Phosphoproteome Study Reveals Hsp27 as a Novel Signaling Molecule Involved in GDNF-Induced Neurite Outgrowth Zhen Hong,†,# Qun-Ye Zhang,‡,# Jun Liu,†,# Zhi-Quan Wang,§ Yu Zhang,† Qin Xiao,† Jing Lu,‡ Hai-Yan Zhou,† and Sheng-Di Chen*,†,§ Department of Neurology & Institute of Neurology, Ruijin Hospital affiliated to Shanghai Jiaotong University School of Medicine, Shanghai, 200025 China, State Key Laboratory of Medical Genomics, Ruijin Hospital Affiliated to Shanghai Jiaotong University School of Medicine, Shanghai, 200025 China, and Institute of Health Science, Shanghai Institutes of Biological Sciences (SIBS), Chinese Academy of Sciences (CAS) & Shanghai Jiaotong University School of Medicine, Shanghai, 200025 China Received December 8, 2008
Glial-cell-line-derived neurotrophic factor (GDNF) is a most potent survival factor for dopaminergic neurons. In addition, GDNF was also found to promote neurite outgrowth in dopaminergic neurons. However, despite the potential clinical and physiological importance of GDNF, its mechanism of action is unclear. Therefore, we employed a state-of-the-art proteomic technique, DIGE (Difference in twodimensional gel electrophoresis), to quantitatively compare profiles of phosphoproteins of PC12-GFRR1RET cells (that stably overexpress GDNF receptor R1 and RET) 0.5 and 10 h after GDNF challenge with control. A total of 92 differentially expressed proteins were successfully identified by mass spectrometry. Among them, the relative levels of phosphorylated Hsp27 increased significantly both in 0.5 and 10 h GDNF-treated PC12-GFRR1-RET cells. Confocal microscopy and Western blot results showed that the phosphorylation of Hsp27 after GDNF treatment was accompanied by its nuclear translocation. After the mRNA of Hsp27 was interfered, neurite outgrowth of PC12-GFRR1-RET cells induced by GDNF was significantly blocked. Furthermore, the percentage of neurite outgrowth induced by GDNF was also reduced by the expression of dominant-negative mutants of Hsp27, in which specific serine phosphorylation residues (Ser15, Ser78 and Ser82) were substituted with alanine. Our data also revealed that p38 MAPK and ERK are the upstream regulators of Hsp27 phosphorylation. Hence, in addition to the numerous novel proteins that are potentially important in GDNF mediated differentiation of dopaminergic cells revealed by our study, our data has indicated that Hsp27 is a novel signaling molecule involved in GDNF-induced neurite outgrowth of dopaminergic neurons. Keywords: GDNF • neurite outgrowth • proteomics • phosphorylation • Hsp27 • dopaminergic neuron
Introduction Glial-cell-line-derived neurotrophic factor (GDNF) is a distant member of the transforming growth factor β superfamily and a founding member of the GDNF family ligands (GFLs). GDNF was originally discovered because of its ability to promote the survival of the embryonic dopaminergic neurons of the midbrain, that is, those neurons that degenerate in Parkinson disease (PD).1 Its ability to rescue dopaminergic (DA) neurons supported the idea that GDNF might ameliorate degeneration of DA neurons in * Corresponding author: Sheng-Di Chen, M.D., Ph.D., Department of Neurology & Institute of Neurology, Ruijin Hospital affiliated to Shanghai Jiaotong University School of Medicine, No. 197 2nd Ruijing Road, Shanghai 200025, China. Phone and fax number: +86-21-6445-7249. E-mail: chen_sd@ medmail.com.cn. † Department of Neurology & Institute of Neurology, Ruijin Hospital affiliated to Shanghai Jiaotong University School of Medicine. ‡ State Key Laboratory of Medical Genomics, Ruijin Hospital affiliated to Shanghai Jiaotong University School of Medicine. # Authors contributed equally to this work. § Institute of Health Science, Shanghai Institutes of Biological Sciences (SIBS).
2768 Journal of Proteome Research 2009, 8, 2768–2787 Published on Web 03/16/2009
patients with PD.2,3 GDNF is also a potent survival factor for spinal motoneurons and so could have clinical importance for the treatment of amyotrophic lateral sclerosis.4 In addition, GDNF regulates the differentiation of both DA neurons and many peripheral neurons, such as sympathetic, parasympathetic, sensory and enteric neurons.5-7 All these findings strengthened the point that GDNF is functionally important, especially in nervous system. The neurotrophic and morphogenic activities of GDNF are mediated by its interaction with a multicomponent receptor complex formed by a common transmembrane signaling component, the RET receptor tyrosine kinase8-10 and a ligand-binding receptor, which belongs to a family of glycosyl phosphatidylinositol-anchored membrane proteins, called GDNF family receptor R1 (GFRR1).11-14 Binding of GDNF to RET and GFRR1 induces RET phosphorylation.7 After phosphorylation, RET induces the activation of several intracellular pathways, among which the MAPK and the PI3K are of particular interest.15,16 However, the detail of GDNF signaling is still obscure. 10.1021/pr801052v CCC: $40.75
2009 American Chemical Society
Hsp27 Involved in GDNF-Induced Neurite Outgrowth Recently, two-dimensional gel electrophoresis (2-DE) has been applied to analyze intracellular signaling events. However, due to insufficient resolution power, the dense spots of components for cytoskeleton or housekeeping metabolic enzymes have frequently shadowed low-abundance proteins such as factors involved in signal transduction. To overcome this problem, prefractionation of proteins is highly desirable. Since phosphorylation is a basic and dominant mechanism of information transfer inside the cell and also plays a key role in the GDNF signaling, we focus on the phosphoproteins which are involved in the GDNF signal transduction pathway. A prefractionation procedure, phosphoprotein purification, is suitable for the identification of the phosphorylation status of all proteins at a given time of cell life, since it enriches phosphorylated substrates while reducing the number of protein species, which greatly facilitates protein identification. Difference in two-dimensional gel electrophoresis (DIGE) is a powerful tool employed in monitoring the complex differences in proteomic profiles between cells in different functional states.17,18 The original DIGE reagents were based on minimal labeling.17 In reconstruction experiments with standard proteins, DIGE saturation labeling at cysteine residues produced as much as a 10-fold gain in detection sensitivity compared to “minimal labeling” at lysine residues.19 Therefore, as a part of our ongoing effort in defining the mechanisms related to the GDNF induced protection and differentiation on dopaminergic neurons,20-22 we have utilized rat pheochromocytoma (PC12) cells as a model since they possess much of the biochemical machinery associated with dopaminergic neurons and could be induced to undergo differentiation upon treatment of GDNF,23 to globally identify the phosphorylated factors involved in the GDNF cascades. In this study, we combined three key methodologies, enrichment of phosphorylated proteins, DIGE saturation labeling, and MALDI-TOF/TOF mass spectrometric identification of proteins, to quantitatively compare profiles of phosphoproteins of PC12 cell after GDNF challenge. We have identified 92 phosphoproteins, which displayed significant changes in relative abundance after GDNF exposure, as candidates for targets of phosphorylation in response to GDNF treatment in PC12 cells. Furthermore, we indubitably provided evidence that Hsp27 was a novel component of the GDNF signaling pathways and played a key role in the GDNF-induced differentiation of PC12 cells.
Materials and Methods Reagents. Tissue culture media and supplements were obtained from Gibco BRL (Grand Island, NY). CyDye DIGE fluor saturation dyes Cy3 and Cy5, Immobilized pH gradient (IPG) strips, Pharmalytes, DeStreak reagent and rehydration solution were purchased from GE Healthcare (Amersham, Bucks, U.K.). Sequencing-grade trypsin was obtained from Promega (Madison, WI). All chemical reagents were obtained from SigmaAldrich (St. Louis, MO) unless otherwise specified. The antiHsp27 antibody was from Stressgen (San Diego, CA). The phospho-Hsp27 (ser 82) monoclonal antibody was from Epitomics (Burlingame, CA). Monoclonal antibody against β-actin was from Sigma-Aldrich (St. Louis, MO). Horseradish peroxidase (HRP)-conjugated secondary antibodies were from Jackson (West Grove, PA). Fluorescent labeled secondary antibodies were from Molecular Probes (Eugene, OR). OptiMEM and LIPOFECTAMINE 2000 was from Invitrogen (Rockville, MD). GDNF human recombinant was from ProSpec-Tany Technogene Ltd. (Israel). PrimeScript first strand cDNA Synthesis Kit
research articles and PCR Enzymes (LaTaq and PrimeSTAR HS DNA Polymerase) were from TakaRa (Japan). Cell Culture. Previously established stably transfected PC12GFRR1-RET cells which were the generous gift from Prof. Cheng He’s laboratory were grown in DMEM supplemented with 5% FBS and 5% heat-inactivated horse serum.24 Hygromycin and G418 were needed to maintain the extraneous protein expression in PC12-GFRR1-RET. The cultures were maintained at 37 °C in a humidified atmosphere of 5% CO2/95% air, and the medium was changed every 2 days. Cells were plated onto polyL-lysine-precoated 24-well plates or 100-mm dishes (cell density for both, 1.5 × 104/cm2) 24 h before treatment. Differentiation Assay. Differentiation assay of PC12 cells was performed as described before.25 Briefly, PC12 cells were stimulated with or without GDNF for the indicated time. For some experiments, pEGFP-Hsp27 series, siRNA of Hsp27, or scramble siRNA was added 24 h (in case of pEGFP-Hsp27 series) or 72 h (in case of siRNA) before GDNF addition. Cells with longer neurites than the cell radius were counted as positive.26 The percentage of neurite-positive cells was measured under an inverted microscope. Cells were counted in six fields throughout the entire culture dish. At least 400 cells were counted per sample. Experiments were repeated at least three times. Protein Preparation and Phosphoprotein Enrichment. Phosphoprotein enrichment was carried out using Qiagen columns according to the manufacturer’s instructions with minor modifications. The protein concentration was determined using the Bio-Rad RC DC protein assay kit (Bio-Rad, Hercules, CA). Saturation Labeling for DIGE. Labeling optimization was carried out as recommended in the manufacturer’s protocol in order to determine the optimal concentrations of TCEP (Tris [2 carboxyethyl] phosphine hydrochloride) (Molecular Probes, Eugene, OR) and dye required for this sample. The amounts of 2 nmol TCEP and 4 nmol dye were determined to be optimal for this sample (data not shown). From each sample, 5 µg of protein was reduced with 2 nmol TCEP for 1 h at 37 °C. Subsequently, 4 nmol Cy5 saturation dye freshly dissolved in anhydrous Dimethylformamide (DMF) was added and incubated at 37 °C for a further 30 min. The labeling reaction was quenched by the addition of an equal volume of 2× lysis buffer (7 M urea, 2 M thiourea, 4% CHAPS, 2 mg/mL DTT, 2% (v/v) Pharmalytes pH 3-10). For the internal standard, equal aliquots (5 µg) of each sample were pooled and labeled with Cy3 saturation dye. Prior to electrophoretic separation, 5 µg of the Cy3-labeled pooled internal standard was added to each of the Cy5-labeled 5 µg analytical samples. Protein Separation and Analysis. IPG strips, pH 4-7 (linear) 24 cm, were rehydrated in De-Streak rehydration solution containing 1% (v/v) Pharmalytes overnight. Isoelectric focusing (IEF) was performed using the Ettan IPGphor IEF system (GE Healthcare) and following a separation protocol of five phases of graduated voltages from 300 to 8000 V with a total focusing time of 72 000 Vh. Strips were equilibrated in equilibration buffer (6 M urea, 100 mM Tris, pH 8.0, 30% glycerol, 2% SDS, 30 mM DTT) for 15 min prior to second-dimension SDS-PAGE. Strips were then immediately applied to 12.5% SDS polyacrylamide gels. A 0.5% agarose overlay containing 0.001% bromophenol blue was applied and second-dimension separation was performed at 4 W per gel overnight until the blue dye front migrated off the bottom of the gel. Journal of Proteome Research • Vol. 8, No. 6, 2009 2769
research articles Gels were scanned using a FLA-5100 imaging system (FuJiFilm, Japan). For the Cy3 image, the 532-nm excitation laser and LPG filter were used, and for Cy5, the 635-nm excitation laser and LPR filter were used. The images were imported into DeCyder differential analysis software v5.0 (GE Healthcare) for analysis. The differential in-gel analysis module was used to pair the Cy5 analytical image to its Cy3 internal standard, to assign spot boundaries and to calculate parameters such as normalized spot volumes or protein abundance. The gel with the greatest number of spots was automatically assigned as the master gel. All images were matched to the standards and the master image, and statistical analyses were then performed. The Student’s t test was performed for every matched spot set comparing the average and SD of protein abundance for a given spot between the control and GDNF treated groups. The results were then filtered to include only those proteins in which the abundance differed significantly between the two groups on the basis of a nominal Bonferroni-type correction based on a spot map of approximately 2500 proteins, with p e 0.05 being considered significant. Preparative Gels, In-Gel Digestion, and Mass Spectrometry (MS). To obtain sufficient material for MALDI-TOF identification of the proteins of interest, preparative gels were run with 450 µg of protein from pooled analytical samples. In total, 150 µg of protein each from control, 0.5 h GDNF group, and 10 h GDNF group was pooled and reduced with 200 nmol TCEP for 1 h at 37 °C. Subsequently, 400 nmol Cy3 saturation dye in anhydrous DMF was added and labeling was carried out for 30 min at 37 °C. Following first- and second-dimension separation, silver poststaining was performed after gel scanning using a MS-compatible protocol as described elsewhere.27 The protein spots of interest were cut out of 2-D gels using Gelpix Spot-Excision Robot (Genetix, Hampshire, U.K.) and transferred into a 96-well plate and incubated in silver destaining solution (equal volume of 30 mM potassium ferricyanide and 100 mM sodium thiosulfate) at room temperature, followed by washing with Milli-Q water, 25 mM ammonium bicarbonate/50% acetonitrile (ACN), and 100% ACN. When the gel pieces were dried in a vacuum, the proteins were digested overnight in 10 µL of trypsin (4 ng/µL, mass spectrometry grade, Promega, Madison, WI) in 25 mM ammonium bicarbonate at 30 °C. The reaction was terminated with 2 µL of 1% of TFA, and the peptide fragments were enriched and desalted with ZipTip pipet (Millipore, Billerica, BA) tips according to the manufacturer. Tryptic peptides were lyophilized and resuspended in 1 µL of matrix solution containing 10 mg/mL R-cyano-4-hydroxycinnamin acid prepared in 50% ACN/0.1% TFA. The samples were spotted onto the sample target plate. Peptide mass spectra were obtained on a matrix-assisted laser-desorption/ionization time-of-flight/time-of-flight (MALDI-TOF-TOF) mass spectrometer (4700 Proteomics Analyzer, Applied Biosystem, Foster City, CA). The peptide mass fingerprints were obtained in the mass range between 800 and 4000 Da with ca. 5000 laser shots. Trypsin autolytic peaks were used for internal calibration of the mass spectra. Up to 5 of the most intense peaks, excluding the known background peaks or keratin peaks, were selected for subsequent MS/MS data acquisition. Collision induced energy issued from atmosphere was adjusted to 5 × 10-7 Torr for MS/MS spectra acquisition. Protein identification was processed and analyzed by searching the Swiss-Prot protein database using the MASCOT search engine of Matrix Science that integrated in the Global Protein Server (GPS) Workstation. The mass tolerance, the most 2770
Journal of Proteome Research • Vol. 8, No. 6, 2009
Hong et al. important parameter, was limited to 50 ppm. The results from both the MS and MS/MS spectra were accepted as a good identification when the GPS score confidence was higher than 95%. Data Sorting. To determine protein functional relationships within and between each data set, we used the Database for Annotation, Visualization and Integrated Discovery (DAVID) (http://david.abcc.ncifcrf.gov/). DAVID is a Web-based, client/ server application that allows users to access a relational database of functional annotation derived primarily from LocusLink at National Center of Biotechnology Information (NCBI). Through DAVID, proteins identified with a LocusLink number can be grouped by Gene Ontologies and annotated with gene names and aliases as well as functional summaries. Database, Literature Searching, and Prediction of Phosphoproteins. To increase confidence in the assignment of phosphorylated proteins and reduce the risk of contamination from unphosphorylated proteins during phosphoprotein purification, all identified proteins were checked by Phospho.ELM (version 7.0), a database of experimentally verified phosphorylation sites in eukaryotic proteins. We also conducted PubMed literature searches on each protein to determine whether some proteins were reported to be phosphorylated. For further investigation of these identified proteins, and phosphorylation sites of which were not reported previously, the proteins were checked to meet the criteria set by the Eukaryotic Linear Motif resource for Functional Sites in Proteins (ELM) server, a resource for predicting functional sites in eukaryotic proteins. Western Blotting. A total of 20 µg of proteins from each sample was loaded onto 12% sodium dodecyl sulfate polyacrylamide gels (SDS-PAGE) for electrophoresis as described previously in detail.28 Following separation, the proteins were transferred to PVDF membranes (Bio-Rad Laboratories, Hercules, CA), and probed overnight at 4 °C with primary antibodies (1:2000 for Hsp27, 1:1000 for phospho-Hsp27 (Ser82), 1:10 000 for β-actin in bovine serum album). After washing with TBS-T (0.1% Tween 20 in TBS), HRP-conjugated secondary antibodies were added for 2 h at room temperature, and detections were carried out by ECL advanced Western Blotting Detection kit (Amersham, U.K.). Band intensities were quantified by densitometric analyses using an AxioCam digital camera (ZEISS, Germany) and the KS400 photo analysis system (Version 3.0). Immunocytochemistry. Immunostaining was performed as described previously.28,29 The 4% paraformaldehyde-fixed cells were blocked and incubated overnight at 4 °C with primary antibodies diluted in antibody diluents (anti-Hsp27, 1:100), followed by fluorescent labeled secondary antibodies, that is, goat anti-rabbit (Alexa Fluor 594, Molecular Probes, Eugene, OR). Images were recorded using a laser scanning confocal microscope (Carl Zeiss, LSM510). Semiquantitative RT-PCR. The first strand of cDNA was prepared with 5 µg of RNA using the PrimeScript first strand cDNA Synthesis Kit (TakaRa, Japan). In the preliminary experiments, the relative amounts of cDNA and the range of PCR cycles that permitted the linear amplification of Hsp27 and β-actin were determined. The primers were Hsp27 sense, 5′CGCGTGCCCTTCTCGCTACTG-3′; antisense, 5′-AATTTGGGCACGGGCCTCGAAAG-3′; and β-actin sense, 5′-CTGGGACGATATGGAGAAGATTTG-3′; antisense 5′-GACAGTGAGGCCAGGATAGAGC-3′. The PCR conditions for rat Hsp27 and β-actin were 95 °C for 30 s, 50 °C for 30 s, followed by 72 °C for 45 s. The number of cycles for Hsp27 was 30, and for β-actin
Hsp27 Involved in GDNF-Induced Neurite Outgrowth
research articles
Figure 1. The ability of GDNF to induce differentiation of PC12-GFRR1-RET cells. (A) Cells were stimulated with 100 ng/mL GDNF for different times (a, 0 h; b, 10 h; c, 24 h; d, 72 h) and photographed using a phase-contrast light microscope. (B) Dose-dependent response in neurite outgrowth of PC12-GFRa1 cells to GDNF at 72 h. (C) Time-response curve for neurite outgrowth of PC12-GFRR1-RET cells to 100 ng/mL GDNF. Scale bar in (A) represents 100 µm.
25. With the use oof the NIH image (version 6; NIH, Bethesda, MD), each band was surveyed for its relative band intensity. Within the range of the linear PCR amplification, the relative expression of the Hsp27 message was evaluated by calculating the band intensity ratio of Hsp27/β-actin. Plasmid Constructs. Full-length human wild-type Hsp27 (Hsp27-WT), nonphosphorylation mutant (Hsp27-Ala), and a phosphorylation mutant form of human Hsp27 (Hsp27-Asp) were generous gifts from Prof. Kyungjin Kim (the School of Biological Sciences, Seoul National University, Seoul, Korea). The Hsp27-Ala contains a form of human Hsp27 wherein the three serines known to be phosphorylated (Ser15, Ser78, and Ser82) have been mutated to alanine residues; the Hsp27-Asp contains serine to Aspartic acid mutations at Ser15, Ser78, and Ser82. To construct the GFP infusion protein Hsp27 pEGFPC2-Hsp27 series, we performed PCR and subcloned full-length
Hsp27 series into pEGFP-C2 vector from Clontech. All of the constructs were fully sequenced before used for transformation or transfection. Inhibition of Hsp27 Expression by RNA Interference. For RNA interference experiments, the pZ-OFF EGFP vector was the generous gift from Dr. Craig Garner’s laboratory. The following oligos for the small interfering RNA (siRNA) constructs (Genscript, Piscataway, NJ) were cloned into the pZOFF EGFP vector using BglII and SalI sites. Hsp27 siRNA, TTAACTGTGAGCTCCTCAGGA; scrambled siRNA, CGGCTTGACCATCATGAGCTA. Transfection. For transfections, the PC12 cells were seeded at a density of 1.5 × 104 cells/well in a 12-well plate. The expression vectors of pEGFP-C2 were transfected at culture days 3-5 using OptiMEM and LIPOFECTAMINE 2000 according to manufacturer’s instructions (Invitrogen, Rockville, MD). Journal of Proteome Research • Vol. 8, No. 6, 2009 2771
research articles
Hong et al.
The cultures were terminated for immunofluorescent microscopy analyses at various time points for assays. Statistical Analysis. Grouped data were expressed as a mean ( SEM. Changes between groups were analyzed by Student t test or ANOVA using Origin 6.1. Repeated measures were performed at least three times in all experiments. p < 0.05 was accepted as significant.
Results Glial Cell Line-Derived Neurotrophic Factor Induces Differentiation of PC12-GFRr1-RET Cells. Since there is no endogenous GFRR1 and little RET in wild-type PC12 cells, we used a previously established PC12-GFRR1-RET cell line,30 PC12 cells that stably express GDNF receptors GFRR1 and RET. GDNF could induce the neurite outgrowth, which is one of the indicators for neuronal differentiation, and promote the survival of PC12-GFRR1-RET cells at low concentrations.30 The previously reported ability of these cells to respond to GDNF implies that they express additional downstream components of the signaling pathway(s). To establish conditions for functional proteomics analysis, we first tested the ability of recombinant GDNF to induce differentiation of PC12-GFRR1-RET cells. As shown in Figure 1A, by 72 h of culture, GDNF had produced a clear increase in numbers of differentiated PC12-GFRR1-RET cells. GDNF promoted the neurite outgrowth of PC12-GFRR1-RET cells in a dose-dependent manner over the concentration range from 0.01 to 1000 ng/mL with an EC50 of approximately 4.7 ng/mL (Figure 1B). In addition, the time-response curve for the cells exposed to 100 ng/mL GDNF is shown in Figure 1C. The differentiated PC12-GFRR1-RET cells were increased with an increased time of GDNF bath application. With 100 ng/mL GDNF less than 12 h, there was not any morphological change in the cells. Up to 72 h of culture with GDNF, nearly 82 ( 2.5% cells differentiated. We assume that the proteins in response to GDNF in the early stage are more involved in the signal transduction, and therefore, 0.5 and 10 h of 100 ng/mL GDNF treatment (in comparison with untreated cells) are taken as time points for the proteomics assay to detect GDNF-specific targets. Proteomic Analysis of Phosphorylated Proteins in Early Stage of PC12-GFRr1-RET Cells Treated with GDNF. Known that GDNF can induce the differentiation of PC12-GFRR1-RET cells in the time-dependent pattern, we went ahead to investigate the factors that might be involved in the early response of GDNF before differentiation of PC12 cells. To accomplish this goal, we combined three key methodologies, enrichment of phosphorylated proteins, DIGE saturation labeling, and mass spectrometric identification of proteins (Figure 2). In the preliminary experiments, we first used a wide pH interval of 3-10 for IPG strips to give an overview of phosphorylated protein-enriched fractions from the three lysates (control, 100 ng/mL GDNF for 0.5 h and for 10 h groups) distributions. The results showed that most protein spots were found to be located in the middle part of the gel (data not shown). Thus, IPG strips with narrower pH range of 4-7 were used to achieve a better separation of these proteins. Phosphoprotein-enriched fractions from the three lysates were labeled with cyanine dye Cy5, respectively, and the pooled internal standard was labeled with Cy3. The Cy3 and Cy5 labeled samples were combined and run on the same gel. The experimental design allowed direct comparison between GDNF treated 0, 0.5 or 10 h groups and the pooled internal standard. 2772
Journal of Proteome Research • Vol. 8, No. 6, 2009
Figure 2. A flowchart of proteomic analysis.
This standard, being a pool of equal amounts of all samples within the experiment, should be representative of every protein present and is the same across all gels. The standard provides an average image against which all other gel images are normalized, thus, removing much of the experimental variation and reducing gel-to-gel variation. In total, 2117 protein spots were present on the master gel as determined by the DeCyder Differential Analysis Software (see Figures 3 and 4). In Figure 3A,B, a representative 2-DE gel of phosphoproteome of GDNF 0.5 h treated PC12 cells (Cy5, blue pseudocolor) versus internal standard (Cy3, red pseudocolor) is shown. Some evident differences were immediately detectable from superimposition of images (Figure 3C,D) of the 2-DE maps of internal standard (Figure 3A) and treated (Figure 3B) samples labeled with Cy3 and Cy5. The blue spots visible in Figure 3D indicating the presence of phosphorylated proteins were readily distinguished from the pooled internal standard, which also means relative abundance of the proteins in the three groups. Specifically, we investigated those phosphorylated proteins that were present either in higher or lower amounts after 0.5 or 10 h of GDNF treatment. Combining three independent experiments together, 125 differentially expressed protein spots (p < 0.05) with at least a 1.3-fold up- or down-regulation were evaluated by DeCyder software in GDNF treated 0.5 and 10 h phosphoproteome (Figure 4). Spots that were up- or downregulated >1.3-fold were subjected to protein identification by mass spectrometry analysis as described under Materials and Methods. Among them, 92 spots were identified and listed in Table 1. To provide some evidence about the phosphorylation
Hsp27 Involved in GDNF-Induced Neurite Outgrowth
research articles
Figure 3. Representative gel images. Example showing the Cy3-labeled standard gel image (A) and gel image of Cy5-labeled GDNF 0.5 h sample (B). Cy3 or Cy5 signals are exhibited as red and blue pseudocolors, and a merge of the two images is shown in (C). (D) A magnified view of the boxed region in (C) showing candidate spots. (E) The 3-D plots output from DeCyder showing increased abundance in this sample (b) compared to the internal standard (a). Mr, molecular weight.
Figure 4. Results of gel image analysis using DeCyder DIA module. Spot differences between untreated control group and GDNF treated 2-D gels are shown. (A) Green circles highlight spots that detected by DeCyder DIA module. (B) Green circles highlight spots showing increase or decrease at least by 1.3-fold in abundance after GDNF treatment.
of the proteins, phosphorylation sites were checked in Phospho.ELM, a database of experimentally verified phosphorylation sites in eukaryotic proteins. As the database is currently incomprehensive, literature searches through the NCBI PubMed database were also performed to determine the modification of phosphorylation of these proteins. We listed references to some proteins phosphorylated into one column of Table 1. As expected, most proteins have been recognized as phosphorylated protein in the database and literature. Simultaneously, the proteins, phosphorylation of which is not included in the Phospho.ELM database, were further validated for the presence of phosphorylation sites or motifs using the ELM server, a resource for predicting functional sites in eukaryotic proteins. All the proteins identified in Table 1 were successfully predicted
according to their respective phosphorylation sites, except for one unknown error. Expression Patterns after GDNF Exposure. To identify coordinately regulated proteins at the different time points of GDNF exposure, a hierarchical clustering method was applied. From the cluster analysis of the differentially expressed protein spots, more proteins increased and fewer proteins decreased after GDNF treatments (Figure 5). For 75 protein spots, a higher protein expression (positive expression ratio > 1.3-fold) was detected in both of the GDNF proteome (58 protein spots were identified, which represent 47 genes), whereas 20 protein spots showed a higher abundance in normal control group (negative expression ratio > 1.3-fold) (14 spots, which represent 13 genes, were identified). In addition, 19 protein spots (13 spots which Journal of Proteome Research • Vol. 8, No. 6, 2009 2773
research articles
Hong et al. a
Table 1. Total Proteins Identified by MALDI-TOF-TOF
no.
1 2 3 4 5 6 7 8 9 10 11 12 13
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
2774
protein name
DNA repair protein RAD50 T-complex protein 1 subunit epsilon Galectin-7 Fructose-bisphosphate aldolase A HSP27 Heterogeneous nuclear ribonucleoprotein K Calreticulin precursor Peptidyl-prolyl cis-trans isomerase A Plectin-1 (PLTN) (PCN) HSP27 Tropomyosin beta chain Structural maintenance of chromosomes protein 1A 1-phosphatidylinositol4,5-bisphosphate phosphodiesterase gamma 2 F-actin capping protein subunit alpha-1 Tyrosine 3-monooxygenase ATP-dependent RNA helicase DDX39 Heat shock protein HSP 90-beta 28S ribosomal protein S7, mitochondrial precursor Tyrosine 3-monooxygenase Tubulin beta-2 chain ATPase family AAA domain-containing protein 1 BAG family molecular chaperone regulator 5 Pyp protein Translationally controlled tumor protein Fatty acidbinding protein, heart (H-FABP) Tubulin beta-5 chain Tropomyosin beta chain Stress-70 protein, mitochondrial precursor Tubulin alpha-6 chain Tubulin alpha-1 chain Synaptonemal complex protein 1 Prohibitin Glyceraldehyde3-phosphate dehydrogenase Dihydropyrimidinase-related protein 2 Alpha Crystallin B chain Annexin A2 COP9 signalosome complex subunit 4 Dihydropyrimidinase-related protein 2
Uniprot accession no.
0.5 h/control
10 h/control
protein score
Phospho.ELM database or ELM server
protein score C.I.%
Proteins Up-Regulated by Both 0.5 and 10 h GDNF Treatments Q9JIL8 2.146 2.1545 89 100 S635, S640, T690 (H)
Pubmed literature
58-60
Q68FQ0
1.0476
2.1526
59
99.343
P97590 P05065
2.049 2.0006
1.9713 1.9658
268 73
100 99.973
S6, S12 (H) S46, T65 (H), S36 (M)
61 62
P42930 P61980
1.1107 1.3426
1.9014 1.8541
232 92
100 100
S15, S78, S82 etc. (H) S284 (R)
56, 63 64
P18418
1.5608
1.8494
61
Yes
65
P62937
1.8272
1.7546
174
100
T157 (H)
66
P30427 P42930 P58775
3.13 1.4446 1.4039
1.7289 1.6852 1.6821
48 99 69
91.336 100 99.925
S125, S149. S720 etc. (H) S15, S78, S82 etc. (H) S283 (O)
60, 66 56, 63 67
Q9Z1M9
1.078
1.5978
66
99.863
S360, S957, S960 etc. (H)
58, 68
P24135
1.4579
1.5798
68
99.913
Y753, Y759, Y1197 etc. (R)
69-71
Q3T1K5
1.57
1.5377
302
100
Yes
P04177 Q5U216
1.0451 1.001
1.531 1.5237
62 118
99.678 100
S8, S19, S31 etc. (R) Yes
72, 73
P34058
1.0684
1.5205
147
100
S255 (R)
60, 74
Q5I0K8
1.5179
1.5195
57
98.958
Yes
P04177
1.2181
1.4867
67
99.891
S8, S19, S31 etc. (R)
72, 73
Q6P9T8
1.9514
1.4727
83
99.997
Y36 (H), S172, S382 (M)
75, 76
Q505J9
1.6205
1.4687
50
94.006
S322 (M)
77
Q5QJC9
1.4156
1.4626
55
98.349
S20 (H)
44
Q499R7 P63029
1.7107 1.4626
1.4319 1.4079
387 107
Yes S46, S64 (H)
78
P07483
1.6336
1.3793
81
Y20 (R)
79
P69897
2.0014
1.3775
182
P58775
1.4368
1.3697
51
P48721
1.8826
1.362
154
Q6AYZ1
1.6497
1.3515
65
99.815
Yes
P68370 Q03410
1.7191 2.3109
1.3472 1.3272
72 55
99.968 98.231
Yes Yes
P67779 P04797
1.6228 1.5167
1.3262 1.3262
153 181
100 100
S252, S254 (H) S83 (H)
60 60
P47942
1.4772
1.314
150
100
T509, T515, S518 etc. (H)
81, 82
P23928
1.4504
1.2705
197
100
S19, S21, S43 etc. (H)
83
Q07936 Q68FS2
1.4701 1.66
1.2689 1.26
197 57
100 98.831
T3, S12, S18 etc. (H) T459, S463 (M)
60, 84 77
P47942
1.93
1.2473
114
100
S465, T509, T514 (M)
62
Journal of Proteome Research • Vol. 8, No. 6, 2009
99.513
100 100 99.995 100 95.239 100
Yes
Yes S283 (O)
80
S650 (M)
77
research articles
Hsp27 Involved in GDNF-Induced Neurite Outgrowth Table 1a. Continued
no.
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
59 60 61 62 63 64 65 66 67 68 69 70 71
72 73 74 75 76 77 78
protein name
Plectin-1 Alpha-enolase Protein Wnt-4 precursor 3′ histone mRNA exonuclease 1 Annexin A1 Heterogeneous nuclear ribonucleoprotein K Nucleolar GTP-binding protein 1 Actin, cytoplasmic 2 Protein disulfideisomerase A6 precursor Heat shock cognate 71 kDa protein Tubulin alpha-6 chain Triosephosphate isomerase HSP27 Tubulin alpha-6 chain DNA repair protein RAD50 Eukaryotic translation initiation factor 3 subunit 2 NSFL1 cofactor p47 Hexokinase-2 Actin, cytoplasmic 2 Pleckstrin homology-like domain family A member 3
Uniprot accession no.
0.5 h/control
10 h/control
protein score
Phospho.ELM database or ELM server
protein score C.I.%
Proteins Up-Regulated by Both 0.5 and 10 h GDNF Treatments P30427 3.3764 1.2415 79 99.992 P04764 3.3803 1.1772 110 100 Q9QXQ5 1.4924 1.1391 53 97.384
S135, S149, S720 etc. (H) Y25, Y44 (M) Yes
Q5FVR4
1.4175
1.1374
63
99.726
P07150 P61980
1.9123 1.7782
1.1334 1.1182
218 75
100 99.981
Q99P77
1.5115
1.1168
53
97.384
P63259 Q63081
1.5078 1.6268
1.1045 1.1042
129 344
P63018
1.5184
1.104
57
Q6AYZ1
2.2981
1.0981
51
P48500
1.441
1.068
144
100
S21 (H)
P42930 Q6AYZ1
1.5419 1.7166
1.04 1.0198
321 122
100 100
S15, S78, S82 etc. (H) Yes
Q9JIL8
1.3481
1.0144
67
Q5XIG8
1.4589
1.0122
187
O35987 P27881 P63259 Q5PQT7
1.4186 2.3789 1.4444 1.3825
1.011 1.0092 1.0053 1.0025
285 62 88 50
58, 60 75
Yes Y21 (R), S27 (M) S284 (R)
85 86
S468, S470, S472 (H)
60
Y166, Y218 (M) S428
75 58
98.909
Y15 (M), T477 (H)
58, 75
95.557
Yes
100 100
99.888
Proteins Up-Regulated by 0.5 h GDNF Treatment, but Down-Regulated GABA-B receptorQ30A01 8.5491 -1.007 79 interacting scaffolding protein RNA-binding Q3B7D9 1.3856 -1.044 48 protein C2orf38 homologue Spliceosome Q63413 5.0001 -1.048 153 RNA helicase bat1 Glutathione S-transferase P P04906 1.4399 -1.098 302 Tubulin beta-2C Q6P9T8 1.8417 -1.139 102 chain PhosphatidylinositolQ9Z1L0 1.7376 -1.169 83 4,5-bisphosphate 3-kinase catalytic Vascular endothelial O35757 3.0511 -1.178 83 growth factor C precursor Heterogeneous nuclear P61980 4.0544 -1.194 134 ribonucleoprotein K Tubulin beta-5 chain P69897 1.3316 -1.317 126 Fructose-bisphosphate P05065 1.4554 -1.525 138 aldolase A Synaptonemal complex Q03410 1.0703 -1.61 67 protein 1 Methyl-CpG-binding Q00566 1.4514 -1.733 80 protein 2 ATP synthase subunit beta, P10719 1.3202 -1.914 307 mitochondrial precursor
Pubmed literature
60
S635, S640, T690 (H)
58-60
100
Y445 (M)
75
100 99.613 99.999 94.898
S114 (R) Yes Y166, Y218 (M) Yes
86 75
by 10 h GDNF Treatment 99.992 Yes 91.915
Yes
100
Yes
100 100
Y33, T34 (M) Yes
77
99.993
Y772, S1070 (H)
87, 88
99.991
Yes
100
S284 (R)
86
100 100
Yes S36 (M), S46, T65 (H)
60, 62
99.893
Yes
99.994
S80, S166 (H)
100
58, 60
Yes
Proteins Down-Regulated by 0.5 h GDNF Treatment, but Up-Regulated by 10 h GDNF Treatment Dynamin-1 P21575 -1.255 1.8649 53 97.384 S774, S778, S859 etc. (H) HSP27 P42930 -1.2444 1.818 149 100 S15, S78, S82 etc. (H) Hemoglobin subunit P11517 -1.1021 1.7653 83 99.997 Y42, S45 (M) beta-2 Cystatin-B P01041 -1.1651 1.5053 218 100 Yes Protein S100-A8 P50115 -1.1758 1.3851 116 100 S103 (H) 40S ribosomal P29314 -1.3859 1.154 64 99.744 Yes protein S9 14-3-3 protein epsilon P62260 -1.8466 1.067 208 100 ELM server - error
58, 75 77
58
Journal of Proteome Research • Vol. 8, No. 6, 2009 2775
research articles
Hong et al.
Table 1b. Continued
no.
79 80 81 82 83 84 85 86 87 88 89 90 91 92
protein name
Peroxisomal bifunctional enzyme 14-3-3 protein zeta/delta Secretogranin-1 precursor Regulator of G-protein signaling 8 Peptidyl-prolyl cis-trans isomerase A Heat shock protein HSP 90-beta DNA repair protein RAD50 Rab GTPase-binding effector protein 1 Heat shock 70 kDa protein 4 Peroxiredoxin-2 Heat shock 70 kDa protein 4 40S ribosomal protein SA Myelin transcription factor 1-like protein Hemoglobin subunit alpha
Uniprot accession no.
0.5 h/control
10 h/control
protein score
Proteins Down-Regulated by Both 0.5 and 10 h GDNF Treatments P07896 -1.7037 -1.001 51 95.557 P63102
-1.4656
-1.061
140
O35314 P49804
-1.529 -1.5873
-1.065 -1.085
71 66
P10111
-1.5105
-1.091
109
P34058
-1.3068
-1.21
189
Q9JIL8
-1.2258
-1.363
53
O35550
-1.5091
-1.391
62
O88600
-1.6792
-1.443
95
P35704 O88600
-1.3329 -1.3655
-1.511 -1.612
P38983 P70475
-1.1143 -1.5896
P01946
-1.8776
Phospho.ELM database or ELM server
protein score C.I.%
100
Pubmed literature
Yes S58 (H)
89
S149 (H) S168 (H)
90 91
100
T157 (H)
66
100
S255 (R)
86
97.323
S635, S640, T690 (H)
58
99.622
S410 (H, M)
60
100
Y89, Y336 (H)
75
281 92
100 100
Yes Y89, Y336 (H)
75
-1.695 -1.733
126 62
100 99.613
Y139 (H, M) Yes
-1.844
294
100
Yes
99.958 99.869
75
a Identified spots are listed together with their protein names, accession numbers, MS/MS scores and Protein score C.I.%. GDNF stimulation of cells for 0.5 and 10 h caused up- (positive ratios) or down-regulation (negative ratios) of proteins. The results were pooled from three independent experiments; the cut-off p-value was
