, Paolo Arcari a, g, * a

Department of Molecular Medicine and Medical Biotechnologies, University of Naples Federico II, Naples, Italy Department of Environmental, Biological and Pharmaceutical Sciences and Technologies, Second University of Naples, Caserta, Italy Institute of Biostructures and Bioimaging-IBB, CNR, Naples, Italy d IRCCS Multimedica, Milan, Italy e Laboratory of Biotechnology, Department of Anesthesia, Surgical and Emergency Sciences, Second University of Naples, Via Costantinopoli 16, I-80138 Naples, Italy f Department of Motor Sciences and Wellness, University of Parthenope, Naples, Italy g CEINGE, Advanced Biotechnology Scarl, Via Gaetano Salvatore 486, I-80145 Naples, Italy b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 20 June 2014 Accepted 10 August 2014 Available online 17 August 2014

Gastrokine 1 (GKN1) is a stomach-specific protein expressed in normal gastric tissue but absent in gastric cancer. GKN1 plays a major role in maintaining gastric mucosa integrity and is characterized by the presence of a BRICHOS domain consisting of about 100 amino acids also found in several unrelated proteins associated with major human diseases like BRI2, related to familial British and Danish dementia and surfactant protein C (SP-C), associated with respiratory distress syndrome. It was reported that recombinant BRICHOS domains from BRI2 and SP-C precursor (proSP-C) prevent fibrils formation of amyloid-beta peptide (Ab), that is the major component of extracellular amyloid deposits in Alzheimer's disease. Here we investigated on the interaction between human recombinant GKN1 (rGKN1) and Ab peptide (1e40) that derives from the partial hydrolysis of the amyloid precursor protein (APP). GKN1 prevented amyloid aggregation and fibrils formation by inhibiting Ab(1e40) polymerization, as evaluated by SDS-PAGE, thioflavin-T binding assay and gel filtration experiments. Mass spectrometry showed the formation of a prevailing 1:1 complex between GKN1 and Ab(1e40). SPR analysis of GKN1/Ab interaction led to calculate a dissociation constant (KD) of 34 mM. Besides its interaction with Ab(1e40), GKN1 showed also to interact with APP as evaluated by confocal microscopy and Ni-NTA pull-down. Data strongly suggest that GKN1 has anti-amyloidogenic properties thus functioning as a chaperone directed against unfolded segments and with the ability to recognize amyloidogenic polypeptides and prevent their aggregation. te  française de biochimie et biologie Mole culaire (SFBBM). All rights © 2014 Elsevier B.V. and Socie reserved.

Keywords: Gastrokine 1 Gastric cancer BRICHOS domain Ab peptide APP Amyloidogenesis

1. Introduction

Abbreviations: Ab(1e40), amyloid-beta peptide 1-40; GKN1, gastrokine 1; flGKN1, full length GKN1; rGKN1, recombinant GKN1; proSP-C, prosurfactant protein C; Ni-NTA, nickel-nitrilotriacetic acid; APP, amyloid precursor protein. * Corresponding author. Department of Molecular Medicine and Medical Biotechnologies, University of Naples Federico II, Via S. Pansini 5, I-8031 Naples, Italy. Tel.: þ39 0817463120; fax: þ39 0817463653. ** Corresponding author. Tel.: þ39 0817463119; fax: þ39 0817463653. E-mail addresses: [email protected] (E. Rippa), [email protected], arcari@ unina.it (P. Arcari).

Gastrokine 1 (GKN1), a tissue-specific protein, is expressed in the human stomach of healthy individuals but is absent in gastric adenocarcinoma tissues [1e3]. Individuals with a low expression of the protein have an increased risk to develop gastric diseases [4,5]. The protein is, in fact, downregulated in gastric mucosa samples infected by Helicobacter pylori, considered as one of the leading cause for gastric cancer development [2]. Evidence suggests that GKN1 is involved in filling the lumen of the surface layer of epithelial cells to maintain the integrity of the mucosa and to regulate cell proliferation and differentiation [1,6]. Human GKN1 is made of 185 amino acids containing a 20 amino acid extracellular

http://dx.doi.org/10.1016/j.biochi.2014.08.004  te  française de biochimie et biologie Mole culaire (SFBBM). All rights reserved. 0300-9084/© 2014 Elsevier B.V. and Socie

92

F. Altieri et al. / Biochimie 106 (2014) 91e100

signal peptide localized in the N-terminal region [1]. The protein is characterized by the presence of a central structural BRICHOS domain [7] of about 100 amino acids containing two conservative cysteine residues most likely involved in disulfide bridges [8]. The association of GKN1 BRICHOS domain with biological functions has been proposed but not yet ascertained. In fact, the BRICHOS domain has been found in proteins associated with a wide range of functions and diseases. These include BRI2, which is related to familial British and Danish dementia (FBD and FDD); Chondromodulin-I (ChM-I), related to chondrosarcoma and lung surfactant protein C (SP-C), related to respiratory distress syndrome (RDS) [7]. Although GKN1 seems to play an important role in the physiology of gastric mucosa and in suppressing carcinogenic processes, a full characterization of its structural and biological activities is still lacking. To accomplish this deficiency, we have recently reported the characterization and secondary structural properties of human recombinant GKN1 (rGKN1). Using bio-informatics tools, we found that GKN1 BRICHOS domain has structural features similar to those endowed by BRICHOS domain-containing protein family [8]. In particular, predicted GKN1 3D models suggested a structural organization of its BRICHOS domain resembling that of the corresponding BRICHOS domain of prosurfactant protein C (proSP-C), a trans membrane (TM) protein expressed in epithelial type II cells. proSP-C contains in the TM region a polypeptide segment defined “discordant a-helix” because its structure, according to secondary structure prediction should form instead a b-strand [9]. This region can misfold and form amyloid fibrils associated with pulmonary disease [10,11]. It has also been shown that the C-terminal domain (CTC) of proSP-C and its BRICHOS domain protect the TM part of proSP-C from aggregation into amyloid [12,13]. Moreover, it has been shown that recombinant BRICHOS domains from BRI2, as reported by Peng et al., 2010 [14], prevents fibrils formation of amyloid-beta peptide (Ab), the major component of extracellular amyloid deposits in Alzheimer's disease [15]. Similar property has also been shown by the precursor of lung surfactant protein C (proSP-C) [13]. These findings led us to explore whether also rGKN1 was endowed with a chaperone-like activity toward amyloidogenic peptides. Using biochemical, spectroscopic and mass spectrometry investigations, we demonstrate that rGKN1 affects protein aggregation by interacting with and preventing fibrils formation of Ab (1e40) peptide. Moreover, using SPR technology we have determined the affinity constant between rGKN1 and Ab(1e40) peptide. The uncovering of this specific property of GKN1 might provide a valid contribution for developing new strategies to prevent protein misfolding. 2. Materials and methods

previously [8]. Stock solutions with concentrations of 313.3 mM rGKN1 were used for the experiments. 2.3. Cell culture, transfection and western blot analysis Human gastric adenocarcinoma cell line (AGS) and human neuroblastoma cell line (SH-SY5Y) were grown in DMEM-F12 (Dulbecco's modified Eagle medium-Cambrex) and DMEM, respectively, supplemented with heat inactivated FBS, 100 U/ml penicillin, 100 mg/mL streptomycin, 1% L-glutamine at 37  C in a 5% CO2 atmosphere and transfected with 4 mg of vector pcDNA 3.1 or pcDNAGKN1 encoding the full length GKN1 (flGKN1, containing the first 20 amino acids leader peptide and His6-Tag sequence at the Cterminal) as already described [16]. The efficiency of transfection of gastric cancer cells with flGKN1 was always evaluated by a parallel transfection using EGFP vector as control. In general, after transfection, the average value of the ratio between number of green fluorescent cells/total number of cells was 0.5 ± 0.1. All experiments were performed in both AGS and SH-SY5Y cells and the results obtained were comparable. Cells were scraped, washed twice in cold phosphate buffered saline (PBS) and resuspended in 20e40 mL of lysis buffer (50 mM TriseHCl pH 7.4, 1% NP40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mg/ml aprotinin, leupeptin, pepstatin, 1 mM Na3VO4, 1 mM NaF) for 30 min on ice and centrifuged at 14,000 g for 20 min at 4  C. Protein extract from cell lines was prepared as already described [2]. Protein concentration was determined by a modified Bradford method [17], using the Bio-Rad (Milan, Italy) protein assay and compared with BSA standard curve. Equal amounts of cytosolic proteins (20 mg) were separated by SDSePAGE, electrotransferred to PVDF membrane and reacted with the different antibodies. Blots were then developed using enhanced chemiluminescence detection reagents (SuperSignal West Pico, Pierce, Rockford, IL, USA.) and exposed to X-ray film. All films were analyzed using Adobe photoshop and Image J software. 2.4. Ni-NTA agarose pull-down Transfected AGS cell extracts (500 mg) were incubated with 50 mL of Ni-NTA Agarose (Qiagen) pre-equilibrated with binding buffer (50 mM NaH2PO4, 300 mM NaCl pH 8.0) for 16 h at 4  C. After incubation the resin was washed 4 times with binding buffer containing 10 mM imidazole to reduce nonspecific bound proteins, resuspended in 30 mL of SDS loading buffer, heated to 95  C for 5 min and subjected to Western blot analysis as described above using specific anti-APP antibody (Acris, San Diego, CA).

2.1. Materials Amyloid b-peptide (1e40) (DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVV) A was purchased from Abnova (Cambridge, UK) and stored as lyophilized powder at 80  C until its use. To obtain monomeric starting solutions, the peptide was dissolved in dimethyl sulfoxide (DMSO, Merck, Sweden) at a concentration of 138.5 mM before being diluted in experimental buffers. Chicken cystatin was purchased from SigmaeAldrich (Milan, Italy) and dissolved at a concentration of 76.8 mM before being diluted in experimental buffer. CM5 sensor chips and other reagents and buffers for SPR analyses by BIAcore were from GE Healthcare (Milan, Italy). 2.2. GKN1 expression and purification Human recombinant GKN1 (rGKN1) lacking the first 20 amino acids leader peptide was expressed and purified as described

2.5. Ab(1e40) aggregation and fibrils formation Ab(1e40) (MW 4329.9) was dissolved in DMSO at a final concentration of 138.5 mM. Experiments were performed by coincubating Ab(1e40) (17.3 mM) with rGKN1 (1.7 mM) at 37  C in 10 mM sodium phosphate (NaP) (pH 7.0) and 150 mM sodium chloride (NaCl) with 10% (v/v) DMSO, under agitation. At various time points, samples were removed to determine the level of aggregation. The samples were centrifuged for 6 min at 16,000 g (14,000 rpm), and the supernatants were removed and centrifuged for an additional 2 min at the same speed. The supernatant from the last centrifugation was then analyzed by SDS-PAGE on 16% TrisTricine gels under nonreducing conditions and stained with Coomassie. As controls, Ab(1e40) was incubated with 1.7 mM chicken cystatin (MW 13.3) [18] and analyzed under the conditions described above.

F. Altieri et al. / Biochimie 106 (2014) 91e100

2.6. Thioflavin T assay for Ab aggregation

93

SEC was performed on an AKTA Purifier System using a Superdex 75 prepacked gel filtration column for high-resolution (GE Healthcare). The column was equilibrated and eluted at a flow rate of 0.5 mL/min with 100 mM NaCl and 50 mM NaP (pH 7.0) recording the absorbance at 280 nm. The elution volumes of the following standard proteins were used for column calibration: bovine serum albumin (BSA) (66 kDa), carbonic anhydrase (30 kDa) and cytochrome C (12 kDa). rGKN1 (30 mg in 10 mM NaP pH 7.0 and 150 mM NaCl 10% (v/v) DMSO) was incubated in the presence of Ab(1e40) in a 1:1 and 1:3 molar ratios. Before loading, all the samples were incubated at 37  C for 1 h.

CM5 sensor chip via the primary amine groups (amine coupling kit; GE Healthcare). The carboxymethylated dextran surface was activated by the injection of a mixture of 0.2 M N-ethyl-N0 - (diethylamino-propyl)carbodiimide and 0.05 M N-hydroxysuccinimide (EDC/NHS chemistry) according to the manufacturer's instructions [21]. The immobilization of ligand was efficiently performed at 15.0 mg/mL in 10 mM sodium acetate buffer (NaOAc) pH 4.0 containing 2.5% DMSO deriving from the peptide stock solution. The remaining N-hydroxysuccinimide esters were blocked by injection of 1 M ethanolamine hydrochloride (O-AEA) pH 8.5. A reference channel was prepared and used as control blank. All immobilization steps were performed at a flow rate of 5 mL/min using 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, 0.005% P20 (HBS-EP, GE Healthcare) (pH 7.4) as running buffer. rGKN1 was injected at different concentrations ranging from 7.5 mM to 130 mM. After each injection the surface was regenerated with pulses of a solution of 10 mM sodium hydroxide (NaOH). The activity of the immobilized peptide was not affected by the regeneration conditions employed and the chip was reusable to achieve two independent sets of binding experiments. Analyses were performed at 25  C at a flow rate of 20 mL/min in HBS-EP buffer (GE Healthcare). In all binding experiments, association phases ran for 180 s and dissociation phases for 300 s. Nonspecific binding from the reference channel was subtracted from the working channels using the BIAevaluation analysis package (version 4.1, GE Healthcare). Data were fitted using the software GraphPad Prism 5, versus 5.0 (GraphPad Software).

2.8. GKN1/Ab(1e40) interaction by MALDI-TOF mass spectrometry

2.10. Confocal laser scanning microscopy experiment

MALDI-MS experiments were performed in positive linear mode on a MALDI-TOF micro MX (Waters Co.), equipped with a pulsed nitrogen laser (l ¼ 337 nm). The instrument was calibrated using a three-point external calibration using a mixture (10 pmol/mL) of trypsinogen (24 kDa), cytochrome C (12 kDa) and insulin (5.8 kDa) as standard proteins (SigmaeAldrich) using a polynomial equation, as suggested by the manufacturer. All spectra were processed and analyzed using the MassLynx 4.0 software. The instrument source voltage was set to 12 kV. The pulse and detector voltages were optimized at 1999 V and 5200 V, respectively. Measurements were performed in the mass range m/z 3000e30000 with a suppression mass gate set to m/z 1000 and an extraction delay of 600 nsec [19]. Data were recorded by accumulating and averaging at least 10 spectra randomly acquired over the well surface. After averaging, spectra were processed for peak smoothing. To observe at best the rGKN1$Ab(1e40) complex, we performed scouting experiments by changing the amount and type of matrices and the protein/peptide ratio [20]. The complex was observed at best using sinapinic acid (solution 10 mg/mL in ethanol/NH4HCO3 at 1:1 molar ratio, v/v) as the ionizing matrix and at 1:1 molar ratio of rGKN1:Ab. Optimized spectra were therefore acquired on a solution containing 10 pmol/ mL rGKN1 in 5 mM NaP and 10 pmol/mL Ab in 1% DMSO. All spectra were acquired by spotting 1 mL of matrix solution on the target plate dried at room temperature. Then, protein/peptide samples (1 mL) were applied on top of the matrix crystal layer and dried again. A second matrix layer (1 mL), was formed on top of the sample layer preparation resulting in a matrix-sample-matrix sandwich. Spectra were collected after complete solvent evaporation. MALDI-MS experiments were also performed separately on 10 pmol/mL rGKN1 and Ab peptide solutions, under the same conditions.

SH-SY5Y cells were fixed with glutaraldehyde 2.5% in PBS for 20 min. After washing three times with PBS, cells were permeabilized with 0.1% Triton-X100 in PBS for 10 min and then washed three times with PBS. Subsequently, cells were blocked with 1% BSA in PBS for 20 min. After appropriate washes, cells were incubated with purified rabbit anti-amyloid precursor protein, Cterminal (Sigma Aldrich, USA) and mouse anti-GKN1 antibody (M01, Clone 2E5, Abnova, Heidelberg, Germany) diluted 1:200 in 1% BSA for 1 h. After washes with PBS, cells were incubated for 1 h with the appropriate secondary antibodies conjugated to fluorochromes and diluted 1:1000 in 1% BSA. Nuclear staining was obtained by using TO-PRO 3 Iodide (Invitrogen, Molecular Probes Eugene, Oregon, USA) diluted 1/1000 after 1 h of contact. Cells were washed three times with PBS and then observed with a Nikon Confocal Microscope C1 equipped with a EZ-C1 Software for data acquisition with 60 oil immersion objective.

2.9. Surface Plasmon Resonance analysis

SDS-PAGE was first utilized to study the effects of rGKN1 on Ab aggregation. Ab peptide (1e40) was incubated with rGKN1 and with the control protein chicken cystatin at 5:1 molar ratio. As reported in Fig. 1, at time zero incubation, peptide amounts were quite similar in all samples. Following incubation for 25 h, we observed a

Thioflavin T binding was assessed on Ab(1e40) incubated in the presence or absence of rGKN1 at 5:1 molar ratio in 10 mM NaP (pH 7.0) and 150 mM NaCl 10% (v/v) DMSO at 37  C under agitation. At various time points, 100 mL aliquots were removed and 160 mL of 50 mM thioflavin T were added in 10 mM NaP (pH 7.0) and 150 mM NaCl. Sample with rGKN1 alone was also included as control experiment. The samples were incubated for 5 min in the dark before fluorescence was measured. The wavelengths for excitation and emission were 450 and 482 nm, respectively. Each sample was measured in duplicate. 2.7. Size-exclusion chromatography (SEC)

Surface Plasmon Resonance (SPR) measurements were performed using a BIAcore 3000 instrument (GE Healthcare). Ab peptide (1e40) was covalently immobilized to the dextran matrix of a

2.11. Bioinformatic analysis Protein secondary structure prediction was done using PSIPRED (Protein Structure Prediction Server, UCL Department of Computer Science, USA). Protein sequences and alignment were obtained from UniProt 2002e2012 (Universal Protein Resource Consortium, Cambridge, United Kingdom; Geneva, Switzerland; Washington, United States of America). Prediction Trans membrane regions and orientation was performed at TMpre, Swiss Institute of Bioinformatics [22]. 3. Results 3.1. rGKN1 prevents the aggregation of Ab(1e40)

94

F. Altieri et al. / Biochimie 106 (2014) 91e100

Fig. 1. rGKN1 prevents aggregation of Ab(1e40). Panel A. (a) 17.3 mM Ab(1e40) incubated up to 24 h in the absence (B) or (b) in the presence (:) of 1.7 mM chicken cystatin (Cyst) (-) or (c) in the presence (▵) 1.7 mM rGKN1 (C). Panel B. (g) 17.3 mM Ab(1e40) incubated up to 7 days in the absence (B) or (h) in the presence (:) of 1.7 mM chicken cystatin(Cyst) (-) or (i) in the presence (▵) 1.7 mM rGKN1 (C). (d, e, f, k, l, m) densitometric evaluation of gel band intensities.

strong reduction of the Ab band in samples were the peptide was untreated or incubated with cystatin, whereas no reduction in the amount of Ab was observed when rGKN1 was added to the peptide samples (Fig. 1, panel A). Similar results were also observed after 7 days incubation. In fact, a drastic reduction of the amount of Ab was observed for the samples corresponding to Ab alone or the sample co-incubated with chicken cystatin. The sample containing rGKN1 showed instead amounts of soluble Ab almost similar to those observed at the beginning of incubation (Fig. 1, panel B). These results suggest that rGKN1 was able to prevent Ab aggregation. The ability of rGKN1 to prevent Ab peptide aggregation was also analyzed by Thioflavin T binding assay. According to data obtained with SDS-PAGE, binding of Thioflavin T to rGKN1 co-incubated with Ab was strongly suppressed compared to Ab peptide alone (Fig. 2), thus again supporting the view that GKN1 can prevent the polymerization of Ab(1e40).

the absence of Ab(1e40) showed a pattern compatible with the presence of a monomeric form of the protein whereas Ab(1e40) showed a pattern compatible with the presence of prefibrillar soluble monomeric form of the peptide. Both rGKN1 and Ab(1e40) eluted at a volume corresponding to that of an MW of about 20,000 Da and 5000 Da (Fig. 3A), respectively as determined by column calibration (not shown). A mixture of Ab(1e40) and GKN1 (3:1) was incubated at 37  C for 1 h followed by gel filtration (Fig. 3B), SDS-PAGE and Western blot of the collected fractions (Fig. 3C). The results showed that in the presence of rGKN1, no detectable change in the elution profile of the protein was observed and almost all Ab(1e40) was still found in the monomeric form with only traces or perhaps no Ab(1e40) co-eluted with monomeric rGKN1, as also confirmed by western blot analysis of column fractions (Fig. 3C). Similar results were obtained when rGKN1/Ab (1e40) were incubated at a 1:1 molar ratio (not-shown).

3.2. Size exclusion chromatography

3.3. GKN1/Ab interaction by MALDI-TOF mass spectrometry

The interaction between rGKN1 and Ab(1e40) was analyzed by gel filtration. The 280 nm absorbance elution profile for rGKN1 in

The non-covalent interaction of rGKN1 with Ab peptide (1e40) was also investigated by MALDI-MS under non-denaturing

F. Altieri et al. / Biochimie 106 (2014) 91e100

95

conditions. To observe at best the rGKN1$Ab complex, MS analyses were performed under different experimental conditions by changing the amount and type of matrices and the protein/peptide ratio. The optimized conditions for detecting the rGKN1$Ab complex included the use of sinapinic acid (solution 10 mg/mL in ethanol/NH4HCO3 1:1, v:v) as ionizing matrix and a 1:1 molar ratio of rGKN1:Ab peptide, as described in the Methods section. The MS spectrum of Ab(1e40) is shown in Fig. 4A. Only a mass peak corresponding to the peptide (m/z 4331.0, theoretical average [MþH]þ ¼ 4330.8 Da) was detected. Similarly, when rGKN1 was analyzed in isolation (Fig. 4B), a peak corresponding to the monomeric protein was expectedly observed at m/z 19406.0 (theoretical average [MþH]þ ¼ 19408.6 Da). By analyzing the mixture of rGKN1 and Ab(1e40) (Fig. 4C), an ion peak at m/z 23734.7 was clearly revealed, attesting the occurrence of a 1:1 rGKN1$Ab complex (theoretical average [MþH]þ ¼ 23738.45 Da). 3.4. SPR analysis

Fig. 2. GKN1 inhibits Ab(1e40) polymerization. Thioflavin T fluorescence binding was assessed at different times of incubation (0e4 h) of 8.5 mM Ab(1e40) alone (C) and in the presence of 1.7 mM rGKN1 (B). A sample of 1.7 mM rGKN1 alone was also included as control (:). Experiments were performed in duplicate with similar results.

To probe the specificity of the interaction between rGKN1 and Ab(1e40) and to determine their affinity constant, SPR analysis was performed. The Ab peptide was successfully immobilized on the dextran matrix of a CM5 sensor chip at about 1400 RU density. Fig. 5A depicts the binding sensograms of freshly prepared rGKN1 in HBS buffer injected at concentrations ranging between 7.5 and

Fig. 3. Gel filtration of rGKN1/Ab(1e40) mixture. (A) Elution profile of rGKN1 (continuos line) and Ab(1e40) alone (dashed line). (B) Elution profile of the rGKN1/Ab(1e40) mixture after incubation at 1:3 molar ratio for 1 h at 37  C. (C) Western blot with anti-GKN1 (upper panel) and anti-Ab(1e40) (lower panel) antibodies of the rGKN1/Ab(1e40) fractions.

96

F. Altieri et al. / Biochimie 106 (2014) 91e100

Fig. 4. GKN1/Ab(1-40) interaction by MALDI-TOF mass spectrometry. (A) MALDI-TOF spectra of Ab(1e40), (B) rGKN1 and (C) rGKN1 in the presence of Ab(1e40). A magnification of the ion peak at m/z 23734.7 corresponding to the 1:1 rGKN1$Ab(1e40) complex (theoretical average [MþH]þ ¼ 23738.45 Da is reported in C.

130 mM on the immobilized Ab surface. The specificity of binding was verified achieving saturation at the highest tested concentration of 130 mM. A dissociation constant (KD) of 3.4 ± 0.7  105 M was extrapolated by data fitting of a plot of RUmax values from each binding determination against protein concentration, using a nonlinear regression analysis (Fig. 5B). Cystatin, used as negative control, did not show any reliable and measurable binding response to Ab(1e40) when tested at 60 mM (Fig. 5C), further confirming the specificity of binding.

flGKN1 (green fluorescence) and endogenous APP (red fluorescence) showed a prominent localization within the cellular membrane. The superimposition of the two panels showed a more intense merge signal in regions corresponding to cellular membrane. Taking advantage of the presence in the pCDNA3.1-GKN1 vector of the (His)6-tag sequence [16], a pull-down was performed after transfection of the AGS cells with flGKN1. As reported in Fig. 7, in cells transfected with flGKN1, APP was specifically pulled-down by flGKN1 as detected by anti-APP antibody, thus confirming its interaction with the amyloid precursor.

3.5. Interaction between GKN1 and APP 4. Discussion Using Confocal Laser Scanning Microscope we investigated the possible colocalization between amyloid precursor protein and flGKN1 within SH-SY5Y cells. As reported in Fig. 6, transfected

Proteins containing the BRICHOS domain like proSP-C and BRI2 are membrane proteins associated to degenerative and proliferative

F. Altieri et al. / Biochimie 106 (2014) 91e100

97

Fig. 5. Surface Plasmon Resonance analyses of the binding between GKN1 and Ab(1e40). (A) Sensorgrams of the binding of rGKN1 to Ab(1e40) immobilized on a CM5 Biacore sensor chip, tested at concentrations between 7.5 and 30 mM. Experiments were carried out at 25  C, at a constant flow rate of 20 mL/min using HBS as running buffer. (B) Plot of RUmax from duplicate binding experiments versus concentration (mM). We determined a KD ¼ 3.4 ± 0.7  105 M by fitting data with a nonlinear regression algorithm (GraphPad Prism 4). (C) Sensorgram obtained following the injection of Cystatin at the concentration of 60 mM onto immobilized Ab(1e40) peptide showing that no interaction occurred between the protein and the immobilized ligand.

diseases such as lung fibrosis and British dementia, respectively [7]. These proteins contain an amino acidic region with a tendency to form b-sheet structures that can be proteolytically removed with mechanisms that are not yet fully understood. The recombinant BRICHOS domains of these proteins exert a chaperon-like function by preventing fibrils formation of amyloid-beta peptides (1e40) and (1e42) [11,15]. In this work, we demonstrate for the first time that also human recombinant gastrokine 1 (rGKN1), a stomach specific protein belonging to the family of BRICHOS domain containing proteins [7], is able in vitro to prevent the aggregation of the Alzheimer disease-associated peptide Ab(1e40) released from the trans membrane region of the amyloid precursor protein (APP) [23]. Both SDS-PAGE and thioflavin T assays show that the protein is

able to exert its anti-Ab aggregation function at sub-stoichiometric concentrations with respect to Ab(1e40) (1:10 rGKN1/Ab molar ratio), in agreement with similar data reported for proSP-C and BRI2 [18,24]. In our experimental conditions, the aggregation of pure Ab(1e40) was reduced in the presence of rGKN1 by about 40% within 24 h and by about 80% after one week, as evaluated from the densitometric analysis of the gels (Fig. 1). This behavior was similar to that reported for recombinant proSP-C even though in this case the incubation was prolonged up to 20 days [18]. Also thioflavin T assay showed that rGKN1 blocked aggregation of Ab(1e40). In this case, using 5 times less rGKN1 with respect to Ab(1e40), no thioflavin T-positive species were observed up to 4 h incubation [18], suggesting the occurrence of nonfibrillogenic and soluble

98

F. Altieri et al. / Biochimie 106 (2014) 91e100

Fig. 6. Colocalization of flGKN1 and APP in SH-SY5Y cells by confocal microscopy. (A) GKN1 immunostaining. (B) APP immunostaining. (C) Nuclear staining. (D) Merged images.

aggregates only. Gel filtration was used in order to detect the formation of a possible complex between rGKN1 and Ab(1e40). Samples containing rGKN1 and Ab(1e40) at 1:1 and 1:3 M ratios showed that Ab(1e40) remained mainly in the stable monomeric form, as suggested by the unchanged elution profile. However, none or only trace amounts of the peptide were detected by Western blotting within the column fractions eluting together with rGKN1, a result almost identical to that obtained with the isolated BRICHOS domains of BRI2 and proSP-C on Ab(1e40) and Ab(1e42) aggregation [24]. To further investigate the biochemical interaction between GKN1 and Ab(1e40), we analyzed by MALDI-TOF mass spectrometry samples containing the preformed complex. Mass peaks clearly indicated the formation of 1:1 complexes between Ab(1e40) and rGKN1. Though the MALDI MS technique is not generally believed to provide quantitative data, we noticed that the

Fig. 7. Pull-down experiment. AGS cells extracts were pulled-down with Ni-NTA agarose and washed with 10 mM imidazole. The Ni-NTA resin was directly loaded on SDS-PAGE and analyzed by western blot by using an anti-APP antibody (upper panel) and an anti-GKN1 antibody (lower panel). (Lane 1) AGS cells transfected with pCDNA3.1-GKN1 encoding flGKN1; (lane 2) AGS cells transfected with pCDNA 3.1; (lane 3) Ni-NTA alone; (lane 4) control APP expression in AGS cells transfected with pCDNA3.1-GKN1 encoding flGKN1.

mass peak corresponding to the 1:1 complex between rGKN1 and Ab(1e40) (m/z 23734.7) was strikingly much less intense (about 10%) compared to those obtained with isolated rGKN1 and Ab(1e40) or those recorded in the complex mixture. These data suggested that rGKN1 and Ab(1e40) have a rather scarce tendency to form a stable molecular complex or that only a fraction of Ab is available for binding. These hypotheses reconcile the results obtained by mass spectrometry and gel filtration, because we can reasonably expect that very labile GKN1Ab complexes could be readily dissociated during the chromatographic separation. We additionally speculate that the apparent very poor GKN1 binding to Ab, could instead reflect a condition in which only properly folded Ab monomers, likely those detached from and in equilibrium with large insoluble or soluble oligomers, are recognized by the BRICHOS domain of the protein. Such a scenario would explain why only substoichiometric amounts of GKN1 are required to suppress amyloid fibrils and why only a fraction of the Ab monomers are found in complex with the GKN1. This hypothesis is also consistent with the supposed chaperon-like properties of GKN1, which, by catalyzing an Ab conformational transition, would contribute to disrupt aggregates by transforming it in more soluble monomers or oligomers. To further clarify this aspect and to measure the strength of the interaction between GKN1 and Ab(1e40), the affinity constant (KD) was determined using the SPR technology. As expected, the KD value obtained (3.4  105 M) indicated a scarce affinity and was about one order of magnitude higher than that determined for HspB8 (1.19  106 M), a member of sHsp family, also named Hsp21 of H11, that possesses a chaperon activity and plays a role in regulating Ab aggregation [25]. Our result was strongly supported by comparable kD values determined in titrating recombinant proSP-C BRICHOS domain with peptides designed to model amyloid fibril formation [26]. Remarkably, no other KD values for

F. Altieri et al. / Biochimie 106 (2014) 91e100

99

Fig. 8. Amino acid sequence alignment of the BRICHOS domains of GKN1, pro-SP-C and BRI2. Common secondary structural elements, predicted at PSIPRED server, are highlighted in colors. Identities and similarities are marked by asterisks and dots, respectively.

BRICHOS domain-containing proteins towards Ab(1e40) have been determined so far. These results suggest that like BRI2 and proSP-C, GKN1 has the ability to suppress Ab(1e40) aggregation by a mechanism of direct interaction of its BRICHOS domain, which most likely interferes with the equilibrium of Ab(1e40) nucleation process [23]. The similar behavior shared by these proteins is very intriguing if one considers the low sequence identity among the different BRICHOS domains (less than 10%) (Fig. 8). In fact, the sequence alignment showed the presence of very few identities, including, for example, D79, C95 and C156 [7]. Remarkably, in sharp contrast with the very poor sequence identity, these BRICHOS domains show broad common secondary structural features (a-helix and b-sheet), as predicted by bioinformatic PSIPRED server. The interaction between GKN1 BRICHOS domain and Ab(1e40) in its extended bstructure could be mediated by the formation of hydrogen bonds with the hydroxylated amino acids Tyr 66, Thr73, Tyr148 present on face A of the b-sheet segments of GKN1, thus preventing the formation of aggregate species. This hypothesis was supported by the observation that in several BRICHOS protein families, there is a very strong tendency for aromatic residues on face A of the b-sheets that are fully or partially conserved within BRICHOS domains [27]. In proSP-C, these residues are Tyr104, Tyr106, Tyr113, Tyr122, and Tyr195, whereas in BRI2, Tyr156 and Tyr229 as numbered in Fig. 8. These conserved aromatic residues display structural feature evocative of polyphenolic compounds able to retard Ab aggregation [27,28] and that might participate in binding and stabilization of bstrand regions of the target peptide [23]. It was shown that the dementia gene BRI2 was an inhibitor of APP processing [29,30] and that BRI3, a member of the same gene family, was able to bind APP and to inhibit the production of Ab [31]. BRI gene family products that comprise also BRI1, are type II transmembrane (TM) proteins containing a BRICHOS domain. Also flGKN1 is predicted as type II TM protein (N-terminus inside and 2 strong transmembrane helices F3-A20 and L158-V174). We therefore investigated whether also flGKN1 was able to interact with APP. Confocal microscopy performed in SH cells transfected with flGKN1 and pull-down for GKN1 in transfected AGS cells clearly indicated an interaction between the two proteins. The biological reason why GKN1 showed anti-amyloidogenic properties is still unclear. The presence of the BRICHOS domain might be associated to complex post-translational processing of the protein such as intramolecular chaperon [7]. The GKN1 TM N-terminal hydrophobic region is known to constitute a signal peptide [1] whereas no data are yet available regarding a possible activity of

GKN1 BRICHOS domain as intramolecular chaperon toward the bstrand propensity of its C-terminal sequence. Beside the proposed important role in gastric mucosal protection [1,5], in gastric cells GKN1 might also exert other additional functions, such as that of being a molecular chaperon that delays amyloid aggregation by physically interacting with Ab within and outside the context of APP, thus possibly exploiting the twofold mechanism of disrupting Ab aggregates and inhibiting secretase processing [32]. A deeper characterization of its putative regulation activity of APP cleavage is needed, as it would have outstanding physiological implications and relevant therapeutic applications. 5. Conclusions In conclusion, this is the first report showing that GKN1, a stomach specific protein, owns the capability to prevent Ab(1e40) amyloid aggregation. Most likely, this activity can be ascribed to the GKN1 BRICHOS domain present in the central region of the protein. Given its protective role toward gastric tissue and gastric mucosal integrity [1,5], the presence of the BRICHOS domain might be important for GKN1 to exert a chaperon-like activity for protecting and avoiding peptide aggregation of the cysteine-rich C-terminal region or for transport or secretion of amyloid fibrils [21,23,27]. The interaction of GKN1 with APP might also be important for the regulation of APP processing perhaps through the competition with proteolytic enzymes by blocking the access of a-, b- and g-secretases to APP. The production of the isolated rGKN1 BRICHOS domain is an important future step to better highlight these results and to open new pharmacological perspective against amyloid diseases. Conflict of interest All the authors that participated to the present work declare that there are no conflict of interest. Contributors FA, CSDS and both GM performed all the biochemical characterization of the molecular interaction between the protein and the peptide. VS, ADM and AC were involved in mass spectrometry analysis. AS and RM gave a contribution in the biacore measurements. VQ was involved in confocal microscopy analysis. MM cured the SEC separations. ER contributed with the aspects regarding cell

100

F. Altieri et al. / Biochimie 106 (2014) 91e100

culture and transfection. PA overviewed the work and prepared the manuscript. Acknowledgments This work was supported by funds from Programmi di Ricerca Scientifica di Rilevante Interesse Nazionale (2008BKRFBH_003), from FIRB N RBNE08NKH7_003 and from PON Ricerca e Competitivit a 2007-2013 (PON01_02782, PON01_01602, PON01_02342). AS is fully supported by FIRB N RBNE08NKH7_003. References [1] T.E. Martin, C.T. Powell, Z. Wang, S. Bhattacharyya, M.M. Walsh-Reitz, K. Agarwal, F.G. Toback, A novel mitogenic protein that is highly expressed in cells of the gastric antrum mucosa, Am. J. Physiol. Gastrointest. Liver Physiol. 285 (2) (2003) G332eG343. [2] G. Nardone, E. Rippa, G. Martin, A. Rocco, R.A. Siciliano, A. Fiengo, G. Cacace, A. Malorni, G. Budillon, P. Arcari, Gastrokine 1 expression in patients with and without Helicobacter pylori infection, Dig. Liver Dis. 39 (2) (2007) 122e129. [3] G. Nardone, G. Martin, A. Rocco, E. Rippa, G. La Monica, F. Caruso, P. Arcari, Molecular expression of gastrokine 1 in normal mucosa and in Helicobacter pylori-related preneoplastic and neoplastic lesions, Cancer Biol. Ther. 7 (12) (2008) 1890e1895. [4] S.F. Moss, J.W. Lee, E. Sabo, A.K. Rubin, J. Rommel, B.R. Westley, F.E. May, J. Gao, P.A. Meitner, R. Tavares, M.B. Resnick, Decreased expression of gastrokine 1 and the trefoil factor interacting protein TFIZ1/GKN2 in gastric cancer: influence of tumor histology and relationship to prognosis, Clin. Cancer Res. 14 (13) (2008) 4161e4167. [5] K. Shiozaki, S. Nakamori, M. Tsujie, J. Okami, H. Yamamoto, H. Nagano, K. Dono, K. Umeshita, M. Sakon, H. Furukawa, M. Hiratsuka, T. Kasugai, S. Ishiguro, M. Monden, Human stomach-specific gene, CA11, is downregulated in gastric cancer, Int. J. Oncol. 19 (2001) 701e707. [6] K.A. Oien, F. McGregor, S. Butler, R.K. Ferrier, I. Downie, S. Bryce, S. Burns, W.N. Keith, Gastrokine 1 is abundantly and specifically expressed in superficial gastric epithelium, down-regulated in gastric carcinoma, and shows high evolutionary conservation, J. Pathol. 203 (3) (2004) 789e797. nchez-Pulido, D. Devos, A. Valencia, BRICHOS: a conserved domain in [7] L. Sa proteins associated with dementia, respiratory distress and cancer, Trends Biochem. Sci. 27 (7) (2002) 329e332. [8] L.M. Pavone, P. Del Vecchio, P. Mallardo, F. Altieri, V. De Pasquale, S. Rea, N.M. Martucci, C.S. Di Stadio, P. Pucci, A. Flagiello, M. Masullo, P. Arcari, E. Rippa, Structural characterization and biological properties of human gastrokine 1, Mol. Biosyst. 9 (2013) 412e421. [9] Y. Kallberg, M. Gustaffson, B. Persson, J. Thyberg, J. Johansson, Prediction of amyloid fibril-forming proteins, J. Biol. Chem. 276 (16) (2001) 12945e12950. [10] R.G. Warr, S. Hawgood, D.I. Buckley, T.M. Crisp, J. Schilling, B.J. Benson, P.L. Ballard, J.A. Clements, R.T. White, Low molecular weight human pulmonary surfactant protein (SP5): isolation, characterization, and cDNA and amino acid sequences, Proc. Natl. Acad. Sci. U S A 84 (22) (1987) 7915e7919. [11] H. Willander, G. Askarieh, M. Landreh, P. Westermark, K. Nordling, H. Ker€ anen, E. Hermansson, A. Hamvas, L.M. Nogee, T. Bergman, A. Saenz, C. Casals, € rnvall, H. Berglund, J. Presto, S.D. Knight, J. Johansson, HighJ. Åqvistg, H.J. o resolution structure of a BRICHOS domain and ist implications for antiamyloid chaperone activity on lung surfactant protein C, Proc. Natl. Acad. Sci. U S A 109 (7) (2012) 2325e2329. [12] H. Johansson, K. Nordling, T.E. Weaver, J. Johansson, The Brichos domaincontaining C-terminal part of pro-surfactant protein C binds to an unfolded poly-val transmembrane segment, J. Biol. Chem. 281 (2006) 21032e21039.

[13] C. Nerelius, E. Martin, S. Peng, M. Gustafsson, K. Nordling, T. Weaver, J. Johansson, Mutations linked to interstitial lung disease can abrogate antiamyloid function of prosurfactant protein C, Biochem. J. 416 (2008) 201e209. €rnvall, J. Johansson, The extracellular domain of Bri2 [14] S. Peng, M. Fitzen, H. Jo (ITM2B) binds the ABri peptide (1e23) and amyloid b-peptide (Ab1e40): Implications for Bri2 effects on processing of amyloid precursor protein and Ab aggregation, Biochem. Biophys. Res. Commun. 393 (2010) 356e361. [15] H. Willander, E. Hermansson, J. Johansson, J. Presto, BRICHOS domain associated with lung fibrosis, dementia and cancer e a chaperone that prevents amyloid fibril formation? FEBS J. 278 (20) (2011) 3893e3904. [16] E. Rippa, G. La Monica, R. Allocca, M.F. Romano, M. De Palma, P. Arcari, Overexpression of gastrokine 1 in gastric cancer cells induces fas-mediated apoptosis, J. Cell. Physiol. 226 (2011) 2571e2578. [17] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. 72 (1976) 248e254. [18] C. Nerelius, M. Gustafsson, K. Nordling, A. Larsson, J. Johansson, Anti-amyloid activity of the c-terminal domain of proSP-c against amyloid b-peptide and Medin, Biochemistry 48 (2009) 3778e3786. [19] R. Tamburino, V. Severino, A. Sandomenico, M. Ruvo, A. Parente, A. Chambery, et al., De novo sequencing and characterization of a novel Bowman-Birk inhibitor from Lathyrus sativus L. seeds by electrospray mass spectrometry, Mol. Biosyst. 8 (12) (2012) 3232e3241. [20] T.B. Farmer, R.M. Caprioli, Determination of protein-protein interactions by matrix-assisted laser desorption/ionization mass spectrometry, J. Mass Spectrom. 33 (8) (1998) 697e704. [21] B. Johnsson, S. Lofas, G. Lindquist, Immobilization of proteins to a carboxymethyldextran-modified gold surface for biospecific interaction analysis in surface plasmon resonance sensors, Anal. Biochem. 198 (1991) 268e277. [22] K. Hofmann, W. Stoffel, TMbase e a database of membrane spanning proteins segments, Biol. Chem. Hoppe-Seyler 374 (1993) 166. [23] D. Selkoe, R. Kopan, Notch and Presenilin: regulated intramembrane proteolysis links development and degeneration, Annu. Rev. Neurosci. 26 (2003) 565e597. [24] H. Willander, J. Presto, G. Askarieh, H. Biverstål, B. Frohm, S.D. Knight, J. Johansson, S. Linse, BRICHOS domains efficiently delay fibrillation of amyloid b-peptide, J. Biol. Chem. 287 (37) (2012) 31608e31617. €ller, B. Kamps, B. Kusters, M.L. Maat[25] M.M. Wilhelmus, W.C. Boelens, I. Otte-Ho Schieman, R.M. de Waal, M.M. Verbeek, Small heat shock protein HspB8: its distribution in Alzheimer's disease brains and its inhibition of amyloid-beta protein aggregation and cerebrovascular amyloid-beta toxicity, Acta Neuropathol. 111 (2006) 139e149. [26] M. Fitzen, G. Alvelius, K. Nordling, H. Jornvall, T. Bergman, J. Johansson, Peptide-binding specificity of the prosurfactant protein C Brichos domain analyzed by electrospray ionization mass spectrometry, Rapid Commun. Mass Spectrom. 23 (2009) 3591e3598. [27] S.D. Knight, J. Presto, S. Linse, J. Johansson, The BRICHOS domain, amyloid fibril formation, and their relationship, Biochemistry 52 (2013) 7523e7531. [28] Y. Porat, A. Abramowitz, E. Gazit, Inhibition of amyloid fibril formation by polyphenols: structural similarity and aromatic interactions as a common inhibition mechanism, Chem. Biol. Drug. Des. 67 (2006) 27e37. [29] A. Fotinopoulou, M. Tsachaki, M. Vlavaki, A. Poulopoulos, A. Rostagno, B. Frangione, J. Ghiso, S. Efthimiopoulos, BRI2 interacts with amyloid precursor protein (APP) and regulates amyloid b (Ab) production, J. Biol. Chem. 280 (35) (2005) 30768e30772. [30] M. Tsachaki, A. Fotinopoulou, N. Slavi, V. Zarkou, J. Ghiso, S. Efthimiopoulos, BRI2 interacts with BACE1 and regulates its cellular levels by promoting its degradation and reducing its mRNA levels, Curr. Alzheimer Res. 10 (5) (2013) 532e541. [31] S. Matsuda, Y. Matsuda, L. D'Adamio, BRI3 inhibits amyloid precursor protein processing in a mechanistically distinct manner from its homologue dementia gene BRI2, J. Biol. Chem. 284 (23) (2009) 15815e15825. [32] J. Eder, A.R. Fersht, Pro-sequence-assisted protein folding, Mol. Microbiol. 16 (4) (1995) 609e614.

Anti-amyloidogenic property of human gastrokine 1.

Gastrokine 1 (GKN1) is a stomach-specific protein expressed in normal gastric tissue but absent in gastric cancer. GKN1 plays a major role in maintain...
2MB Sizes 0 Downloads 8 Views