Centrosomal aggregates and Golgi fragmentation disrupt vesicular trafficking of DAT

Neurobiology of Aging 33 (2012) 2462–2477 www.elsevier.com/locate/neuaging

Centrosomal aggregates and Golgi fragmentation disrupt vesicular trafficking of DAT Francisco J. Diaz-Corralesa,*, Ikuko Miyazakib, Masato Asanumab, Diego Ruanoc, Rosa M. Riosa,* b

a Departamento de Señalización Celular, Centro Andaluz de Biología Molecular y Medicina Regenerativa (CABIMER), Seville, Spain Department of Brain Science, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan c Departamento de Bioquímica y Biología Molecular, Facultad de Farmacia, Universidad de Sevilla, Seville, Spain

Received 19 August 2011; received in revised form 8 November 2011; accepted 10 November 2011

Abstract Lewy bodies containing the centrosomal protein ␥-tubulin and fragmentation of Golgi apparatus (GA) have been described in nigral neurons of Parkinson’s disease (PD) patients. However, the relevance of these features in PD pathophysiology remains unknown. We analyzed the impact of proteasome inhibition in the formation of ␥-tubulin-containing aggregates as well as on GA structure. SH-SY5Y cells were treated with the proteasome inhibitor Z-Leu-Leu-Leu-al (MG132) to induce centrosomal-protein aggregates. Then, microtubules (MTs) and Golgi dynamics, as well as the vesicular transport of dopamine transporter (DAT) were evaluated both in vitro and in living cells. MG132 treatment induced ␥-tubulin aggregates which altered microtubule nucleation. MG132-treated cells containing ␥-tubulin aggregates showed fragmentation of GA and perturbation of the trans-Golgi network. Under these conditions, the DAT accumulated at the centrosomal-Golgi region indicating that the vesicular transport of DAT was disrupted. Thus, centrosomal aggregates and fragmentation of GA are 2 closely related processes that could result in the disruption of the vesicular transport of DAT toward the plasma membrane in a model of dopaminergic neuronal degeneration. © 2012 Elsevier Inc. All rights reserved. Keywords: Aggresome; Centrosome; Dopamine transporter; ␥-Tubulin; Golgi apparatus; Inclusion bodies; Lewy bodies; Microtubules; Neurodegeneration; Parkinson’s disease; Proteasome; Protein aggregation

1. Introduction Parkinson’s disease (PD) is the second most common age-related neurodegenerative disorder after Alzheimer’s disease (de Lau and Breteler, 2006). Neuropathological lesions are characterized by severe loss of pigmented-dopaminergic neurons mainly from the sustantia nigra pars compacta (SNc). Nevertheless, other brain stem structures, * Corresponding author at: Avda. Américo Vespucio s/n, Edif. CABIMER, 41092-Sevilla, Spain. Tel.: ⫹34 954468004; fax: ⫹34 954461664. E-mail addresses: [email protected] (F.J. Díaz Corrales). *Alternate corresponding author at: Avda. Américo Vespucio s/n, Edif. CABIMER, 41092-Sevilla, Spain. Tel.: ⫹34 954468004; fax: ⫹34 954461664. E-mail addresses: [email protected] (R.M. Rios). 0197-4580/$ – see front matter © 2012 Elsevier Inc. All rights reserved. 10.1016/j.neurobiolaging.2011.11.014

spinal cord, or cortical regions are also affected (Lees et al., 2009; Levy et al., 2009). Degeneration of dopaminergic neurons leads to a progressive loss of dopaminergic terminals in the striatum and a decrease on dopamine (DA) levels with the consequent appearance of the main motor symptoms (Levy et al., 2009). Simultaneously to neuronal cell loss, protein aggregates appear forming intracytoplasmic inclusions known as Lewy bodies (LBs), pale bodies, and Lewy neuritis (Lees et al., 2009; Shults, 2006). LBs are usually 1 or more eosinophilic spherical structures with a dense core surrounded by a halo and mainly located in the cytoplasm of remaining neurons (Shults, 2006; Wakabayashi et al., 2007). LBs are considered as the hallmark of PD pathology, although they may be present in other neurodegenerative diseases. The main component of LBs is an aggregated form of ␣-synuclein (Spillantini et al.,

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

1997); but a large number of other proteins have also been identified into these inclusion bodies, including centrosomal proteins such as ␥-tubulin and pericentrin (McNaught et al., 2002b; Shults, 2006). Thus, although ␣-synuclein is the main component of LBs, the consequences of protein aggregation in LBs of other key proteins for cell functioning should be studied in the context of PD pathophysiology. Neuronal loss in neurodegenerative diseases including PD has been associated with potential common pathogenic mechanism such as oxidative stress, mitochondrial dysfunction, neuroinflammatory response, etc. (Jellinger, 2009; Levy et al., 2009). Growing evidence based on postmortem, genetic, and experimental research have suggested that ubiquitin-proteasome system (UPS) failure and protein aggregation could also play a key role in the etiopathogenesis of both familial and sporadic forms of PD (Jellinger, 2009; Levy et al., 2009; McNaught et al., 2002a). The UPS is the main nonlysosomal system responsible for the degradation of short-lived and unwanted proteins (Betarbet et al., 2005). It has been hypothesized, based on morphological and biochemical data, that LBs in PD are formed through an aggresome-related process (McNaught et al., 2002b; Olanow et al., 2004). Aggresomes are intracellular accumulations of misfolded, aggregated proteins which form when the capacity of the UPS is overwhelmed either by excessive formation of unwanted proteins or by impaired protein degradation (Johnston et al., 1998; Olanow et al., 2004). In response to this proteolytic unbalance, unwanted proteins are retrograde transported through microtubules (MTs) from the cell periphery toward the centrosome where they fuse into an aggregate surrounded by intermediate filaments known as aggresome (Johnston et al., 1998). Therefore, the presence of ␥-tubulin and pericentrin into LBs suggest that they could form through an aggresome-related process, and in this case the abnormally folded proteins could accumulate at the neuronal centrosome. The centrosome is a nonmembrane-bound organelle composed of 2 orthogonally arranged centrioles surrounded by the pericentriolar matrix (Nigg and Raff, 2009) that is responsible for MT nucleation. This function is accomplished through the ␥-tubulin ring complex, a protein complex that contains ␥-tubulin (Raynaud-Messina and Merdes, 2007). Accumulation of centrosomal proteins such as ␥-tubulin or pericentrin and impairment of MT nucleation have been described in osteosarcoma cells treated with proteasome inhibitors (Didier et al., 2008). Thus, centrosomal structure and functioning seem to be a target of proteasome inhibition, suggesting that this organelle could be affected during the neurodegenerative process induced by proteasome inhibitors. In nonpolarized mammalian cells, the centrosome exhibits a close structural and functional relationship with the Golgi apparatus (GA). It is well-known that both the integrity and the pericentrosomal position of the Golgi ribbon

2463

depends on MTs (Thyberg and Moskalewski, 1999). In the absence of MTs, the GA fragments into many elements that appear dispersed throughout the cytoplasm. Interestingly, fragmentation of GA has been widely described in several neurodegenerative diseases, which are also characterized by the presence of abnormal protein aggregates (Fan et al., 2008; Gonatas et al., 2006). Indeed, fragmentation of GA has been observed in SNc of PD brain samples into the same neurons in which LBs and pale bodies were also identified (Fujita et al., 2006). In addition, it has been reported that ␣-synuclein aggregation induces fragmentation of GA in cell culture systems (Gosavi et al., 2002). This evidence might suggest that protein aggregation and fragmentation of GA could be common and related features of degenerative neuronal cells. However, the exact mechanisms that trigger or contribute to the fragmentation of GA in neurodegenerative diseases including PD as well as the relationship of centrosomal-protein aggregates with fragmentation of GA and the consequences that might arise from such morphological disturbances are still unknown. In mammalian cells, the GA is a single organelle composed of hundreds of stacks of cisternae connected by tubular bridges forming a reticular network that is known as the Golgi ribbon. The GA is a polarized structure in which proteins arriving from the endoplasmic reticulum (ER) enter the organelle by the cis-face, go through the medial cisternae, where they are modified, and leave the organelle by the trans-face. Finally, the proteins within transport vesicles are either inserted into membrane organelles, including the plasma membrane, as transmembrane proteins are secreted to the extracellular space. The dopamine transporter (DAT) is a transmembrane protein exclusively expressed in dopaminergic neurons. DAT represents the most important mechanism to regulate DA levels in the synaptic cleft (Storch et al., 2004; Zahniser and Sorkin, 2009). DAT is post-translationally N-glycosylated at the second extracellular loop in the GA, and this modification seems to be important for both the integration of DAT into the cell membrane and DA uptake (Li et al., 2004; Zahniser and Sorkin, 2009). Therefore, changes in GA structure might disrupt vesicular transport of DAT, thus affecting the synaptic transmission of dopaminergic neurons. In this work, the dynamics of centrosomal aggresome formation together with the GA fragmentation process were analyzed by live imaging in a dopaminergic neuroblastoma cell line treated with Z-Leu-Leu-Leu-al (MG132), which is a reversible proteasome inhibitor widely used to mimic parkinsonian features in vitro and in vivo experimental models of PD (Sun et al., 2006; Xie et al., 2010). We also studied the effect of such fragmentation on the vesicular transport of DAT. In MG132-treated cells, large centrosomal-protein aggregates and fragmentation of GA were observed. In addition, cells with centrosomal aggregates showed abnormal MT nucleation. The vesicular transport of DAT was disrupted on those cells, leading to an accumula-

2464

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

tion of DAT signal into the centrosomal-GA region. Our results might be relevant for understanding the effects exerted by centrosomal-protein aggregates and fragmentation of GA on vesicular trafficking of DAT in degenerative dopaminergic neurons as well as they could also reveal new therapeutic targets to ensure an adequate DAT trafficking to the dopaminergic terminals of SNc neurons in PD patients.

2. Materials and methods 2.1. Cell culture The human-derived dopaminergic neuroblastoma SHSY5Y cell line (American Type Culture Collection [ATCC] number CRL-2266) were cultured in Dulbecco’s modified eagle medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA) and supplemented with 10% (vol/vol) fetal bovine serum (FBS), 4 mM L-glutamine, 200 U/mL penicillin and 200 ␮g/mL streptomycin. SH-SY5Y cells were plated at a density of 1 ⫻ 105 cells/cm2 onto 6- or 96-well culture plates and 60- or 100-mm culture dishes (Sarstedt, Inc., Newton, NC, USA). Cells were continuously cultured at 37 °C under 5% CO2 95% air gas mixture. After 24-hour incubation, the cells were exposed to 5, 50, and 500 nM MG132 (Sigma-Aldrich) diluted in dimethyl sulfoxide (DMSO) during 24 or 48 hours. Up to 10 ␮M MG132 was also tested during 12 or 24 hours. Control group was treated with vehicle alone (final concentration of DMSO: 1% vol/ vol). 2.2. Proteasome inhibition assay To measure proteasome activity, SH-SY5Y cells were plated in 6-well plates and incubated with MG132 as described above. Following MG132 treatment, the cells were washed and lysed by rapid freezing and thawing 3 times in extraction buffer (10 mM Tris-HCl, pH 7.8, 0.5 mM 1,4dithiothreitol, 5 mM adenosine-5=-triphosphate (ATP), 0.03% Triton X-100, and 5 mM MgCl2). The lysates were centrifuged at 13000g at 4 °C for 20 minutes. Protein concentrations were adjusted to 0.5 ␮g/␮L. Individual wells of a black 96-well plate were filled with 82 ␮L assay buffer (50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), pH 7.4, 100 mM NaCl, 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 10 mM 1,4-dithiothreitol, 0.1 mM ethylenediaminetetraacetic acid (EDTA), 10% glycerol), 8 ␮L cell lysates, and 10 ␮L of 1 of 3 different fluorogenic proteasome peptide substrates (Peptides International, Louisville, KY, USA): Suc-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (MCA) to monitor chymotrypsin-like activity, Boc-Leu-Arg-ArgMCA to measure trypsin-like activity, or Z-Leu-Leu-GluMCA to follow peptidylglutamyl peptide hydrolyzing (PGPH) activity. A fluorometer (Varioskan Flash, Thermo Electron Corp., Vantaa, Finland) was used to determine enzyme activity (excitation/emission: 380/460 nm). The re-

sults were normalized to give activities relative to the controls as percentages of proteasome activity. 2.3. Cell viability assay Cell viability was determined by quantitative colorimetric assay WST-1 (Clontech Laboratories, Inc., Mountain View, CA, USA). Cells were ground in 96-well culture plates, and after MG132 treatment the cells were incubated with 10 ␮L of premixed WST-1 (Clontech Laboratories, Inc.) during 1 hour at 37 °C. Then, the absorbance at 440 nm was measured using a multiwell plate reader (Varioskan Flash; Thermo Electron Corporation). Cell viability was estimated as percentage of living cells relative to the number of vehicle-treated cells. The results were also confirmed by trypan blue-exclusion assay. 2.4. DNA constructs and transfection The transmembrane domain of galactosyl–transferase (GT) fused to green fluorescent protein (GFP) and cloned into the vector pIRESneo was previously produced in our laboratory (Cardenas et al., 2009). Human DAT cDNA in the bicistronic vector pCIN4 was kindly supplied by Dr. Maarten E.A. Reith (New York University, School of Medicine) (Liang et al., 2009). Specific primers were designed using the pCIN4-DAT sequence, and KpnI specific restriction sites were incorporated at the 5= and 3= ends. Primer sequences were as follows: DAT forward 5=-GATCGGTACCATGAGCAAGTCCAAATGC-3= and DAT reverse 5=-GATCGGTACCTCACACCTTCAGCCAGTG-3=. The amplified fragment was cloned into the vector pDsRedMonomer-Hyg-C1 (pDsRed) (Clontech Laboratories, Inc.) which was previously dephosphorylated. The construct was verified by sequencing. For stable expression or transient expression of pDsRed-DAT cells were transfected using Lipofectamine2000 (Invitrogen, Carisbad, CA, USA) reagent according to the manufacturer’s instructions. Stably transfected cells expressing the pIRESneo-GT-GFP construct were selected with 1000 ␮g/mL G418 sulfate (InvivoGen, San Diego, CA, USA) during 7 days. 2.5. Immunofluorescence and live-cell imaging SH-SY5Y cells were grown in 12 mm round coverslips and fixed in 100% methanol at ⫺20 °C for 6 minutes or 4% paraformaldehyde in 10 mM phosphate-buffered saline (PBS) pH 7.4 for 10 minutes and then labeled with appropriate antibodies. Centrosomes were vizualized by mouse monoclonal anti-␥-tubulin (clone GTU-88) (Abcam, Cambridge, UK). MTs were labeled by rat monoclonal anti-␣-tubulin (Abcam) and acetylated MTs were visualized by mouse monoclonal antiacetylated ␣-tubulin (Sigma-Aldrich). Mouse monoclonal anti-GM130 (BD Biosciences, San Jose, CA, USA), rabbit polyclonal antiGMAP210 (Infante et al., 1999), and human autoimmune sera recognizing GMAP210 (Rios et al., 1994) were used

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

2465

as cis-Golgi network markers. Mouse monoclonal antiGolgin 97 (Invitrogen) and human autoimmune sera recognizing Golgin 245 were used as trans-Golgi network markers. DAT was labeled by polyclonal rabbit anti-DAT (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Appropriate secondary antibodies DyLight 488, DyLight 549 (Jackson ImmunoResearch Laboratories, Inc., Baltimore, PA, USA) and Alexa Fluor 633 (Molecular Probes, Inc., Eugene, OR, USA) were used for double or triple Immunofluorescence (IF). 4=,6-Diamidino-2-phenylindole dihydrochloride (DAPI) dye (Roche, IN, USA) was used to visualize the cell nucleus. Cells were examined under a motorized upright wide-field microscope (Leica DM6000B, Wetzlar, Germany), and acquisition was performed at room temperature using an oil immersion objective (63⫻ APO PL HCX 1.4 NA; Leica) and a cooled charge-coupled device (CCD) camera (Leica DFC350FX). Confocal images were captured by a confocal Leica TCS SP5 with an HCX PL APO Lambda blue 63 ⫻ 1.4 OIL objective at 22 °C. Pericentriolar-matrix area in thresholded images, line scan, and gray level average of GMAP210 and Golgin 245 signals as well as overlapping coefficients or GMAP210 with DAT signal were carried out with MetaMorph 7.1.7.0 software (Universal Imaging Co, PA, USA). Adobe (San Jose, CA, USA) Photoshop CS3 software was used for digital amplification, input level adjustments, and final-image edition. For live imaging of the fluorescent proteins in living cells, transfected cells with pIRESneo-GT-GFP or pDsRed-DAT constructs were grown on 35-mm glass bottom culture dishes (MatTek Corporation, Ashland, MA, USA). To avoid background emission, dishes were filled with 2 mL 10% FBS DMEM-F12 without phenol red (Sigma-Aldrich) and placed in a 37 °C incubation chamber attached to a Leica DMI6000 motorized microscope. Acquisition was performed using an oil immersion objective

Fig. 1. Proteasome inhibition decreases cell viability and induces insoluble ␥-tubulin aggregates in a dopaminergic neuroblastoma cell line. Proteasome inhibition quantified by peptide degradation of 3 different fluorogenic-proteasome substrates (A) and dose-dependent cell death measured by WST-1 (Clontech Laboratories, Inc., Mountain View, CA, USA) assay in SH-SY5Y cells treated with Z-Leu-Leu-Leu-al (MG132) during 24 hours (B). Confocal microscopy images of SH-SY5Y cells immunostained with anti-␥-tubulin antibodies (C–D) and treated with vehicle (C) or 500 nM MG132 during 24 hours (D). Normal centrosomes observed in control group (C, insets) and large and round ␥-tubulin aggregates in MG132treated cells (D, insets). Increase in pericentriolar-matrix area (E) as well as in the detergent-insoluble fraction of ␥-tubulin protein (F) in MG132treated cells. Results in (A) and (B) are expressed as percentages and represent means ⫾ standard error of the mean (SEM) of 3 independent experiments. Results in (E) are expressed as pixels and represent means ⫾ SEM of 5 random images; more than 100 cells were evaluated in each group. * p ⬍ 0.05, ** p ⬍ 0.01, # p ⬍ 0.001, analysis of variance (ANOVA) followed of Dunnett test. The images shown are representatives of 5 photos taken in each group. Scale bars represent 20 ␮m. PGPH, peptidylglutamyl peptide hydrolyzing.

2466

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

Fig. 2. Centrosomal-protein aggregates alter microtubule (MT) nucleation in SH-SY5Y cells, and Z-Leu-Leu-Leu-al (MG132) treatment increases MT acetylation. Confocal images of MT regrowth assay (A–J’’) in SH-SY5Y cells treated with vehicle (A–E’’) or 500 nM MG132 (F–J’’). Cells were double immunostained with anti-␣-tubulin (A–E and F–J) and anti-␥-tubulin antibodies (A’–E’ and F’–J’). The nuclei were visualized by

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

(63 ⫻ APO PL HCX, 1.4 NA) and a cooled change: Hamamatsu ORCA-ER camera (Hamamatsu Photonics, Hamamatsu, Japan). After addition of 500 nM MG132 to the cell culture medium, the frames were acquired each 10 or 30 minutes during 24 hours. Leica Application Suite AF software (v. 1.8.2) was used to perform blind 3-D deconvolution (refractive index 1.518) and final edition of cell-live images were carried out with ImageJ software (National Institutes of Health, Bethesda, MD, USA). 2.6. Western blot (WB) Cell cultures on the 6-well culture plates were washed with cold PBS, and then lysed in 150 ␮L of Tris-ID buffer (Tris 50 mM, pH 7.4, NaCl 150 mM, 1% ipegal) supplemented with protease inhibitors. After incubation on ice for 60 minutes, the homogenates were centrifuged (16,000g for 20 minutes at 4 °C), and supernatants were collected (detergent-soluble fraction). The pellet (detergent-insoluble fraction) was resuspended in 60 mM TrisHCl, 2% sodium dodecyl sulfate (SDS), and 2.5% 2-mercaptoethanol and sonicated for 10 seconds. Total protein concentrations of cell lysates from both fractions were determined by Bio-Rad DC protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA). WB was performed using 10 ␮g of total cell lysate and proteins were separated through a 10% SDS-polyacrylamide gel. Transfer of proteins to nitrocellulose membranes and antibody incubation were performed following the standard protocols. Stripping buffer was used to reprove nitrocellulose membranes. 2.7. MT regrowth assay and GA reassembly experiments In control- and MG132-treated cells, MTs were depolymerized with 10 ␮M nocodazole (Sigma-Aldrich) for 2 hours. Then, cells were rinsed 5 times with fresh DMEM medium, and the coverslips were fixed in methanol at several time points after nocodazole washout (0, 1, 5, 10, 15, and 120 minutes). MT regrowth in fixed cells were observed by IF using anti-␥-tubulin and anti-␣-tubulin antibodies. To evaluate GA reassembly after nocodazole washout, cells were also stained with anti-GMAP210 antibodies. 2.8. Vesicular stomatitis virus glycoprotein transport assay SH-SY5Y cells on 60-mm culture dishes were transiently transfected with the GFP-vesicular stomatitis virus glycoprotein (VSVG) temperature-sensitive mutant cDNA

2467

construct kindly supplied by Dr. Gustavo Egea (Barcelona University, Medical School). Then, cells were rinsed with fresh DMEM medium containing vehicle or 500 nM MG132 and incubated at a nonpermissive temperature (40 °C) during 20 hours to accumulate GFP-VSVG in the ER. Cells were maintained for 1 additional hour at a nonpermissive temperature in the presence of cycloheximide (100 ␮g/mL). To allow surface transport, cells were shifted to a permissible temperature (32 °C) and coverslips were fixed in 4% paraformaldehyde at several time points (0, 15, 30, 60, and 120 minutes) and processed for IF. 2.9. Determination of intracellular DA and their metabolites SH-SY5Y cells were treated with vehicle or 500 nM MG132 during 24 hours and then homogenized with 5 volumes of 200 mM ice-cold perchloric acid containing 10 mM EDTA. After centrifugation (11,750g, for 20 minutes at 4 °C), the supernatant was filtered (0.45 ␮m) and then injected directly into a high-performance liquid chromatography (HPLC) with electrochemical detectors (HPLC-ECD; Tosoh, Co., Tokyo, Japan). Regional concentrations of DA and their metabolites 3,4-dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA) were determined. The HPLC system consisted of a delivery pump (PX-8020, Tosoh, Co.) and an analytical column (EICOMPAK SC5ODS, 3.0 mm ⫻ 150 mm; Eicom, Co., Kyoto, Japan). An electrochemical detector (EC-8020, Tosoh, Co.) with glassy carbon was used with a voltage setting of 700 mV and an Ag/AgCl reference electrode. A mobile phase containing 0.1-M citrate-sodium acetate buffer (pH 3.5), methanol (17% vol/vol), EDTA-2Na, and sodium 1-octanesulfonate was infused at a flow rate of 0.6 mL/minute. Protein concentration was determined with a protein assay kit (Bio-Rad Laboratories). 2.10. Statistical analysis Values were expressed as mean ⫾ the standard error of the mean (SEM). Differences between groups were examined for statistical significance using 1-way analysis of variance (ANOVA) followed by post hoc Dunnett’s test using the control group as reference. The statistical significance of intracellular DA and their metabolites was estimated by unpaired 2-sample t tests. Yates’ ␹2 test was used to evaluate the statistical association between ␥-tubulin aggregates and fragmentation of GA. Odds ratios and 95% confidence intervals were estimated for these 2

DAPI dye (A’’–E’’ and F’’–J’’, blue). Characteristic MT-aster observed after 5 minutes of nocodazole washout in vehicle-treated cells (C’’, arrow). Abnormal MT nucleation and curved MTs in MG132-treated group (H and J, arrowheads). Confocal images of SH-SY5Y cells (K–L’’) treated with vehicle (K–K’’) or 500 nM MG132 during 24 hours (L–L’’). Cells were double immunostained with anti-␣-tubulin (K–L) and antiacetylated ␣-tubulin antibodies (K’–L’). Right panels represent merged images (K’’–L’’). Complex perinuclear network of acetylated MTs in MG132-treated cells (L’). Increase of acetylated ␣-tubulin protein in 500 nM MG132-treated group (M). The images shown are representatives of 5 photos taken in each group. Scale bars represent 10 ␮m (A’’–D’’, F’’–I’’, and K–L’’) and 5 ␮m (E’’and J’’).

2468

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

Fig. 3. Proteasome inhibition induces fragmentation of GA without perturbing its pericentrosomal distribution. Confocal microscopy images (A–D and F–I) of SH-SY5Y cells treated with vehicle (A–B and F–G) or 500 nM Z-Leu-Leu-Leu-al (MG132) during 24 hours (C–D and H–I). The Golgi apparatus (GA) was immunostained by anti-GMAP210 antibodies (A, C, F’, and H’). Centrosomes or ␥-tubulin aggregates were visualized by anti-␥-tubulin antibodies (F and H). The nucleus was labeled by DAPI dye (A’–B, C’–D, F’’–G, and H’’–I). Merged (A’, C’, F’’’, and H’’’) and amplified images (B, D, G, and I) are shown. Fragmentation of GA was observed in MG132-treated cells (C and D, arrow). The Golgi fragments were distributed around the ␥-tubulin aggregates (H–I). Deformation of nuclear envelope was observed in MG132-treated cells (H’’, arrowhead). Normal distribution of GA next to the centrosome in control cells (F–G). Quantification of GA morphology in SH-SY5Y cells treated with different doses of MG132 during 24 hours (E). Results in (E) are expressed as percentages and represent means ⫾ standard error of the mean (SEM) of 5 random images; more than 100 cells were counted in each group. * p ⬍ 0.05, ** p ⬍ 0.01, # p ⬍ 0.001, analysis of variance (ANOVA) followed of Dunnett test. The images shown are representatives of 5 photos taken in each group. Scale bars represent 20 ␮m (A’–B, C’–D, F’’’, and H’’’) and 5 ␮m (G and I). ND, nondeterminate.

qualitative variables. A p value less than 0.05 denoted the presence of a statistically significant difference. 3. Results 3.1. Proteasome inhibition decreases cell viability and induces insoluble ␥-tubulin aggregates in a dopaminergic neuroblastoma cell line In order to induce cell death and aggregation of ␥-tubulin protein we have treated SH-SY5Y cells with the proteasome inhibitor MG132. First, we determined the dose and exposure time to MG132 necessary to induce appropriate proteasome inhibition and cell death levels. Proteasome inhibition was measured by the degradation of 3 different fluorogenic-peptide substrates. Treatment with 50 nM MG132 during 24 hours

significantly inhibited chymotripsin-like and PGPH activities, whereas 500 nM-MG132 dose inhibited all 3 enzymatic activities in the same period of time (Fig. 1A). WST-1 assay (Clontech Laboratories, Inc.) showed a significant decrease in cell viability estimated at 82% and 51% after 24-hour exposure to 50 nM or 500 nM MG132, respectively (Fig. 1B). When exposure time was prolonged until 48 hours, cell viability decreased until 14% in the 500-nM MG132-treated group (Supplementary Fig. 1A). Highest doses were also tested but severe cell death was observed making difficult the morphological evaluation of treated cells. Thus, treatment with 500 nM MG132 during 24 hours inhibited proteasome activity to decrease cell viability around 50%. We then evaluated the effect of MG132 exposure on centrosomal morphology in SH-SY5Y cells by immuno-

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477 Table 1 Contingency table showing the total number and percentages of SH-SY5Y cells with or without centrosomal-protein aggregates and fragmentation of GA in 500 nM MG132-treated group during 24 hours Fragmentation of GA Yes (%) No (%) Total (%)

b

Centrosomal-protein aggregates

a

Yes (%)

No (%)

Total

137 (83.0) 8 (4.9) 145 (87.9)

11 (6.7) 9 (5.5) 20 (12.1)

148 (89.7) 17 (10.3) 165 (100)

Key: GA, Golgi apparatus; MG132, Z-Leu-Leu-Leu-al (MG132). a Centrosomal-protein aggregates were visualized by immunostaining with anti-␥ tubulin antibodies. b GA structure was evaluated by anti-GMAP210 anitibodies. Yates’ ␹2 test ⫽ 2553; p ⬍ 0.001; odds ratio ⫽ 14.01; 95% confidence interval, 4.51– 43.57.

staining with anti-␥-tubulin antibodies (Fig. 1C and D). In control cells, centrosomes were visualized as a single or pair of dots adjacent to the nucleus with a diameter of around 1 ␮m (Fig. 1C, inset, and Supplementary Fig. 1B–C’). On the contrary, in cells exposed to 500 nM MG132 for 24 hours the ␥-tubulin signal appeared in large aggregates (Fig. 1D, inset) with diameter ranging from 3 to 10 ␮m. Larger aggregates were observed by IF after 48 hours of exposure in the remaining cells (Supplementary Fig. 1D–E’’). Quantification of the ␥-tubulin aggregate diameter showed a 4-fold increase in 500 nM MG132-treated cells with respect to control cells (Fig. 1E). MG132 treatment also affected the solubility of ␥-tubulin protein. It is known than in mammalian cells ␥-tubulin is distributed in 2 different fractions, the soluble cytosolic form and the insoluble centriole-associated form (Moudjou et al., 1996). WB showed that MG132 treatment increased the concentration of ␥-tubulin protein in the detergent insoluble fraction in a dose dependent manner, while no changes were observed in the soluble fraction (Fig. 1F). Thus, the presence of a large ␥-tubulin aggregate is associated with an increase in the insoluble fraction of this protein in agreement with the previously reported insolubility of these aggregates observed in primary cell cultures of mesencephalic neurons treated with rotenone, a mitochondrial Complex I inhibitor (Diaz-Corrales et al., 2005), and in epithelial cell lines exposed to proteosome inhibitors (Didier et al., 2008). 3.2. Centrosomal-protein aggregates alter MT nucleation in SH-SY5Y cells, and MG132 treatment increases MT acetylation To evaluate whether MT nucleation is affected in SH-SY5Y cells containing ␥-tubulin aggregates, we performed MT repolymerization experiments in control(Fig. 2A–E’’) and MG132-treated cells (Fig. 2F–J’’). Cells were treated with nocodazole for 2 hours in order to completely depolymerize MTs in SH-SY5Y cells (Fig. 2A and F). After nocodazole washout, cells were

2469

fixed at 1, 5, 10, and 15 minutes and then stained with anti-␣-tubulin to label MTs (Fig. 2A–E and F–J) and anti-␥-tubulin antibodies to visualize centrosomes or the ␥-tubulin aggregates (Fig. 2A’–E’ and F’–J’). In control cells, MT repolymerization from the centrosome was detected 1 minute after nocodazole washout (Fig. 2B– B’’). After 5 minutes, the characteristic MT-aster was clearly observed (Fig. 2C–C’’, arrow). MT elongation continued over the time (Fig. 2D–D’’), and a radial array was observed 15 minutes after washout of the drug (Fig. 2E–E’’). On the contrary, MG132-treated cells showed an aberrant MT nucleation pattern (Fig. 2G–J’’) with the MTs growing from different points of the ␥-tubulin aggregates and without showing a clear radial array formation after nocodazole washout (Fig. 2H–J’’). In addition to this aberrant MT nucleation pattern, MTs nucleated from ␥-tubulin aggregates were more curved and thicker than those in control group (Fig. 2H and J, arrowheads). Given that curvature is a morphological characteristic of acetylated MTs (Piperno et al., 1987), we then wondered if proteasome inhibition could also affect MT acetylation. To investigate this possibility, we performed IF and WB of acetylated ␣-tubulin in MG132 treated cells (Fig. 2K–L’’ and M). A larger number of MTs containing acetylated ␣-tubulin were observed in MG132-treated cells when compared with nontreated cells, and these acetylated MTs formed a complex network in the perinuclear region (Fig. 2L’). Accordingly, WB showed that acetylated ␣-tubulin increased in the 500 nM MG132-treated group (Fig. 2M). Thus, although large ␥-tubulin aggregates exhibit some MT nucleation capacity, the nucleation process was aberrant in cells containing ␥-tubulin aggregates. Interestingly, MTs growing from these ␥-tubulin aggregates were strongly acetylated in MG132-treated cells. 3.3. Proteasome inhibition induces fragmentation of GA without perturbing its pericentrosomal distribution We then analyzed the morphology of GA by confocal microscopy in control and MG132-treated SH-SY5Y cells. Immunostaining of GA was performed with an antibody against GMAP210, a cis-Golgi and MT-binding protein required for Golgi ribbon formation (Ríos et al., 2004). Most of control cells exhibited a compact-Golgi ribbon located near the nucleus and around the centrosome (Fig. 3A–B and F–G). The GA appeared intact in 78.9% of control cells and only 9.3% showed some degree of fragmentation (Fig. 3E). On the contrary, the 94.2% of 500 nM MG132-treated cells displayed significant fragmentation of GA and just 3.8% of cells showed an intact Golgi ribbon (Fig.3C–D, arrow, and E). In the 50-nM MG132-treated group 27.5% of cells exhibited fragmentation of GA (Fig. 3E). These Golgi fragments did not disperse but remained distributed at the periphery of the large centrosomal aggregate (Fig. 3H–I and Supplementary Video 1). These ␥-tubulin aggregates deformed the nuclear envelope (Fig. 3H’’, arrowhead). After 24 hours of exposure to 500 nM

2470

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

Fig. 4. Proteasome inhibition differentially affected Golgi compartments. Confocal microscopy images (A–B’’) of SH-SY5Y cells treated with vehicle (A–A’’) or 500 nM Z-Leu-Leu-Leu-al (MG132) during 24 hours (B–B’’). The cis-Golgi face was labeled by anti-GMAP210 antibodies (A–B), and the trans-Golgi face was immunostained by anti-Golgin 245 antibodies (A’–B’). The right panels represent merged images (A’’–B’’). Gray level average of random line scans (C) drawn on the normal Golgi ribbon (A’’, blue line) or Golgi fragments obtained after MG132 treatment (B’’, blue line). Gray level of GMAP210 (C, red) and Golgin 245 signals (C, green) are represented. Confocal images of SH-SY5Y cells treated with 500 nM MG132 during 24 hours (D–G). Cells were triple immunostained with anti-GMAP210 (D, F), anti-GM130 (D’), anti-Golgin 97 (F’) and anti-Golgin 254 antibodies (D’’ and F’’). Merged (D’’’–F’’’) and amplified images (E and G) are shown. Disassembly of trans-Golgi markers from the Golgi fragments in MG132 treated cells (D’’’–F’’’, insets). WB of GMAP210, GM130, Golgin 245, and Golgin 97, ␣-tubulin was used as loading control (H). The images shown are representatives of 5 photos taken in each group. Scale bars represent 5 ␮m (A’’, B’’, E, and G) and 10 ␮m (D’’’ and F’’’).

MG132 around 83% of remaining cells showed a large ␥-tubulin aggregate and fragmentation of GA in the same cell. Quantification of these phenotypes is shown in Table 1. The statistical analysis rejected the null hypothesis that both variables are independent. Thus, in MG132-treated cells both ␥-tubulin aggregates and fragmentation of GA are associated.

We observed that in control cells, only a few acetylated MTs colocalized with the GA ribbon (Supplementary Fig. 2A–A’’), whereas GA fragments were embedded in the large network of acetylated MTs that surround the large centrosomal aggregate in MG132-treated cells (Supplementary Fig. 2B–B’’, arrow). To evaluate whether the distribution of

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

2471

Fig. 5. Proteasome inhibition interrupts the vesicular stomatitis virus glycoprotein (VSVG) transport from the Golgi apparatus (GA) to the plasma membrane. SH-SY5Y cells were transfected with a green fluorescent protein (GFP)-VSVG temperature-sensitive mutant cDNA construct and treated with vehicle or 500 nM Z-Leu-Leu-Leu-al (MG132). Cells were incubated at a nonpermissive temperature (40 °C) during 20 hours to accumulate GFP-VSVG in the endoplasmic reticulum (ER) and maintained for 1 additional hour in the presence of cycloheximide to block protein synthesis. Cells were shifted to a permissible temperature (32 °C) to allow surface transport of VSVG, and coverslips were fixed at several time points (0, 15, 30, 60, and 120 minutes) and processed for immunofluorescence (IF) as shown in (A). Confocal microscopy images (B–I’’) of SH-SY5Y expressing the GFP-VSVG construct (B–E and F–I) and treated with vehicle (B–E’’) or 500 nM MG132 (F–I’’). The GA was immunostained by anti-GMAP210 antibodies (B’–E’ and F’–I’). Merged images are shown (B’’–E’’and F’’–I’’). Dashed lines represent the plasma membrane. Normal kinetics of VSVG transport from the ER to GA and expression on the plasma membrane (B’’–E’’, insets). Abnormal VSVG transport in MG132-treated cells with accumulation of VSVG signal into the Golgi fragments and without reaching the cell surface (F’’–I’’, insets). Expression in the plasma membrane of VSVG signal in control group (D’’–E’’, arrows) and a large aggregate of VSVG signal in MG132-treated cell (I’’, arrowhead). The images shown are representatives of 5 photos taken in each group. Scale bars represent 10 ␮m.

GA fragments around the centrosomal aggregate depends on MTs, SH-SY5Y cells were treated with nocodazole for 2 hours to depolymerize MTs and then allow to recover for an additional 2-hour period after washout of the drug. As expected, in the presence of nocodazole MTs were depolymerized and the GA was fragmented and dispersed throughout the cytoplasm both in control and treated cells (Supplementary Fig. 2C–C’’’ and E–E’’’). In the same way, MTs were completely repo-

lymerized after 2 hours of nocodazole washout and GA reassembled next to the centrosome or to the centrosomal aggregates in control or in MG132 treated cells, respectively (Supplementary Fig. 2D–D’’’ and F–F’’’, arrowhead). These results indicated that the capacity of nocodazole-generated Golgi ministacks to be concentrated at the minus ends of MTs persist even when centrosomal structure and MT nucleation are strongly perturbed. However, the fusion of these elements

2472

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

Fig. 6. Vesicular transport of dopamine transporter (DAT) is disrupted in Z-Leu-Leu-Leu-al (MG132)-treated cells with centrosoma-proteins aggregates and fragmentation of Golgi apparatus (GA). Confocal microscopy images (A–C’’’) of SH-SY5Y cells treated with vehicle (A–A’’’) or 500 nM MG132 during 24 hours (B–C’’’). Centrosomes or the ␥-tubulin aggregates were visualized by anti-␥-tubulin antibodies (A–C). The GA was immunostained by

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

into a single structure was completely inhibited in MG132treated cells, and the Golgi fragments remained distributed at the periphery of ␥-tubulin aggregates. In order to gain further insights into the MG132induced GA fragmentation, we performed live-cell imaging in SH-SY5Y cells stably expressing the transmembrane domain of the Golgi enzyme GT fused to GFP. Before MG132 exposure, SH-SY5Y cells showed a normal Golgi ribbon next to the nucleus (Supplementary Video 2, Supplementary Fig. 3). At 1 hour after treatment, the Golgi ribbon was already fragmented in several large and small elements (Supplementarty Video 2, Supplementary Fig. 3, arrow). By 2 hours the GA became completely fragmented in small elements that remained together and close to the nucleus. At this same time, deformation of the nuclear envelope started indicating the formation of a big aggregate around which Golgi elements were distributed (Supplementary Video 2, Supplementary Fig. 3, arrowhead). Finally Golgi elements acquired a spherical distribution similar to that observed by IF (Supplementary Video 2, Supplementary Fig. 3). Fragmentation of GA persisted until the collapse and cell death. These in vivo results indicate that fragmentation of GA and formation of perinuclear aggregates are early events, at the morphological level, occurring after proteasome inhibition. 3.4. Proteasome inhibition differentially affected Golgi compartments In order to investigate whether the stacking of cisternae was also perturbed by proteasome inhibition, we analyzed the distribution of several Golgi compartment markers in control and MG132-treated cells. We compared the distribution of 2 cis-Golgi proteins (GMAP210 and GM130) and 2 trans-Golgi proteins (Golgin 245 and Golgin 97) by confocal microscopy. In control cells showing a compact and tangled Golgi ribbon, the cis- and trans-Golgi faces positioned next to each other (Fig. 4A–A’’). A line scan drawn on the Golgi ribbon showed the juxtaposition of GMAP210 and Golgin 245 labelings representing the cis- and trans-Golgi faces, respectively (Fig. 4A’’, blue lines, and C). Notably, in the 500 nM MG132-treated cells Golgin 245 labeling almost com-

2473

pletely disappeared (Fig. 4B–B’’, blue line, and C). The cis-Golgi markers, GMAP210 and GM130, remained mostly unchanged in the GA fragments of MG132treated cells (Fig. 4D–D’ and F), whereas the trans-Golgi markers, Golgin 245 and Golgin 97, almost completely disappeared from the Golgi fragments (Fig. 4D’’and F’– F’’). WB analysis revealed an increase of both GMAP210 and GM130 in 500 nM-MG132-treated cells, whereas neither the concentration of Golgin 245 nor that of Golgin 97 was affected (Fig. 4H). These results suggest that the loss of Golgin 245 and Golgin 97 labeling in IF studies is not due to degradation of these proteins but either to their dissociation of Golgi membranes or even to the disassembly of the trans-Golgi network. Altogether these results indicate that proteasome inhibition differentially affected the different compartment. Thus, the trans-Golgi network was more affected that the cis-face of the GA in MG132-treated cells. The reason for this differential response to proteasome inhibition remains unknown. 3.5. Proteasome inhibition interrupts the VSVG transport from the GA to the plasma membrane In order to evaluate the secretory pathway in MG132treated cells, we follow the transport from the ER to the plasma membrane of the VSVG, a widely used assay to study the kinetic of protein trafficking through membrane compartments. Misfolded form of VSVG is retained in the ER at 40 °C, but when temperature is shifted to 32 °C, VSVG synchronously moves to the Golgi complex before being transported to the plasma membrane (Hirschberg et al., 1998). Control and MG132-treated cells were transfected with the GFP-VSVG construct and then incubated at different temperatures as detailed in the scheme (Fig. 5A). In control and MG132-treated cells GFP-VSVG signal was retained into the ER before the shift to a permissible temperature (Fig. 5B–B’’and F–F’’). The transport of VSVG signal from the ER to GA occurred quickly during the first 15 minutes. After 30 minutes, the VSVG signal was accumulated at the GA of control cells (Fig. 5C–C’’) or at the Golgi fragments in the MG132-treated group (Fig. 5G–G’’). Thus, the protein trafficking from the ER to GA was not affected.

anti-GMAP210 antibodies (A’–C’). Distribution of DAT signal was observed by anti-DAT antibodies (A’’–C’’). Merged (A’’’–C’’’) and amplified images (C–C’’’) are shown. Small amount of dopamine transporter (DAT) signal reaching the plasma membrane in MG132-treated cells (B’’, filled arrow). Large ␥-tubulin aggregate, fragmentation of GA and accumulation of DAT signal in the GA fragments of MG132-treated cells (B–B’’’, insets and C–C’’’, white arrowheads). Overlap coefficients of DAT and GMAP210 signals (D). Time-lapse frames of SH-SY5Y cells permanently expressing the pIRESneogalactosyl-transferase (GT)-green fluorescent protein (GFP) construct (E–H) and cotransfected with the pDsRed-DAT construct (E’–H’). Cells were treated with 500 nM MG132 during 24 hours. Merged images (E’’–H’’) and amplification of GA region (I, K, M, and O) and plasma membrane (J, L, N, and P) are shown. Colocalization of Golgi fragments with DAT signal after 7.5 hours of initiated MG132 treatment (M, empty arrows). Measurement of intracellular dopamine (DA) and its metabolites by high-performance liquid chromatography (HPLC) in vehicle and MG132-treated cells (Q). Results in (D) are expressed as arbitrary units and represent means ⫾ standard error of the mean (SEM) of 5 random images. # p ⬍ 0.001, analysis of variance (ANOVA) followed of Dunnett test. Results in (Q) are expressed as pmol/mg protein and represent means ⫾ SEM of 4 independent experiments. * p ⬍ 0.05 and ** p ⬍ 0.01, unpaired 2-sample t tests. The images shown are representatives of 5 photos taken in each group. Scale bars represent 10 ␮m (A–B’’’), 5 ␮m (C–C’’’), 20 ␮m (E–H’’), and 5 ␮m (I–P). DOPAC, 3,4-dihydroxyphenylacetic acid; HVA, homovanillic acid.

2474

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

Between 60 and 120 minutes, the VSVG signal reached the plasma membrane in vehicle-treated cells (Fig. 5D– E’’, arrows). On the contrary, the VSVG signal was still detected in the Golgi elements in MG132-treated cells at the same time point, and it was almost complete absent from the plasma membrane (Fig. 5H–H’’). After 120 minutes most of VSVG signal was found in the perinuclear region of MG132-treated cells where protein aggregates are present (Fig. 5I–I’’, arrowhead). We conclude that vesicular traffic of VSVG from the GA to the plasma membrane was interrupted under these conditions resulting in the accumulation of VSVG signal in the perinuclear region. Due to VSVG being probably not the best marker for analyzing perturbation of trafficking in polarized cells, we could not determine whether the VSVG transport from the GA to the neurites (axon- or dendritelike) was preferentially affected. 3.6. Vesicular transport of DAT is disrupted in MG132treated cells with centrosoma-protein aggregates and fragmentation of GA To more specifically evaluate the consequence of the MG132-induced vesicular transport inhibition in the trafficking of proteins relevant to PD, we follow the transport of endogenous DAT in SH-SY5Y cells treated with vehicle or with 500 nM MG132. Before adding the drug, endogenous DAT was detected at the plasma membrane and no colocalization with GMAP210 at the GA was observed (Fig. 6A–A’’’). By contrast, in MG132-treated cells DAT signal clearly colocalized with Golgi fragments that surrounded the centrosomal aggregate (Fig. 6B–B’’’). Although some DAT signal seemed to reach the plasma membrane in MG132-treated cells (Fig. 6B’’, filled arrow), most DAT signal accumulated in the centrosomal-GA region (Fig. 6C–C’’’, arrowheads). Moreover, colocalization analysis of DAT and GMAP210 signals revealed a significant increase in the overlapping coefficient in MG132-treated cells when compared with control group (Fig. 6D), which confirmed the accumulation of DAT into the GA fragments. To analyze the sequence of events that drives the accumulation of DAT signal in the centrosomal-GA region, live-cell imaging was carried out in GT-GFPSH-SY5Y cell line transiently transfected with the pDsRed-DAT construct (Supplementary Video 3, Fig. 6E– P). Before MG132 exposure, Golgi ribbon was clearly observed, and DAT reached the plasma membrane (Supplementary Video 3, Fig. 6E–E’’, I, and J). From 5 to 13 hours after MG132 treatment, DAT signal started to accumulate in the perinuclear region surrounded by Golgi fragments (Supplementary Video 3, Fig. 6F–G’’, and K–N). Colocalization of DAT signal with Golgi fragments was observed in several frames (Supplementary Video 3, Fig. 6M, empty arrows). In parallel to this progressive accumulation of DAT signal, the expression of DAT in the plasma membrane decreased (Supplemen-

tary Video 3, Fig. 6J, L, N, and P). The number of DAT-containing vesicles moving away from the GA progressively decreased until its complete disappearance (Supplementary Video 3). After 21 hours, cell processes began to retract and DAT signal was no longer detected at the cell surface (Supplementary Video 3, Fig. 6H–H’’, O, and P). High resolution images of Golgi area in pDsRed-DAT transfected cells (Supplementary Fig. 4) showed partial colocalization of DAT and Golgi staining in membrane elements surrounding the centrosomal aggregate (Supplementary Fig. 4B’). In addition, a few DAT-containing structures were visible away from the central area, close to the neuritis (Supplementary Fig. 4B’’). These results suggest that both exit from the GA and transport to the cell surface was perturbed in MG132treated cells. Therefore, IF and time-lapse results demonstrate that MG132 treatment not only induces fragmentation of GA but also blocks the GA to the plasma membrane transport step. DAT transport was blocked at this point of the secretory pathway resulting in a decrease of DAT expression in the plasma membrane. Finally, we wanted to determine whether disruption of vesicular transport of DAT modified the concentration of intracellular DA and its metabolites (DOPAC and HVA). To do this, samples from control and MG132-treated cells were collected and analyzed by HPLC. Unfortunately, we detected that intracellular DA and HVA levels were significantly higher in 500 nM MG132-treated that in control cells even in the absence of exogenously added DA (Fig. 6Q). These results indicated that MG132 treatment affected the DA metabolism itself, probably by increasing DA synthesis or decreasing DA release, and made difficult the evaluation of DAT functionality.

4. Discussion Protein aggregation and formation of inclusion bodies as well as fragmentation of GA are common characteristics presented in several neurodegenerative diseases including PD (Gonatas et al., 2006; Jellinger, 2009; McNaught and Olanow, 2006). The study of the molecular mechanisms involved in fragmentation of GA and formation of proteins aggregates as well as the effects they produce on cell functioning might provide new missing parts in the puzzle of PD. In this article, we demonstrated that proteasome inhibition induces large centrosomal protein aggregates, fragmentation of GA, and disruption of DAT trafficking toward the plasma membrane. The relevance of these results to PD pathophysiology arises from the potential connection between 3 morphological and neuroimaging features observed in nigral neurons of PD patients, such as inclusion bodies, fragmentation of GA, and the early reduction of DAT expression in their synaptic terminals at the striatum. As mentioned above, there is growing evidence supporting that UPS failure might contribute to the degen-

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

eration of dopaminergic neurons in PD as well as that LBs in PD might be formed by a process similar to the formation of aggresomes in cell culture systems (Ardley et al., 2003; McNaught et al., 2002b; Olanow et al., 2004; Olzmann et al., 2008; Waxman et al., 2009). Aggresomes can be experimentally induced by proteasome inhibition or by overexpression of prone-aggregate proteins (Johnston et al., 1998). In this study, the formation of aggresomes was induced effectively and exclusively by a low concentration of MG132 using a dopaminergic cell line. Although the formation of aggresomes has been considered as a cytoprotective mechanism of cells to recruit potentially toxic misfolded proteins at the centrosome to facilitate their degradation and elimination by autophagy (Olzmann et al., 2008; Tanaka et al., 2004), we observed that ␥-tubulin aggregates produced by proteasome inhibition affected MT nucleation in SH-SY5Y cells suggesting that aggresomes are not as innocuous as believed. Indeed, we already described that rotenone, a mitochondria Complex I inhibitor, induces also insoluble ␥-tubulin aggregates in primary cell cultures of dopaminergic neurons and astrocytes. Neurons with disorganized centrosomes exhibited neurite retraction and MT destabilization, and astrocytes showed disturbances of mitotic spindles (Diaz-Corrales et al., 2005). Thus, if LBs are bona fide aggresomes, MT nucleation might be perturbed in nigral neurons of PD patients as a consequence of protein aggregation at the neuronal centrosome. MG132-treated cells with enlarged centrosomes curiously displayed an increase in acetylated MTs. It is well known that MTs and their posttranslational modifications are essential for the development and maintenance of neuronal polarization as well as organization of axons and dendrites (Conde and Cáceres, 2009; Fukushima et al., 2009). Therefore, any change in MT dynamics might rebound in the synaptic transmission. There are several lines of evidence indicating a relationship between acetylated MTs and the aggresome. It has been reported that histone deacetylase 6 (HDAC6), the enzyme responsible for MT deacetylation, is essential for the formation of aggresomes. HDAC6 links polyubiquitinated proteins to the dynein motor complex in order to be retrogradely transported toward the centrosome (Kawaguchi et al., 2003). In addition, HDAC6 accumulates in aggresomes induced by proteasome inhibition and also in LBs of PD patients (Kawaguchi et al., 2003). Thus, if HDAC6 is sequestered into the inclusion bodies, it could indirectly affect the deacetylation of MTs. This could account for the higher number of acetylated MTs in MG132-treated cells. Whether acetylated MTs are increased in PD brain samples, and the effects of HDAC6 overexpression or interference should be tested in future experiments. It is well known that centrosome and GA are 2 closely related organelles mainly through the MT network (Thyberg and Moskalewski, 1999). It is known that alteration on MT nucleation activity of the centrosome directly affect the structural integrity of GA. As we mentioned above, fragmentation of GA has been observed in neurons with inclu-

2475

sion bodies in PD as well as in several neurodegenerative diseases. This suggests that both morphological features coexist during the neurodegenerative process. We described that both ␥-tubulin aggregates and fragmentation of GA were present into the same cell in around 80% of MG132treated cells. Our in vivo studies indicate that fragmentation of GA and formation of perinuclear aggregates are early morphological changes detected in MG132-treated cells. It has been reported that mice expressing the mutant transgene for Cu, Zn superoxide dismutase, an animal model of amyotrophic lateral sclerosis, display a severe fragmentation of GA which occurs even before the occurrence of motor symptoms (Mourelatos et al., 1996). It suggests that fragmentation of GA might be an early step during the neurodegenerative process. We also demonstrated that distribution of Golgi fragments around the ␥-tubulin aggregates is an MT-dependent process that can take place in these perturbed cells whereas the fusion of these Golgi elements into a single structure is completely blocked. Therefore, it is the latter step in the Golgi biogenesis process that is specifically inhibited by proteasome inhibition. We consider that other molecular mechanisms might also contribute to the fragmentation of GA observed during the neurodegenerative process. Indeed, it has been described that protein kinases antagonists decreases neuronal cell death exposed to apoptotic stimulus (Nakagomi et al., 2008). Thus, protein aggregation and cytoskeleton modifications as well as high kinases activity might interact to produce fragmentation of GA in neurodegeneration. Protein trafficking from the GA to the plasma membrane is an essential mechanism of neuronal cells to develop neurite outgrowth and synaptogenesis (Sytnyk et al., 2004; Tang, 2001). In this work, we demonstrate that proteasome inhibition profoundly perturb membrane trafficking from the GA to the plasma membrane. Interestingly, the localization of some trans-Golgi proteins was modified by MG132 treatment. Why trans-Golgi proteins exhibit a different behavior than cis-Golgi proteins have to be clarified, but a simple explanation could be the lack of membrane input delivered from the cis-to the trans-Golgi due to a disruption of intra-GA transport. The trans-Golgi network is responsible for sorting and dispatching of mature cargos into membrane carries to the plasma membrane, endosomes, secretory granules, back to the ER or even GA as well as it also receives endosomal cargo (Lieu et al., 2008). Thus, perturbation of the trans-Golgi network in MG132-treated cells could explain why vesicular transport of VSVG and DAT was disrupted with the subsequent accumulation into the centrosomal-GA region of both proteins. Recently, it has been reported that intracerebroventricular injection of 6-hydroxydopamine in rats, a widely used experimental model of PD, induces a transient accumulation of DAT into the ER-Golgi compartment (Afonso-Oramas et al., 2010). Although DAT transport was not evaluated in those 6-hydroxydopamine treated rats, the accumulation of DAT sig-

2476

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477

nal in GA could represent a deficiency on DAT transport similar to the findings of our experiments in a cellular system. On the other hand, it has been reported that overexpression of wild type or A53T mutant ␣-synuclein delayed ER to Golgi transport (Thayanidhi et al., 2010). In addition, it has been described that DAT trafficking is affected in ␣-synuclein overexpressing cells (Gosavi et al., 2002). Those results coupled with our experimental data suggest that disruption in protein trafficking through different cellular compartments might be an important mechanism in PD pathophysiology. Positron emission tomography images from asymptomatic members of autosomal dominant PD linked to LRRK2 mutations and early symptomatic patients with the sporadic form of PD have showed reduction of DAT binding greater than DA storage or even than vesicular monoamine transporter 2 density (Adams et al., 2005; Lee et al., 2000). Those observations suggest that DAT reduction is more evident than the loss of other markers for dopaminergic terminals. This nonparallel reduction of DAT expression with the loss of dopaminergic terminals has been explained as a compensatory mechanism of degenerative neurons to maintain appropriate extracellular DA levels through downregulation of DAT. However, whether DAT is preferentially targeted in the early progress of PD could not be discarded (Adams et al., 2005). We demonstrated that in MG132-treated SH-SY5Y cells, vesicular transport of DAT is disrupted. The simplest explanation is that this perturbation on vesicular transport of DAT is secondary to changes in GA structure and/or in ␥-tubulin aggregation. Because these 2 phenotypes, ␥-tubulin aggregation in LBs and GA fragmentation are present in PD, we speculate that a similar mechanism could account for the early reduction of striatal DAT observed in positron emission tomography studies of PD patients. It is possible that SNc neurons in PD patients are suffering a chronic condition of proteolytic stress caused for several reasons such as excessive formation of unwanted proteins (mutant, misfolded, or denatured proteins) or by impaired protein degradation (Olanow et al., 2004), which could lead to the formation of centrosomal protein aggregates and perturbation of Golgi structure and functioning. This could eventually result in a block of membrane transport toward the plasma membrane, thus affecting the turnover of important membrane proteins of dopaminergic neurons such as DAT. If this hypothetical sequence of events and connection between them could be confirmed in PD patients, this would boost the study of new therapeutic targets to keep the structural integrity of GA ensuring an adequate transport of DAT and a balance in the dopaminergic synapses. Disclosure statement The authors disclose no conflicts of interest. Acknowledgements The authors thank Paloma Domínguez Giménez and Jesus Cardenas for excellent technical assistance. Francisco J.

Díaz Corrales was supported by the fellowship JAE-Doc from the “Consejo Superior de Investigaciones Científicas” (CSIC), Spain. This work was supported in part by “Consejería de Innovación Ciencia y Empresa, Proyecto de Excelencia” P07-CVI-03199, Spain; “Ministerio de Ciencia e Innovación” BFU2009 – 07182, Spain; and Consolider CSD2009 – 00016, Spain; and Grants in Aid for Scientific Research (C) from the Japanese Ministry of Education, Culture, Sports, Science and Technology. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.neurobiolaging. 2011.11.014. References Adams, J.R., van Netten, H., Schulzer, M., Mak, E., McKenzie, J., Strongosky, A., Sossi, V., Ruth, T.J., Lee, C.S., Farrer, M., Gasser, T., Uitti, R.J., Calne, D.B., Wszolek, Z.K., Stoessl, A.J., 2005. PET in LRRK2 mutations: comparison to sporadic Parkinson’s disease and evidence for presymptomatic compensation. Brain 128, 2777–2785. Afonso-Oramas, D., Cruz-Muros, I., Barroso-Chinea, P., Alvarez de la Rosa, D., Castro-Hernandez, J., Salas-Hernandez, J., Giradlez, T., Gonzalez-Hernandez, T., 2010. The dopamine transporter is differentially regulated after dopaminergic lesion. Neurobiol. Dis. 40, 518 –530. Ardley, H.C., Scott, G.B., Rose, S.A., Tan, N.G., Markham, A.F., Robinson, P.A., 2003. Inhibition of proteasomal activity causes inclusion formation in neuronal and non-neuronal cells overexpressing Parkin. Mol. Biol. Cell 14, 4541– 4556. Betarbet, R., Sherer, T.B., Greenamyre, J.T., 2005. Ubiquitin-proteasome system and Parkinson’s diseases. Exp. Neurol. 191, S17–S27. Cardenas, J., Rivero, S., Goud, B., Bornens, M., Rios, R.M., 2009. Golgi localisation of GMAP210 requires two distinct cis-membrane binding mechanisms. BMC Biol. 7, 56. Conde, C., Cáceres, A., 2009. Microtubule assembly, organization and dynamics in axons and dendrites. Nat. Rev. Neurosci. 10, 319 –332. de Lau, L.M., Breteler, M.M., 2006. Epidemiology of Parkinson’s disease. Lancet Neurol. 5, 525–535. Diaz-Corrales, F.J., Asanuma, M., Miyazaki, I., Miyoshi, K., Ogawa, N., 2005. Rotenone induces aggregation of gamma-tubulin protein and subsequent disorganization of the centrosome: relevance to formation of inclusion bodies and neurodegeneration. Neuroscience 133, 117–135. Didier, C., Merdes, A., Gairin, J.E., Jabrane-Ferrat, N., 2008. Inhibition of proteasome activity impairs centrosome-dependent microtubule nucleation and organization. Mol. Biol. Cell 19, 1220 –1229. Fan, J., Hu, Z., Zeng, L., Lu, W., Tang, X., Zhang, J., Li, T., 2008. Golgi apparatus and neurodegenerative diseases. Int. J. Dev. Neurosci. 26, 523–534. Fujita, Y., Ohama, E., Takatama, M., Al-Sarraj, S., Okamoto, K., 2006. Fragmentation of Golgi apparatus of nigral neurons with alpha-synuclein-positive inclusions in patients with Parkinson’s disease. Acta Neuropathol. 112, 261–265. Fukushima, N., Furuta, D., Hidaka, Y., Moriyama, R., Tsujiuchi, T., 2009. Post-translational modifications of tubulin in the nervous system. J. Neurochem. 109, 683– 693. Gonatas, N.K., Stieber, A., Gonatas, J.O., 2006. Fragmentation of the Golgi apparatus in neurodegenerative diseases and cell death. J. Neurol. Sci. 246, 21–30. Gosavi, N., Lee, H.J., Lee, J.S., Patel, S., Lee, S.J., 2002. Golgi fragmentation occurs in the cells with prefibrillar alpha-synuclein aggregates and precedes the formation of fibrillar inclusion. J. Biol. Chem. 277, 48984 – 48992.

F.J. Diaz-Corrales et al. / Neurobiology of Aging 33 (2012) 2462–2477 Hirschberg, K., Miller, C.M., Ellenberg, J., Presley, J.F., Siggia, E.D., Phair, R.D., Lippincott-Schwartz, J., 1998. Kinetic analysis of secretory protein traffic and characterization of golgi to plasma membrane transport intermediates in living cells. J. Cell Biol. 143, 1485–1503. Infante, C., Ramos-Morales, F., Fedriani, C., Bornens, M., Rios, R.M., 1999. GMAP-210, A cis-Golgi network-associated protein, is a minus end microtubule-binding protein. J. Cell Biol. 145, 83–98. Jellinger, K.A., 2009. Recent advances in our understanding of neurodegeneration. J. Neural Transm. 116, 1111–1162. Johnston, J.A., Ward, C.L., Kopito, R.R., 1998. Aggresomes: a cellular response to misfolded proteins. J. Cell Biol. 143, 1883–1898. Kawaguchi, Y., Kovacs, J.J., McLaurin, A., Vance, J.M., Ito, A., Yao, T.P., 2003. The deacetylase HDAC6 regulates aggresome formation and cell viability in response to misfolded protein stress. Cell 115, 727–738. Lee, C.S., Samii, A., Sossi, V., Ruth, T.J., Schulzer, M., Holden, J.E., Wudel, J., Pal, P.K., de la Fuente-Fernandez, R., Calne, D.B., Stoessl, A.J., 2000. In vivo positron emission tomographic evidence for compensatory changes in presynaptic dopaminergic nerve terminals in Parkinson’s disease. Ann. Neurol. 47, 493–503. Lees, A.J., Hardy, J., Revesz, T., 2009. Parkinson’s disease. Lancet 373, 2055–2066. Levy, O.A., Malagelada, C., Greene, L.A., 2009. Cell death pathways in Parkinson’s disease: proximal triggers, distal effectors, and final steps. Apoptosis 14, 478 –500. Li, L.B., Chen, N., Ramamoorthy, S., Chi, L., Cui, X.N., Wang, L.C., Reith, M.E., 2004. The role of N-glycosylation in function and surface trafficking of the human dopamine transporter. J. Biol. Chem. 279, 21012–21020. Liang, Y.J., Zhen, J., Chen, N., Reith, M.E., 2009. Interaction of catechol and non-catechol substrates with externally or internally facing dopamine transporters. J. Neurochem. 109, 981–994. Lieu, Z.Z., Derby, M.C., Gleeson, P.A., 2008. Structure and domain organization of the trans-Golgi network, in: Mironov, A.A., Pavelka, A.A. (Eds.), The Golgi Apparatus—State of the Art 110 Years After Camilo Golgi’s Discovery. Springer, Wien, Austria, pp. 358 –374. McNaught, K.S., Mytilineou, C., Jnobaptiste, R., Yabut, J., Shashidharan, P., Jennert, P., Olanow, C.W., 2002a. Impairment of the ubiquitin-proteasome system causes dopaminergic cell death and inclusion body formation in ventral mesencephalic cultures. J. Neurochem. 81, 301–306. McNaught, K.S., Olanow, C.W., 2006. Protein aggregation in the pathogenesis of familial and sporadic Parkinson’s disease. Neurobiol. Aging 27, 530 –545. McNaught, K.S., Shashidharan, P., Perl, D.P., Jenner, P., Olanow, C.W., 2002b. Aggresome-related biogenesis of Lewy bodies. Eur. J. Neurosci. 16, 2136 –2148. Moudjou, M., Bordes, N., Paintrand, M., Bornens, M., 1996. gammaTubulin in mammalian cells: the centrosomal and the cytosolic forms. J. Cell Sci. 109, 875– 887. Mourelatos, Z., Gonatas, N.K., Stieber, A., Gurney, M.E., Dal Canto, M.C., 1996. The Golgi apparatus of spinal cord motor neurons in transgenic mice expressing mutant Cu,Zn superoxide dismutase becomes fragmented in early, preclinical stages of the disease. Proc. Natl. Acad. Sci. U. S. A. 93, 5472–5477. Nakagomi, S., Barsoum, M.J., Bossy-Wetzel, E., Sütterlin, C., Malhotra, V., Lipton, S.A., 2008. A Golgi fragmentation pathway in neurodegeneration. Neurobiol. Dis. 29, 221–231. Nigg, E.A., Raff, J.W., 2009. Centrioles, centrosomes, and cilia in health and disease. Cell 139, 663– 678.

2477

Olanow, C.W., Perl, D.P., DeMartino, G.N., McNaught, K.S., 2004. Lewybody formation is an aggresome-related process: a hypothesis. Lancet Neurol. 3, 496 –503. Olzmann, J.A., Li, L., Chin, L.S., 2008. Aggresome formation and neurodegenerative diseases: therapeutic implications. Curr. Med. Chem. 15, 47– 60. Piperno, G., LeDizet, M., Chang, X.J., 1987. Microtubules containing acetylated alpha-tubulin in mammalian cells in culture. J. Cell Biol. 104, 289 –302. Raynaud-Messina, B., Merdes, A., 2007. Gamma-tubulin complexes and microtubule organization. Curr. Opin. Cell Biol. 19, 24 –30. Ríos, R.M., Sanchís, A., Tassin, A.M., Fedriani, C., Bornens, M., 2004. GMAP-210 recruits gamma-tubulin complexes to cis-Golgi membranes and is required for Golgi ribbon formation. Cell 118, 323–335. Rios, R.M., Tassin, A.M., Celati, C., Antony, C., Boissier, M.C., Homberg, J.C., Bornens, M., 1994. A peripheral protein associated with the cis-Golgi network redistributes in the intermediate compartment upon brefeldin A treatment. J. Cell Biol. 125, 997–1013. Shults, C.W., 2006. Lewy bodies. Proc. Natl. Acad. Sci. U. S. A. 103, 1661–1668. Spillantini, M.G., Schmidt, M.L., Lee, V.M., Trojanowski, J.Q., Jakes, R., Goedert, M., 1997. Alpha-synuclein in Lewy bodies. Nature 388, 839 – 840. Storch, A., Ludolph, A.C., Schwarz, J., 2004. Dopamine transporter: involvement in selective dopaminergic neurotoxicity and degeneration. J. Neural Transm. 111, 1267–1286. Sun, F., Anantharam, V., Zhang, D., Latchoumycandane, C., Kanthasamy, A., Kanthasamy, A.G., 2006. Proteasome inhibitor MG-132 induces dopaminergic degeneration in cell culture and animal models. Neurotoxicology 27, 807– 815. Sytnyk, V., Leshchyns’ka, I., Dityatev, A., Schachner, M., 2004. TransGolgi network delivery of synaptic proteins in synaptogenesis. J. Cell Sci. 117, 381–388. Tanaka, M., Kim, Y.M., Lee, G., Junn, E., Iwatsubo, T., Mouradian, M.M., 2004. Aggresomes formed by alpha-synuclein and synphilin-1 are cytoprotective. J. Biol. Chem. 279, 4625– 4631. Tang, B.L., 2001. Protein trafficking mechanisms associated with neurite outgrowth and polarized sorting in neurons. J. Neurochem. 79, 923–930. Thayanidhi, N., Helm, J.R., Nycz, D.C., Bentley, M., Liang, Y., Hay, J.C., 2010. Alpha-synuclein delays endoplasmic reticulum (ER)-to-Golgi transport in mammalian cells by antagonizing ER/Golgi SNAREs. Mol. Biol. Cell 21, 1850 –1863. Thyberg, J., Moskalewski, S., 1999. Role of microtubules in the organization of the Golgi complex. Exp. Cell Res. 246, 263–279. Wakabayashi, K., Tanji, K., Mori, F., Takahashi, H., 2007. The Lewy body in Parkinson’s disease: molecules implicated in the formation and degradation of alpha-synuclein aggregates. Neuropathology 27, 494 –506. Waxman, E.A., Covy, J.P., Bukh, I., Li, X., Dawson, T.M., Giasson, B.I., 2009. Leucine-rich repeat kinase 2 expression leads to aggresome formation that is not associated with alpha-synuclein inclusions. J. Neuropathol. Exp. Neurol. 68, 785–796. Xie, W., Li, X., Li, C., Zhu, W., Jankovic, J., Le, W., 2010. Proteasome inhibition modeling nigral neuron degeneration in Parkinson’s disease. J. Neurochem. 115, 188 –199. Zahniser, N.R., Sorkin, A., 2009. Trafficking of dopamine transporters in psychostimulant actions. Semin. Cell Dev. Biol. 20, 411– 417.