Journal of Integrated Omics *Corresponding author: PD Dr. Martin von Bergen. UFZ, Helmholtz-Centre for Environmental Research, Department of Proteomics, Permoserstr. 15, 04318 Leipzig, Germany. Fax: +49-341-2351786. Email Address: Martin.vonbergen@ufz.de. | DOI: 10.5584/jiomics.v1i1.50 Franziska Dautel et al., 2010 | Journal of Integrated Omics 170-179:171 dealing with gel-to-gel discrepancies, labeling inefficiencies, and dyeand batch effects. Gel-to-gel discrepancies arise from run-time differences, variances in the loaded protein amounts or dye-front deformations [4]. Accounting for these differences is important for both 2-DE and 2-D DIGE. However, the dye-effect is specific for DIGE-projects, as the application of three different fluorophores can cause preferential dye-protein binding, variances in the fluorescent signal and background and differences in gel migration of the labeled proteins. As a result, protein abundances are not directly comparable when the proteins are labeled differently in various samples [5-7]. In addition, the experimental execution for a large number of samples is often divided into several batches of 6 or 12 gels. As a consequence, results of protein expression often cluster with the performed batches rather than with the individual samples and replicates. The goal is to identify spots that are truly differentially expressed, while accounting for statistical issues such as the multiple testing problem. This multiple testing problem states the accumulation of false positives as a general property of confidence-based statistical tests. These tests are applied across multiple features such as individual spots in DIGE to detect significantly altered changes in protein abundance [8]. This study reports an experimental design for a 2-factor analysis (time and concentration): murine hepatoma cells (Hepa1c1c7) were treated with the procarcinogen benzo-apyrene (B[a]P) and protein concentrations were quantified using 2-D DIGE (Fig. 1). Differential protein expression induced by B[a]P (or active B[a]P-metabolites) has previously been studied in different cellular models using one incubation time point and several B[a]Por B[a]P-metabolite concentrations [9-13]. In contrast, this B[a]P-protein expression analysis sampled four incubation time points at one toxic (5 μM) and one sub-acute B[a]P-concentration (50 nM), which required the processing of 36 samples in total [14]. In order to process the data originating from these experiments, a statistical analysis pipeline was developed to account for dyeand batch effects and to extract concentrationand time-dependent protein profiles. 2. Material and Methods 2.1 Cell culture and BaP exposure Murine hepatoma cells (Hepa1c1c7, ATCC No. CRL-2026; LGC Promochem, Wesel, Germany) were cultured as described elsewhere [15]. The cells were exposed to 50 nM B[a]P (Sigma-Aldrich, Steinheim, Germany), 5 μM B[a]P or DMSO for 2, 4, 12 and 24 h. Three independent biological replicates of all treatments were prepared. 2.2 DIGE and Data Analysis 2.2.1 Difference gel electrophoresis Cells were washed and lysed according to the procedure previously described [16]. Protein extracts were prepared and labeled according to manufacturer’s recommendations (GE Healthcare, Uppsala, Sweden). A Cy2-labeled common internal standard for all gels was prepared from a mixture of all samples IPG strips (24 cm, pH range 3-10 NL; GE Healthcare, Freiburg, Germany), which were rehydrated overnight and focused for 100,000 Vhrs using an Ettan IPGphor 3 isoelectric focusing unit (GE Healthcare, Freiburg, Germany) as described earlier [17]. Second dimension separation was performed using an Ettan DALTtwelve electrophoresis system (GE Healthcare, Uppsala, Sweden) on 12 % SDS-PAGE gels. The gels were scanned using the Ettan DIGE Imager Scanner (GE Healthcare, Uppsala, Sweden).
Read full abstract
Nov 26, 2011
Ilona Schreck + 4
Open Access