Two-dimensional gel electrophoresis: an overview

Transcription

1 Two-dimensional gel electrophoresis: an overview Reprinted from trends in analytical chemistry volume 22 issue 5 may 2003 By David E. Garfin Courtesy of Bio-Rad Laboratories

2 Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 Trends Two-dimensional gel electrophoresis: an overview David E. Garfin Two-dimensional polyacrylamide gel electrophoresis of proteins is a robust and reproducible technique. It is the most widely used separation tool in proteomics. Current efforts in the field are directed at development of tools for expanding the range of proteins accessible with two-dimensional gels. # 2003 Published by Elsevier Science B.V. David E. Garfin* Life Science Group, Bio-Rad Laboratories, 2000 Alfred Nobel Drive, Hercules, California 94547, USA *Corresponding author; dave_garfin@bio-rad. com 1. Introduction Proteomics was built around the twodimensional gel. The idea that multiple proteins can be analyzed in parallel grew from two-dimensional gel maps [1^3]. Two-dimensional gels have provided much valuable information and they will continue to be an integral part of proteomics research for the foreseeable future [4,5]. Several books [6^10] and journal issues (e.g., Electrophoresis 22 (14) 2001) are devoted to the role of the two-dimensional gel in proteomics. 2. The two-dimensional polyacrylamide gel From the point of view of separations technology, two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) has not changed much since its inception. It is an orthogonal separation technique in that proteins are separated by two di erent physicochemical principles. Proteins (polypeptides) are rst separated on the basis of their (ph-dependent) net charges by isoelectric focusing (IEF) and further separated on the basis of their molecular masses by electrophoresis in the presence of sodium dodecyl sulfate (SDS). Both procedures are carried out in polyacrylamide gels (Fig. 1). IEF and SDS- PAGE are both high-resolution techniques. An outstanding feature of 2-D PAGE is that the resolution obtained during the rst separation step is not lost when the IEF gel is joined to the SDS- PAGE gel. It is this feature that imparts exceptional resolution to 2-D PAGE and distinguishes it from other separation methods [1]. In IEF, proteins are separated by electrophoresis in a ph gradient. Each type of protein molecule accumulates, or focuses, into a sharp band at its characteristic isoelectric point (pi), de ned as the ph at which it carries no net electrical charge. The initial separation for 2-D PAGE by IEF was originally done in capillary gels with ph gradients generated by carrier ampholytes (synthetic amphoteric compounds) [11]. The daunting task of learning to handle imsy capillary gels plus the poor reproducibility of carrier ampholyte IEF led to the wide acceptance of immobilized ph gradients (IPGs) for IEF. The ph gradients of IPGs are generated by means of bu ering compounds that are covalently bound into porous, polyacrylamide gels [12]. The ph gradients, xed as they are in IPG gels, remain stable over extended run times at very high voltages, a requisite for high-resolution separations. IPGs are cast on plastic backing sheets and are cut into mechanically stable strips that are easily manipulated. With a few notable exceptions (e.g., [13,14]), 2-D PAGE is now done almost exclusively with IPGs as the IEF media. The tools for 2-D PAGE are readily available and reasonably priced. Commercial manufacturers provide nearly every reagent and tool required for successful 2-D PAGE. The commercial availability of quality-controlled IPGs made in a multitude of ph ranges and lengths goes alongwaytowardsensuring exibility and reproducibility of the rst-dimension /03/$ - see front matter # 2003 Published by Elsevier Science B.V. doi: /s (03)

3 Trends Trends in Analytical Chemistry, Vol. 22, No. 5, Sample preparation Figure 1. Steps in 2-D PAGE. Proteins prepared from an extract of a cell or tissue source are separated on the basis of their isoelectric points (pi) by isoelectric focusing (IEF). They are next made ready for the seconddimension separation (equilibrated) and separated on the basis of their molecular weights (MW) by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The first and second dimensions are orthogonal both geometrically and in terms of separation principles. In this example, proteins from a lysate of Escherichia coli were separated in an immobilized phgradient strip (range ph4 7) and then by SDS-PAGE (from about 10 to 150 kda). The two-dimensional protein array was stained and imaged by fluorescence methods. The strip of vertical bands above the two-dimensional gel is the image of a dye-stained IEF strip that was run in parallel. separation. The reproducibility of the second-dimension separation has also been improved by the availability of pre-cast SDS-PAGE gels and (or) devices for casting and running multiple gels. There is su cient standardization in the available systems to enable intralaboratory and interlaboratory data comparisons and collaborations. The actual running of 2-D PAGE (with IPGs) is rather straightforward and easily learned (see, e.g., [7]). However, sample preparation for 2-D PAGE is another matter. For the most part, successful 2-D PAGE depends on e cient extraction and solubilization of proteins. Unlike the situation for DNA, there is no universal sample-preparation method suitable for all proteins. Each source of protein presents its own samplepreparation challenges. Proper sample treatment should begin from the moment the material is collected. Care must be exercised to prevent proteolysis following cellular death [13] (also see [7], p. 67). Tissue samples should be snap frozen and ground under liquid nitrogen in the presence of protease inhibitors. Some established protocols for sample preparation can be found on the internet, e.g., at and their associated links. Once proteins have been extracted from the source material, they must be prepared for 2-D PAGE. The major goal of sample preparation is to solubilize as many proteins as possible and to maintain their solubility throughout the 2-D PAGE process. According to current practice, proteins are denatured to their constituent polypeptide chains so that polypeptide sequences can be matched to their corresponding gene sequences. Secondary sample-preparation concerns are the removal of non-proteinaceous material that may interfere with 2-D PAGE and the prevention of artifactual modi cations of polypeptides. Of course, samplepreparation reagents must be compatible with the IEF stage of 2-D PAGE. This subject has been reviewed [15] (also see [7], p. 9, and [8], p. 9). Proteins are extracted from source material by wellestablished cell-disruption methods then solubilized and denatured by means of chaotropes, detergents, and reducing agents. Protein extraction can be done directly into a solubilization solution or extracted proteins can be diluted into a solubilization solution. Solubilization solution is also incorporated into the matrix of the IPG strip to maintain protein solubility during IEF. Chaotropes urea and thiourea are compatible with IEF. They are employed at high concentrations to disrupt hydrogen-bonded structures in the sample proteins. Non-ionic or zwitterionic detergents are used to disrupt hydrophobic interactions. CHAPS, Triton X-100, and newer detergents, such as sulfobetaine SB3-10 and amidosulfobetaine ASB-14[16], are popular IEF-compatible additives. Reducing agents dithiothrietol (DTT), dithioerythritol, and tributyl phosphine (TBP) reduce disul de bonds (cystine residues) to sulfhydryls (cysteine residues) and one or the other of them, usually DTT or TBP, is included in the solubiliza

4 Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 Trends tion mixture. A good solubilization solution for most purposes consists of 8 M urea, 4% CHAPS, 2 mm TBP, and 0.2% (w/v) carrier ampholytes (plus a trace of bromophenol blue tracking dye). For the purposes of IEF, TBP is preferred over DTT as a reducing reagent; the reaction of TBP with disul des is stoichiometric allowing its use at low concentrations, it is not as sensitive as DTT to dissolved oxygen, and it is uncharged. DTT is charged at alkaline ph and can migrate towards the anode, leaving some protein sulfhydryls prone to reoxidation. The optimum concentration of DTT is about 50 mm. At DTT concentrations below about 25 mm or above about 75 mm, fewer proteins are seen in gels than at 50 mm. The other additives mentioned above are included as needed to increase the solubilizing power of the sample solution. Thiourea, SB 3-10, and ASB-14are particularly e ective in this regard. A cocktail containing all six components (5 M urea, 2 M thiourea, 2% CHAPS, 1% SB 3^10, 1% ASB-14, 2 mm TBP. 0.2% carrier ampholytes) might be the best solubilization solution currently available. However, this cocktail can leave a wide detergent band at the bottom of the gel. SDS and lithium dodecyl sulfate may also be used for protein extraction as long as proteins are available at high enough concentrations to allow them to be diluted (at least 10-fold) into solutions of IEF-compatible detergents [17]. Sequential extraction is an e ective way to compartmentalize proteins in terms of their solubility [18]. It constitutes a third dimension of separation. In this procedure, proteins are extracted into solutions of increasing solubilization power and then separated by 2-D PAGE (Fig. 2). Insoluble proteins remaining from one extraction are treated with the next solution in the sequence. Thus, proteins are treated rst with aqueous bu er, then with urea/chaps/tbp, and next with urea/thiourea/chaps/sb 3-10/TBP. A fourth extraction with urea/thiourea/chaps/asb-14/tbp is also possible. The remaining insoluble material from the nal extraction can be taken up in SDS-PAGE sample solution and run in a one-dimensional gel. Membrane isolation with carbonate extraction has been used to identify and study membrane proteins [19]. Carbonate isolation works well for the isolation of bacterial membranes, but, when used with some eukaryotic systems, may precipitate other proteins as well [20]. Very hydrophobic proteins ^ those with more than about two transmembrane spanning domains ^ do not seem amenable to currently available solubilization procedures for 2-D PAGE [21]. They are often ignored, except in the cases of specialized membrane studies. Nucleic acids interfere with e cient IEF separations by both binding to proteins and increasing the viscosity of the sample solution. The usual way of dealing with Figure 2. Sequential extraction. The flow chart depicts a scheme for partitioning a protein mixture into components of decreasing solubility. The insoluble pellet from each extraction step is further extracted with a stronger chaotrope/detergent solution (after [18]). nucleic acids is to degrade them with nucleases prior to IEF. Non-speci c micrococcal nuclease or a mixture of (pancreatic) RNase and DNase added to the extraction mixture usually degrades nucleic acids su ciently to allow for good IEF. Polysaccharides in the sample solutions are often not severely detrimental to IEF. They are usually ignored, mainly because there is no simple method for treating them. The best way to separate lipids from samples appears to be by precipitating the proteins with organic solvents [22]. More often than not, however, this results in protein precipitates that are di cult to dissolve. 4. Some hints about running 2-D PAGE Apart from proteolysis, the most signi cant artifactual modi cation of proteins is carbamylation. This results from the breakdown of urea in sample solutions. Urea in solution degrades to cyanate that can react with the amino groups of proteins. This makes proteins more acidic by eliminating positive charges. Only fresh solutionsofureashouldbeusedin2-dpage.itisalsowise to limit the temperature reached during IEF to about 30 C to minimize urea breakdown. Some manufacturers of power supplies intended for running IPG strips set a default current limit of 50 ma per strip in order to hold down the running temperature. When 265

5 Trends Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 carbamylation is not a concern, current limits can be overridden. The e ect of salts on IEF in the sample solution is indirectly tied to the carbamylation problem and the current limits of the IEF cells. Salts contribute to the electrical conductivity of the sample solution and result in high current drains. It is best if salts can be eliminated prior to IEF. Nevertheless, IEF can still be performed in those cases where it is not possible or desirable to desalt the samples. With high-salt samples, it takes considerable time to reach high voltages, since the power supply holds the run at the voltage corresponding to the current limit. Contrary to intuition, salts do not clear from solution rapidly when voltage is applied. It can take hours to clear high concentrations of salt from standard-length IPGs when the current is limited to 50 ma per strip. The best advice for samples containing high concentrations of salts or other ions is to be patient during IEF. Electrode wicks should be used as reservoirs to trap the solution ions and the focusing programs should be set for extended run times. For this same reason, it is wise to limit the concentration of carrier ampholytes in the running solution to less than 0.2% (w/v) in order to avoid extended run times. Ampholytes contribute to the initial conductivity of the sample solution and until they begin to focus they limit the voltage that can be attained when the power supply limits the current. Isoelectric focusing is run in the same solubilizing solution as was used to dissolve or extract the proteins (except SDS solutions). Commercial IPGs are supplied as dried gel strips that must be rehydrated in sample solution prior to use. The need for rehydration led to another simpli cation of the method since protein samples can be loaded into IPG strips during the rehydration step (see [7], p. 221). Rehydration loading takes advantage of the fact that, in IEF, proteins can initially be distributed throughout the focusing medium. This technique is very popular because it is very easy and convenient. In addition, very high protein loads (milligrams of total protein) are possible with rehydration loading. If possible with the particular set up, it is wise to blot IPG strips after rehydration loading and before the IEF run. This removes liquid that might provide a parallel current path along the surface of the IPG and it also removes urea that might crystallize on thesurfaceofthestrip. Rehydration loading is done either passively or actively. In active rehydration, a small voltage (about 50 V) is impressed along the IPG strips while they are absorbing sample. Although there then exists a possible parallel current path through unabsorbed sample solution, protein loss appears to be minimal. Large proteins (those larger than about 100 kda) may enter IPG gels better under active rather than passive rehydration. Before the advent of rehydration loading, specially designed cups were used to load samples into rehydrated IPG strips. There are still several occasions when sample-application cups give better results than rehydration loading. At the extremes of the ph scale (e.g., ph 3^6 or ph 7^10), sample loading from cups yields better quality gels than rehydration loading. For best results, loading should be from the anode side of the IPGs for basic strips and from the cathode side of IPGs for acidic ranges. The isoelectric focusing run is best done at the highest possible power conditions available. It is advantageous to take advantage of the current limits of IPG power supplies and rapidly ramp voltages up to the nal desired setting. The time integral of the applied voltage, termed volt-hours (Vh), serves as a bookkeeping device with which to achieve reproducibility between runs. Once good focusing conditions have been established, the same Vh value should be used for all runs with the same protein mixture (start counting Vh with the high-voltage step). Vh values range from about 10 kvh for 7-cm IPGs and simple samples to 60 kvh or more for longer strips. After focusing, it is acceptable to have the power supply hold the voltage at an intermediate level (500^1000 V) until it is convenient to process the IPG strips for the second dimension. Focused strips may also be stored frozen until the second dimension can be run. Focused proteins must be processed in order to get them ready for the second-dimension electrophoresis run. This treatment of the focused proteins, called equilibration, is done while the proteins are still in the IPG strip. It is necessary both to saturate the isoelectrically focused proteins with SDS and to keep sulfhydryl groups reduced and prevent their reoxidation. Proteins at their pis are electrically uncharged and will not move from the IPG into the SDS-PAGE gel unless they are coated with SDS. Reduction and alkylation of polypeptide sulfhydryls are done along with SDS treatment. Equilibration consists of two short steps each of about 10 minutes in which IPGs are soaked in an SDS solution at rst containing DTT or TBP and then with iodoacetamide replacing the reducing agent. Iodoacetamide alkylates sulfhydryls, converting them to carboxyamidocysteine residues so that they cannot recombine to form disul des. Reduction and alkylation are best done between ph 8 and 9 to minimize alkylation of amido groups. Equilibration presents something of a bottleneck in the 2-D PAGE process, so there have been attempts to do equilibration in a single SDS step. Notable schemes toward this goal include the use of TBP and acrylamide together in a single reduction-alkylation step and carrying out reduction and alkylation of proteins in solution prior to IEF. However, the TBP-acrylamide method does not appear to provide alkylation of 266

6 Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 sulfhydryls as complete as the two-step DTT-iodoacetamide treatment [23]. Alkylation of proteins prior to IEF changes their pis and, therefore, also the 2-D PAGE pattern. Moreover, alkylation with iodoacetamide is ine cient in the presence of thiourea [24]. Once proteins have been reduced, alkylated, and saturated with SDS, it is a relatively simple matter to transfer them to the second-dimension gel. The IPG stripissimplyplacedontopofthesds-pagegeland sealed in place with agarose. The SDS-PAGE gel is set up and run, as is routinely done for protein analysis (see [7], p. 245). Electrophoresis drives the SDS-coated polypeptides out of the IPG and into the SDS-PAGE gel. Gels can be made in several sizes ranging from mini gels of about 7 cm 8cmor13cm 9cmandupto largegelsintherangeof25cm 20 cm. Commercial pre-cast gels are available in all of the popular types. In general, the larger the gel the better the resolution and the more protein that can be loaded into the gel. It is good practice to use fast-running, small-size gels while working out the conditions for best sample preparation. Large gels should be used for detailed analyses. In many cases, especially when the proteins of interest resolve well, it is possible to revert to small gels during targeted research. 5. Detecting low-abundance proteins Low-abundance proteins are of great interest in proteomic research and can be studied with 2-D PAGE. However, almost by de nition, the concentrations of low-abundanceproteinsinasamplearenearorbelow the lower detection limits of 2-D PAGE and its associated protein stains. It is necessary to increase the relative amounts of low-abundance proteins in the sample in order to be able to visualize them in the gels. Merely increasing the protein load to the 2-D gel is often insu cient, because high-abundance proteins will dominate the gels and can hide low-abundance proteins. Moreover, at high protein loads, resolution is lost and, therefore, so is the ability to distinguish closely spaced protein spots. Considerable e ort is being devoted to the development of pre-fractionation methods as a means for enriching the content of low-abundance proteins in samples for 2-D PAGE. The basic idea behind pre-fractionation is to segregate sample proteins into distinguishable fractions containing limited numbers of proteins. Sequential extraction, described above, is an example of one type of approach to pre-fractionation. Chromatographic methods have also been exploited. Ion exchange [25] and hydroxyapatite [26] matrices have been used to fractionate complex protein patterns prior to 2-D PAGE. A nity methods can be used either to isolate particular sets of proteins or to deplete samples of annoying high-abundance components (e.g., albumin and IgG in serum (Fig. 3)). Liquid-phase isoelectric focusing devices [27^29] o er another tool for obtaining low abundance proteins. These devices all su er from the protein losses inherent in IEF methods, since many proteins aggregate when they are focused (there is no electrostatic repulsion between protein molecules at their pi). In many cases, losses are an acceptable compromise for the ability to detect low-abundance proteins. Ashasalwaysbeenthecaseinproteinchemistry, proteins are lost with every treatment step. This may be irrelevant when the purpose of a study is establishment of a database of proteins. However, protein loss becomes important when quanti cation is needed. In the latter case, the fewer manipulations employed the better for the study. Narrow-range and micro-range IPGs give access to increased numbers of proteins through increased rst-dimension resolution without requiring pre-fractionation [30]. The ph ranges covered by these IPGs span 1^3 ph units. This serves to spread out the limited number of proteins with pis withinthecoveredrangealongtheentirelengthofthe strips and subsequently across the entire area of the 2-D gel. Closely-spaced proteins become resolved as the range of the IEF separation is decreased from 7 ph units (wide range ph 3^10) to 3 ph and then to 1 ph unit (Fig. 4). In addition, narrow-range and micro-range IPGs allow for increased protein loads because the relevant mass for optimum resolution is that from proteins with pis within the ph range of the IPG strip. Proteins outside the range of the strip migrate to the electrode ends and out of the separation region. Increased protein loads improve the chances for visualizing low-abundance proteins. 6. Detection and analysis of gel-separated proteins Trends Visualization of proteins in gels is accomplished with staining techniques [31] (also see [8], p. 107, and [9], p. 65). Despite the availability of a wide variety of speci c stains, the majority of 2-D PAGE gels are stained with Coomassie Brilliant Blue (CBB), SYPRO Ruby, or some type of silver stain. (SYPRO is a trademark of Molecular Probes, Inc., Eugene, Oregon 97402, USA.) For proteomics work, protein stains must be compatible with mass spectrometry (MS) and that has limited the choice of silver stain to those that do not include gluteraldehyde treatment or oxidation steps [32]. Coomassie Blue and SYPRO Ruby stains are both compatible with MS. When formulated as a colloidal sol [33], colloidal CBB is very easy to use and can be made environmentally benign [34]. Colloidal CBB is essentially an endpoint stain, meaning that gels can be left in 267

7 Trends Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 Figure 3. Affinity pre-fractionation. It is well known that albumin and IgG obscure many other serum proteins in 2-D PAGE. Removing them enables examination of lower abundance proteins that are otherwise obscured. In this example, a mixture of affinity ligands Cibacron Blue and Protein A was used to remove albumin and IgG, respectively, from human serum. By so doing, many proteins found in the effluent (B) became visible as compared to gels made from whole serum (A). Section (C) shows proteins bound and subsequently eluted from the affinity spin column. These bound proteins consist mainly of albumin, IgG, and fragments of albumin (unpublished). The same amount of protein (200 mg) was loaded on each gel. it overnight for convenience. A post-stain water wash is important with colloidal CBB. The wash removes excess colloidal dye particles from the gel surface and also drives the dye molecule into the proteins in the gel. Thus, the wash increases the signal-to-noise ratio of the stained gel. Colloidal CBB is the least sensitive of the three stains mentioned. Its detection limit is about 10 ng of protein per spot. Colloidal CBB stains a wide range of proteins and can respond linearly over two orders of magnitude in protein amount (depending on the proteins). Protein spots turn blue on staining with CBB and gel images can be captured with a scanning densitometer (visible light). SYPRO Ruby is a uorescent stain [35]. It is also an endpoint stain that can detect proteins at about 1 ng per spot. Depending on the proteins involved, SYPRO Ruby can be linear in response over three orders of magnitude in protein amount. A uorescent imager is necessary for visualizing protein spots stained with SYPRO Ruby. Silver staining is the most sensitive non-radioactive method for protein visualization, enabling protein spots containing a bit less than 1 ng to be detected. The range of linearity is less than two orders of magnitude. All silver-staining methods are very temperature dependent. They require several, usually precisely timed, manipulations and are very subjective whenitcomestodecidingwhentostopdevelopment. Thus, silver staining is the least reproducible of the stains. There is no universal stain that will react equivalently with all of the proteins in the gel. It is, therefore, valuable to stain with more than one kind of procedure at some stage of the experimental study. Double staining is possible and can show up di erences in the staining characteristics of the di erent proteins in the gel. This is usually done by following SYPRO Ruby or silver staining with CBB staining. Image-acquisition instruments range from simple cameras and light boxes to sophisticated laser-based uorescence detectors [36,37]. For subsequent digital image analysis, gel images must be captured electronically. The three categories of image-acquisition devices used with 2-D PAGE are document scanners, chargecoupled-device (CCD) cameras, and laser-based detectors. Document scanners (scanning densitometers) operate in visible light and are used when gels have been stained with CBB or silver. CCD cameras can be used with either visible or uorescent stains. High-sensitivity CCDs are cooled to increase their signal-to-noise ratios. They operate with illumination provided by either light boxes (visible or UV irradiation) for transmittance or overhead lamps for epi-illumination. Laser devices are the most sophisticated image-acquisition instruments. They are particularly useful for uorescently-stained gels. Detection is generally accomplished with photomultiplier tubes. With 2-D PAGE gels of any size, image resolution of 100^150 mm is adequate. For best quantitative results, theimagershouldbematchedtotheanalysissoftware. The most convenient arrangement is to use analysis software that controls the imager. Image-analysis software is the heart of proteomics research (see [7], p. 363, and [9], p. 131). It provides the analysis and control functions necessary to integrate and manage the various separation and analysis processes. Although gel images can be examined visually, objective quanti cation and comparison of the large number of protein spots require computer assistance. Software for 2-D PAGE analysis de nes and 268

8 Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 Trends Figure 4. Micro-range IPG. In IEF, narrow phranges favor high resolution. Sometimes, merely narrowing the phrange of the first-dimension separation enables otherwise obscured proteins to be examined. In this example, equal aliquots (300 mg) of protein extracted from the membranes of human blood cells (red and white) were subjected to 2-D PAGE. The gels obtained with a narrow-range, ph4-7, IPG (upper) and with a micro-range, ph , IPG (lower) are compared. The micro-range gradient gives nearly twice as many proteins as the equivalent section (boxed) of the narrow-range gradient. The indicated phand %T values are only approximate in these cropped images. quanti es spots in 2-D gels, removes background patterns, matches images from related gels, compares the intensities of corresponding spots in related gels (quantitative changes in expression), prepares gel data for presentation (reports), and exports gel-image information to databases. Image-analysis software also guides the excision of proteins from gels for further analysis, whether this is done manually or automatically with a spot-excision robot. For example, software comparison of images from related protein samples, such as from experimental and control cells, can highlight di erentiallyexpressed proteins that can be selectively excised from appropriate gels. Alternatively, when the interest is in mapping proteins from a given source, it may be su cient to excise and to analyze only the most abundant proteins in the sample, as represented by the most intense gel spots. Computer-generated cut lists drive spot-excision tools in well-integrated systems and greatly simplify spot collection, management, and record keeping when many spots are excised. Currently, the ultimate analysis tool for gel-separated proteins is the mass spectrometer [38]. In many cases with proteins obtained from gels, peptide-mass ngerprinting (PMF) in a MALDI-ToF spectrometer gives suf- cient information for protein identi cation (see [8], p. 197, [9], p. 151, and [10], p. 77). However, only those proteins with sequences residing in databases can be identi ed by PMF. De-novo sequencing with tandem mass spectrometers must be used to identify proteins not represented in databases. In either case, the better image-analysis software packages will directly annotate gel images with protein-identi cation information. Simply clicking on a spot in an annotated gel image will bring up the identi cation and mass spectrum of the protein. 2-D PAGE is particularly valuable in analyzing di erential expression (Fig. 5). A properly executed di erential analysis utilizes groups of 3-to-6 replicate gels for each growth condition. The software averages gel-togel variations and compares the composite images from each replicate group. Di erential labeling schemes 269

9 Trends Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 Figure 5. Differential expression. 2-D PAGE is particularly valuable in differential expression analysis. Shown here are corresponding sections of gels prepared with proteins extracted from E. coli cultures grown at 30 C (upper) and at 42 C (lower). The arrowheads indicate some of the differences in protein expression between the two growth conditions. Approximately 70 differences between the two cultures were identified by software analysis of the complete images. have been devised to compare test and control samples in the same gel. Methods not involving radioactive amino acids include uorescent labels for amino groups in proteins [39] and mass tags for sulfhydryls (hydrogen versus deuterium) [40]. Among other things, the former scheme su ers from potential over labeling as a consequence of the large numbers of amines in proteins [41] and the latter technique misses proteins lacking cysteines [42]. Furthermore, with both schemes, replicate gels must still be run to obtain the best results. 7. Perspectives Proteomics is in the early stages of its development as a discipline. Present emphasis is on mapping of proteins from various cell types and on identifying di erences in protein expression [43]. 2-DPAGEisparticularlywellsuitedtothesetypesof discovery-phase research. It also lends itself well to targeted research, where the expression of particular proteins is followed during systematic treatment regimes or alterations in growth conditions. It is a mistake, though, to equate proteomics with 2-D PAGE. There are many cases where other separation approaches, particularly chromatography, would be a better choice. All separation techniques that can be applied to proteins should be considered as being mutually complementary. No other technique separates charge and size isomers of polypeptides as well as 2-D PAGE. Hundreds to thousands of polypeptides can be resolved in a single 2-DPAGEgelandeveryoneispureornearlypure. These polypeptides can be quanti ed, probed with antibodies (via blotting), tested for post-translational modi cations, or extracted for MS analysis. Moreover, there is something reassuring about being able to look at a stained gel and see the array of protein spots. It takes only a bit of practice to be able to make a qualitative assessment of the outcome of a 2-D PAGE experiment. Often, mere visual comparison of gel images is enough to show those proteins a ected by di ering growth conditions. While 2-D PAGE is an ideal tool for discovery-phase research, not all expressed proteins can be displayed in a single gel. Low-abundance proteins, very large and very small proteins, basic and acidic proteins, and hydrophobic proteins present their own special challenges for 2-D PAGE. The approaches being taken for low-abundance and large proteins were discussed above. Thus, pre-fractionation methods are being applied to protein mixtures to account for the wide range in concentrations. Large proteins do not enter IPG strips under standard conditions of rehydration loading. Active rehydration at low voltage and cup loading facilitate entry of large proteins into IPGs. Small proteins, those under 10 kda, are lost in the bands of SDS micelles that form behind the bu er fronts in standard SDS-PAGE gels. They are, however, resolvable with use of alternative gel bu er systems; e.g., Tricine gels [44]. Cup loading turns out to be important for separation of basic and acidic proteins. In narrow-range IPGs, such as ph 3^6 or ph 7^10, rehydration loading leads to prominent horizontal streaking. Placement of the sample-application point is perhaps the key factor in successful separation of basic and acidic proteins. Wellformed spots are obtained when the loading cup is placed at the cathode end of acidic, narrow-range IPGs or at the anode end of basic, narrow-range IPGs. IPGs for very basic proteins, with pis in the ph 8^11 range, may require matrices other than polyacrylamide [45]. For reasons that are not clear, proteins with pis greater than about ph 10.5 are extremely di cult to 270

10 Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 focus in IPGs. Fortunately, nature does not seem to provide many highly acidic proteins [46], so the problem of what to do with them is not a major issue. As mentioned above, if proteins are highly hydrophobic, it may not be possible to deal with them with 2-D PAGE [21]. A frequent criticism of 2-D PAGE is that it is a relatively manual technique that does not lend itself easily to automation for high-throughput analyses. Automation for high throughput is of concern mainly to pharmaceutical and contract laboratories continually screening multiple samples. To other kinds of facilities, where a limited number of biological systems are being studied, high throughput may rather connote the means for arriving rapidly and easily at biologically relevant observations (i.e., putting the researchers through to the biology as quickly as possible). ThemeasuredapproacheddepictedinFig.6isarapid route to the biology of many systems. The recommendation is to concentrate e orts initially on the highabundance proteins in the system; ignore low-abundance and refractory proteins at rst. Many of the highabundance proteins can be identi ed in a matter of only a few weeks (a few days for experienced researchers). These identi ed proteins establish a reference map for further 2-D PAGE studies. Several of the high-abundance proteins are sure to be metabolic enzymes or proteins speci c to the type or organism, such as toxicity factors. Thus the rst gels give information on how the cell expends its energy. While one is contemplating the biological signi cance of the initial data, one can begin thinking about how to attack low-abundance proteins (via pre-fractionation) and refractory proteins (via whatever means it takes). Figure 6. A measured approach to proteome analysis. The identities of the most abundant proteins in a 2-D PAGE gel are likely to reveal much about the metabolism and (or) status of the parent cells. High-abundance, housekeeping proteins should be thoroughly studied before heroic measures are undertaken for hard-to-see proteins. The diagram serves to formalize an approach already taken by many proteomics researchers. References Trends [1] N.G. Anderson, A. Matheson, N.L. Anderson, Proteomics 1 (2001) 3. [2] V.C. Wasinger, S.J. Cordwell, A. Cerpa-Poljak, J.X. Yan, A.A. Gooley, M.R. Wilkins, M.W. Duncan, R. Harris, K.L. Williams, I. Humphrey-Smith, Electrophoresis 16 (1995) [3] W.P. Blackstock, M.P. Weir, Trends Biotechnol. 17 (1999) 121. [4 ] J.L. Harry, M.R. Wilkins, B.R. Herbert, N.H. Packer, A.A. Gooley, K.L. Williams, Electrophoresis 21 (2000) [5] T. Rabilloud, Proteomics 2 (2002) 3. [6] M.R. Wilkins, K.L. Williams, R.D. Appel, D.F. Hochstrasser (Eds.), Proteome Research: New Frontiers in Functional Genomics, Springer, Berlin, Germany, [7] A.J. Link (Ed.), 2-D Proteome Analysis Protocols, Humana Press, Totowa, NJ, USA, [8] T. Rabilloud (Ed.), Proteome Research: Two-Dimensional Gel Electrophoresis and Identi cation Methods, Springer, Berlin, Germany, [9] S.R. Pennington, M.J. Dunn (Eds.), Proteomics: From Protein Sequence to Function, BIOS, Oxford, UK, [10] D.C. Liebler, Introduction to Proteomics: Tools for the New Biology, Humana Press, Totowa, NJ, USA, [11] M.G. Harrington, D. Gudeman, T. Zewert, M. Yun, L. Hood, Methods: A Companion to Methods, Enzymology 3 (1991) 98. [12] P.G. Righetti, Immobilized ph Gradients: Theory and Methodology, Elsevier, Amsterdam, The Netherlands, [13] J. Klose, U. Kobalz, Electrophoresis 16 (1995) [14] J.E.Celis,G.Ratz,B.Basse,J.B.Lauridsen,A.Celis,N.A.Jensen, P. Gromov, in: J.E. Celis (Editor), Cell Biology: A Laboratory Handbook, 2nd Ed., Vol. 4, Academic Press, San Diego, California, USA, 1998, p [15] T. Rabilloud, Electrophoresis 17 (1996) 813. [16] M. Chevallet, V. Santoni, A. Poinas, D. Rouquie, A. Fuchs, S. Kie er, M. Rossignol, J. Lunardi, J. Garin, T. Rabilloud, Electrophoresis 19 (1998) [17] M. Fountoulakis, B. Taka cs, Electrophoresis 22 (2001) [18] M.P.Molloy,B.R.Herbert,B.J.Walsh,M.I.Tyler,M.Traini, J.-C.Sanchez,D.F.Hochstrasser,K.L.Williams,A.A.Gooley, Electrophoresis 19 (1998) 837. [19] M.P. Molloy, B.R. Herbert, M.B. Slade, T. Rabilloud, A.S. Nouwens, K.L. Williams, A.A. Gooley, Eur. J. Biochem. 267 (2000) [20] E. Olivieri, B. Herbert, P.G. Righetti, Electrophoresis 22 (2001) 560. [21] K. Bu«ttner, J. Bernhardt, C. Scharf, R. Schmid, U. Ma«der, C. Eymann, H. Antelman, A. Vo«lker,U.Vo«lker, M. Hecker, Electrophoresis 22 (2001) [22] R. Mastro, M. Hall, Anal. Biochem. 273 (1999) 313. [23] J.X. Yan, W.C. Kett, B.R. Herbert, A.A. Gooley, N.H. Packer, K.L. Williams, J. Chromatogr. A 813 (1998) 187. [24] B. Herbert, M. Galvani, M. Hamdan, E. Olivieri, J. MacCarthy, S. Pedersen, P.G. Righetti, Electrophoresis 22 (2001) [25] A. Butt, M.D. Davison, G.J. Smith, J.A. Young, S.J. Gaskell, S.G. Oliver, R.J. Beynon, Proteomics 1 (2001) 42. [26] M. Fountoulakis, M.-F. Taka cs, P. Berndt, H. Langen, B. Taka cs, Electrophoresis 20 (1999) [27] B. Herbert, P.G. Righetti, Electrophoresis 21 (2000) [28] X. Zuo, D.W. Speicher, Anal. Biochem. 284(2000) 266. [29] C.L. Nilsson, P. Davidsson, Mass Spectrom. Rev. 19 (2000) 390. [30] J.A. Westbrook, J.X. Yan, R. Wait, S.Y. Welson, M.J. Dunn, Electrophoresis 22 (2001) [31] T. Rabilloud, Anal. Chem. 72 (2000) 48A

11 Trends Trends in Analytical Chemistry, Vol. 22, No. 5, 2003 [32] J.X. Yan, R. Wait, T. Berkelman, R.A. Harry, J.A. Westbrook, C.H. Wheeler, M.J. Dunn, Electrophoresis 21 (2000) [33] V. Neuho, N. Arnold, D. Taube, W. Ehrhardt, Electrophoresis 9 (1988) 255. [34] H. Nivinskas, K.D. Cole, BioTechniques 20 (1996) 380. [35] W.F. Patton, Electrophoresis 21 (2000) [36] W.F. Patton, J. Chromatogr. A 698 (1995) 55. [37] M.D. MillerJr., R.A. Acey, L.Y.-T. Lee, A.J. Edwards, Electrophoresis 22 (2001) 791. [38] J.R. YatesIII, J. Mass Spectrom 33 (1998) 1. [39] M. Unlu, M.E. Morgan, J.S. Minden, Electrophoresis 18 (1997) [40] M.Smolka,H.Zhou,R.A.Aebersold,Mol.CellProteomics1 (2002) 19. [41] W.F. Patton, J.M. Beechem, Curr. Opin. Chem. Biol. 6 (2002) 63. [42] W.F. Patton, B. Schulenberg, T.H. Steinberg, Curr. Opin. Biotechnol. 13 (2002) 321. [43] R.A. Van Bogelen, E.E. Schiller, J.T. Thomas, F.C. Neidhardt, Electrophoresis 20 (1999) [44] H. Scha«gger, G. von Jagow, Anal. Biochem. 166 (1987) 368. [45] M.P. Molloy, N.D. Phadke, H. Chen, R. Tyldesley, D.E. Gar n, P.C. Andrews, J.R. Maddock, Proteomics 2 (2002) 899. [46] R. Schwartz, C.S. Ting, J. King, Genome Res. 11 (2001) 703. David Gar n is Proteomics Applications Manager in the Life Science Group of Bio-Rad Laboratories. Among other things, he helps to develop novel proteomic applications for Bio-Rad s large line of products for the separation and analysis of biomolecules. He received BS and MSEE degrees from the University of Minnesota, Minneapolis, USA, and a PhD (Biophysics) degree from the University of California, Berkeley, USA. He is a Councilor of the American Electrophoresis Society