doi: 10.1161/01.RES.0000201280.20709.26
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2006 American Heart Association, Inc.
Reviews |
Proteomics of the Heart
Unraveling Disease
From the BioScience Communications (E.M.), London, UK; and Proteome Research Centre (M.J.D.), UCD Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Ireland.
Correspondence to Professor Michael J. Dunn, Proteome Research Centre, UCD Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland. E-mail michael.dunn{at}ucd.ie
This Review is part of a thematic series on Proteomics, which includes the following articles:
Cardiovascular Proteomics: Evolution and Potential
Applied Proteomics: Mitochondrial Proteins and Effect on Function
Organelle Proteomics: Implications for Subcellular Fractionation in Proteomics
Identification of Novel Signaling Complexes By Functional Proteomics
Proteomic Approaches to Analyze the Dynamic Relationships Between Nucleocytoplasmic Protein Glycosylation and Phosphorylation
Proteomics of the Heart: Unraveling Disease
Jennifer E. Van Eyk Guest Editor
 |
Abstract |
|---|
Top
Abstract
IntroductionHeart diseases resulting in heart failure are among the leading causes of morbidity and mortality in developed countries. Underlying molecular causes of cardiac dysfunction in most heart diseases are still largely unknown but are expected to result from causal alterations in gene and protein expression. Proteomic technology now allows us to examine global alterations in protein expression in the diseased heart and can provide new insights into cellular mechanisms involved in cardiac dysfunction. The majority of proteomic investigations still use 2D gel electrophoresis (2-DE) with immobilized pH gradients to separate the proteins in a sample and combine this with mass spectrometry (MS) technologies to identify proteins. In spite of the development of novel gel-free technologies, 2-DE remains the only technique that can be routinely applied to parallel quantitative expression profiling of large sets of complex protein mixtures such as whole cell lysates. It can resolve >5000 proteins simultaneously (
2000 proteins routinely) and can detect <1 ng of protein per spot. Furthermore, 2-DE delivers a map of intact proteins, which reflects changes in protein expression level, isoforms, or post-translational modifications. The use of proteomics to investigate heart disease should result in the generation of new diagnostic and therapeutic markers. In this article, we review the current status of proteomic technologies, describing the 2-DE proteomics workflow, with an overview of protein identification by MS and how these technologies are being applied to studies of human heart disease.
Key Words: proteomics • two-dimensional gel electrophoresis • mass spectrometry • cardiomyopathy • heart failure
|
Introduction |
|---|
In 2001, a major milestone was reached with the publication of the draft sequence of the human genome.1,2 The human genome was revealed to contain fewer open reading frames (ORFs;
30 000 ORFs) encoding functional proteins than was generally predicted and, like all other completed genomes, was shown to contain many "novel" genes with no ascribed functions. Moreover, it is now apparent that one gene does not encode a single protein because of processes such as alternative mRNA splicing, RNA editing, and post-translational protein modification. Therefore, the functional complexity of an organism far exceeds that indicated by its genome sequence alone. Therefore, it is clear that the "omic" approaches to the global study of the products of gene expression, including transcriptomics, proteomics, and metabolomics, will play a major role in elucidating the functional role of the many novel genes and their products and in understanding their involvement in biologically relevant phenotypes in health and disease. Powerful techniques such as cDNA and oligonucleotide microarrays and serial analysis of gene expression make it possible to undertake rapid, global transcriptomic profiling of mRNA expression. However, there is often a poor correlation between mRNA abundance and the quantity of the corresponding functional protein present within a cell.3,4 Additionally, cotranslational and post-translational modification (PTM) events can occur and result in a diversity of protein products from a single ORF.5 These modifications include phosphorylation, sulfation, glycosylation, hydroxylation, N-methylation, carboxymethylation, acetylation, prenylation, and N-myristohylation. Events such as processing of mRNA transcripts and PTMs of proteins are processes that cannot be examined at the level of mRNA, although a recently developed method of intron-specific microarrays does make it possible to examine RNA splicing.6 In addition, protein maturation and degradation are dynamic processes that can control the amount of functionally active protein within a cell.
Because of the evidence above, there is now a compelling justification for direct and large-scale analysis of proteins. The concept of mapping the human complement of protein expression was first proposed more than 20 years ago7,8 with the development of a technique in which large numbers of proteins could be separated simultaneously by 2D polyacrylamide gel electrophoresis (2-DE).9,10 However, it was not until 1995 that the term "proteome," defined as the protein complement of a genome, was first coined by Wilkins working as part of a collaborative team at Macquarie (Australia) and Sydney universities (Australia).11,12
|
From Protein Separation to Protein Identification |
|---|
Sample Preparation
Perhaps the most important step in a proteomics experiment is sample preparation. Any artifacts introduced at this stage can often be magnified, thus potentially impairing the validity of the results. No single method for sample preparation can be applied universally because of the diverse nature of samples that are analyzed by 2-DE,13 but some general considerations can be mentioned. The high-resolution capacity of 2-DE allows detection of subtle PTMs such as phosphorylation, but it will also readily reveal artifactual modifications such as protein carbamylation that can be introduced by heating protein extracts in the presence of urea. In addition, proteases present within samples can readily result in artifactual spots. Therefore, all samples, including heart, should be subjected to minimal handling, kept cold at all times, and processed in the presence of protease inhibitors.
Solubilization
The first requirement for proteome analysis is the solubilization of complex protein mixtures such as are represented by total protein extracts of human heart tissue. Ideal protein solubilization for 2-DE would result in the disruption of all noncovalently bound protein complexes and aggregates into a solution of individual polypeptides.13 Failure to achieve this will result in persisting protein complexes within the sample, resulting in new, invalid spots in the 2-DE profile, with a concomitant reduction in the intensity of those spots representing the single polypeptides. Solubilization must also permit the removal of substances, such as salts, lipids, polysaccharides, and nucleic acids, that can interfere with the 2-DE separation. Finally, the sample proteins must remain soluble during the 2-DE process. For the abovementioned reasons, sample solubility is one of the most critical factors for successful protein separation by 2-DE.
The most popular method for protein solubilization for 2-DE remains that originally described by O’Farrell10 using a mixture of 9.5 mol/L urea, 4% wt/vol of the nonionic detergent Nonidet P-40, or the zwitterionic detergent CHAPS (3[(cholamidopropyl)dimethylammonio]-1-propane sulfonate), 1% wt/vol of the reducing agent dithiothreitol (DTT), and 2% wt/vol of synthetic carrier ampholyte of the appropriate pH range (so-called "lysis buffer"). Although this method works well for many types of sample, it is not universally applicable, with membrane proteins representing a particular challenge.14 Variations in solubilization buffer constituents using newly developed detergents such as sulfobetaines,15 additional denaturing agents such as thiourea,16 and alternative reducing agents (eg, trubutyl phosphine)17 can help to improve protein solubilization and hence the concentration of extracted protein for certain sample types. It must be stressed that the choice of solubilization buffer must be optimized for each sample type to be analyzed by 2-DE. In our hands, the following solubilization buffer, 9.5 mol/L urea, 2% (wt/vol) CHAPS, 1% (wt/vol) DTT, 0.8% (v/v) Pharmalyte pH 3 to 10, and protease inhibitors, continues to give efficient and reproducible lysis of human heart; Stanley et al provide an example of optimizing the solubilization of human myocardium.18
Two-Dimensional Gel Electrophoresis
The basic technique of 2-DE in which proteins are separated in the first dimension according to their charge properties (isoelectric point [pI]) under denaturing conditions, followed by their separation in the second dimension according to their relative molecular mass (Mr) by SDS-PAGE, was developed 30 years ago.9,10 Nevertheless, it remains the core technology of choice for the majority of applied proteomic projects13,19,20 because of its ability to separate simultaneously thousands of proteins and to indicate PTMs that result in alterations in protein pI or Mr. Additional advantages are the high-sensitivity visualization of the resulting 2D separations, compatibility with quantitative computer analysis to detect differentially regulated proteins, and the relative ease with which proteins from 2-DE gels can be identified and characterized by mass spectrometry (MS).
For proteome analysis, it is essential that 2-DE should generate highly reproducible 2D protein separations. Developments over the last few years, particularly those involving the use of immobilized pH gradients (IPGs) for the first-dimension isoelectric focusing (IEF) separation, have resulted in the current 2-DE method that combines increased resolving power and high reproducibility with relative simplicity of use. Details of these developments can be found in recent reviews19,21,22 and have been used recently in a study evaluating the capabilities and limitations of several 2-DE techniques for separating proteins from rat heart tissue.23
Increasing Proteomic Coverage by 2-DE
Large-format 2D gels using wide pH 3 to 10 gradients are able to separate 2000 proteins from a complex sample such as a total myocardial protein lysate (Figure 1). However, it is clear that this gives incomplete proteomic coverage for tissues such as the human heart, which may be expressing >10 000 proteins at any given time. This inability to cope with the enormous diversity of cellular proteins often results in several protein species comigrating in the same spot on a 2-DE gel.24 However, intermediate (eg, pH 4 to 7, 6 to 9) and narrow (eg, pH 4.0 to 5.0, 4.5 to 5.5, 5.0 to 6.0, 5.5 to 6.7)-range IPG IEF gels are now available and have the capability of "pulling apart" this protein profile and increasing resolution in particular regions. This "zoom gel" approach results in enhanced proteomic coverage, with the added advantage that narrow range IPGs can tolerate higher protein loading.25,26 However, the disadvantage of this approach is the increased workload if large numbers of samples are to be investigated.
|
Another problem associated with 2-DE protein coverage is the difficulty of obtaining good separations of basic proteins that represent a significant proportion of the total predicted proteome of most eukaryotic organisms. Basic proteins have a tendency to form pronounced streaks rather than discrete spots attributable to electroendosmotic effects, the migration of reducing agents such as DTT and the potential hydrolysis of acrylamide at basic pH values.27 These effects can be partially overcome using DTT replenishment during IEF combined with the inclusion of glycerol or isopropanol to suppress electroendosmosis28,29 or using the reagent hydroxyethyldisulphide that prevents the reformation of disulphide bonds.30,31
Increasing Proteomic Coverage by Cellular, Subcellular, and Protein Fractionation
A major limitation to the capacity of 2-DE to display complete proteomes is the very high dynamic range of protein abundance, estimated at 106 for cells and tissues32 and 1012 for plasma.33 This is beyond the dynamic range of 2-DE, with an estimated maximum dynamic range of 104.19 Reproducible sample fractionation methods will therefore be essential to enrich low-abundance proteins present in biological samples such as the heart.
Two general strategies can be used: cellular/subcellular fractionation and protein fractionation. The former methods include immunoisolation, electromigration (eg, free flow electrophoresis), flow cytometry, density gradient isolation of organelles such as mitochondria,34,35 isolation of membranes,20 and sequential differential extraction using a series of reagents with increasing solubilizing power.36 A particular problem in proteomic analysis of the heart is the diversity of cell types that are present. Proteomic profiles of total myocardial lysates are dominated by the proteins present in cardiac myocytes, but such samples will also contain lower amounts of proteins derived from other cell types, including fibroblasts, smooth muscle cells, and endothelial cells. Recently, we started to apply the technique of laser capture microdissection (LCM), in which a laser beam is used to isolate specific regions of interest from microscope sections of tissue. Although this technique generally results in the isolation of relatively small amounts of material, it has been shown to be possible to perform proteomic studies of the resulting protein samples.37–39 In preliminary studies, we have been able to generate sufficient material by LCM of human cardiac tissue sections to produce large-format 2D gels of proteins from isolated cardiac myocytes and microvessels40 (Figure 2).
|
Methods for fractionation of solubilized proteins include electrophoretic separation of the proteins in solution by techniques including continuous free flow electrophoresis,41 recycling IEF,42 and IEF using multicompartment electrolysers.43 The latter approach seems particularly promising for use in conjunction with narrow-range pH gradient IPG IEF for 2-DE.44 Protein mixtures can also be subfractionated by traditional chromatographic techniques such as reverse phase, ion exchange, size exclusion, affinity chromatography, and chromatofocusing.
Alternatives to 2-DE
The limitations inherent in the 2-DE approach have resulted in significant efforts to develop alternative approaches that do not depend on 2-DE and in particular avoid the use of IEF with its attendant problems of protein solubility. The simplest approach is the use of 1D SDS-PAGE in conjunction with protein identification by MS/MS so that several proteins comigrating in a single band can be identified. However, this method is limited by the complexity of the protein mixture that can be analyzed and has been most successfully applied to the study of protein complexes that can be isolated by techniques such as immunoprecipitation.45 Other approaches have been designed to dispense with the need for a gel-based separation and to take advantage of the automation, throughput, and sensitivity that can be provided by techniques based on a combination of liquid chromatography (LC) and MS. In these so-called "shotgun" approaches, the complex protein sample is initially digested with an enzyme (typically trypsin). The resulting peptide fragments are then separated by one or more dimensions of LC to reduce the complexity of peptide fractions that are then introduced (either online or offline) into a tandem mass spectrometer for sequence-based identification. For example, the so-called "MudPIT" approach of Yates46 combining multidimensional LC, tandem MS, and database searching has been shown to be capable of detecting and identifying 1500 yeast proteins in a single analysis.47 More recently, a comprehensive study investigating normal and diseased heart tissue has used the MudPIT technique, readily achieving functional enrichment, high-confidence identification, and relative quantification of hundreds of organelle and tissue-specific proteins, including detection of low-abundance transcriptional regulators, signaling factors, and proteins linked to cardiac disease.48 A single-dimension LC separation, combined with the very high resolution of a Fourier-transform ion cyclotron resonance mass spectrometer, was able to identify nearly 1900 proteins (>61% of the predicted proteome) of the bacterium Deinococcus radiourans.49 Although these approaches are certainly very powerful, it is important to realize that they generate raw lists of proteins present in a sample, with no information on the relative quantitative abundance of the individual protein components and give no indication as to whether the proteins identified are subject to PTM.
This problem is currently being addressed by the development of MS-based techniques in which stable isotopes are used to differentiate between two populations of proteins. This approach consists of four steps: (1) differential isotopic labeling of the two protein mixtures, (2) digestion of the combined labeled samples with a protease such as trypsin or Lys-C, (3) separation of the peptides by multidimensional LC, and (4) quantitative analysis and identification of the peptides by MS/MS. The most widely used method that is now commercially available is the isotope-coded affinity tag (ICAT) method based on the labeling of cysteine residues,50,51 limiting its application to proteins containing these residues. Recently, there has been a proliferation of different isotopic labeling methods for MS-based quantitative proteomics.52,53 These include quantitative methods such as iTRAQ,53,54 whereby the N terminus of peptides, generated from two or more protein digests to be compared, are chemically tagged, combined, fractionated, and analyzed by tandem MS for protein identification and quantitation. The advantages of this approach are increased sensitivity and the fact that PTMs can be analyzed; however, the major limitation of the iTRAQ method is that MS/MS analysis has to be performed on all eluting peptides to identify only a small set.55 Novel labeling techniques are reviewed in Julka and Regnier.55 It seems likely that this group of methods will be increasingly used in proteomic investigations, but their quantitative reproducibility remains to be established. Moreover, the dynamic range of the ICAT technique has been reported to be no better than 2-DE56 and, therefore, can be complementary to a 2-DE approach in identifying a different subset of proteins from a given sample.22 Finally, there is much interest in the development of antibody and protein arrays for quantitative expression profiling,57–59 but considerable work remains to be performed before this approach can be routinely used in proteomic investigations. However, a recent study by Horn et al profiled the autoantibody repertoire of plasma from dilated cardiomyopathy (DCM) patients against a human protein array consisting of 37 200 redundant, recombinant human proteins. They were able to perform qualitative and quantitative validation of these putative autoantigens on protein microarrays to then be able to identify novel putative DCM specific autoantigens. As well as analyzing the whole IgG autoantibody repertoire, Horn et al also analyzed the IgG3 antobody repertoire in the same plasma samples to gain an understanding of the characteristics of IgG3 subclass antibodies. By combining screening a protein expression library with protein microarray technology, Horn et al were able to detect 26 proteins by the IgG antibody repertoire and 6 proteins bound by the IgG3 subclass. Several of the autoantobodies identified in the plasma of DCM patients are known to be associated with heart failure.60
It is therefore clear that until these alternative approaches mature into robust techniques for quantitative protein expression profiling, 2-DE will remain the separation workhorse in many proteomic investigations. This technique has the capacity to support the simultaneous analysis of the changes in expression of hundreds to thousands of proteins, and, as such, it remains unrivaled as an "open"’ protein expression profiling approach.
Protein Detection and Visualization
After 2-DE, the separated proteins must be visualized at high sensitivity, but there is a need for the detection method to combine the properties of an extended dynamic range, a linear staining response, and, if possible, to be compatible with downstream protein identification by MS. Silver staining with its high sensitivity (0.1 ng protein) has, until recently, been the method of choice, but its limited dynamic range and restricted quantitative capacity have driven the development of alternative detection methods based on the use of fluorescent compounds. The SYPRO dyes, in particular SYPRO Ruby, are currently the most appropriate postelectrophoretic stains because these combine high sensitivity with an extended dynamic range for improved quantitation with MS compatibility.22,61
A pre-electrophoretic fluorescent staining method based on the labeling of protein samples with N-hydroxy succinimidyl ester derivatives of fluorescent cyanine (Cy) dyes and known as 2D difference gel electrophoresis62,63 is currently being widely used. This approach has the advantage that a pair of protein samples can be labeled separately with Cy3 and Cy5 derivatives. The two samples can be mixed and then separated together on the same 2D gel. The resulting 2D gel is then scanned to acquire the Cy3 and Cy5 images separately. Improved quantitative accuracy of comparison of multiple pairs of samples can be achieved using a pooled internal standard labeled with a third dye: Cy2.64,65 Recently, saturation labeling with cysteine-reactive cyanine fluorescent dyes has been described.66 This technique provides increased sensitivity for expression profiling of scarce samples such as laser-microdissected clinical specimens.67
In addition to the aforementioned dyes, a new range of fluorescent dyes has recently become popular. It is now possible to stain 2D gels for specific proteins such as those that are in a phosphorylated/hyper-phosphorylated or glycosylated state. Pro-Q Diamond dye is a new fluorescent phosphosensor technology suitable for the detection of phosphoserine-, phosphothreonine- and phosphotyrosine-containing proteins directly in IEF gels, SDS–polyacrylamide gels, and 2-DE gels. In addition, Pro-Q Diamond can be used for the detection of phosphoproteins or phosphopeptides arrayed on protein chips or affixed to beads.68 Pro-Q Emerald 300 fluorescent stain can be used for the detection of glycoproteins in polyacrylamide gels with as little as 2 to 4 ng of lipopolysaccharide being detectable in contrast to 250 to 1000 ng required for detection with conventional silver staining.69 In both cases, 2D gels can be poststained with SYPRO Ruby dye, allowing sequential two-color detection of either phosphorylated and unphosphorylated proteins or glycosylated and nonglycosylated proteins.
Protein Identification
MS has become the technique of choice for protein identification because these methods are very sensitive, require small amounts of sample (femtomole to attomole concentrations), and have the capacity for high sample throughput.70–72 The primary tool for protein identification is typically peptide mass fingerprinting (PMF). This technique is based on a set of peptide masses obtained by MS analysis of a protein digest (usually trypsin) that provides a characteristic mass fingerprint of a protein. The protein is then identified by comparison of the experimental mass fingerprint with theoretical peptide masses generated in silico using protein and nucleotide sequence databases. This approach proves very effective when trying to identify proteins from species whose genomes are completely sequenced but is not so reliable for organisms whose genomes have not been completed. This has been a problem in the past for proteomic studies of some animal models of heart disease, for example, those involving rats, dogs, pigs, and cows, but has been shown to be overcome effectively by improving PMF by adopting an orthogonal approach combined with amino acid compositional analysis.73
If it proves impossible to identify a protein based on PMF alone, it is then essential to obtain amino acid sequence information. This can be generated by conventional automated chemical Edman microsequencing but is most readily accomplished using tandem MS (MS/MS). MS/MS takes advantage of two-stage MS instruments, matrix-assisted laser desorption ionization (MALDI)–MS with postsource decay, MALDI–time-of-flight (TOF) -TOF-MS/MS, or electrospray ionization (ESI)-MS/MS triple-quadropole, ion-trap, or Q-TOF machines, to induce fragmentation of peptide bonds. One approach is to generate a short partial sequence or "tag," which is used in combination with the mass of the intact parent peptide ion to provide significant additional information for the homology search.74 A second approach uses a database-searching algorithm SEQUEST75 to match uninterpreted experimental MS/MS spectra with predicted fragment patterns generated in silico from sequences in protein and nucleotide databases.
Bioinformatics
Once protein data has been generated, using the methods described previously, it needs to be collated in a systematic way. Bionformatics plays a central role in doing this. It is a fundamental tool for quantitative analysis of differential patterns of protein expression in 2D gels,76 and there are a variety of bioinformatic tools for identifying proteins based on MS and other chemical data. Most of these tools with their associated databases are available on the Internet through the World Wide Web (WWW) and can be accessed through the ExPASy proteomics server. A major problem in proteomics will be integrating the output from large-scale proteomic studies. This is essential if we are to exploit the information on changes in the expression of what can be large numbers of proteins of diverse functions to understand their role in biological systems and in dysfunction in disease. This can be driven in an hypothesis-driven way by investigating individual (or groups) proteins in more detail. However, it is unlikely that a full understanding will be generated at the level of individual proteins and that this will only be achieved by understanding how proteins act within the context of subcellular, cellular, tissue, organ, and whole animal structures. Thus, it will be essential to compute these interactions and networks to determine the functional system involved in health and disease. This has been termed "systems biology," and a recent article77 illustrates the development of this approach in the case of the heart. In addition, the Human Proteome Organization Proteomics Standards Initiative, established four years ago, aims to define community standards for data representation in proteomics, thus facilitating data comparison, exchange, and verification.78,79
|
Heart 2-DE Protein Databases |
|---|
There are four gel protein databases of cardiac proteins, established by three independent groups, that can be accessed via the WWW (Table). These databases facilitate proteomic research into heart diseases containing information on several hundred cardiac proteins that have been identified by protein chemical methods. They all conform to the rules for federated 2-DE protein databases.80 In addition, 2-DE protein databases for other mammals, such as the mouse, rat,81 dog,82 pig, and cow, are also under construction to support work on animal models of heart disease and heart failure.
|
|
Dilated Cardiomyopathy |
|---|
DCM is a disease of unknown etiology. It is a severe disease characterized by impaired systolic function resulting in heart failure. To date, proteomic investigations into human heart disease have centered around DCM. The known contributory factors of DCM are viral infections, cardiac-specific autoantibodies, toxic agents, genetic factors, and sustained alcohol abuse. The expression of as many as 100 cardiac proteins has been observed to significantly alter in their expression in DCM, with the majority of these proteins being less abundant in the diseased heart. This has been reported in numerous studies.81,82,86–90 Identification of many of these proteins has been via chemical methods such as MS,86,91–93 classifying them into three broad functional classes: cytoskeletal and myofibrillar proteins; proteins associated with mitochondria and energy production; and proteins associated with stress responses.94
Investigating the contribution of these changes to altered cellular function underlying cardiac dysfunction is now a major challenge. This has already begun. For example,59 isoelectric isoforms of 27-kDa heat shock protein (HSP27) have been observed to be present in human myocardium using traditional 2-DE large format gels. Twelve of these protein spots in the pI range of 4.9 to 6.2 and mass range of 27 000 to 28 000 Da were significantly altered in intensity in myocardium taken from patients with DCM. Ten of these protein spots were significantly changed in myocardium taken from patients with ischemic heart failure.90
More recently, because of advances in the characterization of the phosphoproteome, Fernando et al were able to map the signal perturbations, in mouse postnatal hearts, that were associated with activation of a myopathic cascade as mediated by the mitogen activated protein kinase mitogen-activated protein kinase kinase (MKK) 6. They demonstrated that MKK6 promoted the development of cardiomyopathy through multiple substrate interactions.95
|
Subproteomics of the Heart: Protein Kinase C Signal Transduction Pathways |
|---|
Protein kinase signal transduction pathways have been studied extensively and characterized in the myocardium. Much work has focused on studying the role of individual kinases involved in ischemic preconditioning (IP). IP describes the reduction in susceptibility to myocardial infarction that follows brief periods of sublethal ischemia.96 This reduction can manifest itself as a 4-fold reduction in infarct size, this being secondary to a delay in the onset and rate of cell necrosis during the subsequent lethal ischemia.97
The involvement of protein kinase C (PKC) in preconditioning was first suggested by Downey et al.98 They demonstrated that pharmacological inhibition of PKC blocked IP and that the infarct-sparing effects of IP could be mimicked by activating PKC using phorbol esters. However, although many groups have repeated this experiment,99–102 it is still uncertain which PKC isoform(s) is/are responsible.
PKC-
activation has been shown to play a significant role in protection using isoform-specific inhibitory peptides, which are able to abolish protection in response to IP.103,104 The binding of a specific PKC isoform is thus prevented from localizing with its substrate(s), and there is a loss of function. However, the selectivity of this approach has been brought into question. Another approach to validating PKC- as having a cardioprotective role is to use animal models in which the gene of interest, in this case that encoding PKC-, is lacking.105
A functional proteomic approach to investigate PKC-–mediated cardioprotection has been adopted by Ping et al.106 In their approach, immunoprecipitation, using PKC- monoclonal antibodies, was performed to isolate PKC- complexes. This "subproteome" is then separated out using 1D/2D electrophoresis and putative candidate proteins, which associate with PKC-, are then identified using Western blotting/MS–based techniques. To validate this proteomic data, the colocalization of candidate proteins with the PKC- complex is established using PKC-–glutathione-S-transferase (GST) affinity pull-down assays. Expression of these candidate proteins in cardiac cells is then confirmed using isolated mouse cardiac myocytes.106 Using this approach, Ping et al found that within the "PKC- subproteome," PKC- forms complexes with at least 93 proteins in the mouse heart.107–111 The identified proteins can be separated into six different classes of molecule incorporating structural proteins, signaling molecules, metabolism-related proteins, stress-activated proteins, transcription- and translation-related proteins, and PKC- binding domain containing proteins.108,110 More recently, the same research group has demonstrated for the first time assembly of PKC--Akt–endothelial NO synthase (eNOS) signaling modules in vitro and in the mouse heart. These modules are implicated in cardiac protection and provide a mechanistic link between PKC-, Akt, and eNOS, associated previously with such protection.112
|
Subproteomics of the Heart: Mitochondria |
|---|
Mitochondria are involved in a variety of cellular processes. Their primary role is the production of ATP, but they are also involved in ionic homeostasis, apoptosis, oxidation of carbohydrates and fatty acids, and a variety of other catabolic and anabolic pathways. As a consequence of this functional diversity, mitochondrial dysfunction is involved in a variety of diseases including heart disease.113 Characterization of the mitochondrial proteome could thus provide new insights into cardiac dysfunction in heart disease.114
It has been predicted that the human mitochondrial proteome comprises 1500 distinct proteins.115 This level of complexity should be capable of being addressed using the currently available panel of proteomic technologies. Providing that sufficient tissue is available, mitochondria can be purified relatively easily by differential centrifugation114 and the mitochondrial proteins separated by 2-DE and identified by PMF. The most comprehensive report using this approach is a study of rat liver mitochondria in which broad- and narrow-range IPG 2-DE gels resulted in the identification of 192 different gene products from a total of 1800 protein spots, of which 70% were enzymes with a broad spectrum of catalytic activities.34 A similar study identified 185 different gene products from 600 protein spots in the mitochondrial proteome of the neuroblastoma cell line IMR-32.116 This approach has not been systematically applied to the study of the mitochondrial proteome of the heart, although the 2-DE/PMF approach has been applied in differential expression studies of hearts from knockout mouse strains deficient in creatine kinase35 and mitochondrial superoxide dismutase.114 More recently, Liu et al used a proteomics approach to investigate changes in cardiomyocyte mitochondrial protein expression in a model of chronic restraint stress (inducing cardiac dysfunction as well as cardiomyocytes injury) in the rat. Compared with control, 11 protein spots were found to alter in their expression after chronic restraint stress. Seven of these proteins were identified by MALDI-TOF MS. Five of these proteins involved in the Krebs cycle and lipid metabolism in mitochondria decreased after chronic restraint stress. They were identified as carnitine palmitoyltransferase2, mitochondrial acyl-CoA thioesterase 1, isocitrate dehydrogenase 3 (NAD+) 
, fumarate hydratase 1, and pyruvate dehydrogenase ß. Two of the identified proteins, creatine kinase and prohibitin, increased after chronic restraint stress.117
Proteomics based on the 2-DE approach experiences problems associated with the analysis of membrane proteins so that such mitochondrial proteins are poorly represented on 2D profiles. In addition, many mitochondrial proteins are more basic than cytosolic proteins, mitochondria are rich in low molecular weight (<10 kDa) proteins, and mitochondrial proteins are poorly described in databases.118 In an attempt to overcome some of these hurdles, Pflieger et al119 separated the proteins from isolated yeast mitochondria by 1D SDS-PAGE to overcome the problems associated with the IEF dimension of 2-DE. The SDS gel was then cut into 27 slices of 2 mm, and tryptic digests of the proteins contained in these bands were analyzed by LC-MS/MS. This approach resulted in the identification of 179 gene products (ie, similar to the number identified from isolated rat liver mitochondria by 2-DE/MS).34 However, these proteins represented a broader range of proteins than covered by the 2-DE–based approach, with their physicochemical properties spanning a wide range of pI, Mr and hydrophobicity.
An alternative approach to increasing proteomic coverage is based on the analysis of isolated intact mitochondrial protein complexes. One strategy is based on the use of sucrose density gradients to separate intact mitochondrial complexes solubilized with N-dodecyl-ß-D-maltoside.120 Initially, the proteins from the individual fractions were analyzed by 2-DE. However, subsequently, this approach has been coupled with 1D SDS-PAGE and MALDI PMF analysis of tryptic digests of excised protein bands.115 When applied to human heart mitochondria, this approach resulted in the identification of 615 bona fide or potential mitochondrial proteins, many of which had not been reported previously using 2-DE.115 Proteins with a wide range of pI, Mr, and hydrophobicities were reported with a high coverage of the known subunits of the oxidative machinery of the inner mitochondrial membrane. A significant proportion of the identified proteins are associated with signaling, RNA, DNA, and protein synthesis, ion transport, and lipid metabolism.115 In a recent study of complex I purified from bovine heart mitochondria, three independent separation methods (1-D SDS-PAGE, 2-DE, and reverse-phase high-performance LC) combined with MALDI-PMF and ESI-MS/MS were used, and the intact enzyme was shown to be an assembly of 46 different proteins.121
Another approach to overcoming the limitations inherent in the IEF dimension of 2-DE is to use alternative types of 2D separations. 2D blue native (BN) electrophoresis122 can be used to separate membrane and other functional protein complexes as intact, enzymatically active complexes in the first dimension. This is followed with a second-dimension separation by Tricine–SDS-PAGE to separate the complexes into their component subunits. This method, combined with protein identification by MALDI PMF, has been applied to several studies of the mitochondrial proteome.123,124 In a study of human heart mitochondria using BN/SDS-PAGE, the individual subunits of all five complexes of the oxidative phosphorylation system were represented, and a novel variant of cytochrome c oxidase subunit Vic was reported.125
Additional 2D systems have been optimized for the analysis of membrane proteins. Almost 10 years ago, Hartinger et al adapted a previously published procedure126 for the separation of integral membrane proteins. Discontinuous gel electrophoresis in an acidic buffer system using the cationic detergent benzyldimethyl-n-hexadecylammonium chloride was used in the first dimension followed by discontinuous SDS-PAGE in the second dimension. They were able to demonstrate that complex membrane protein mixtures can be resolved with a resolution 
5-fold higher than that of 1D SDS-PAGE.127
Plasma membrane proteins from the yaest Saccharomyces cerevisiae were successfully isolated and analyzed using a deoxycholate stripping protocol to remove cytosolic proteins from a plasma membrane fraction and sucrose gradient flotation to remove ribosomal proteins. Plasma membrane proteins were then resolved by 2-DE using the cationic detergent cetyl tromethyl ammonium bromide in the first and SDS in the second dimension.128 More recently, Rais et al varied parameters such as acrylamide concentration, urea content, and the trailing ion used for SDS gels to modify protein mobilities. Coupling two SDS gels to a 2-DE system, hydrophobic (membrane) proteins were successfully separated from water soluble (cytosolic) proteins, allowing identification by MS of previously unaccessible hydrophobic proteins.129
|
Animal Models of Heart Disease |
|---|
Investigations of human diseased tissue samples can be compromised by factors such as the disease stage, tissue heterogeneity, genetic variability, and the patient’s medical history/therapy. Avoiding any of the above complications when working with human samples can prove to be extremely difficult. An alternative approach is to apply proteomics to appropriate models of human disease, animal models being an attractive alternative.
There are several models of cardiac hypertrophy, heart disease, and heart failure in small animals, particularly the rat. Proteomic analysis of these models has generally focused on changes in cardiac proteins in response to alcohol130,131 and lead132 toxicity.
More recently, Fernando et al used a gel-based kinase assay coupled to MS identification as an approach to map global kinase activity in the context of cardiomyopathy in the postnatal mouse heart.95 Faber et al133 performed a differential proteomic profiling study on right ventricular hypertrophy using a rat model of pulmonary artery banding. The cytoplasmic fraction showed altered expression of metabolic proteins indicative of a shift from fatty acid to glucose as a substrate for energy supply and upregulation of three isoforms of HSP27, suggesting an altered stress response during hypertrophy. The myofilament fraction showed upregulation of desmin and -B-crystallin. Using rabbits, White et al134 used proteomics to characterize global changes in cardiac protein expression in response to ischemia/reperfusion injury. A total of 53 protein spots were affected, and the identified proteins were from four functional classes: (1) the sarcomere and cytoskeleton, (2) redox regulation, (3) energy metabolism, and (4) the stress response.
Unfortunately, the cardiac physiology of small animal models and their normal pattern of gene expression (eg, isoforms of the major cardiac contractile proteins) differ from that in larger mammals such as humans. Therefore, investigations have moved into higher mammals, and two proteomic studies of heart failure in large animals have been published. One study investigated pacing-induced heat failure in the dog,135,136 whereas the second, based in our laboratory, investigated bovine DCM.137 Both studies demonstrated shared similarities with the proteome analysis of human DCM, with the majority of changes involving reduced protein abundance in the diseased heart.
Identifying altered canine and bovine proteins has proved to be particularly challenging because these species are poorly represented in current genomic databases. As a result of this, new bioinformatic tools (MultiIdent) have had to be developed to facilitate cross-species protein identification.138 The most significant change observed for bovine DCM was a 7-fold increase in the enzyme ubiquitin carboxyl-terminal hydrolase (UCH).137 This could potentially facilitate increased protein ubiquitination in the diseased state, leading to proteolysis via the 26S proteosome pathway. Interestingly, there is evidence to suggest that inappropriate ubiquitination of proteins could contribute to the development of heart failure.139
More recently, we investigated whether the ubiquitin-proteosome system is perturbed in the heart of human DCM patients.140 As in bovine DCM, expression of the enzyme UCH was >8-fold elevated at the protein level and >5-fold elevated at the mRNA level in human DCM. Moreover, this increased expression of UCH was shown by immunocytochemistry to be associated with the myocytes that do not exhibit detectable staining in control hearts. Overall, protein ubiquitination was increased 5-fold in DCM relative to control hearts, and using a selective affinity purification method, we were able to demonstrate enhanced ubiquitination of a number of distinct proteins in DCM hearts. We identified a number of these proteins by MS. Interestingly, many of these proteins were the same proteins that we found previously to be present at reduced abundance in DCM hearts.86 This new evidence strengthens our hypothesis that inappropriate ubiquitin conjugation leads to proteolysis and depletion of certain proteins in the DCM heart and may contribute to loss of normal cellular function in the diseased heart.
|
Proteomics of Cultured Cardiac Myocytes |
|---|
Cell culture systems are attractive models for proteomic analysis because they should provide defined systems with much lower inherent variability between samples, particularly if established cell lines are used. Cell culture systems are therefore ideal for detailed proteomic investigations of responses in protein expression to controlled stimuli. However, it is clear that cells maintained in culture respond by alterations in their pattern of gene, and consequently protein, expression such that this can be quite different from that found in vivo. This process can occur quite rapidly in primary cultures of cells established from tissue samples and is even more profound in cells maintained long term, particularly where transformation has been used to establish immortal cell lines. The situation is even worse in the case of cardiac myocytes. Although neonatal cardiac myocytes can be maintained and grown in vitro, adult cells are terminally differentiated and can be maintained for relatively short times in vitro but are not capable of cell division.
In fact, there are relatively few published proteomic investigations of isolated cardiac myocytes. In a study of beating neonatal rat cardiac myocytes, 2-DE was used to investigate the regulation of protein synthesis by catecholamines.141,142 Marked changes in protein expression were observed in response to treatment with norepinephrine.142 The use of the -adrenoceptor blocker prazosin allowed a clear classification of - and non- (probably ß)–adrenoceptor-mediated catecholamine effects on protein expression.141 Unfortunately, none of the proteins could be identified at the time this studied was carried out.
Arnott et al143 used phenylephrine-treated neonatal rat cardiac myocytes as a model of cardiac hypertrophy for proteomic analysis. In this 2-DE–based study, 11 protein spots were found to display statistically significant changes in expression on induction of hypertrophy. Of these, 8 showed higher expression, and 3 were decreased in abundance in hypertrophied cells. All of these proteins were successfully identified by a combination of PMF by MALDI-TOF and partial sequencing by LC-MS/MS. The atrial isoforms of myosin light chains (MLCs) 1 (2 spots) and 2 were increased, as was the ventricular isoforms of MLC2 (2 spots). Other proteins that increased were chaperonin cofactor a, nucleoside diphosphate kinase a, and HSP27. The proteins that were decreased were identified as mitochondrial matrix protein p1 (2 spots) and NADH ubiquinone oxidoreductase 75-kDa subunit. The changes in expression of MLC isoforms are consistent with previous studies of the expression of MLC isoforms in cardiac hypertrophy by Northern blot analysis of mRNA and by immunofluorescence. In contrast, the changes in expression of the other proteins had not been shown previously to be associated with cardiac hypertrophy.143
In a similar study, endothelin (ET) was used to induce hypertrophy in neonatal rat cardiac myocytes. ET treatment was found by 2-DE to result in a 2-fold decrease in 21 proteins compared with the levels in untreated cells.144 The ET-induced hypertrophy was accompanied by a 30% increase in MLC1 and MLC2.
Isolated adult rabbit cardiac myocytes have been used in a proteomic study of myocardial IP, induced pharmacologically with adenosine.145 Here, a subproteomic approach was used in which cytosolic and myofilament-enriched fractions were analyzed by 2-DE. Various adenosine-mediated changes in protein expression were detected in the cytosolic fraction, but these proteins were not subsequently identified. The most striking finding was novel PTM of MLC1 that was shown to be attributable to phosphorylation at two sites.145 The functional significance of this finding to preconditioning remains to be established. In a recent study of adult human cardiac myocytes from patients with end-stage heart failure, two protein spots associated with MLC1 were observed on 2-DE gels.146 These two spots may represent phosphorylated and nonphosphorylated MLC as described by Arrell et al,145 but the amount of the putatively phosphorylated form was not found to differ between failing and control (transplant donor) samples.
|
Proteomic Characterization of Cardiac Antigens in Heart Disease and Transplantation |
|---|
Proteomics can be used to identify cardiac-specific antigens that elicit antibody responses in heart disease and after cardiac transplantation. This approach makes use of Western blot transfers of 2D gel separations of cardiac proteins. These are probed with patient serum samples and developed using appropriately conjugated antihuman immunoglobulins. This strategy has revealed several cardiac antigens that are reactive with autoantibodies in DCM147,148 and myocarditis.149 As discussed previously, using human protein arrays, the autoantibody repertoire of plasma from DCM patients has been profiled and novel DCM specific autoantigens identified.60 Cardiac antigens that are associated with antibody responses after cardiac transplantation have also been characterized and may be involved in acute150 and chronic151 rejection. Recently, we used a proteomic approach to identify whether specific protective proteins are expressed in the hearts of long-term transplant patients who remain free of chronic rejection. Our results suggest that expression of a specific diphosphorylated form of the small HSP27 is associated with healthy blood vessels.152
More recently, a proteomics approach was used to identify markers of cardiac allograft rejection in human serum. Cardiac biopsies were taken from nonrejecting and rejecting patients and analyzed initially by 2-DE. Thirteen proteins, found to be upregulated during rejection, were identified as either cardiac specific or heat shock proteins. Two of these proteins (B-crystallin and tropomyosin) were further investigated by developing an in-house ELISA. Serum levels of B-crystallin and tropomyosin significantly higher in sera associated with rejection, confirming 2-DE data.153
|
Concluding Remarks |
|---|
Established proteomic technologies, together with the new and alternative strategies currently under development, are now making it possible to address important issues in biomedicine. Here, we attempted to illustrate how proteomics is being used to characterize protein expression in the human heart and to investigate changes in protein expression associated with cardiac dysfunction in disease. The proteomics studies that have been performed on cardiac tissue from both human patients and appropriate animal models are providing new insights into the cellular mechanisms involved in cardiac dysfunction. In addition, they should result in the discovery of new diagnostic or prognostic biomarkers and the identification of potential drug targets for the development of new therapeutic approaches for combating heart disease.
|
Acknowledgments |
|---|
M.J.D. is recipient of a Science Foundation Ireland Research Professorship and work in his laboratory is supported by SFI under grant no. 04/RPJ/B499.
|
Footnotes |
|---|
Original received September 14, 2005; revision received December 2, 2005; accepted December 14, 2005.
|
References |
|---|
- Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, et al; International Human Genome Sequencing Consortium. Initial sequencing and analysis of the human genome. Nature. 2001; 409: 860–921.[CrossRef][Medline] [Order article via Infotrieve]
- Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, Smith HO, Yandell M, Evans CA, Holt RA, et al. The sequence of the human genome. Science. 2001; 291: 1304–1351.
[Abstract/Free Full Text] - Anderson L, Seilhamer J. A comparison of selected mRNA and protein abundances in human liver. Electrophoresis. 1997; 18: 533–537.[CrossRef][Medline] [Order article via Infotrieve]
- Gygi SP, Rochon Y, Franza BR, Aebersold R. Correlation between protein and mRNA abundance in yeast. Mol Cell Biol. 1999; 19: 1720–1730.
[Abstract/Free Full Text] - Gooley AA, Packer NH. The importance of protein co- and post-translational modifications in proteome projects. In: Wilkins MR, Williams KL, Appel RD, Hochstrasser DF, eds. Proteome Research: New Frontiers in Functional Genomics. Berlin: Springer; 1997.
- Clark TA, Sugnet CW, Ares M Jr. Genomewide analysis of mRNA processing in yeast using splicing-specific microarrays. Science. 2002; 296: 907–910.
[Abstract/Free Full Text] - Anderson NG, Matheson A, Anderson NL. Back to the future: the human protein index (HPI) and the agenda for post-proteomic biology. Proteomics. 2001; 1: 3–12.[CrossRef][Medline] [Order article via Infotrieve]
- Anderson NG, Anderson L. The human protein index. Clin Chem. 1982; 28: 739–748.[Medline] [Order article via Infotrieve]
- Klose J. Protein mapping by combined isoelectric focusing and electrophoresis of mouse tissues. A novel approach to testing for induced point mutations in mammals. Humangenetik. 1975; 26: 231–243.[Medline] [Order article via Infotrieve]
- O’Farrell PH. High resolution two-dimensional electrophoresis of proteins. J Biol Chem. 1975; 250: 4007–4021.
[Abstract/Free Full Text