S156
Electrophoresis 2009, 30, S156–S161
Celebrating 30 years Jonathan S. Minden1 Susan R. Dowd1 Helmut E. Meyer2 Kai Stu¨hler2 1
Department of Biological Sciences, Carnegie Mellon University, Pittsburgh, PA, USA 2 Medizinisches Proteom-Center, Ruhr-Universita¨t Bochum, Germany
Received February 17, 2009 Revised March 10, 2009 Accepted March 11, 2009
Difference gel electrophoresis Difference gel electrophoresis (DIGE) was invented to circumvent the inherent variability of 2-DE. This variability is a natural consequence of separating thousands of proteins over a large space, such as a 15 20 cm slab of polyacrylamide gel. The originators of 2-DE envisioned being able to compare cancerous cells and normal cells to understand what makes these cells different. Gel-to-gel variability made this an extremely difficult task. We reasoned that if both samples could be run on the same gel, then the inherent variability would be obviated. Thus, we created matched sets of fluorescent dyes that allows one to compare two or three protein samples on a single gel. In the 12 years since the description of DIGE first appeared in Electrophoresis, this founding paper has been cited over 660 times. This review highlights some of the improvements and applications of DIGE. We hope these examples are illustrative of what has been done and where the field is headed.
Keywords: 2-D-Difference gel electrophoresis / Biomarker / Differential proteome analysis / Fluorescence dye saturation labeling / Microdissection DOI 10.1002/elps.200900098
1 Introduction The original idea for difference gel electrophoresis (DIGE) was born from a desire to detect protein changes caused by mutational events. How does a cell’s physiology and biochemistry change in response to a mutational change? At the time of its inception, the state-of-the-art method for maximal separation of proteins was 2-DE [1–3]. Given the complexity of cellular extracts and the variability of 2-DE, it was reasoned that side-by-side comparisons would be extremely difficult to interpret and that a more fruitful approach would be to develop an internally controlled method where both the sample and its control are run on the same gel. This internally controlled scheme would require a method for differentially labeling all proteins within a cellular extract. To make the method more accessible to the average laboratory, fluorescent labeling reagents were sought. Four design rules were adopted for the development of the fluorescent, DIGE dyes: (i) each matched set of dyes must react with the same amino acid
Correspondence: Dr. Jonathan S. Minden, Carnegie Mellon University – Biological Sciences, 4400 Fifth Avenue, Pittsburgh, PA 15213, USA E-mail:
[email protected] Fax: +1-412-268-7129
Abbreviation: DIGE, difference gel electrophoresis
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
residues, (ii) they must preserve the charge of the target amino acid, (iii) the dyes must have similar molecular masses, and (iv) they must have distinct fluorescent characteristics. At the time of the idea’s inception, suitable fluorescent dyes were not available. Thus the DIGE dyes, which are based on a cyanine-dye backbone, were designed to meet these criteria. J. M. and Dr. Alan Waggoner, the father of applying cyanine dyes to biological research, designed the DIGE dyes over lunch one day. The napkin ¨ nlu ¨, containing the dye structures was handed to Mustafa U a Carnegie Mellon graduate student, who along with Mary Morgan, a Carnegie Mellon undergraduate student, synthesized the first set of DIGE dyes. The first 2-D-DIGE, gelimaging system was constructed by J. M. using a liquidcooled CCD camera mounted on a kitchen cabinet purchased from IKEA. The commercial development of the DIGE dyes (CyDyesTM) was subsequently done in collaboration with Amersham Biosciences (now GE Healthcare). The first published report of DIGE appeared in ¨ nlu ¨ et al. [4]. Electrophoresis in 1997 by U
2 Protein detection using fluorescent dyes As most covalent modifications affect protein migration in 2-DE gels, we sought to minimize these effects. Preserving the charge of the target amino acid residue maintains the isoelectric point of the labeled proteins. Dye labeling of www.electrophoresis-journal.com
Electrophoresis 2009, 30, S156–S161
proteins unavoidably affects their migration in the molecular mass dimension of 2-DE gels. This apparent mass perturbation has more to do with how the dyes affect SDS binding than directly altering the mass of the protein. To minimize this apparent mass perturbation effect, two levels of protein labeling were found to be acceptable: either minimal labeling, where each protein carries a maximum of one dye molecule, or saturation labeling, where 100% of all reactive residues are dye-coupled. Labeling between these two extremes generates size heterogeneity in the protein populations. Two matched set of DIGE dyes have been developed: one set is lysine-reactive and the other is cysteine-reactive. A typical 50 000 Da protein has about 30 lysine residues. Attempts to label all lysine residues in a protein led to a large increase in the apparent molecular mass of the proteins and caused protein precipitation. We thus turned to minimal labeling of proteins with the lysinereactive dyes (referred to as minimal-labeling Cy2, Cy3 and Cy5). To ensure that only a single lysine residue was labeled per protein, conditions were established where about one in 20 proteins had one lysine residue labeled, the rest of the proteins had no label. A single dye molecule has negligible impact on the apparent molecular mass of proteins over 30 kDa, relative to un-labeled proteins. Minimally labeled proteins under 30 kDa run slightly slower than their unlabeled counterparts, but this shift is usually less than one half diameter of the protein spot. Post-electrophoresis staining of proteins is often used to direct the excision of protein spots for MS analysis [5]. The second set of DIGE dyes react with cysteine residues. Because there are generally fewer cysteines per protein and the cysteine-reactive dyes are highly soluble zwitterions, these dyes are suitable for saturation labeling of all cysteine residues. As a consequence, sensitivity for protein detection is improved and enables a successful 2-DE analysis of samples with low protein concentration. In recent years, various studies have demonstrated the successful application of saturation labeling to detect protein differences in scarce samples derived from 1000 to 5000 cells [6–10]. These two DIGE labeling options provide rapid methods for preparing differentially labeled samples for fluorescencebased proteome comparisons. At the time of its creation, there were no suitable fluorescent gel imaging systems for DIGE gels. Two imaging approaches have been developed to capture the fluorescent signals from DIGE gels. The first is a CCDbased system that employs wide-field illumination and an open-gel configuration that allows direct access for a spotcutting robot. The second imaging system utilizes laser scanning of the gel sandwiched between two, low-fluorescence, glass plates. Both systems provide high-quality, wide dynamic range images of DIGE gels. The laser scanning method requires a second device for cutting protein spots from the gel for MS analysis. With the advent of suitable DIGE dyes and sensitive imaging systems, the most crucial aspect of performing informative DIGE-based proteomics studies is experimental design and analysis. & 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Proteomics and 2-DE
S157
3 Design and statistical analysis of 2-DDIGE experiments 2-D-DIGE was developed to determine protein expression changes in proteomes of biological samples with higher accuracy. The introduction of an internal standard led to an enormous increase in accuracy and multiplexing of samples using different fluorescent dyes reduced the systematic variation [4, 11]. Once the accuracy and performance of DIGE were confirmed, methods for data processing and statistical analysis emerged to fully exploit the potential of DIGE. Solutions from transcriptome studies utilizing multiplexed microarrays were considered, but could not be directly applied to DIGE because the data exhibited variability due to: (i) false-positive spots (detection errors, ‘‘weak spots’’); (ii) non-matched spots (missing data) and (iii) different sources of background fluorescence. Data obtained from biological experiments comprises systematic and biological variation. Before one can learn about the biological variation in the samples-of-interest, one must understand and correct for sources of systematic variation. One source of systematic variation using DIGE comes from the different extinction coefficients of the dyes. Because Cy5 has a higher extinction coefficient than Cy3, it has been shown that the Cy5 dye causes higher volume values than the Cy3 dye under similar illumination settings [12]. To calibrate the spot volumes, a scaling factor and a background offset were used to adjust for the dye-specific gain [13, 14]. The background offset is used because the different dyes result in different background fluorescences. A detailed examination of imager-based systematic error taking several sources of fluorescent signal and imager response into account revealed that many of these errors can be accurately modeled and effectively corrected [15]. It is necessary to apply these corrections to the raw data before assessing the biological differences in the samples. Several statistical approaches have been applied for the analysis of DIGE data. These include the classical threshold approach relying on a threshold above which an expression change should be significant [13], univariate methods or multivariate methods [16]. Univariate methods e.g. Student’s ttest or Mann–Whitney U-test performing a so-called spot-byspot analysis to detect significant changes in expression for individual spots. Multivariate analysis, e.g. principal component analysis, utilizes all the protein spot data simultaneously, to look for patterns in expression changes. The univariate approach gives strong candidates that have had significant changes in expression, while the multivariate approaches can detect subtler changes across sets of proteins that work in cohort such as those in a pathway. Therefore, the Student’s t-test found a broad application for the analysis of DIGE data. But, for the application of statistical analysis specific assumptions have to be considered. For instance, the Student’s t-test (parametric test) is based on the assumption that the analyzed data set (protein spots quantities) follows a normal distribution, whereas the Mann–Whitney U-test (nonparametric tests) does not depend on the normality assumption. www.electrophoresis-journal.com
S158
Electrophoresis 2009, 30, S156–S161
J. S. Minden et al.
The analysis of spot intensities from DIGE experiments revealed a lognormal distribution of the spot volumes [12, 13]. Moreover, the variance of the spot volumes was dependent on the mean of the spot volumes. To utilize the Student’s t-test, the raw data require a second transformation using an arsinh transformation for normalization and variance stabilization. The dye-specific gain for the calibrated and transformed spot volumes was minimized and there was no more dependence between variance and mean. This arsinh transform reduces the bias for low-volume spots introduced by the logarithmic transformation, which is still used in the majority of software systems. For the application of multivariate analysis to compare multiple 2-D-DIGE gels, missing data represent major problem since most statistical methods rely on the assumption that the data set to be analyzed is complete. However, in 2-D-DIGE studies with gel replicates between 10 and 30% of the values are missing [17]. This is due to the fact that gel-to-gel variation is high for 2-DE gels. Several different methods have been applied to handle missing data. The ‘‘row-mean method’’ uses the average of all available measurements in the row where the value is missing as estimator for this missing value, while the ‘knearest neighbor method’ uses a more sophisticated method that assumes a relationship between the expression profiles of some proteins [17]. Pedreschi et al. have reported that Bayesian principal component analysis exhibits more consistent performance [18]. However, there is a risk that a Bayesian approach will bias the statistical analysis by introducing additional, possibly erroneous information. A frequently asked question regarding DIGE experimental design concerns the number of replicates required for significance. Since DIGE greatly reduces systematic variance, one only needs to focus on estimating the sufficient number of biological replicates [14]. To estimate the sufficient number of replicates, a statistical test based on the sample size, the magnitude of expression change one wants to detect as well as on the coefficient of variation for the data can be utilized [19, 20]. Relying on the DIGE-inherent systematic variance of o20%, three to five biological replicates are sufficient to determine a fold change in the range of 1.5 and 2.0 [21]. In cases where the biological variance exceeds the systematic variance, the number of biological replicates has to be appropriately increased [20, 22]. Karp et al. reported a bias in the p value distribution in the threedye DIGE experimental setup [20]. This bias was due to the use of a Cy2-labeled internal standard, which produces a weak fluorescent signal relative to Cy3 and Cy5. These investigators suggest using a two-dye (Cy3 and Cy5) setup along with advanced statistical tools to handle this methodological bias. This example and the open question of ‘‘What is a good strategy for missing data?’’ show that there is room for improvement of the statistical analysis of DIGE data.
4 DIGE applications DIGE has been used to assess proteome changes in a wide variety of circumstances, from animal development to the & 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
effects of disease on tissues and bodily fluids and from subcellular compartments to the development of novel proteome separation schemes. In this section we provide a very brief overview of the uses of DIGE starting with whole organisms, then proceeding to tissues, sub-cellular structures and bodily fluids. Finally, we will discuss emerging methods to increase the resolution of protein separation for DIGE analysis.
4.1 Model organisms The first application of DIGE to developmental biology was the study of ventral furrow formation in Drosophila melanogaster embryos [23]. This study showed that many aspects of cell physiology change as cells change developmental states. Validation of the observed proteome changes revealed an unusual dependency between the proteosome and iron homeostasis to change many aspects of the ventral furrow cell’s proteome [24]. DIGE has since been used to study the development of zebrafish comparing different days post fertilization [25]. Zebrafish has also been used to assess the effects of alcohol and the flame retardant, TBBPA, on development [26, 27]. DIGE has been used to analyze the development of several non-animal species, including Arabidopsis [28–35], rice [36], flax [37], strawberry [38] and yeast [39–44]. These studies asked a variety of questions ranging from developmental stages and ripening to senescence and a number of environmental responses to cold, CO2, metals, brassinosteriods and oligogalacturonides. Plants have also been used to assess DIGE technology and compare it with other proteome comparison schemes [32, 33]. Similar questions were asked about the effect of growth conditions, oxidative stress and metal challenge to several types of yeast, as well as the comparison of different proteome analysis methods [39–41, 43, 44]. One study dispelled the notion that electromagnetic fields affect cell physiology [42]. These organism-level studies have revealed many interesting insights into the normal and perturbed development of model organisms. A key next step will be to extend these studies to understand the molecular connection between the sets of candidate difference-proteins found in these diverse studies.
4.2 Tissue proteomics The direct analysis of clinical tissue specimens exhibits the advantage of unbiased interrogation of the underlying pathobiological processes and correlation of the proteome with clinical parameters. Tissue proteomics offers the opportunity to identify new biomarkers for immunohistochemistry and even in vitro diagnostics. Human tissue samples contain candidate biomarkers, if existent, in higher concentrations that can be analyzed with highly sensitive proteomics techniques. Since human tissues are quite heterogeneous, cell populations of interest should be isolated utilizing specific www.electrophoresis-journal.com
Proteomics and 2-DE
Electrophoresis 2009, 30, S156–S161
methods to avoid cell dilution by counterpart cells, stroma cells, vascular endothelial cells and inflammatory cells [22]. In order to select for appropriate cell types, various techniques, for instance, manual microdissection [45] or laser beam processing [46] can be applied. Normally, microdissection results in a small amount of sample (1000–10 000 cells) and therefore sensitive MS- or gel-based methods, allowing high-resolution proteomics to be established [22]. DIGE saturation labeling has been successfully applied in different studies [6, 10, 47, 48]. Kondo et al. were the first to report the application of DIGE saturation-labeling analysis of 6.6 mg from microdissected mice adenoma tissue [10]. Sitek et al. developed a protocol for DIGE saturation labeling, allowing a protein profiling of 1000 microdissected cells from normal pancreatic ducts and different PanIN grades revealing about 2500 protein spots. This protocol comprised various different processing steps, including tissue staining, optimization of protein labeling and evaluation of a reference proteome for protein identification [49]. DIGE saturation-labeling of microdissected tissue has been applied to lung cancer tissues [49, 50], pancreatic cancer tissues [49], gastric cancer tissues [9], esophageal cancer tissues [51], neurons [8], kidney tissues [47] and liver tissues [52, 53]. Using this approach, the candidate biomarkers, tropomyosin and microfibrillar-associated protein 4 (MFAP-4), were identified as difference-proteins in human cirrhotic liver tissue. These were subsequently validated in tissue by immunohistochemistry and in sera of patients suffering from liver fibrosis due to HCV infection. The quantification of microfibrillar-associated protein 4 serum samples by a recently established ELISA allowed a distinction between cirrhotic and non-diseased patients to be made with high confidence (91.6% sensitivity and 95.6% specificity) [53]. In addition, Kondo and colleagues recently identified pfetin as a novel prognostic biomarker of gastrointestinal stromal tumors [54]. Using immunohistochemistry examination of pfetin expression in 210 cases revealed that the 5-year metastasis-free survival rate was 93.9 and 36.2% for patients with pfetin-positive and pfetin-negative primary tumors, respectively. These results are very encouraging that combining microdissection with saturation-labeling DIGE and MS will lead to the discovery of a large array of disease markers for a wide variety of diseases and conditions.
4.3 Subcellular proteomics Given the complexity of whole-cell proteomes, it often desirable to fractionate cell extracts to focus on a particular compartment of interest. DIGE has been used to compare sub-cellular proteomes of several different compartments under a variety of conditions [55]. Mitochondria was the first organelle to be studied using DIGE, where the proteomes of heart mitochondria from wild-type mice were compared with creatine kinase knock-out mice [56]. Single membranebounded organelles, such as lysosomes, peroxisomes and & 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
S159
exosomes, have been analyzed by DIGE [57–60]. Nuclear proteins have also been compared [61, 62]. Two important and technically challenging compartments, integral membrane proteins and the phosphoproteome, have been assessed by DIGE. 2-DE analysis of membrane proteins remains a challenge because of loss of solubility during IEF and difficulty in MS analysis of membrane proteins [31, 63–65]. Phosphoproteome analysis is limited by the enrichment schemes used to capture phosphorylated proteins [31, 66–68]. Another scheme for spreading out the proteome over a larger separation space is to use microscale solution IEF to generate a set of discrete pI samples that are subsequently analyzed on narrow range IEF 2-DE gels [69, 70]. The role of DIGE in these experiments is to provide an internal standard to make comparisons more rigorous. Technical advancements in the resolution of whole-cell or membrane- and phospho-proteins are needed to increase proteome sampling efficiency. Any separation step has the potential to introduce handling artifacts. Therefore, one should fluorescently label the proteins to be analyzed and combine the labeled samples as early in the workflow as possible to limit the appearance of handling artifacts.
4.4 Body fluids One of the ‘‘Holy Grails’’ of proteomics is the discovery of disease markers in bodily fluids [71]. This goal is so prized that many more review and prospectus articles have been written about disease marker discovery than actual research articles about newly discovered disease markers. The largest challenge in analyzing these fluids is the ten orders of magnitude protein concentration range [72]. Despite this technological barrier, DIGE has been used to assess proteome changes in bodily fluids such as serum [73, 74], urine [75], amniotic fluid [76], cervical–vaginal fluid [77] or CSF [78]. There is still a great deal of work that needs to be done to enrich for low abundance proteins that serve as potential disease markers.
5 Concluding remarks DIGE-based comparative proteomics has made a significant impact on our understanding of normal development and the effect that mutations, diseases and environmental stresses have on the proteomes of a range of organisms. It is important to appreciate that no proteomic method is able to sample the entire proteome – we are sampling the most prominent changes associated with the states being compared. The observed proteome changes will always need to be validated and examined for their molecular role in the process of interest. DIGE provides a platform for controlling variation due to sample preparation, protein separation and difference detection. DIGE technology is still in its infancy. One can anticipate that advances in protein www.electrophoresis-journal.com
S160
J. S. Minden et al.
separation, image acquisition and data analysis will further increase the sensitivity to, and resolution of, biologically relevant changes in the proteome under an almost infinite number of circumstances. Jonathan Minden is the inventor of DIGE and receives royalties from GE Healthcare for its sales.
6 References [1] Garrels, J. I., J. Biol. Chem. 1979, 254, 7961–7977. [2] Klose, J., Humangenetik 1975, 26, 231–243. [3] O’Farrell, P. H., J. Biol. Chem. 1975, 250, 4007–4021. ¨ nlu¨, M., Morgan, M. E., Minden, J. S., Electrophoresis [4] U 1997, 18, 2071–2077. [5] Gharbi, S., Gaffney, P., Yang, A., Zvelebil, M. J., Cramer, R., Waterfield, M. D., Timms, J. F., Mol. Cell Proteomics 2002, 1, 91–98. [6] Sitek, B., Luttges, J., Marcus, K., Kloppel, G., Schmiegel, W., Meyer, H. E., Hahn, S. A. et al., Proteomics 2005, 5, 2665–2679. [7] Sitek, B., Potthoff, S., Schulenborg, T., Stegbauer, J., Vinke, T., Rump, L. C., Meyer, H. E. et al., Proteomics 2006, 6, 4337–4345. [8] Wilson, K. E., Marouga, R., Prime, J. E., Pashby, D. P., Orange, P. R., Crosier, S., Keith, A. B. et al., Proteomics 2005, 5, 3851–3858. [9] Greengauz-Roberts, O., Stoppler, H., Nomura, S., Yamaguchi, H., Goldenring, J. R., Podolsky, R. H., Lee, J. R. et al., Proteomics 2005, 5, 1746–1757. [10] Kondo, T., Seike, M., Mori, Y., Fujii, K., Yamada, T., Hirohashi, S., Proteomics 2003, 3, 1758–1766. [11] Alban, A., David, S. O., Bjorkesten, L., Andersson, C., Sloge, E., Lewis, S., Currie, I., Proteomics 2003, 3, 36–44. [12] Jung, K., Gannoun, A., Sitek, B., Meyer, H. E., Stu¨hler, K., Urfer, W., Rev. Stat. 2005, 3, 44–111. [13] Karp, N. A., Kreil, D. P., Lilley, K. S., Proteomics 2004, 4, 1421–1432. [14] Karp, N. A., Spencer, M., Lindsay, H., O’Dell, K., Lilley, K. S., J. Proteome Res. 2005, 4, 1867–1871. [15] Sellers, K. F., Miecznikowski, J., Viswanathan, S., Minden, J. S., Eddy, W. F., Electrophoresis 2007, 28, 3324–3332. [16] Berth, M., Moser, F. M., Kolbe, M., Bernhardt, J., Appl. Microbiol. Biotechnol. 2007, 76, 1223–1243. [17] Jung, K., Gannoun, A., Sitek, B., Apostolov, O., Schramm, A., Meyer, H. E., Stu¨hle, K. et al., Rev. Stat. 2006, 4, 67–80. [18] Pedreschi, R., Hertog, M. L., Carpentier, S. C., Lammertyn, J., Robben, J., Noben, J. P., Panis, B. et al., Proteomics 2008, 8, 1371–1383.
Electrophoresis 2009, 30, S156–S161
[21] Karp, N. A., Feret, R., Rubtsov, D. V., Lilley, K. S., Proteomics 2008, 8, 948–960. [22] Meyer, H. E., Stuhler, K., Proteomics 2007, 7, 18–26. [23] Gong, L., Puri, M., Unlu, M., Young, M., Robertson, K., Viswanathan, S., Krishnaswamy, A. et al., Development 2004, 131, 643–656. [24] Puri, M., Goyal, A., Senutovich, N., Dowd, S. R., Minden, J. S., Mol. Biosyst. 2008, 4, 1126–1135. [25] Lucitt, M. B., Price, T. S., Pizarro, A., Wu, W., Yocum, A. K., Seiler, C., Pack, M. A. et al., Mol. Cell Proteomics 2008, 7, 981–994. [26] Damodaran, S., Dlugos, C. A., Wood, T. D., Rabin, R. A., Eur. J. Pharmacol. 2006, 547, 75–82. [27] De Wit, M., Keil, D., Remmerie, N., van der Ven, K., van den Brandhof, E. J., Knapen, D., Witters, E. et al., Chemosphere 2008, 74, 96–105. [28] Amme, S., Matros, A., Schlesier, B., Mock, H. P., J. Exp. Bot. 2006, 57, 1537–1546. [29] Casasoli, M., Spadoni, S., Lilley, K. S., Cervone, F., De Lorenzo, G., Mattei, B., Proteomics 2008, 8, 1042–1054. [30] Deng, Z., Zhang, X., Tang, W., Oses-Prieto, J. A., Suzuki, N., Gendron, J. M., Chen, H. et al., Mol. Cell Proteomics 2007, 6, 2058–2071. [31] Tang, W., Deng, Z., Oses-Prieto, J. A., Suzuki, N., Zhu, S., Zhang, X., Burlingame, A. L. et al., Mol Cell Proteomics 2008, 7, 728–738. [32] Hebeler, R., Oeljeklaus, S., Reidegeld, K. A., Eisenacher, M., Stephan, C., Sitek, B., Stuhler, K. et al., Mol Cell Proteomics 2008, 7, 108–120. [33] Lilley, K. S., Dupree, P., J Exp Bot 2006, 57, 1493–1499. [34] Lilley, K. S., Griffiths, D. R., Brief. Funct. Genomic. Proteomic. 2003, 2, 106–113. [35] Qiu, Q. S., Huber, J. L., Booker, F. L., Jain, V., Leakey, A. D., Fiscus, E. L., Yau, P. M. et al., Photosynth. Res. 2008, 97, 155–166. [36] Komatsu, S., Zang, X., Tanaka, N., J. Proteome Res. 2006, 5, 270–276. [37] Hotte, N. S., Deyholos, M. K., BMC Plant Biol. 2008, 8, 52. [38] Bianco, L., Lopez, L., Scalone, A. G., Di Carli, M., Desiderio, A., Benvenuto, E., Perrotta, G., J. Proteomics 2008. [39] Weeks, M. E., Sinclair, J., Butt, A., Chung, Y. L., Worthington, J. L., Wilkinson, C. R., Griffiths, J. et al., Proteomics 2006, 6, 2772–2796. [40] Hu, Y., Wang, G., Chen, G. Y., Fu, X., Yao, S. Q., Electrophoresis 2003, 24, 1458–1470. [41] Kolkman, A., Dirksen, E. H., Slijper, M., Heck, A. J., Mol. Cell Proteomics 2005, 4, 255–266. [42] Sinclair, J., Weeks, M., Butt, A., Worthington, J. L., Akpan, A., Jones, N., Waterfield, M. et al., Proteomics 2006, 6, 4755–4764. [43] Lis, M., Bobek, L. A., FEMS Immunol. Med. Microbiol. 2008, 54, 80–91.
[19] Stuhler, K., Pfeiffer, K., Joppich, C., Stephan, C., Jung, K., Muller, M., Schmidt, O. et al., Proteomics 2006, 6, 4899–4913.
[44] Valerius, O., Kleinschmidt, M., Rachfall, N., Schulze, F., Lopez Marin, S., Hoppert, M., Streckfuss-Bomeke, K. et al., Mol. Cell Proteomics 2007, 6, 1968–1979.
[20] Karp, N. A., McCormick, P. S., Russell, M. R., Lilley, K. S., Mol. Cell Proteomics 2007, 6, 1354–1364.
[45] Whetsell, L., Maw, G., Nadon, N., Ringer, D. P., Schaefer, F. V., Oncogene 1992, 7, 2355–2361.
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
Proteomics and 2-DE
Electrophoresis 2009, 30, S156–S161
[46] Emmert-Buck, M. R., Bonner, R. F., Smith, P. D., Chuaqui, R. F., Zhuang, Z., Goldstein, S. R., Weiss, R. A. et al., Science 1996, 274, 998–1001. [47] Sitek, B., Potthoff, S., Schulenborg, T., Stegbauer, J., Vinke, T., Rump, L. C., Meyer, H. E. et al., Proteomics 2006, 6, 4506–4513.
S161
[62] Tan, H. T., Zubaidah, R. M., Tan, S., Hooi, S. C., Chung, M. C., J. Proteome Res. 2006, 5, 1098–1106. [63] Alfonso, P., Canamero, M., Fernandez-Carbonie, F., Nunez, A., Casal, J. I., J. Proteome Res. 2008, 7, 4247–4255. [64] Hagner-McWhirter, A., Winkvist, Marouga, R., J. Vis. Exp. 2008.
M.,
Bourin,
S.,
[48] Fujii, K., Kondo, T., Yokoo, H., Yamada, T., Iwatsuki, K., Hirohashi, S., Proteomics 2005, 5, 1411–1422.
[65] Schindler, J., Jung, S., Niedner-Schatteburg, G., Friauf, E., Nothwang, H. G., J. Neural Transm. 2006, 113, 995–1013.
[49] Poschmann, G., Sitek, B., Sipos, B., Ulrich, A., Wiese, S., Stephan, C., Warscheid, B. et al., Mol. Cell Proteomics 2009.
[66] Morandell, S., Stasyk, T., Grosstessner-Hain, K., Roitinger, E., Mechtler, K., Bonn, G. K., Huber, L. A., Proteomics 2006, 6, 4047–4056.
[50] Seike, M., Kondo, T., Fujii, K., Okano, T., Yamada, T., Matsuno, Y., Gemma, A. et al., Proteomics 2005, 5, 2939–2948.
[67] Stasyk, T., Morandell, S., Bakry, R., Feuerstein, I., Huck, C. W., Stecher, G., Bonn, G. K. et al., Electrophoresis 2005, 26, 2850–2854.
[51] Hatakeyama, H., Kondo, T., Fujii, K., Nakanishi, Y., Kato, H., Fukuda, S., Hirohashi, S., Proteomics 2006, 6, 6300–6316.
[68] Stasyk, T., Schiefermeier, N., Skvortsov, S., Zwierzina, H., Peranen, J., Bonn, G. K., Huber, L. A., Mol. Cell Proteomics 2007, 6, 908–922.
[52] Orimo, T., Ojima, H., Hiraoka, N., Saito, S., Kosuge, T., Kakisaka, T., Yokoo, H. et al., Hepatology 2008, 48, 1851–1863.
[69] Han, M. J., Herlyn, M., Fisher, A. B., Speicher, D. W., Electrophoresis 2008, 29, 695–705.
[53] Molleken, C., Sitek, B., Henkel, C., Poschmann, G., Sipos, B., Wiese, S., Warscheid, B. et al., Hepatology 2008. [54] Suehara, Y., Kondo, T., Seki, K., Shibata, T., Fujii, K., Gotoh, M., Hasegawa, T. et al., Clin. Cancer Res. 2008, 14, 1707–1717. [55] Lilley, K. S., Dupree, P., J. Exp. Bot. 2006, 57, 1493–1499. [56] Kernec, F., Unlu, M., Labeikovsky, W., Minden, J. S., Koretsky, A. P., Physiol. Genomics 2001, 6, 117–128. [57] Della Corte, A., Maugeri, N., Pampuch, A., Cerletti, C., de Gaetano, G., Rotilio, D., Platelets 2008, 19, 43–50. [58] Gustin, C., Delaive, E., Dieu, M., Calay, D., Raes, M., Arterioscler. Thromb. Vasc. Biol. 2008, 28, 491–497. [59] Islinger, M., Luers, G. H., Zischka, H., Ueffing, M., Volkl, A., Proteomics 2006, 6, 804–816. [60] Schmidt, H., Gelhaus, C., Nebendahl, M., Lettau, M., Watzl, C., Kabelitz, D., Leippe, M. et al., Proteomics 2008, 8, 2911–2925. [61] Bernocco, S., Fondelli, C., Matteoni, S., Magnoni, L., Gotta, S., Terstappen, G. C., Raggiaschi, R., Proteomics 2008, 8, 930–938.
[70] Han, M. J., Speicher, D. W., Methods Mol. Biol. 2008, 424, 241–256. [71] Pitteri, S. J., Hanash, S. M., Expert Rev. Proteomics 2007, 4, 589–590. [72] Anderson, N. L., Anderson, N. G., Mol. Cell Proteomics 2002, 1, 845–867. [73] Byrne, J. C., Downes, M. R., O’Donoghue, N., O’Keane, C., O’Neill, A., Fan, Y., Fitzpatrick, J. M. et al., J. Proteome Res. 2008. [74] Issaq, H. J., Veenstra, T. D., Electrophoresis 2007, 28, 1980–1988. [75] Simon, S. L., Lamoureux, L., Plews, M., Stobart, M., Lemaistre, J., Ziegler, U., Graham, C., et al., Proteome Sci. 2008, 6, 23. [76] Michaels, J. E., Dasari, S., Pereira, L., Reddy, A. P., Lapidus, J. A., Lu, X., Jacob, T. et al., J. Proteome Res. 2007, 6, 1277–1285. [77] Pereira, L., Reddy, A. P., Jacob, T., Thomas, A., Schneider, K. A., Dasari, S., Lapidus, J. A. et al., J. Proteome Res. 2007, 6, 1269–1276. [78] Gao, W. M., Chadha, M. S., Berger, R. P., Omenn, G. S. et al., J. Neurotrauma 2007, 24, 43–53.
Jonathan Minden is a Professor in the department of Biological Sciences at Carnegie Mellon University. In addition to inventing 2-D-DIGE, the Minden lab uses 2-D-DIGE in conjunction with fluorescent probe development and time-lapse microscopy to study two central questions in the developmental biology of the fruit fly embryo: (1) Tissue morphogenesis-how do the cells in fruit fly embryos change their shape to drive tissue morphogenesis?; and (2) Programmed cell death-how do cells decide to die in order to maintain tissue homeostasis?
& 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.electrophoresis-journal.com
