Discovering robust protein biomarkers for disease from relative expression reversals in 2-D DIGE data

This study assesses the ability of a novel family of machine learning algorithms to identify changes in relative protein expression levels, measured using 2-D DIGE data, which support accurate class prediction. The analysis was done using a training set of 36 total cellular lysates comprised of six normal and three cancer biological replicates (the remaining are technical replicates) and a validation set of four normal and two cancer samples. Protein samples were separated by 2-D DIGE and expression was quantified using DeCyder-2D Differential Analysis Software. The relative expression reversal (RER) classifier correctly classified 9/9 training biological samples (p<0.022) as estimated using a modified version of leave one out cross validation and 6/6 validation samples. The classification rule involved comparison of expression levels for a single pair of protein spots, tropomyosin isoforms and alpha-enolase, both of which have prior association as potential biomarkers in cancer. The data was also analyzed using algorithms similar to those found in the extended data analysis package of DeCyder software. We propose that by accounting for sources of within- and between-gel variation, RER classifiers applied to 2-D DIGE data provide a useful approach for identifying biomarkers that discriminate among protein samples of interest.     

Functional proteomic view of metabolic regulation in "Aromatoleum aromaticum" strain EbN1

The denitrifying "Aromatoleum aromaticum" strain EbN1 utilizes a wide range of aromatic and nonaromatic compounds under anoxic and oxic conditions. The recently determined genome revealed corresponding degradation pathways and predicted a fine-tuned regulatory network. In this study, differential proteomics (2-D DIGE and MS) was used to define degradation pathway-specific subproteomes and to determine their growth condition dependent regulation. Differential protein profiles were determined for cultures adapted to growth under 22 different substrate and redox conditions. In total, 354 different proteins were identified, 199 of which displayed significantly changed abundances. These regulated proteins mainly represented enzymes of the different degradation pathways, and revealed different degrees of growth condition specific regulation. In case of three substrate conditions (e.g. phenylalanine, anoxic), proteins previously predicted to be involved in their degradation were apparently not involved (e.g. Pdh, phenylacetaldehyde dehydrogenase). Instead, previously not considered proteins were specifically increased in abundance (e.g. EbA5005, predicted aldehyde:ferredoxin oxidoreductase), shedding new light on the respective pathways. Moreover, strong evidence was obtained for thus far unpredicted degradation pathways of three hitherto unknown substrates (e.g. o-aminobenzoate, anoxic). Comparing all identified regulated and nonregulated proteins provided first insights into regulatory hierarchies of special degradation pathways versus general metabolism in strain EbN1.     

Selective Labelling of Cell-surface Proteins using CyDye DIGE Fluor Minimal Dyes

Surface proteins are central to the cell''s ability to react to its environment and to interact with neighboring cells. They are known to be inducers of almost all intracellular signaling. Moreover, they play an important role in environmental adaptation and drug treatment, and are often involved in disease pathogenesis and pathology (1). Protein-protein interactions are intrinsic to signaling pathways, and to gain more insight in these complex biological processes, sensitive and reliable methods are needed for studying cell surface proteins. Two-dimensional (2-D) electrophoresis is used extensively for detection of biomarkers and other targets in complex protein samples to study differential changes. Cell surface proteins, partly due to their low abundance (1 2% of cellular proteins), are difficult to detect in a 2-D gel without fractionation or some other type of enrichment. They are also often poorly represented in 2-D gels due to their hydrophobic nature and high molecular weight (2). In this study, we present a new protocol for intact cells using CyDye DIGE Fluor minimal dyes for specific labeling and detection of this important group of proteins. The results showed specific labeling of a large number of cell surface proteins with minimal labeling of intracellular proteins. This protocol is rapid, simple to use, and all three CyDye DIGE Fluor minimal dyes (Cy 2, Cy 3 and Cy 5) can be used to label cell-surface proteins. These features allow for multiplexing using the 2-D Fluorescence Difference Gel Electrophoresis (2-D DIGE) with Ettan DIGE technology and analysis of protein expression changes using DeCyder 2-D Differential Analysis Software. The level of cell-surface proteins was followed during serum starvation of CHO cells for various lengths of time (see Table 1). Small changes in abundance were detected with high accuracy, and results are supported by defined statistical methods.Open in a separate windowClick here to view.^{(76M, flv)}     

Data variance and statistical significance in 2D-gel electrophoresis and DIGE experiments: comparison of the effects of normalization methods

Identifying changes in the relative abundance of proteins between different biological samples is often confounded by technical noise. In this work, we compared eight normalization methods commonly used in two-dimensional gel electrophoresis and difference gel electrophoresis (DIGE) experiments for their ability to reduce noise and for their influence on the list of proteins whose difference in abundance between two samples is determined to be statistically significant. With respect to reducing noise we find that, while all methods improve upon unnormalized data, cyclic linear normalization is the least well suited to gel-based proteomics and the performances of the other methods are similar. We also find in DIGE data that the choice of normalization method has less of an impact on the noise than does the decision to use an internal reference in the experimental design and that both normalization and standardization using the internal reference are required to maximally reduce variance. Despite the similar noise reduction achieved by most normalization methods, the list of proteins whose abundance was determined to differ significantly between biological groups differed depending on the choice of normalization method. This work provides a direct comparison of the impact of normalization methods in the context of common experimental designs.     

Statistical analysis of the experimental variation in the proteomic characterization of human plasma by two-dimensional difference gel electrophoresis

The complexity of human plasma presents a number of challenges to the efficient and reproducible proteomic analysis of differential expression in response to disease. Before individual variation and disease-specific protein biomarkers can be identified from human plasma, the experimental variability inherent in the protein separation and detection techniques must be quantified. We report on the variation found in two-dimensional difference gel electrophoresis (2-D DIGE) analysis of human plasma. Eight aliquots of a human plasma sample were subjected to top-6 highest abundant protein depletion and were subsequently analyzed in triplicate for a total of 24 DIGE samples on 12 gels. Spot-wise standard deviation estimates indicated that fold changes greater than 2 can be detected with a manageable number of replicates in simple ANOVA experiments with human plasma. Mixed-effects statistical modeling quantified the effect of the dyes, and segregated the spot-wise variance into components of sample preparation, gel-to-gel differences, and random error. The gel-to-gel component was found to be the largest source of variation, followed by the sample preparation step. An improved protocol for the depletion of the top-6 high-abundance proteins is suggested, which, along with the use of statistical modeling and future improvements in gel quality and image processing, can further reduce the variation and increase the efficiency of 2-D DIGE proteomic analysis of human plasma.     

Growth phase‐dependent global protein and metabolite profiles of Phaeobacter gallaeciensis strain DSM 17395, a member of the marine Roseobacter‐clade

The marine heterotrophic roseobacter Phaeobacter gallaeciensis DSM 17395 was grown with glucose in defined mineral medium. Relative abundance changes of global protein (2‐D DIGE) and metabolite (GC‐MS) profiles were determined across five different time points of growth. In total, 215 proteins were identified and 147 metabolites detected (101 structurally identified), among which 60 proteins and 87 metabolites displayed changed abundances upon entry into stationary growth phase. Glucose breakdown to pyruvate apparently proceeds via the Entner–Doudoroff (ED) pathway, since phosphofructokinase of the Embden–Meyerhof–Parnas pathway is missing and the key metabolite of the ED‐pathway, 2‐keto‐3‐desoxygluconate, was detected. The absence of pfk in other genome‐sequenced roseobacters suggests that the use of the ED pathway is an important physiological property among these heterotrophic marine bacteria. Upon entry into stationary growth phase (due to glucose starvation), sulfur assimilation (including cysteine biosynthesis) and parts of cell envelope synthesis (e.g. the lipid precursor 1‐monooleoylglycerol) were down‐regulated and cadaverine formation up‐regulated. In contrast, central carbon catabolism remained essentially unchanged, pointing to a metabolic “stand‐by” modus as an ecophysiological adaptation strategy. Stationary phase response of P. gallaeciensis differs markedly from that of standard organisms such as Escherichia coli, as evident e.g. by the absence of an rpoS gene.     

Proteome analysis of human colon cancer by two-dimensional difference gel electrophoresis and mass spectrometry

Two-dimensional difference gel electrophoresis (2-D DIGE) coupled with mass spectrometry (MS) was used to investigate tumor-specific changes in the proteome of human colorectal cancers and adjacent normal mucosa. For each of six patients with different stages of colon cancer, Cy5-labeled proteins isolated from tumor tissue were combined with Cy3-labeled proteins isolated from neighboring normal mucosa and separated on the same 2-D gel along with a Cy2-labeled mixture of all 12 normal/tumor samples as an internal standard. Over 1500 protein spot-features were analyzed in each paired normal/tumor comparison, and using DIGE technology with the mixed-sample internal standard, statistically significant quantitative comparisons of each protein abundance change could be made across multiple samples simultaneously without interference due to gel-to-gel variation. Matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) and tandem (TOF/TOF) MS provided sensitive and accurate mass spectral data for database interrogation, resulting in the identification of 52 unique proteins (including redundancies due to proteolysis and post-translationally modified isoforms) that were changing in abundance across the cohort. Without the benefit of the Cy2-labeled 12 sample mixture internal standard, 42 of these proteins would have been overlooked due to the large degree of variation inherent between normal and tumor samples.     

Investigating sample pooling strategies for DIGE experiments to address biological variability

If biological questions are to be answered using quantitative proteomics, it is essential to design experiments which have sufficient power to be able to detect changes in expression. Sample subpooling is a strategy that can be used to reduce the variance but still allow studies to encompass biological variation. Underlying sample pooling strategies is the biological averaging assumption that the measurements taken on the pool are equal to the average of the measurements taken on the individuals. This study finds no evidence of a systematic bias triggered by sample pooling for DIGE and that pooling can be useful in reducing biological variation. For the first time in quantitative proteomics, the two sources of variance were decoupled and it was found that technical variance predominates for mouse brain, while biological variance predominates for human brain. A power analysis found that as the number of individuals pooled increased, then the number of replicates needed declined but the number of biological samples increased. Repeat measures of biological samples decreased the numbers of samples required but increased the number of gels needed. An example cost benefit analysis demonstrates how researchers can optimise their experiments while taking into account the available resources.     

A novel experimental design for comparative two-dimensional gel analysis: two-dimensional difference gel electrophoresis incorporating a pooled internal standard

The comparison of two-dimensional (2-D) gel images from different samples is an established method used to study differences in protein expression. Conventional methods rely on comparing images from at least 2 different gels. Due to the high variation between gels, detection and quantification of protein differences can be problematic. Two-dimensional difference gel electrophoresis (Ettan trade mark DIGE) is an emerging technique for comparative proteomics, which improves the reproducibility and reliability of differential protein expression analysis between samples. In the application of DIGE different samples are labelled with mass and charge matched spectrally resolvable fluorescent dyes and are then separated on the same 2-D gel. Using an Escherichia coli lysate "spiked" with varying amounts of four different known proteins, we have tested a novel experimental design that exploits the sample multiplexing capabilities of DIGE, by including a standard sample in each gel. The standard sample comprises equal amounts of each sample to be compared and was found to improve the accuracy of protein quantification between samples from different gels allowing accurate detection of small differences in protein levels between samples.     

Proteomics strategy based on liquid-phase IEF and 2-D DIGE: application to bone marrow mesenchymal progenitor cells

Global comparative proteomics is a promising new approach with broad application in basic and clinical biological science. Recent advances include the development of 2-D DIGE, a proteomic equivalent to mRNA differential display, in which differentially labeled samples are multiplexed and analyzed by high-resolution 2-DE. This study presents a new 2-D DIGE protocol, in which complex protein samples from normal and leukemic human bone marrow mesenchymal progenitor cells were used as model samples for a novel combination of liquid-phase IEF with 2-D DIGE. Using liquid-phase IEF, the normal and leukemic cells were pre-fractionated into five subproteomes after multiplexing but prior to DIGE. Under these conditions, 2-D DIGE resolved >5000 protein-containing spots within the pH range 4.6-7.0. Differential labeling combined with subsequent MALDI-MS/MS identified proteins that were differentially expressed in leukemic cells. This analysis mapped protein identities to 128 mesenchymal progenitor cell proteins with at least one unique peptide match at >95% confidence. Of these proteins, 72 (56%) were expressed more than 1.25-fold higher or lower in leukemic cells compared with normal cells (p<0.05). These data were used to infer gene ontology biological processes that may be altered in leukemic bone marrow mesenchymal progenitor cells.     

Evaluation of the combination of bead technology with SELDI-TOF-MS and 2-D DIGE for detection of plasma proteins

Today biomarker discovery is one of the most active aspects of proteomic investigations. However, the wide dynamic range of plasma proteins makes the analysis very challenging because high abundance proteins tend to mask those of lower abundance. Using a large bead-based library of combinatorial peptide ligands (Equalizer beads or ProteoMiner), the dynamic range of the protein concentration is compressed, the high abundance proteins present in the sample are reduced and the low abundance proteins are enriched, while retaining representatives of all proteins within the sample. In the present study, the combination of beads with surface enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF-MS) and two-dimensional differential gel electrophoresis (2-D DIGE) technology were evaluated considering efficiency, reproducibility, sensitivity, and compatibility. The bead technology is easily compatible with both SELDI-TOF-MS and 2-D DIGE and the samples can be analyzed directly without any processing of the sample. The use of the beads prior SELDI-TOF-MS and 2-D DIGE enabled detection of many new protein spots/peaks and increased resolution and improved intensity of low abundance proteins in a reproducible fashion compared with the depletion technique. Several proteins have been identified by the combination of beads, 2-D DIGE and MS for example different kinds of complement factors and cytoskeletal proteins. Our data suggest that integration of the bead technology with our current proteomic technologies will enhance the possibility to deliver new peptide/protein biomarker candidates in our projects.     

A Comparison of Protein Extraction Methods Suitable for Gel-Based Proteomic Studies of Aphid Proteins

Protein extraction methods can vary widely in reproducibility and in representation of the total proteome, yet there are limited data comparing protein isolation methods. The methodical comparison of protein isolation methods is the first critical step for proteomic studies. To address this, we compared three methods for isolation, purification, and solubilization of insect proteins. The aphid Schizaphis graminum, an agricultural pest, was the source of insect tissue. Proteins were extracted using TCA in acetone (TCA-acetone), phenol, or multi-detergents in a chaotrope solution. Extracted proteins were solubilized in a multiple chaotrope solution and examined using 1-D and 2-D electrophoresis and compared directly using 2-D Difference Gel Electrophoresis (2-D DIGE). Mass spectrometry was used to identify proteins from each extraction type. We were unable to ascribe the differences in the proteins extracted to particular physical characteristics, cell location, or biological function. The TCA-acetone extraction yielded the greatest amount of protein from aphid tissues. Each extraction method isolated a unique subset of the aphid proteome. The TCA-acetone method was explored further for its quantitative reliability using 2-D DIGE. Principal component analysis showed that little of the variation in the data was a result of technical issues, thus demonstrating that the TCA-acetone extraction is a reliable method for preparing aphid proteins for a quantitative proteomics experiment. These data suggest that although the TCA-acetone method is a suitable method for quantitative aphid proteomics, a combination of extraction approaches is recommended for increasing proteome coverage when using gel-based separation techniques.     

Profiling the metabolic proteome of bovine mammary tissue

2-DE and MALDI mass fingerprinting were used to analyse mammary tissue from lactating Friesian cows. The goal was detection of enzymes in metabolic pathways for synthesis of milk molecules including fatty acids and lactose. Of 418 protein spots analysed by PMF, 328 were matched to database sequences, resulting in 215 unique proteins. We detected 11 out of the 15 enzymes in the direct pathways for conversion of glucose to fatty acids, two of the pentose phosphate pathway enzymes and two of the enzymes for lactose synthesis from glucose. We did not detect enzymes that catalyse the first three reactions of glycolysis. Our results are typical of enzyme detection using 2-DE of mammalian tissues. We therefore advocate caution when relating enzyme abundances measured by 2-DE to metabolic output as not all relevant proteins are detected. 2-D DIGE was used to measure interindividual variation in enzyme abundance from eight animals. We extracted relative protein abundances from 2-D DIGE data and used a logratio transformation that is appropriate for compositional data of the kind represented in many proteomics experiments. Coefficients of variation for abundances of detected enzymes were 3-8%. We recommend use of this transformation for DIGE and other compositional data.     

High-throughput proteomic analysis of human infiltrating ductal carcinoma of the breast

Large-scale proteomics will play a critical role in the rapid display, identification and validation of new protein targets, and elucidation of the underlying molecular events that are associated with disease development, progression and severity. However, because the proteome of most organisms are significantly more complex than the genome, the comprehensive analysis of protein expression changes will require an analytical effort beyond the capacity of standard laboratory equipment. We describe the first high-throughput proteomic analysis of human breast infiltrating ductal carcinoma (IDCA) using OCT (optimal cutting temperature) embedded biopsies, two-dimensional difference gel electrophoresis (2-D DIGE) technology and a fully automated spot handling workstation. Total proteins from four breast IDCAs (Stage I, IIA, IIB and IIIA) were individually compared to protein from non-neoplastic tissue obtained from a female donor with no personal or family history of breast cancer. We detected differences in protein abundance that ranged from 14.8% in stage I IDCA versus normal, to 30.6% in stage IIB IDCA versus normal. A total of 524 proteins that showed > or = three-fold difference in abundance between IDCA and normal tissue were picked, processed and identified by mass spectrometry. Out of the proteins picked, approximately 80% were unambiguously assigned identities by matrix-assisted laser desorbtion/ionization-time of flight mass spectrometry or liquid chromatography-tandem mass spectrometry in the first pass. Bioinformatics tools were also used to mine databases to determine if the identified proteins are involved in important pathways and/or interact with other proteins. Gelsolin, vinculin, lumican, alpha-1-antitrypsin, heat shock protein-60, cytokeratin-18, transferrin, enolase-1 and beta-actin, showed differential abundance between IDCA and normal tissue, but the trend was not consistent in all samples. Out of the proteins with database hits, only heat shock protein-70 (more abundant) and peroxiredoxin-2 (less abundant) displayed the same trend in all the IDCAs examined. This preliminary study demonstrates quantitative and qualitative differences in protein abundance between breast IDCAs and reveals 2-D DIGE portraits that may be a reflection of the histological and pathological status of breast IDCA.     

Correlation between mRNA and protein abundance in Desulfovibrio vulgaris: a multiple regression to identify sources of variations

Parallel profiling of mRNA and protein on a global scale and integrative analysis of these two data types could provide additional insights into the metabolic mechanisms underlying complex biological systems. However, because mRNA and protein abundance are affected by many cellular and physical processes, there have been conflicting results on their correlation. Using whole-genome microarray and LC-MS/MS proteomic data collected from Desulfovibrio vulgaris grown under three different conditions, we systematically investigate the relationship between mRNA and protein abundance by a multiple regression approach, in which some of the key covariates that may affect mRNA-protein relationship were included. The results showed that mRNA abundance alone can explain only 20-28% of the total variation of protein abundance, suggesting mRNA-protein correlation can not be determined by mRNA abundance alone. Among various covariates, analytic variation of protein abundance is the major source for the variation of mRNA-protein correlation, which contributes to 34-44% of the total variation of mRNA-protein correlation. The cellular functional category of genes/proteins contributes 10-15% of the total variation of mRNA-protein correlation, with a more pronounced correlation of the two properties was observed for "central intermediary metabolism" and "energy metabolism" categories. In addition, protein stability also contributes 5% of the total variation of mRNA-protein correlation. The study presents the first quantitative analysis of the contributions of various biochemical and physical sources to the correlation of mRNA and protein abundance in D. vulgaris.     

Biological variation of the platelet proteome in the elderly population and its implication for biomarker research

Knowledge about the extent of total variation experienced between samples from different individuals is of great importance for the design of not only proteomics but every clinical study. This variation defines the smallest statistically significant detectable signal difference when comparing two groups of individuals. We isolated platelets from 20 healthy human volunteers aged 56-100 years because this age group is most commonly encountered in the clinics. We determined the technical and total variation experienced in a proteome analysis using two-dimensional DIGE with IPGs in the pI ranges 4-7 and 6-9. Only spots that were reproducibly detectable in at least 90% of all gels (n = 908) were included in the study. All spots had a similar technical variation with a median coefficient of variation (cv) of about 7%. In contrast, spots showed a more diverse total variation between individuals with a surprisingly low median cv of only 18%. Because most known biomarkers show an effect size in a 1-2-fold range of their cv, any future clinical proteomics study with platelets will require an analytical method that is able to detect such small quantitative differences. In addition, we calculated the minimal number of samples (sample size) needed to detect given protein expression differences with statistical significance.     

Overcoming technical variation and biological variation in quantitative proteomics

Quantitative proteomics investigates physiology at the molecular level by measuring relative differences in protein expression between samples under different experimental conditions. A major obstacle to reliably determining quantitative changes in protein expression is to overcome error imposed by technical variation and biological variation. In drug discovery and development the issue of biological variation often rises in concordance with the developmental stage of research, spanning from in vitro assays to clinical trials. In this paper we present case studies to raise awareness to the issues of technical variation and biological variation and the impact this places on applying quantitative proteomics. We defined the degree of technical variation from the process of two-dimensional electrophoresis as 20-30% coefficient of variation. On the other hand, biological variation observed experiment-to-experiment showed a broader degree of variation depending upon the sample type. This was demonstrated with case studies where variation was monitored across experiments with bacteria, established cell lines, primary cultures, and with drug treated human subjects. We discuss technical variation and biological variation as key factors to consider during experimental design, and offer insight into preparing experiments that overcome this challenge to provide statistically significant outcomes for conducting quantitative proteomic research.     

激素型肾阳虚动物肝线粒体蛋白质组与能量代谢相关性

应用凝胶内差异显示电泳技术研究肾阳虚大鼠肝线粒体蛋白质组,并从肝线粒体蛋白质组角度阐述肾阳虚与能量代谢的关系.8个分别来自于肾阳虚大鼠和正常大鼠的肝线粒体蛋白质样品(各4个)分别用荧光染料Cy3、Cy5标记,以及8个样品等量混合物用Cy2标记作为内标,每一Cy3、Cy5标记样品与Cy2标记的内标等量混合后在同一胶中进行电泳分离,经不同光激发后扫描得到不同样品的蛋白质组图谱.经DeCyder软件结合内标分析,以肾阳虚组动物与正常组动物肝线粒体蛋白质相差1.2倍以上的蛋白作为差异蛋白,实验共获得16个差异蛋白质,经质谱测定和与蛋白质文库比对,鉴定11个蛋白质.其中,肾阳虚动物热休克蛋白60和70、肌氨酸脱氢酶、氨甲酰磷酸合成酶、亚硫酸盐氧化酶、ATP合酶、醛脱氢酶和NADH脱氢酶表达量增加,而丙酮酸脱氢酶、α酮戊二酸脱氢酶、脂酰辅酶A脱氢酶和鸟氨酸氨基转移酶表达量降低.实验表明,肾阳虚动物能量代谢相关酶的变化与肾阳虚的临床虚寒症状有关.