Biochimica et Biophysica Acta 1784 (2008) 983–985
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b b a p a p
The proteome of the human neuroblastoma cell line SH-SY5Y: An enlarged proteome Kambiz Gilany a, Roos Van Elzen a, Kim Mous b, Edmond Coen b, Walter Van Dongen c, Stefaan Vandamme c, Kris Gevaert d,e, Evy Timmerman d,e, Joël Vandekerckhove d,e, Sylvia Dewilde a, Xaveer Van Ostade b, Luc Moens a,⁎ a
Department of Biomedical Sciences, Protein Chemistery, University of Antwerp, Universiteitsplein 1, B-2610 Antwerp, Belgium Department of Biomedical Sciences, Functional Proteomics, University of Antwerp, Universiteitsplein 1, B-2610 Antwerp, Belgium Department of Chemistry, Nucleoside Research and Mass Spectrometry Unit University of Antwerp, B-2610 Antwerp, Belgium d Department of Medical Protein Research, VIB, B-9000 Ghent, Belgium e Department of Biochemistry, Ghent University, B-9000 Ghent, Belgium b c
a r t i c l e
i n f o
Article history: Received 10 September 2007 Received in revised form 7 February 2008 Accepted 5 March 2008 Available online 19 March 2008 Keywords: Proteome analysis N-terminal COFRADIC 2D-LC-MALDI-TOF/TOF-MS SDS-PAGE nano-LC–MS/MS 2D-PAGE
a b s t r a c t The human neuroblastoma cell line SH-SY5Y (ATCC: CRL-2266) is widely used as a neural cellular model system. The hitherto existing proteome data (115 proteins) are here extended. A total of 1103 unique proteins of this cell line were identiﬁed using 2D-LC combined with MALDI-TOF/TOF-MS, SDS-PAGE with nano-LC–MS/ MS, N-terminal COFRADIC analysis with nano-LC–MS/MS and 2D-PAGE with MALDI-TOF/TOF-MS peptide mass ﬁngerprinting. The obtained proteome proﬁle of this cell line is discussed. © 2008 Elsevier B.V. All rights reserved.
The human neuroblastoma cell line SH-SY5Y resembles immature sympathetic neurons/neuroblasts in cell culture [1,2]. The cells are typically locked in an early neuronal differentiation stage, characterized by low levels of neuronal markers. Upon exposure to appropriate growth conditions, SH-SY5Y cells can be driven towards different mature, neuronal (noradrenergic or cholinergic) phenotypes. Due to these multipotent characteristics, the SH-SY5Y cell system is a common model to study the stem cell character of neuroblastoma tumorigenesis, as well as growth and differentiation processes . Additionally, SH-SY5Y cells are characterized by prominent sensitivity to oxidative stress, which may result from a relative low content of certain antioxidants and an increased radical formation associated with dopamine synthesis. Exposure to oxidative stress is well known as a pivotal contributor to acute damage caused by cerebral ischemic-reperfusion and many progressive neurodegenerative diseases such as Parkinson's disease, Alzheimer's disease and amyotrophic lateral sclerosis. The neuronal properties and pronounced sensitivity to oxidative stress made the neuroblastoma cell system an excellent model to study a number of neurological pathologies on the molecular, morphological and physiological level . A search in literature shows 2347 hits (Jan
⁎ Corresponding author. Tel.: +32 3 8202323; fax: +32 3 8202248. E-mail address: [email protected]
(L. Moens). 1570-9639/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.bbapap.2008.03.003
2008) on the human neuroblastoma cell line SH-SY5Y. Surprisingly, no attempts have been done to undertake comprehensive analysis of its proteome. We have used this cell line in our hypoxia/re-oxygenation and cellular oxidative stress experiments . To support our study and those of others on the SH-SY5Y cell line, we here undertook the analysis of its proteome. Several gel-based and gel-free approaches exist for proteomics [6–10]. In this study, we combined four different approaches to obtain a higher proteome coverage of the SH-SY5Y cell line. This proteome database of the human neuroblastoma SH-SY5Y cell line is expected to be a powerful tool for neuroscientists. Table 1 comprises our four different approaches to protein or peptide separation and identiﬁcation: 2D-PAGE coupled to MALDI-TOF/ TOF-MS, SDS-PAGE coupled to nano-LC–ESI-MS/MS, 2D-LC coupled to MALDI-TOF/TOF-MS and N-terminal COFRADIC coupled to nano-LC– ESI-MS/MS (for experimental section and data, see supplementary material). The combination of these techniques resulted in the identiﬁcation of 1103 proteins, which, to our best knowledge, comprises the most complete proteome analysis of a neuronal cell type hitherto reported. In previous proteome studies on the SH-SY5Y cell line, 54 proteins were identiﬁed in a cytosolic fraction (http://www.proteome. jp/2D/XML/SH-SY5Y/shsy5y_menu.html) whereas an additional set of 61 proteins was identiﬁed in isolated mitochondria by 2DE . Our proteome analysis completely covers the two previous works. We also compared our results to a proteome analysis of the human neuroblastoma cell line SK-N-BE2 (621 unique proteins), which possesses
K. Gilany et al. / Biochimica et Biophysica Acta 1784 (2008) 983–985
Table 1 Experimental conditions Method
Mass Material Unique Identiﬁed spectrometry used peptide protein 2′
SDS-PAGE MALDI-TOF/ TOF 2 1-DE SDS-PAGE RP ESI 3 2D-LC SCX RP MALDI-TOF/ TOF 4 N-terminal RP RP ESI COFRADIC
100 µg 100 µg
243 (22%) overlapping proteins with the SH-SY5Y cell line. Additionally, we compared the overlap with a proteome analysis of the HEK293 cell line (970 unique proteins), which is morphologically similar to the SH-SY5Y cell line. These cell lines have a 215 (20%) proteins overlap. The overlapping identiﬁed proteins are (as expected) essential and high abundant proteins such as HSP 90-beta, ﬁlamin A, malate dehydrogenase and 60 kDa heat shock protein . The here identiﬁed proteins were functionally categorized based on Gene Ontology (GO) annotation terms  using the DAVID program package [14,15]. Amongst other, we found that 777 proteins were annotated with GO cellular component terms. The cellular components containing the greatest number of proteins in the SH-SY5Y data set include intracellular organelle (56.6%), intracellular membranebound organelle (42.6%), cytoplasm (42.6%), nucleus (28.3%) and intracellular non-membrane-bound organelle (20.8%) (GOTERM_CC: level 3). A similar analysis for molecular function was conducted; here 760 of the input proteins were classiﬁed. The major molecular functions include purine nucleotide binding (16.4%), RNA binding (12.1%), hydrolase activity (8.1%), unfolded protein binding (4.7%) and cytoskeletal protein binding (4%) (GOTERM_MF: level 3). In order to determine whether the cellular component and the molecular function mentioned above represent functions enriched in the SH-SY5Y data set relative to the “theoretical human proteome”, we used DAVID to calculate probable overrepresentations of protein classiﬁcations relative to the annotated human proteome and thus excluded common protein functions from skewing the functional analysis of the SH-SY5Y subcellular functions. DAVID furthers assigns a statistical signiﬁcance indicator to protein functions and classiﬁcations. The cellular component and molecular function showing the greatest enrichment are shown in Table 2. It seems that cytoplasmic proteins (P b 5.1E–52) are greatly enriched in our dataset. Furthermore, RNA binding (P b 1.8E–18), unfolded protein binding (P b 3.8E–18) and hydrolase activity (P b 4.8E–13) were showing the greatest enrichment. A catalogue of expressed proteins is useful for surveying the characteristics of SH-SY5Y cells and/or to search for candidates for molecular markers of this cell type. This study adds a very large set of proteins to the limited number described in two previous publications. Our study however also suggests that the total number of proteins expressed in the SH-SY5Y cell line is higher than the number of proteins identiﬁed thus far. Therefore, a complete description of the SH-SY5Y proteome still requires more extensive analyses coupled with for instance a sub-fractionation of cellular components (e.g. organelles) or by the application of techniques alternative to the ones used here. Although parallel application of four different methods on the same proteome enables comparison between these methods, it must be mentioned that many instruments and database searching parameters differed and, furthermore, that the amount of starting material is different for the various methods applied. However, we believe that some general conclusions can be drawn and that the combination of methods clearly gives a much better coverage of the SH-SY5Y proteome compared to existing data .
Table 2 Signiﬁcantly over-represented Gene Ontology Consortium (GO) cellular component and molecular function terms in the human neuroblastoma cell line SH-SY5Y Cellular component
Cytoplasma Intracellular organelle Intracellular non-membrane-bound organelle Ribonucleoprotein complex Cytosolic small ribosomal subunit Ribosome Proteasome complex Cytoskeleton Heterogenous nuclear ribonucleoprotein complex Intracellular membrane-bound organelle
42.6 56.6 20.8 10.6 1.2 6.2 1.9 10.4 1 42.6
5.1E–52 4.9E–24 2.4E–23 4.5E–23 4.9E–13 1.0E–10 2.4E–10 7.7E–9 3.5E–7 9.7E–6
Molecular function RNA binding Unfolded protein binding Hydrolase activity, actin on acid anhydrides Purine nucleotide binding Translation factor activity, nucleic acid binding Ligase activity forming carbon-oxygen bonds Translation initiation factor activity Cytoskeletal protein binding Protein domain speciﬁc binding ATP-dependent helicase activity
12.1 4.7 8.1 16.4 2.2 1.4 1.5 4 1.3 1.7
1.8E–39 3.8E–18 4.8E–13 1.6E–10 3.7E–6 5.8E–6 2.1E–5 4.9E–5 1.6E–4 4.5E–4
Tabulated are the top ten cellular components and the molecular functions with the greatest statistical signiﬁcance for enrichment in the SH-SY5Y proteome data set (GOTERM: level 3). The percentage is: involved proteins divided by total proteins multiplied by one hundred. The enrichment P-value (compared to the theoretical human proteome) is calculated based on EASE Score, a modiﬁed Fisher Exact Test. It ranges from 0 to 1. Fisher Exact P-value = 0 represent perfect enrichment. Usually the P-value must be equal to or smaller than 0.05 to be considered strongly enriched in the annotation categories. The closer the value is to zero, the more enriched is the category.
In the ﬁrst approaches, following colloidal Coomassie staining of 2DE separated protein spots, the ImageMaster software detected more than 1000 stained spots. Those spots with the highest intensity (here 400) were selected and in-gel digested with trypsin. Subsequently, analysis by MALDI-TOF/TOF-MS resulted in the identiﬁcation of 173 different proteins. For the second approach, the SH-SY5Y proteome was separated onto a 15% polyacrylamide gel (SDS-PAGE) and protein bands were visualized by Coomassie Brilliant Blue. This separated proteome was excised into ten equally sized gel slices, in-gel digested with trypsin and the tryptic peptides were subjected to LC–ESI-MS/MS analysis. Using this approach, 222 different proteins were identiﬁed. Thirdly, the proteome was immediately digested with trypsin and subjected to on-line 2D liquid chromatography (SCX combined with RP), coupled to automated MALDI-TOF/TOF-MS/MS system. Following database searching, 188 non-redundant proteins were identiﬁed. As a last approach, the N-terminal peptide COFRADIC technology was used in combination with a nano-LC–ESI-MS/MS leading to the identiﬁcation of 782 different proteins using 1083 identiﬁed peptide MS/MS spectra.
Fig. 1. Venn diagram of the number of unique identiﬁed proteins by the four methods used and their overlap.
K. Gilany et al. / Biochimica et Biophysica Acta 1784 (2008) 983–985
Combining all the identiﬁed proteins, we mapped 1103 unique proteins to the proteome of SH-SY5Y cells. Table 1 shows that SDS-PAGE coupled to the nano-LC–ESI-MS/MS enabled the identiﬁcation of a greater number of proteins in comparison with the 2D-PAGE analysis. This could indicate that the former samples in a relatively unbiased manner low-abundance proteins, high molecular weight proteins and proteins with extreme pIs which were not detected on the 2D-gel (see supplementary material). Fig. 1 shows that the overlap between the two gel-based methods is merely 48 proteins. N-terminal COFRADIC analysis had a better performance compared to 2D-LC coupled to MALDI-TOF/TOF-MS/MS (Table 1). It must be mentioned that the 2D LC-separated samples were collected on only four MALDI target plates (192 spots per target), whereas a previous study showed the necessity of a much larger set of spotted peptide samples and thus increased peptide segregation prior to proteome analysis by 2D-LC–MALDI-TOF/TOF-MS/MS . Clearly, reducing the sample's complexity by N-terminal COFRADIC leads to an increased coverage of the analysed proteome. However, the overlap of identiﬁed proteins between the gel-free methods is merely 73 proteins (Fig. 1). We believe that this may be due to the different ionization methods, to the amount of proteome starting material used and to general undersampling [17,18]. As expected, the overall number of proteins identiﬁed by the gelfree methods is much higher than that of the gel-based methods and the protein overlaps between all methods consist of just 21 proteins (Fig. 1). The relatively high number of proteins uniquely identiﬁed by a given method and, consequently, the low number of overlapping proteins suggest that saturation has not been reached and thus that more proteins from the SH- SY5Y proteome “await identiﬁcation”. Our ﬁndings therefore suggest that the information generated by each of the four methods used here is complementary and that the Nterminal COFRADIC method shows the best performance. It must further be mentioned that both 2DE-PAGE based protein identiﬁcation and the N-terminal COFRADIC were able to identify the isomer forms, which shows the strength of these methods. Acknowledgement SD and KM is respectively a post- and pre-doctoral researcher of the FWO (Fund for Scientiﬁc Research- Flanders). RVE is a beniﬁciant of a BOF-UA research grant. The lab in Ghent acknowledges the support of the EC grant “Interaction Proteome” within the 6th Framework Program. We thank Elisabeth Noergaard Nielsen for proof reading of the manuscript. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbapap.2008.03.003. References [19–26] can also be found there. References  J.L. Biedler, L. Helson, B.A. Spengler, Morphology and growth, tumorigenicity, and cytogenetics of human neuroblastoma cells in continuous culture, Cancer Res. 33 (11) (1973) 2643–2652.  W.G. Conroy, D.K. Berg, Neurons can maintain multiple classes of nicotinic acetylcholine receptors distinguished by different subunit compositions, J Biol Chem 270 (9) (1995) 4424–4431.
 A. Jogi, I. Ora, H. Nilsson, A. Lindeheim, Y. Makino, L. Pellinger, H. Axelson, S. Pahlman, Hypoxia alters gene expression in human neuroblastoma cells toward an immature and neural crest-like phenotype, Proc. Natl. Acad. Sci. U. S. A. 99 (10) (2002) 7021–7026.  V. Schaeffer, C. Patte-Mensah, A. Eckert, A.G. Mensah-Nyagan, Selective regulation of neurosteroid biosynthesis in human neuroblastoma cells under hydrogen peroxide-induced oxidative stress condition, Neuroscience 151 (3) (2007) 758–770.  E. Fordel, L. Thijs, W. Martinet, D. Schrijvers, L. Moens, S. Dewilde, Anoxia or oxygen and glucose deprivation in SH-SY5Y cells: a step closer to the unravelling of neuroglobin and cytoglobin functions, Gene 398 (1–2) (2007) 114–122.  J. Klose, U. Kobalz, Two-dimensional electrophoresis of proteins: an updated protocol and implications for a functional analysis of the genome, Electrophoresis 16 (1995) 1034–1059.  D. Pﬂieger, J.P. Le Caer, C. Lemaire, B.A. Bernard, G. Dujardin, J. Rossier, Systematic identiﬁcation of mitochondrial proteins by LC–MS/MS, Anal. Chem. 74 (2002) 2400–2406.  Y. Zhen, N. Xu, B. Richardson, R. Backlin, J.R. Savage, K. Balke, J.M. Peltier, Development of an LC–MALDI method for the analysis of protein complexes, J. Am. Soc. Mass Spectrom. 15 (2004) 803–822.  M.P. Washburn, D. Wolters, J.R. Yates III, Large-scale analysis of the yeast proteome by multidimensional protein identiﬁcation technology, Nat. Biotechnol. 19 (2001) 242–247.  K. Gevaert, M. Goethals, L. Martens, J. Van Damme, A. Staes, G.R. Thomas, J. Vandekerckhove, Exploring proteomes and analyzing protein processing by mass spectrometric identiﬁcation of sorted N-terminal peptides, Nat. Biotechnol. 21 (2003) 566–569.  N.K. Scheffer, S.W. Miller, A.K. Carroll, C. Anderson, R.E. Davis, S.S. Ghosg, B.W. Gibson, Two-dimensional electrophoresis and mass spectrometric identiﬁcation of mitochondrial proteins from an SH-SY5Y neuroblastoma cell line, Mitochondrion 1 (2001) 161–179.  M. Schirle, M.A. Heurtier, B. Kuster, Proﬁling core proteomes of human cell lines by one-dimensional PAGE and liquid chromatography–tandem mass spectrometry, Mol. Cell. Proteomics 2 (2003) 1297–1305.  M. Ashburner, C.A. Ball, J.A. Blake, D. Bostein, H. Butler, J.M. Cherry, A.P. Daivs, K. Dolinski, S.S. Dwight, J.T. Eppig, et al., Gene Ontologoy: tool for the uniﬁcation of biology, Nat. Genet. 25 (2000) 25–29.  G.J.R. Dennis, B.T. Sherman, D.A. Hosack, J. Yang, W. Gao, H.C. Lane, R.A. Lempicki, DAVID: Database for Annotation, Visulatization, and Intergrated Discovery, Geneome Biol. 4 (2003) R:60.  D.A. Hosack, G. Dennis Jr., B.T. Sherman, H.C. Lane, R.A. Lempicki, Identifying biological themes within lists of genes with EASE, Geneome Biol. 4 (2003) R:70.  S.J. Hattan, J. Marchese, N. Khainovski, S. Martin, P. Juhasz, Comparative study of [Three] LC–MALDI workﬂows for the analysis of complex proteomics samples, J. Proteome Res. 4 (2005) 1931–1941.  H. Lim, J. Eng, J.R. Yates III, S.L. Tollaksen, C.S. Giometti, J.F. Holden, M.W. Adams, C.I. Reich, G.J. Olsen, L.G. Hays, Identiﬁcation of 2D-gel proteins: a comparison of MALDI/ TOF peptide mass mapping to µLC–ESI tandem mass spectrometry, J. Am. Soc. Mass Spectrom. 14 (2003) 957–970.  W.M. Bodnar, R.K. Blackburn, J.M. Krise, M.A. Moseley, Exploiting the complementry nature of LC/MALDI/MS/MS and LC/ESI/MS/MS for increased proteome coverage, J. Am. Soc. Mass Spectrom. 14 (2003) 971–979.  K. Tilleman, I. Stevens, K. Spittaels, C.V. Haute, S. Clerens, G. Van Den Bergh, H. Geerts, F. Van Leuven, F. Vandesande, L. Moens, Differential expression of brain proteins inglycogen synthase kinase-3 transgenic mice: a proteomics point of view, Proteomics 2 (2002) 94–104.  V. Neuhoff, N. Arold, D. Taube, W. Ehrhardt, Improved staining of proteins in polyacrylamide gels including isoelectric-focusing gels with clear background at nanogram sensitivity using Coomassie Brilliant Blue G-250 and R-250, Electrophoresis 9 (1988) 255–262.  A. Shevchenko, M. Wilm, O. Vorm, M. Mann, Mass spectrometricsequencing of proteins silver-stained polyacrylamide gels, Anal.Chem. 68 (1996) 850–858.  L.J. Licklider, C.C. Thoreen, J. Peng, S.P. Gygi, Automation of nanoscale microcapillary liquid chromatography–tandem mass spectrometry with a vented column, Anal. Chem 74 (2002) 3076–3083.  R.J. Simpson, L.N. Connolly, J.S. Edders, J.J. Pereira, R.L. Mortiz, G.E. Ried, Proteomics analysis of the human colon carcinoma cell line (LIM 1215): development of a membrane protein database, Electrophoresis 21 (2000) 1707–1732.  P. Van Damme, L. Martens, J. Van Damme, K. Hugelier, A. Staes, J. Vandekerckhove, K. Gevaert, Caspase-speciﬁc and nonspeciﬁc in vivo protein processing during Fasinduced apoptosis, Nat. Methods 2 (2005) 771–777.  L. Martens, J. Vandekerckhove, K. Gevaert, DBToolkit: processing protein databases for peptide-centric proteomics, Bioinformatics 21 (2005) 3584–3585.  O. Carrette, P.R. Burkhard, D.F. Hochstrasser, J.C. Sanchez, Age-related proteome analysis of the mouse brain: a 2-DE study, Proteomics 6 (2006) 4940–4949.