High resolution two-dimensional polyacrylamide gel electrophoresis. I. Methodological procedures

Electrophoresis 1983, 4, 97-1 16

Two-dimensional gel electrophoresis I

97

Review Michael J. Dunn and Arthur H. M. Burghes Jerry Lewis Muscle Research Centre, DeDartment of Paediatrics and Nebnatal Medicine, Royal Postgraduate Medical School, London

High resolution two-dimensional polyacrylamide gel electrophoresis 1, Methodological procedures

Contents

proteins represented by most cellular and subcellular subfractions. The combination of two different gel electrophoretic techniques to produce a two-dimensional (2-D) separation procedure can potentially increase resolving power by several orders of magnitude and make the simultaneous separation and analysis of complex protein mixtures a practical possibility. Several types of 2-D electrophoretic techniques have been described in the literature, but true high-resolution two-dimensional gel electrophoresis (2-D PAGE) was made practical by the pioneering work of O’Farrell [I]. In the first part of this review we will describe the development of highresolution 2-D PAGE and discuss methodological procedures currently in use in various laboratories. In the second part of this review* we will discuss methods of analysis of 2-D gels and describe some applications of 2-D PAGE which demonstrate the potential of the technique for the investigation of many biological problems.

Introduction ........................... 97 Early history of 2-D PAGE . . . . . . . . . . . . . . . 97 The O’Farrell system .................... 98 Order of dimensions ..................... 98 Sample solubilization .................... 99 IEFdimension ......................... 102 TimeofIEF ........................... 102 Effects of urea and the gel medium .......... 102 Non-equilibrium methods (NEPHGE) . . . . . . . . 103 Slab gels for the first dimension . . . . . . . . . . . . 105 Monitoring ofpH gradient and carrier ampholytes 106 Carrier ampholytes ...................... 107 Buffer focusing and other alternatives ........ 108 Equilibration and transfer between dimensions . 108 SDSdimension.. ....................... 109 Stacking .............................. 109 Gel size .............................. 109 Features of SDS gels .................... 110 -2 Early history of 2-D PAGE Gradient engineering ..................... 110 Gradient casting ........................ 110 The history of 2-D electrophoretic separation of proteins Molecular weight standards . . . . . . . . . . . . . . . 111 dates back over a quarter of a century. In 1955, Smithies [21 Multiple gel techniques . . . . . . . . . . . . . . . . . . . 111 described a method of zone electrophoresis of serum proteins Binding gels to supports . . . . . . . . . . . . . . . . . .112 using starch gels which resulted in good resolution due to Estimates of resolution . . . . . . . . . . . . . . . . . . . 112 sieving of protein molecules by the pores of the support meHeterogeneity and artefacts . . . . . . . . . . . . . . . 112 dium. Smithies and Poulik [31 realised that a combination of Conclusion ............................ 113 two qualitatively different electrophoretic processes at rightReferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 angles should yield greater resolution than is possible with either technique separately and demonstrated the practicability of this approach by a 2-D separation of human 1 Introduction serum proteins. Electrophoretic separation of proteins received a considerable boost with the advent of polyacrylElegant and powerful techniques are available for the struc- amide as a support medium [41. The subsequent developtural analysis of DNA and the study of genome organization. ment of methods using discontinuous (multi-phasic) buffer However, until recently techniques were not available for the systems [5,6] in which the sample is loaded onto a large-pore separation, identification and measurement of the several stacking gel and proteins concentrated into narrow zones thousand gene products which can be synthesized by any cell before separation on a small-pore resolving gel resulted in type. Traditional one-dimensional methods of polyacryl- further improvements in resolution. Although there is, theoamide gel electrophoresis (PAGE) can resolve at best about retically, a single gel concentration which is optimal for the 100 proteins from any sample of cellular proteins and are resolution of any two proteins, it was soon realised that there therefore unsuitable for analysing the complex mixtures of can be no single gel concentration which will give maximum separation of all the components in a complex protein mixCorrespondence: Dr. M. J. Dunn, Jerry Lewis Muscle Research Centre, ture. This was clearly demonstrated in a 2-D separation of Royal Postgraduate Medical School. DuCane Road, London W12 OHS, serum components [71. Resolution and sharpness of protein bands can, therefore, be enhanced with gradient polyacrylUK amide gels [8- 101 in which proteins are driven through pores Abbreviations: IEF: Isoelectric focusing; NEPHGE: Non-equilibrium of progressively decreasing size until they are brought to a

1 2 3 4 5 6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 7 8 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 9 10 11 12

~~

~

pH gradient electrophoresis; PAGE: Polyacrylamide gel electrophoresis; PI: Isoelectric point; SB: Sulfobetaine; SDS: Sodium dodecyl sulfate; TCA: Trichloroacetic acid; Vh: Volt hours; Wh: Watt hours; 2-D: Twodimensional OVerlag Chemie GmbH, D-6940 Weinheim, 1983

* To be published in the next issue of Elecrrophoresis 0 173-0835/83/0204-0097 $02.50/0

98

M. J. Dunn and A. H. M.Burghes

near dead stop according to their size. Resolution was further increased by using low concentration polyacrylamide gels in the first dimension which separates proteins predominantly by mobility [ 101. Convex polyacrylamide gradient gels, which separate proteins according to size, were used in the second dimension. This system therefore approached the optimum of the proteins being separated in the two dimensions according to independent parameters. Kaltschmidt and Wittman [ l 1, 121 developed a method for the separation of ribosomal proteins using a discontinuous ouffer system in the first dimension, followed by a continuous gel system in the second dimension. The sample cast in an agarose gel [ 131 was applied to the middle of the first dimension gel so that proteins migrating to the cathode and anode could be resolved simultaneously. This system separated by size in both dimensions, but as different pH’s were used the proteins were separated according to different charged groups in each dimension. Inclusion of Triton X- 100 in this system resulted in separation of proteins by their ability to bind detergent as well as by their mobility [141. In the presence of reducing agents, proteins are dissociated by the anionic detergent, sodium dodecyl sulfate (SDS), into single polypeptide chains without breaking covalent linkages other than disulfide bonds [151. It was found that polyacrylamide gels containing SDS produced excellent separations [15,161 which were dependent on polypeptide molecular weight [ 17, 181. The original methods of SDS PAGE utilized continuous buffer systems. However resolution was increased when a discontinuous buffer system was used 1191. This method was soon applied to 2-D separations of ribosomal proteins where a second dimension separation dependent on polypeptide molecular weight followed a first dimension based mainly on charge differences [20, 2 11. Isoelectric focusing (IEF) is a technique which separates proteins solely on the basis of charge. The theoretical foundation of IEF was developed by Svensson [22,231, but its application became practicable with the synthesis of low molecular weight carrier ampholytes [241. A separation based purely on charge could now be used for the first dimension, in combination with either gels of a single polyacrylamide concentration [25,261 or gradient polyacrylamide gels [27,281. The use of IEF in combination with SDS PAGE allowed the separation of proteins according to two truly independent parameters, i. e. charge and size 1291. The methods of IEF which had been used up to this time were only suitable for use with soluble proteins, and, therefore, modified procedures had to be developed if 2-D protein mapping was to be more generally applicable. Inclusion of urea in the first dimension IEF gels was found to be beneficial for 2-D separations of non-histone nuclear proteins 130-321 and EDTA-extractable erythrocyte membrane proteins [331. Further work in this area was stimulated by investigators interested in hydrophobic membrane proteins. They found that the presence of nonionic detergent, usually Triton X-100, in addition to 8 M urea in the first dimension IEF gels improved 2-D separations of these proteins 134-361. Thus, by 1975 a sophisticated system of 2-D PAGE involving charge separation by IEF in the first dimension, followed by SDS PAGE in the second dimension had evolved. It is this stage of 2-D PAGE development that facilitated high resolution mapping studies of protein components of various tissues [37-401.

Electrophoresis 1983, 4, 97-1 16

3 The O’Farrell system In 1975, O’Farrell [ l ] described a technique of 2-D PAGE including several modifications designed to optimize resolving power of the system for the separation of Escherichia coli (E. coli) proteins. This method is the foundation of 2-D PAGE technology as practised today. O’Farrell was aware that each dimension of a 2-D separation should separate proteins according to truly independent parameters so that proteins are distributed across the entire surface of the gel rather than across a diagonal. A combination of IEF in cylindrical 4 %T 5 %C gels containing 9 M urea and 2 % w/v NP-40 with the discontinuous SDS gradient PAGE slab system of Laemmli [ 191 was chosen because of the high resolution of each system, and because each procedure separates proteins according to different properties, i. e. charge and molecular weight. Several important innovations were made in the technique which resulted in a significant increase in resolution compared with previously published methods. Some of these innovations are listed in Table 1 and compared with those features which we believe to be desirable for optimal resolution in 2-D PAGE. The increased resolution obtained with this system has resulted in what has been termed a “protein explosion” [41]. However, it is clear from Table 1 that the original O’Farrell method does not satisfy certain of our criteria for optimal resolution. In the following sections some aspects of current methodology are discussed in relation to attempts which have been made to overcome some of the major problems in 2-D PAGE.

4 Order of dimensions In most procedures the first dimension is an IEF gel which is then applied to an SDS gel as in the original method [ll. However, some procedures have been described that use SDS gels in the first dimension followed by IEF in a flat-bed gel containing 8 M urea and non-ionic detergent [42-461. In some of these procedures [42,44,461a restricted molecular weight region is excised from the SDS gel and run into an IEF gel. This enables particular regions to be analysed without the necessity for complicated 2-D maps, and in addition allows multiple samples to be run easily. The disadvantages of such systems is that SDS is present during isoelectric focusing and the proteins must be electrophoresed out of a relatively restrictive gel. Other methods using SDS PAGE in the first dimension 143,451 have been designed to produce 2D maps of total cellular components and are therefore more comparable with the original system. In one procedure [431 the proteins were electrophoresed first on a SDS gradient gel. The sample strip was then cut out and polymerized into a 4.5 % polyacrylamide gel into which the proteins were transferred by steady-state stacking. This was necessary due to the severe sieving properties of the gradient gel. The strip of stacking gel containing the proteins was then polymerized into the focusing gel and run on a flat-bed IEF system [431. In an alternative procedure 1451, the proteins were first electrophoresed on a single percentage SDS gel, the sample strip cut out, equilibrated, and focused on a flat-bed apparatus. Systems utilizing SDS in the first dimension were developed to overcome problems of solubility of proteins. It is assumed

Electrophoresis 1983,4, 97-1 16

Two-dimensional gel electrophoresis I

99

Table 1. Features and disadvantages of the original O’Farrell 2-D PAGE procedure [ l ] compared with features required for optimal resolution. Features of the O’Farrell system

Optimal features

Disadvantages of the O’Farrell system

Separation by charge and molecular weight

Separation by two different physicochemical parameters

Hydrophobic properties not analysed

Samples solubilized using 9.5 M urea and 2 % NP-40

Complete solubilization and disruption of all non-covalent interactions Complete entry of sample into both dimensions Solubilizing agents should be stable and have low viscosity

Incomplete solubilization and disaggregation Incomplete entry of sample into gels Urea is viscous and problems with carbamylation

Polyacrylamide support medium

Stable and inert support medium over a wide pH range No sieving during charge separation Good sieving properties for molecular weight separation

Deamidation at extreme alkaline pH Sieving of high molecular weight proteins in IEF

IEF run for 30 to 59.2 Vhlcm’. Long focusing time to get sharper spots

Stable pH gradients True equilibrium and maximum resolution Separate systems to detect mobility variants, i e. amino acid substitutions involving PKa = PI

Cathodic drift. Final pH gradient does not extend above pH 7 Equilibrium not tested

Low field strength used

High field strengths to minimize focusing time and to sharpen bands.

High field strengths cannot be used due to heating and gradient drift

~

~~

~

~~

~

Final gradient pH 4.5 to pH 7.0 Ampholine used.

Optimization of pH gradient to fit distribution of protein PIvalues Maximize number of ampholyte species

Gradient cannot be extended due to excess drift and loss of resolution Only one type of ampholyte used

Equilibration of first dimension gel to allow reaction of proteins with SDS

Minimization of diffusion and protein loss with complete elution of proteins and no streaking

Diffusion and loss of proteins giving reduced resolution

Laemmli stacking gel

Rapid elution with stacking into second dimension

Slow and incomplete elution

Exponential gradient polyacrylamide SDS gels for second dimension

Good resolution of all proteins in second dimension

High molecular weight proteins poorly resolved. Low molecular weight proteins form diffuse spots

Use of large, thin gels t o maximize area of gel used for separation Good reproducibility under standard conditions

High sensitivity using radiolabelled proteins

Good reproducibility t o facilitate inter-gel comparisons

Pattern matching can be difficult

Simultaneous processing of multiple gels

Single gel procedure

High sensitivity of detection with minimum spreading of spots High dynamic range

Spreading of spots. Limited dynamic range

Quantitative evaluation of gels Identification and characterization of separated polypeptides

Limited qualitative and quantitative analysis

that after solubilization in SDS and electrophoresis, the proteins do not reaggregate or become insoluble during subsequent steps in the procedures. There are certain disadvantages to using such a system. Polymerization of a gel strip into another gel [431 may result in loss of protein and this may be particularly severe for basic proteins 1471. SDS gradient polyacrylamide gels give better resolution in the first dimension, but an additional stacking step is necessary to remove the proteins from a restrictive to a non-restrictive gel. This procedure will result in increased lateral diffusion with subsequent loss of resolution compared with the transfer of proteins from a non-restrictive IEF gel to a restrictive SDS gel. Perhaps the major drawback of such systems is that sep-

aration in a restrictive gel is followed by the use of a nonrestrictive gel, with the result that the proteins can undergo more diffusion in the second dimension IEF gel, thus creating larger spots. In addition, the first dimension and stacking gel will contain considerable amounts of salt, which is not necessarily evenly distributed, and this can produce distortion in the electric field during IEF [48,491.

5 Sample solubilization The ideal sample solubilization procedure for 2-D PAGE should result in disruption of all protein complexes and ag-

100

M. J. Dunn and A. H. M. Burghes

gregates into a solution of single polypeptides. It should be remembered that the forces responsible for protein-protein interactions are the same as those for protein folding, so that agents capable of disrupting these interactions are generally denaturing [50]. If polypeptide interactions are incompletely disrupted, then a protein may exist simultaneously in the aggregated state and as a single polypeptide. This will result in the appearance of “artefactual” spots in 2-D maps due to protein complexes, with a concomitant reduction in the intensity of those spots due to single polypeptides. Therefore, solubility and disaggregation are crucial factors in 2-D PAGE. A combination of urea and the non-ionic detergent, NP-40, were originally recommended for sample solubilization[l, 34, 351. However, this procedure does not always result in complete disaggregation of all protein complexes or allow entry of all proteins into the focusing gels, e.g. ribosomal proteins [ll, membrane proteins 15 11, and histones 1521.

Electrophoresis 1983, 4, 97-1 16

100 “C [571. Stirring in SDS for 30 min at 40 OC has been recommended [%I, but it should be remembered that proteolysis can occur using extended reaction at reduced temperatures. The inclusion of 8 M urea with SDS was found to increase solubilization and may be equivalent to heating at 100 “C in the presence of SDS 1561. It has been suggested that autoxidation of membrane lipids can be involved in cross-linking membrane proteins and/or lipids [591. On this basis an inhibitor of autoxidation, butylated hydroxytoluene, was included in a SDS solubilization procedure [601 and was found to improve solubilization and resolution. Proteinnucleic acid interactions can also interfere with sample solubilization. Therefore, a DNA and RNA digestion step can be included in the solubilizatior, prccezure, the method of Garrels [611 resulting in 95 % of the DNA becoming acid soluble. It has also been found that high levels of non-ionic detergent can increase streaking [621.

The use of urea/NP-40 mixtures or low SDS to protein ratios have been found to result in incomplete solubilization of erythrocytes [62,631, with band 3 protein and spectrin missing from the patterns. When SDS was used at a higher SDS to protein ratio [64], the erythrocyte peptides were represented in the 2-D maps and there was less streaking of high molecular weight proteins, particularly when samples were heated at 100OC. However, the inclusion of all the membrane proteins in the patterns can have disadvantages [651. It is interesting that the peripheral erythrocyte membrane protein, spectrin, has been found to be particularly resistant to 2-D analysis. Similarly, the peripheral membrane proteins of soybean root plasma membranes tend to cause streaking and smearing of 2-D patterns 1601. Such peripheral membrane proteins would be expected to be less hydrophobic than integral membrane proteins and therefore more amenable to solubilization by less vigorous procedures. However, it should be remembered that spectrin polypeptides are of high molecular weight [661. In addition, spectrin is a structural protein participating in the cytoskeleton responsible for the maintenance of red cell shape [671. Spectrin is capable of forming stable interactions both with itself and with other membrane components and these complexes are presumably difficult to solubilize. CytoSeveral methods have been described based on solubilization plasmic structural proteins can also be difficult to solubilize in SDS under reducing conditions followed by addition of and Garrels [61] has found that, if low concentrations of high concentrations of urea and NP-40 to “compete” with SDS are used, the cytoskeletal proteins, tubulin and interSDS. These procedures have found particular application to mediate filament protein, are reduced in amount in 2-D patmembrane proteins which are not generally amenable to terns. solubilization in urea and NP-40 mixtures alone. Failure to pellet on centrifugation has often been used as a criterion of In general, then, the inclusion of SDS in the sample preparasolubility in these studies. However, this does not necessarily tion procedure appears to give improved solubilization and indicate complete disaggregation or solubility, and the densi- entry of proteins into 2-D gels. In a comparison of various ty of urea in the solubilization buffers should be taken into solubilization procedures using one-dimensional IEF of skin account [54,551. A more reliable indicator of disaggregation fibroblast proteins we found [681 that essentially similar patand solubility is the amount of sample which can enter the fo- terns were observed with urea/NP-40 mixtures or SDS cusing gel. Optimal solubilization with SDS is usually achiev- methods. However, using SDS less material remained at the ed with heating at 100 OC for 3 min. However, Ames and point of sample applications and the intensity of two bands Nikaido [5 11recommended heating at 70 OC for 30 min and (probably tubulin) were increased. In addition, there was it has been reported that modification of sample can occur at some disturbance in anodic resolution which might be due to 100 O C [561. It was suggested [561 that this might be due to interaction of sulfate with ampholytes as has been described an amino group reaction, but deamidation of asparagine and for sulfated dyes [691. Thus, SDS appears to allow greater glutamine could also occur giving rise to artefactual spots. entry of protein into the gel. Nonetheless, even in these small Deamidation or even breakage of peptide bonds can also oc- amounts, SDS was found to remove all the detergent from cur if the SDS reaction is carried out at elevated pH at the gel due to the formation of mixed micelles [681. Proteins

SDS is known to disrupt the majority of protein interactions [ 15, 17,531, but its anionic nature renders it unsuitable for inclusion in IEF gels. However, O’Farrell [ 1I found that SDS could be used to initially solubilize samples for 2-D PAGE and this resulted in the appearance in the gels of proteins which normally failed to focus. Proteins appeared to migrate according to their native charge, while free SDS formed mixed micelles with NP-40 present in the focusing gels which focused at the extreme anode [ 11. This procedure was further developed by Ames and Nikaido [511 for membrane proteins. Samples were solubilized in SDS and NP-40 was subsequently added. The ratios of SDS to protein (1 :3) and SDS to NP-40 (1 :8) were carefully controlled to maximise protein solubilization while minimising the deleterious effects of high levels of SDS on subsequent IEF. For E . coli, 78 % of total protein was solubilized using this procedure and 80 % of the applied radioactive counts entered the gel. However, this improved solubility was gained at the expense of decreased resolution, since in the presence of SDS the pH of the gels was reduced at the basic end, there were few proteins in the very acid portion of the gels, and the system was prone to horizontal streaking at higher SDS concentrations.

Electrophoresis 1983, 4, 97-1 16

are therefore exposed to an environment containing little or no detergent [68, 701, which can result in their precipitation. Indeed, it is apparent that if erythrocyte membranes solubilized in SDS are focused for long times spots in the final twodimensional patterns tend to appear in an almost vertical line [71I, suggesting that precipitation may have occurred. Some alternative procedures for solubilization have been described and found to be preferable to SDS for certain samples. It was found possible to achieve adequate solubilization of mammalian cell plasma membranes [721 and plant seed proteins 172, 731 using urea and NP-40 under alkaline conditions (pH 10.3) by including potassium carbonate. A similar procecure has been applied to leaf proteins, except that they had to be initially extracted with acetone to remove phenolic pigments and lysine was included in the procedure [741. However, the use of urea at strongly alkaline pH would appear to increase the risk of carbamylation of polypeptides or p-elimination of 0-glycosidically linked oligosaccharides and phosphoserine 1721. Microsomal proteins have been found to be particularly difficult to solubilize and a method using selective solubilization with deoxycholate has been used to analyse separately those proteins not solubilized using an SDS procedure [75]. It appears that some of the proteins, especially cytochrome P450, tend to precipitate during IEF [75,761. This problem can be partially overcome by adjusting the sample to pH 4.3 and applying it at the anode. The insolubility of cytochrome P450 may be due to interaction with DNA which would be reduced at low pH. In addition, due to its basic nature this protein may form ionic interactions with carrier ampholytes similar to heparin and polyarginine 1771. Alternatively, it may interact with alkaline carrier ampholytes due to their hydrophobic nature [781. Detergents such as NP-40 and Triton X- 100 have polyoxyethylene head groups, the hydrophobic nature of which may act to increase the solubility of certain proteins. The zwitterionic, sulfobetaine (SB) type, detergents have been found to be effective solubilizing agents and generally non-denaturing in their action [79, 801. These compounds are compatible with IEF as they behave as “poor carrier ampholytes”, i e. the PI-pKa difference is large, and do not focus into a sharp peak at their isoelectric point, but remain evenly distributed throughout the gel [55,801. SB12 and SB14 have been used to solubilize samples which were subsequently subjected to 2-D PAGE using the O’Farrell system [811. Solubilization was found to be less effective than using the SDS procedure of Ames and Nikaido [ 5 11, but it should be pointed out that zwitterionic detergent was not included in the IEF gels. We have found SB detergents to be less effective than the standard urea/NP40 mixture [681 and they are incompatible with high levels of urea [681.

Two-dimensionalgel electrophoresis I

Chemlcal o r trade name

Head group s t r u c t u r e

Head group nanie

-(O-CHZ-CH2),-OH

Po I yoxyet hy 1ene

101

__ Non-lonlc Lubrol Trlton flercaptans

~IO-CH2-CH2),-OH -S-(O-CH2-CH2

),-OH

Po 1YoxYethy I ene Pol Yoxyethylene

m n y x LO

Amlne oxlde

Rewocld DU 185

AlkYlolamldes

aoctylglycerylether

GI ycero I

-CH-CH-CHz

dH dH AH

Decy 1Q-D-sluco-

suqar

pyranoslde Hb nodecyl-(3-D-

ma1 t o s l d e Dodecyl urea

Amlde H

zwltterlonlc zwlttergent

R

FH3 N+-cH~-cH~-cH~-s-o-

s

CH3

0PhosPhorYlchollne

-O-P-O-CH2-CH2-N

I

0

F? -CH3 I

CH3

Su I phobeta 1ne

Phowhoryl-

chol Ine

Figure I . Possible alternative detergents for sample solubilization.

that ionic interactions must be minimised to allow separation of acidic and basic proteins in IEF. Whatever detergent is used it is important to remember that detergent to protein ratio may be a more important factor for effective solubilization than absolute detergent concentration. Also, enough detergent must be present in the micellar form to dissolve membrane constituents on a stoichiometric basis [551. Few attempts have been made to explore alternative denaturants. This is surprising as increasing the denaturing potential should disrupt more protein interactions. An interesting feature of urea is that it increases the dielectric constant of the solvent [831, which should decrease ionic interactions and promote protein solubilization. Urea presents certain difficulties in handling and until a suitable alternative denaturant is found, ready-made commercial denaturing Some other types of detergent are listed in Fig. 1. Non-ionic IEF gels are not practical. An extensive list of denaturants glucoside detergents can be effective solubilizing agents [821 can be found in the article by Gordon and Jencks [841. and have been used for solubilization of samples prior to 2-D Agents such as dimethylformamide and tetramethylurea are PAGE [811. It is interesting to note that this type of detergent not compatible with the polymerization of polyacrylamide head group does not appear to interact with carrier ampho- [691, while denaturants such as chloral hydrate [851 cannot lytes [681. We have tried to use decyl-P-D-glucoside in the be used for IEF due to their reactivity at alkaline pH. Ames presence of 8 M urea. However, gels cast in this solution con- and Nikaido [5 11 have used guanidinium thiocyanate but tain a white flocculent precipitate of detergent below 37 “C this resulted in loss of protein and the appearance of (A. H. M. Burghes and M. J. Dunn, unpublished observa- extensive charge heterogeneity. A combination of urea and tion). These compounds could also be included in the gel as n-butyl urea has been used as a denaturant for the analysis of they should be compatible with IEF. It should be realised membrane proteins [861.

102

Electrophoresis 1983, 4, 97-116

M. J. Dunn and A. H. M.Burghes

Table 3. Running conditions used for the first dimension in typical NEPHGE procedures.

6 IEF dimension 6.1 Time of IEF

IEF is an equilibrium technique as proteins have no net charge at their isoelectric point (PI) [22,231. The approach of any ampholyte to equilibrium is asymptotic and the number of volt hours (Vh) required for a protein to focus can vary with the conditions used and the proteins investigated 1871. Moreover, very small charge differences will not be resolved until equilibrium is reached. We have observed the number of bands to increase for complex protein mixtures as equilibrium is approached [681. Various methods can be used to determine equilibrium, such as coincidence of bands when samples are comigrated from the anode and cathode, and constancy of pattern over long focusing times [87, 881. In 2-D PAGE, the focusing time is usually reported in terms of Vh [ l , 891. Table 2 lists IEF conditions used in various 2-D PAGE procedures. It can be seen that there is a large variation between different laboratories in the Vh used to obtain equilibrium. In addtition, the length of the IEF gels varies between different systems. It should be realised that the number of Vh necessary for equilibrium varies with the square of the length of gel [901. This is due to the fact that for a longer gel the protein has both a longer distance to migrate and, at the same applied voltage, experiences a lower field strength. Thus, for reproducible conditions it is necessary to carefully control both Vh and gel length. A unit which may be preferable to Vh to describe focusing times may be Vhll’ (where 1 = separation length of the focusing gel in cm). In Table 3 values for Vhll’ have been calculated for various systems. On this basis 0-Farrell’s 111 recommended times are between 30 and 59 Vh/cmz. Interestingly, it can be seen that there is considerable variation between different systems, ranging from 80 Vh/cm’ to as little as 12.9 Vhlcm’.

Table 2. Running conditions used for first dimension IEF in typical 2-D PAGE procedure Gel length (cm)

vh

Vh/cm2

16.5 9.6 8 or 14 12.7 13 16 11 11 32 13 10 10 14 to 15 14 15 20 12 14 to 15 14 to 15 15 10 10

3500 1200 4000-5000 3900 4680 7 200 3650 4000-5000 34500 6000 3600 4000 9000 8680 10000 19000 7000 12000 13000 15000 7000 to 8000 8700

12.9 13 20.4 to 78 24.2 27.7 28.1 30.2 33 to 41.3 34 35.5 36 40 40 to 45.9 44.3 44.4 47.5 48.6 53 to 61.2 57 to 66.3 69 70 to 80 87

Reference

12.5 18 12 12 12 12 12 14 to 15 12

Vh

Vh/cm2

800 3200 1500 1600 to 2000 1600 1800 2000 3200 3000

5.1 9.9 10.4 11 to 14 11.1 12.5 13.9 14.2 to 16.3 21

Reference

In Table 3 values for non-equilibrium (NEPHGE) [911 systems have been calculated and range from 5 to 21 Vhl cm’. Therefore it seems certain that in some cases the IEF gels do not satisfy the equilibrium parameters defined by O’Farrell 111. Attainment of a true steady-state in IEF should allow higher resolution and more reproducible results. Finlayson and Chrambach [881 have investigated conditions for equilibrium in IEF gels containing 8 M urea. They found for the protein investigated that a time equivalent to 33.3 -44.4 Vhlcm’ was required, which is in the range of focusing times most often used in 2-D PAGE separations. In a study 1681 using IEF gels containing 8 M urea and 2 % NP-40 and run on a flat-bed apparatus under conditions where the pH gradient showed increased stability over the pH 3-10 range, we found that to achieve coincidence of proteins migrated from the anode and cathode it was necessary to focus for 87 Vhlcm’. This figure exceeds all values used so far in 2-D PAGE separations, including our own (69 Vh/ cm’) [921. Even utilizing such high Vh/cm2 values one cannot be certain that all of the high molecular weight proteins are at equilibrium. Under denaturing conditions with 8 M urea, artefacts due to conformational states or aggregation should not occur 1931. Therefore, non-equilibrium techniques can be useful as it is the separation of proteins that is of interest and not equilibriumper se (see Section 6.4). However, the non-eqilibrium gel may not show optimal resolution of small charge differences between proteim or be as reproducible. Thus, Vhlcm’ values should be controlled closely between IEF runs to obtain more reproducible results. Using Vh/cm2 as a criterion, however, does not compensate for the resistance of the gel. This has led to the sugges-. tion 1941 that watt hours (Wh) would be a better parameter to use. It is probably better expressed in units of Wh/volume, so that by allowing for changes in gel volume the term can be applied to gels of any size. This parameter can account for factors that affect resistance of the gel such as heat and changes in dielectric constant. However, using neither Whl cm3 or Vhlcm’ are changes in carrier ampholyte concentration taken into account and therefore this factor should be controlled to ensure reproducible results. 6.2 Effects of urea and the gel medium

2-D PAGE is usually carried out under denaturing conditions, with both urea and NP-40 present in the IEF gels [I, 891. IEF gels containing 8 M urea require longer focusing times to reach equilibrium. There are several reasons for this effect. First, the addition of urea to the gel increases the vis-

Electrophoresis 1983, 4, 97-1 16

cosity of the system, thereby retarding the proteins. Reports in the literature suggest that urea leads to more restrictive gels [95, 961, but this can probably be accounted for on the basis of a viscosity effect rather than by any change in pore structure of the gel [681. However, the gel sieving mechanism has been said to be equivalent to viscosity [971 and therefore an increase in viscosity makes the gel more restrictive. Indeed, glycerol I981 and urea 1991 have been added to polyacrylamide gels for the analysis of low molecular weight proteins. The other feature which has been suggested to impede protein migration is that after denaturation in 8 M urea the Stokes radius of the protein is increased [95,1001 so that the unfolded protein will be impeded to a greater extent by the gel matrix. This effect, however, is to a certain extent counteracted by the ability of urea to disrupt protein aggregates and reduce proteins to their constituent polypeptides, thus lowering the effective molecular weight range. In fact, in the presence of non-ionic detergent proteins denatured by urea may be in a similar form to the same proteins solubilized in SDS alone. Polyacrylamide itself is a restrictive matrix and can exert a molecular sieving effect [ 1011. For the IEF dimension it is best to use weak gels (3 % t o 5 %T) which can, however, still restrict the migration of high molecular weight proteins. Attempts have been made to increase gel porosity using gels highly cross-linked with N,N'-methylenebisacrylamide, but this resulted in gels of poor mechanical strength which were also more hydrophobic [ 102- 1041. N,N'-Diallytatrardiamide can be used as an alternative cross-linker, but this appears to lead to incompletely polymerized gels containing agents that can react with proteins [ 102, 105, 1061. Another problem is gradient drift. Acrylamide contains various agents that can give rise to drift: (1) acrylic acid impurities, (2) incorporation of the catalysts, and (3) hydrolysis of amides above pH 10 [1071. To avoid acrylic acid impurities, acrylamide solutions should be freshly prepared and deionized [1081. To overcome the other problems, dimethylaminopropylmethacrylamide can be copolymerized into the gel to charge balance negative groups. This procedure results in gels showing no drift after 5000 Vh [1071. Recently, highly purified agarose [ 109, 1101 and charge-balanced agarose [491 have been introduced for IEF, and these may overcome some of the restrictions of focusing high molecular weight proteins. Hirabayashi [ l l 11 described a 2-D system using agarose IEF in the first dimension. The agarose gel contained a lower concentration of urea, approximately 6 M and the run was performed for 37 Vh/cm2. Although this procedure was developed specifically for high molecular weight proteins, completely satisfactory results were still not obtained which may be due to problems of solubility 1111. It would be preferable to use gels containing 8 M urea and nonionic detergent. To achieve this it was found necessary to use 2 % agarose and longer gelling times in order to obtain a gel with reliable structural integrity [loo]. In addition, urea had to be added to the warm agarose solution, thereby increasing the risk of cyanate formation. Ampholytes or mercaptoethanol can be added as scavangers. Agarose gels containing high urea concentrations (>6 M) vill gel much faster if refrigerated for about 4 h (S. E. Coulson, personal communication). However, the use of these procedures may cause problems with gel adhesion and stability during transfer to the second dimension. The gels were run for

Two-dimensional gel electrophoresis I

103

18 Vh/cm2 which is probably insufficient for the majority of proteins to focus. Run times may be restricted due to difficulties of gradient drift associated with residual electroendosmotic properties of agarose [ 112, 1131. This prblem may be alleviated by using conditions which have been found to enhance gradient stability [ 1121. It may not be possible to use more powerful denaturants to increase sample solubility as these agents may inhibit the hydrogen-bonding of agarose carbohydrate groups, thus preventing gelation.

6.3 Non-equlibrium methods (NEPHGE) One disadvantage of the original 2-D PAGE system [ l l is severe cathodic drift and consequent loss of basic proteins [91, 1141. The pH gradients in this system typically only extend between 4 and 7. Attempts have been made to extend the pH gradient by synthesis of novel carrier ampholytes or by using isotachophoresis [ l , 1141. Proteins which focused in the region above pH 7, even with extended pH gradients, tended to be streaked and badly resolved [1151. Incorporation of arginine and lysine [ 116, 1171 and alkaline carrier ampholytes I1 171 into the gel, and the use of a strong catholyte [ 1171 have facilitated IEF at a basic pH. Despite these modifications, cathodic resolution of protein remained poor. In our laboratory we have modified the IEF procedure and gel composition to achieve a cathodic pH of 10 at equilibrium in rod IEF gels, but the spots were still severely streaked in the cathode region [92, 1181.

In order to overcome these problems of resolution at basic pH a non-equilibrium system, termed non-equilibrium pH gradient electrophoresis (NEPHGE), was developed 19 1I. A typical NEPHGE separation is shown in Fig. 2b. In NEPHGE, proteins are loaded pnto the acid portion of the gel and electrophoresed for a relatively short time (see Table 3) - the time of electrophoresis determining the protein distribution. The carrier ampholytes used were either pH 7-10 or pH 3.5-10 mixtures. The migration rate of acidic proteins during electrophoresis was found to decrease faster than for basic proteins, some of which continued to migrate at an unchanged rate. Consequently some proteins migrated completely out of the gels by 5000 Vh. NEPHGE separates proteins in the presence of a rapidly forming pH gradient. The proteins are separated by a number of parameters: (1) by pH stacking or isotachophoresis, i. e. by their mobility, (2) the mobility of a protein will progressively decrease due to titration of its ionizable groups, and (3) some acidic proteins will focus [911. Reproducibility of NEPHGE has proven very sensitive to experimental conditions, carrier ampholytes, focusing time, gel length and composition of the sample. Therefore, for comparative studies it is necessary to run samples under identical conditions. Acidic proteins are generally more crowded on NEPHGE gels and some fine charge differences could not be distinguished [911. For optimal resolution of any given sample it has been recommended that both an equilibrium IEF and a NEPHGE gel be run using pH 7-10 carrier ampholytes [911. Sanders et al. [1191 found that using conventional solubilization procedures, histones did not enter NEPHGE gels, This was presumed to be due to ionic interactions between DNA and these extremely alkaline proteins. Digestion with S I nuclease followed by treatment with 4 M NaCl and protamine sulfate allowed the quantifiable release of histones

104

M. J. Dunn and A. H.M.Burghes

Electrophoresis 1983,4, 97-1 16

Figure 2. Typical 2-D separations obtained using the O'Farre11 IEF (a) NEPHGE (b) and flat-bed IEF (c) 2-D PAGE techniques. (a) and (b) from [2151.

[1191. This system resolved histones and their phosphorylated variants, and it was claimed that it was possible to detect a single histone phosphorylation [1191. The disadvantage of this sample preparation procedure is the long incubation time which might result in protein modification due either to endogenous proteases or protease contaminants in commercial nuclease preparations [118, 1201. In their work on histones, Kuhn and Wilt [1211 noted that digestion with S , nuclease and inclusion of arginine, lysine and aspartic acid in the gels alleviated the problems of streaking. They concluded that resolution is inherently less in the NEPHGE system since the pH gradient is steeper so reducing separation between bands. They also observed compression of the cathodic portion of the gradient during focusing, which caused physical distortion of the gel. This effect could be minimiz-

ed by submersing that portion ofJhe gel in catholyte solution [1211. Willard etal. 1521 observed that the original NEPHGE procedure 19 1 I did not adequately resolve all the basic proteins from Novikoff hepatoma cells. They noted that a large amount of material, including histones, precipitated at the point of sample application. Their modified solubilization procedure consisted of incubating cells in the presence of 0.5 % phosphatidyl choline and 8 M urea for 1 h. The DNA and phosphatidyl choline were subsequently removed by centrifugation. Other detergents such as cetyltrimethylammonium bromide were not successful. It is interesting to note that the detergent head group is zwitterionic. The spatial orientiation of this headgroup and/or chain length may be crucial factors in disrupting ionic interactions and achieving effective solubilization.

Electrophoresis 1983,4, 97-1 16

Although non-equilibrium gels allow the separation of basic proteins, they have certain disadvantages. Two types of gel are required for the complete analysis of any sample and some types of sample separate poorly using non-equilibrium gels [ 1171. Another disadvantage of non-equilibrium systems is that they do not function purely as charge separations, as the gel exerts a restrictive influence. Thus the two separation parameters used for the 2-D separation are not completely independent [921. It is important to realize that in theory nonequilibrium and equilibrium methods can resolve different types of charge mutations [1221. If an amino acid with a side chain group with a pK, equal to the PIof the protein is substituted for an amino acid with a neutral side chain group, then no net change in the PI of the protein will occur. However, the mobility of a protein could be affected by such a change [122]. Thus NEPHGE, with its ability to detect mobility changes, has the potential of resolving such mutations. An example of this type of modification is acetylation of histones which results in loss of a single charge unit. O’Farrell et al. 1911 commented that acetylated forms of histone were not distinguished by NEPHGE while single charge subsitutions by phosphorylation have been successfully detected [1191. In the case of basic proteins such as histones, the conversion of an amino group to an uncharged moiety would result, at best, in a very small fractional charge change which would not be resolved by IEF or by NEPHGE at alkaline pH. However, if NEPHGE were carried out using acidic carrier ampholytes there would be an increased probability of detecting such a change. The possibility that amino acid alterations involving charged amino acids might not be resolved by IEF [ 1221 should be taken into consideration in 2-D PAGE experiments designed to investigate polymorphic variability.

6.4 Slab gels for the first dimension

Some investigators have used vertical polyacrylamide slab IEF gels instead of the conventional rod gel system [ 5 1,115, 1231. Slab gels have the advantage of being more consistent between individual tracks which overcomes some of the problems associated with intergel comparisons [ 1231. The disadvantage with vertical slabs is the increased gradient drift which occurs unless the apparatus is specially adapted for this purpose [1241. More recently 2-D PAGE systems have been developed in which the first dimension IEF gel is run as a horizontal slab [40, 99, 1251. In the systems described by the latter two groups of investigators, the first dimension gel was a 1 to 2 mm thick slab focused on a flatbed apparatus. Following IEF, each sample lane was cut out, equilibrated and applied to a second dimension SDS gel. A solution of rapidly polymerizing polyacrylamide was used to cement the first dimension strip to the top of the vertical SDS gel. Goldsmith etal. 1991 have described a system for analysis of a large number of protein samples. Their results indicate that IEF in horizontal slab gels produces more reliable results, and facilitates the screening of large numbers of samples. The IEF slab was sliced while cold, the sample strips wrapped individually and stored. Each sample strip was placed on top of vertical SDS slabs, and sealed into place with agarose. The advantages of this horizontal system were summarized as follows [991: good reproducibility, fast and easy to perform, and slight charge variations could be more readily distinguished.

Two-dimensional gel electrophoresis 1

105

Gorg et al. 1261 have designed a 2-D PAGE system using ultrathin horizontal slab gels in both dimensions. The gels are cast on plastic supports and the focusing gels contain 8 M urea but no non-ionic detergent. Thin gels for the first dimension have the advantage that high field strengths and fast equilibration can be used resulting in improved resolution. However, the resolution of 2-D gels using this system is somewhat disappointing, with marked vertical streaking in the second dimension. In these studies, the gels were backed with cellophane which has the disadvantage of being charged. Moreover it can stretch and may not bind polyacrylamide strongly [ 1031. Silanization of glass or polyester sheets to produce a more reliable gel support was developed by Radola [1031. Plastic supports for polyacrylamide gels are commercially available under the trade names Gel Fix (Serva) and Gel Bond PAG (Marine Colloids Division FMC). In our experience, Gel Bond PAG gave the best results, but in the presence of both NP-40 and 8 M urea these plastics do not bind polyacrylamide reliably [68, 1271. In spite of attempts to extend pH gradients in IEF resolution of cathodic proteins remains poor. We have achieved improved resolution of cathodic proteins with one-dimensional horizontal IEF gels, containing 8 M urea and NP-40, cast on silanized glass plates [681. Based on these observations, we have developed a horizontal slab IEF method for use in twodimensional separations. The IEF gels were cast on specially derivitized plastic supports which c6uld be used in the presence of both urea and NP-40 1681. It was found that these gels, run under conditions to enhance pH stability, gave good cathodic resolution, whereas a conventional gel rod system failed to adequately resolve basic proteins [921. A typical gel using this system is shown in Fig. 2c and is compared with a pair of gels obtained using conventional IEF (2a) and NEPHGE (2b) rod gel techniques. It can be seen that, using the flat-bed system, proteins which normally have to be analysed by NEPHGE are resolved in an equilibrium system. It has been noticed that the cathode end of the gel tended to shrink in rod gels [1211 and to become thin in flat-bed gels [681. It appears that proteins, when applied at the cathode, experience difficulty in traversing this region, resulting in streaking of the basic proteins [921. In the flat-bed system the samples can be applied away from the cathode, but using rod gels the sample unavoidably must be applied at the end of the gel. The rod gel system may be improved by treating the tubes to reduce electroendosmosis of the glass walls [ 1181, by reversing the reservoirs so the catholyte is at the bottom, by inclusion of arginine and lysine in the basic reservoir, and by the use of small electrolyte volumes [loll. Sample application from the anode may make it possible to obtain good cathodic resolution in rod gels, but new problems may occur as a result of introducing proteins into the gel through the extremely acidic region. It has been argued on geometrical grounds that flat-bed slabs are more prone to skewed bands. However, in only a few cases have the patterns obtained using rod systems actually been investigated. In fact, skewed bands can be seen in both types of gels [881. Skewing depends, therefore, on the running conditions of the gels [48]. In vertical gel rods, more mechanical instability is experienced which can result in compression of the gel or its simply sliding out of the tube.

106

M. J. Dunn and A. H.M. Burghes

Skewed bands can then result due to physical disturbance of the gel, or to unequal distribution of electrolyte. Kuhn and Wilt 12 1I have observed that the cathode portion of the gel collapses and can result in uneven penetration of the catholyte into the gel giving rise to skewed bands [48]. Horizontal slab gels are not under mechanical stress during focusing and unequal penetration of electrolyte should not occur [481. It is essential, however, that the electrode wicks be straight and parallel to each other. Good electrode contact is crucial, so wicks must be applied carefully, without trapping air bubbles. The electrode itself should be of a reasonably heavy wire or ribbon. In addition, glass fibre wicks containing no charge residues are to be preferred [941. Skewed bands resulting from salt in the sample are unlikely in a rod gel because the salt will be evenly distributed. In a slab gel, however, there are adjacent salt free zones which give rise to electrical disturbances 1481. Thus, better control of sample salt concentration must be exercised with slab gels. Perhaps one of the chief advantages of using a slab gel system is that it can be bound to a plastic support which produces an immobilized gel which will not stretch [921. A 2-D PAGE system with both IEF and SDS gels bound to a support would be ideal. Another factor which affects band shape and resolution is carrier ampholytes, e.g. we have observed that with Ampholine some anodal distortion occurs [92, 1281. However, a mixture of commercial carrier ampholytes, in combination with glass fibre wicks and parallel electrodes will produce bands at least comparable to those obtained with rod gels. In our opinion horizontal slab gels have the advantage of being easier to cast and run than gel rods. There is also greater flexibility in the point of sample application in slabs, although it is more difficult to load large sample volumes onto these gels. However, the success of a flat-bed 2-D PAGE system depends upon reliable binding of the gels to supports. In our opinion the optimum support would involve true covalent binding of polyacrylamide to a thin plastic sheet [921. 6.5 Monitoring of pH gradient and carrier ampholytes

The protein spots on a 2-D gel can be characterized in terms of their pl’s and molecular weights. This procedure requires calibration of both gels. In most published procedures, pH gradients were determined by slicing the focusing gel and eluting in water or 8 M urea. In some instances, 10 mM KCl was added to increase the conductivity of the solution [ 5 11. The use of water has the disadvantage that the concentration of urea in the final solution measured is not known. Measurement of pH in 8 Murea is difficult as urea decreases the activity coefficient of H’ ions [1291. In addition, urea affects the proteolytic group because the solvent is changed [130, 13 11. Such measurements are also carried out on a separate gel and do not allow for variations or stretching. These methods of measurement are, therefore, inherently inaccurate. In an alternative approach Bravo et al. 1132,1331 characterise the focusing dimension by measuring protein mobility relative to p-actin. This type of procedure is also used in normalization procedures for comparing gels and depends on using an invariant protein as the reference spot. Perhaps the best system developed so far for PIcalibration is a set of carbamylation standards derived by heating proteins

Electrophoresis 1983, 4, 97-1 16

in a urea solution for varying lengths of time. Steinberg et al. [1341 used mild carbamylation to calibrate the extent and amount of charge shifts on two-dimensional gels, and Anderson and Hickman [ 1351 further developed carbamylated proteins for calibrating 2-D gels. Carbamylation results in a horizontal row or “train” of spots in the IEF dimension which can be used as markers. The train is numbered from its native, uncarbamylated form (spot 0) and counts directly the charge added to the native molecule. Different proteins will show varying distances between spots and the length of the train will depend on amino acid composition. In NEPHGE gels, the standards are useful for identification and localisation purposes, but they do not indicate the pH at the alkaline end of the gels [1361. These standards are useful for accurate positioning spots on gels, and could be calibrated to account for the effects of urea. Some problems in pH gradient have arisen as a result of carrier ampholytes. When examining urine patterns, Frearson et al. [ 1371 found that Tamm-Horsfall protein could remove basic carrier ampholytes. They also observed a discontinuity in the pH gradient which increased in width proportionately to increasing protein load. There have also been comments in the literature on batch variability of Ampholines [88, 138, 1391. By means of Schlieren optics, ridges have been observed in IEF gels [ 1401 which may represent clumps of carrier ampholytes implying that the distribution of carrier ampholytes is not continuous [1411. Discontinuous zones occur particularly in the alkaline region [141, 1421. The ridges may correspond to the hydration state of the gel [ 1401 and in the alkaline end may be the result of the severe thinning of the gel. Carrier ampholytes from various manufacturers have been compared using carbamylated creatine kinase standards. [1391. It was found that 33 spots were resolved regardless of the commercial carrier ampholyte used. However, the relative distribution of the spots differed with the brand of commercial carrier ampholyte, Pharmalyte yielding the most even spot distribution. Separation patterns and pH gradient profiles obtained using various commercial carrier ampholytes have been compared in our laboratory both in the presence and in the absence of 8 M urea and non-ionic detergent, using polyacrylamide gels [92, 1281 as well as in agarose gels under native conditions [I 121. We found that the pH profile was not the same for each commercial carrier ampholyte, and that this was demonstrated by shifts in the protein profiles in the IEF gels. Allen [ 1431 has also reported variability in ,proteinpatterns using different commercial carrier ampholytes. While making a similar comparison, Taylor et al. [1441 observed that different carrier ampholytes introduced distortions between gels which made computer matching difficult. In addition, different commercial carrier ampholytes are better in different pH regions. Pharmalyte is best in the pH 5-8 region, Servalyt in the acidic region and Ampholine in the alkaline portion of the gel 192, 112, 1281. Establishment of the pH gradient is further complicated by high sample loads, as proteins can themselves act as ampholytes [1391. Low sample loads do not appear to affect the pH gradient significantly. High protein loads at the basic end of the gel have also been shown to result in greater cathodic drift and to distort the pHgradient [1451. However, this may be due to the presence of large amounts of P-mercaptoethanol which becomes ionized above pH 9[ 1461. This can be overcome by applying the sample at the anode and by replacing P-mercaptoethanol with dithiothreitol [ 1461.

Electrophoresis 1983,4, 97-1 16

Two-dimensional gel electrophoresis I

107

Carbamylation standards are excellent for monitoring and acrylic acid had to be removed. Best results were obtained by calibrating pH gradients for comparison of proteins. Howev- reacting (70 "C, 20 h) pentaethylenehexamine, tetramethyler, there are difficulties in using these standards to monitor enepentamine, triethylenetetramine, and hexamethylenethe appropriate distribution or number of carrier ampho- tetramine in the ratio 60:25:10:5 with acrylic acid added lytes. The process of carbamylation decreases the charge of dropwise until the desired N:COOH of 2:l was reached the protein by what has been called a "unit charge" [1351. [ 1531. The carrier ampholytes were then further fractionated Unit shifts in charge will only occur below pH 8.7-9.0 [122, to remove contaminating by-products taking care to avoid 1351. Proteins with amino acid substitutions will have frac- redox reactions caused by exposure to the catholyte or tional charge changes, e. g. aspartic and glutamic acid in the anolyte 11541. acidic region (up to pH 5), arginine and lysine up to 8.7-9.0 in the basic region, and histidine between pH 7 and 6 [ 1221. Alternative syntheses of carrier ampholytes have been develIn addition, various side chains can influence the pK, of the oped [ 155, 1561. Ethyleneimine can be condensed with progroup thus yielding a wider variety of charge changes. pylenediimine to obtain a polyamine. The product is reacted Furthermore, the precise amino acid composition of a pro- with propane sulfone and chloromethylphosphonic acid by tein can place it almost any fractional charge distance from alkylation. These carrier ampholytes are available commeranother protein. The positions of the carbamylated stand- cially (Serva) under the trade name Servalyt [1551. The ards and the magnitude of the charge shifts will depend on amine synthesis used in Servalyt preparation results in more the amino acid composition of the original protein [ 122,135, possible isomers [ 1551. Replacement of acrylic acid with 1361. Since carrier ampholytes are composed of varying itaconic acid produces carrier ampholytes with improved amino and carboxylic acid groups, they can produce a varied charge properties in the pH 5.5-6.5 range [ 1571 which can be distribution in the same way as fractional charge differences mixed with conventional carrier ampholytes to substantially in proteins. Hence, gaps which occur between carbamylation improve the buffering capacity and conductivity of the midstandards do not necessarily indicate a lack of carrier am- pH region. However, to our knowledge this type of carrier pholyte species in these regions, but rather may indicate that ampholyte is not present in commercial preparations. More a large number of carrier ampholytes exist between the two recently, Parmalyte (Pharmacia) has been synthesized by carbamylation standards. The resulting gap could be de- condensation of glycine, glycyl-glycine and amines of scribed as a local pH gradient flattening. That proteins with selected pKas with epichlorohydrin [ 1581. D,L-Isomers of fractional charge differences do exist is evident from Fig. 3 epichlorohydrin and amino acids are incorporated into the and 4 of Tollaksen et al. [I391 where some protein spots are reaction mixture so as to increase the heterogeneity of the present between the carbamylated standard spots. Thus car- mixture. It is interesting to note that Pharmalyte probably bamylated standards are not suitable for optimising the pH contains tertiary amine groups [1591. The main buffering gradient for particular protein mixtures. The protein sample groups for Pharmalyte with their respective buffering pH itself must be used to establish what mixture of carrier am- ranges are: a-aminocarboxylic, pH 2-3 ; glycyl-glycine resipholytes should be used to distribute the spots over the whole dues, pH 3-4; P-hydroxylamines, pH 4-9; and dialkylgel. aminopropyl, pH 9-11 [1591. The heterogeneity of the starting polyamines is important in determining the diversity of the resulting carrier ampholytes. The quality of the IEF separation is inextricably linked to the Charlionet [ 1471, using Vesterberg-type synthesis, has insubstances forming the pH gradient. i. e. carrier ampholytes. creased the diversity of carrier ampholytes by reacting As proteins can vary by fractional charges it is essential to polyamines with 2,3-epoxypropanol- 1,1,2,7,8-diepoxyocmaximise the number of possible carrier ampholyte species tane and N,N'-methylenebisacrylamide and acrylamide - the present in the gel [147, 1481. Highly heterogenous carrier first compound yielding the better results. By this procedure, ampholytes with good batch-to batch reproducibility would primary and secondary amino groups in polyamines are considerably increase the resolution and reproducibility of transferred into secondary and tertiary amines, as well as 2-D PAGE. Vesterberg [241 first described the synthesis of secondary hydroxyl groups. Charlionet [ 1471, by empirically carrier ampholytes by an ingenious procedure involving adapting this scheme, obtained a heterogeneous mixture of coupling of propanoic acid residues to polyethylene poly- modified polyamines, which on reaction with acrylic acid amines 11491. This can be accomplished by reacting poly- produced a highly diverse carrier ampholyte preparation ethylene polyamines with acrylic acid at 70°C [24, 142, with good conductivity, buffering capacity and increased 1501. Other, more acidic and basic carrier ampholytes have resolving power. However, only triethylenetetramine and been synthesized extending the pH gradients from pH 2.5 to tetramethylenepentamine were used in the synthesis. The in11 [ 15 1, 1521. The Vesterberg-type carrier ampholytes (Am- clusion of pentaethylenehexamine now available from pholine, LKB) have been the type most commonly used in 2- Calbiochem (or see [ 1601) would have further increased the D PAGE, although alternative ampholytes have been intro- heterogeneity of the carrier ampholyte. Just [ 16 I] has recentduced recently by other manufactures, e. g. Servalyt (Serva) ly described a method whereby the two reactants penand Pharmalyte (Pharmacia). Vinogradov et al. [ 1501 and taethylenehexamine and acrylic ester were mixed proporRighetti et al. [ 1421 have described the synthesis of carrier tionately with an Ultrograd (LKB) gradient mixer and passampholytes using the same chemical reactions as Vesterberg. ed continually through a reaction coil at 40 "C. The gradient Acrylic acid was reacted with various polyamines, using a shape could be adjusted to obtain the desired ratio. The nitrogen:carboxyl ratio of 2: 1 in order to obtain complete re- methyl ester was used instead of acrylic acid because of its action of the acrylic acid [ 1421. Acidic range carrier ampho- high reaction velocity, and since its excess could easily be lytes could be synthesized separately, but contaminating removed by vacuum distillation. Following synthesis, the es6.6 Carrier ampholytes

108

M.J. Dunn and A. H. M.Burghes

ter was hydrolyzed by heating at 120 “C for 2 h or in a water bath at 40-60 “C until the pH was constant. The physical properties of various commercial carrier ampholytes have been compared [ 1621. Their buffering capacity is not important for analytical IEF separations, and resolution is inversely proportional to 1/4th power of the local conductivity. Therefore, in analytical separations it is most important to have carrier ampholytes with a uniform, low conductivity in the isoelectric state 194, 147, 1621. Pharmalyte has been found to have the most even conductivity [1621. It also appears that the acidic carrier ampholytes in Ampholine are synthesized separately with acetic acid groups and lower polyamines [1621. As can be seen from the preceding discussion, carrier ampholytes are synthesized by different procedures and, therefore, contain different species. It follows that blending of the different carrier ampholyte species should result in a mixture containing more species with different fractional charges and thus incorporate the advantages of each preparation. It is essential, however, to blend carrier ampholytes of the correct pH interval and in the optimal proportion for the protein mixture under investigation. In our own work with whole fibroblast lysates we have demonstrated that mixtures of carrier ampholytes improved the separation of complex IEF patterns, and thus enhanced resolution in the second dimension gel [92, 1281. Clearly it should be possible to improve the diversity of carrier ampholytes by using a mixture including itaconic acid-type carrier ampholytes, carrier ampholytes prepared by the Charlionet procedure [1471 to increase amine diversity, polyamine-derived carrier ampholytes, and by mixing in other types of commercial carrier ampholytes (e. g. Pharmalyte, Servalyt). Further control over the synthesis using methyl ester and the Just [ 1611 procedures may result in improved batch-to-batch reproducibility. In summary, individual carrier ampholyte preparations have not been optimized for either resolution or reproducibility. Therefore, the careful control of carrier ampholyte synthesis and the use of mixtures of the various carrier ampholyte types will produce heterogeneous carrier ampholytes which should increase resolution in the IEF dimension of 2-D separations. 6.7 Buffer focusing and other alternatives

Recently, a technique termed buffer focusing has been described [ 1631, for which a mixture of 47 buffers is particularly suitable [941. The advantages of this mixture are said to be increased gradient stability and absence of binding to proteins [941. In addition, it may also prevent ionic interaction between proteins. This particular mixture has not been used in 2-D PAGE systems, but Polybuffer (Pharmacia) has recently been found to give good results when used in place of carrier ampholytes [1641. This is not surprising since it contains similar groups to Servalyt. Less streaking with Polybuffer was claimed, but streaking in the cathode end is evident in published photographs of gels [ 1641. Spacers can also be added to modify selected regions of pH gradients [ 1651. Arginine and lysine have been used particularly to reinforce the cathode in 2-D gels [ 116, 1661. Likewise, aspartic and glutamic acids have been added to reinforce the acidic portion of the gel [ 1671. Arginine and lysine can be incorporated into the electrode solutions to stabilize the cathode [871.

Electrophoresis 1983, 4, 97-1 16

An interesting recent development has been the introduction of “Immobiline” (LKB) for IEF in immobilised pH gradients. The Immobilines are a series of acrylamide derivatives in which the buffering groups responsible for creating the pH gradient are covalently attached to the polyacrylamide backbone. Tertiary amines are used to produce basic Immobilines and carboxylic groups for the acidic species [1681. Appropriate mixtures of the basic and acidic Immobilines are then incorporated into acrylamide solutions which are used to prepare IEF gels using a gradient casting technique [169, 1701. The Immobilines polymerize with acrylamide via vinyl bonds and in this way a pH gradient of a defined, narrow range can be obtained. Unfortunately, only narrow range pH gradients can be constructed at present. Although this is useful for examining specific regions of 2-D gel patterns, it is not practical for many 2-D PAGE applications. Immobilized pH gradients offer a number of advantages which are particularly attractive for 2-D PAGE, e. g. good reproducibility, and the ability to withstand high voltages and long focusing times without drift. Disadvantages include increased focusing time, high electroendosmotic flow due to H + and OH- imparting a net charge to the medium above pH 9 and below pH 5, longer time for sample entry into the gel which increases the risk of protein precipitation, and ion exchange effects of the medium prior to electrophoresis [ 169, 1701. Some form of charge balancing with non-buffering negative and positive charges in appropriate regions of the gradient may alleviate the problem of electroendosmosis at extremes of pH. If truly immobilized, broad pH gradients could be obtained it would represent a considerable advance for 2-D PAGE technology.

7 Equilibration and transfer between dimensions In the original O’Farrell procedure 11, the IEF gel is equilibrated before application to the second dimension gel. This procedure can result in considerable loss of protein especially when extended times are used. Loss of as much as 5-25 % or more of the original sample has been reported [l, 92, 171, 1721. In addition a serious loss of resolution occurs as the protein bands can broaden by 40 % using a 30 min equilibration time [ 11. Therefore, short equilibration is advantageous. Some investigators routinely omit this step and apply the IEF gel directly to the second dimension slab [70, 1731, the SDS in this case being allowed to concentrate in the stacking gel. However, O’Farrel [ 11 found that omission of the equilibration step resulted in considerable streaking of high molecular weight proteins. In order to reduce diffusion after IEF, Garrels [611 recommends cooling the gels by touching them against dry ice, thus precipating the urea, in combination with a very short (2 min) equilibration step. Although short equilibration times will reduce diffusion and sample loss, it could be argued that short equilibration results in decreased elution of proteins into the second dimension. Fixation with TCA prior to equilibration has been recommended as protein loss during equilibration can make quantitative measurements difficult [ 1111. However, this procedure has not been investigated to ensure that more proteins were not left behind in the IEF gel than by conventional methods. Jackle [ 1471 recommends fixing the gel in methanol/acetic acid and staining with Coomassie Brilliant Blue (R-250) before equilibration and transfer to the second

Electrophoresis 1983, 4, 97-116

dimension. Gorg et al. [ 1751 reported that it is best to utilize Coomassie Blue G250 as R250 leads to streaking. These stains apparently are electrophoretically removed from the proteins but complete elution of proteins from the firstdimension gel should be demonstrated. Coomassie dyes might, however, compete with the same binding sites as SDS and therefore inhibit the interaction of the detergent with the protein. It has also been questioned whether SDS reacts fully with the proteins during equilibration. Using erythrocyte ghosts, it has been found that SDS in the presence of 8 Murea gave the same SDS pattern as ghosts that had been boiled in SDS sample buffer (M. J. Dunn and A. H. M. Burghes, unpublished observation). Hirabayashi [ 1111 has obtained a similar result. It is our feeling that boiling acts to unfold the proteins allowing full reaction with SDS. Urea being a powerful denaturant, probably also acts to unfold the proteins allowing SDS to react fully. In this regard, prior fixation and washing may remove urea, but it is carried out under conditions which may retain the denatured state of the protein. It is interesting to speculate that Coomassie dyes complex with proteins in a similar manner to SDS. The other features of transfer is the cementing of the first dimension IEF gel to the second dimension slab gel. Fastpolymerizing polyacrylamide can be used to embed the IEF gel on top of the SDS-slab gel [701. This process may result in loss of protein [471 presumably due to its reaction with the gel during polymerization. Agarose has been found not to result in retention of protein using Coomassie Brilliant Blue stain [471, but this should be confirmed using more sensitive detection methods. Garrels [6 1I does not cement the gel in place but relies on the fusion of the polyacrylamide during electrophoresis. In most cases where sensitive detection methods such as autoradiography have been used, a considerable amount of labelled material remains at the top of the second dimension gel, e. g. Fig. 3 [611, Fig. 1 [1761; Fig. 3 [921. Whether this is high molecular weight protein or aggregated material is not known. A major danger during transfer of the IEF gel is stretching of the gel during handling procedures. This can result in skewed bands and complicates pattern matching between gels. Binding of focusing gels to plastic sheets should alleviate this problem when using slab gels in the first dimension [921.

8 SDS dimension 8.1 Stacking

In the 2-D PAGE system of O’Farrell [ 11 a stacking gel was used in order to concentrate the proteins from the focusing gel into the second dimension SDS gel. However, Garrels [6 11 has reported that it is unnecessary to use a stacking gel due to the fact that the sample is relatively concentrated. Perhaps more importantly the focusing gel is non-restrictive and therefore itself will serve to concentrate the proteins [6 11. The nature of stacking is somewhat modified in the presence of SDS. As a result of the high charge of SDS and the constant charge to mass ratio of the SDS-coated proteins, all proteins will move with the same mobility. The discontinuous buffer system normally used will concentrate the proteins. In addition, the front boundary will be slowed by the restrictive polyacrylamide allowing further concentration of the bands

Two-dimensional gel electrophoresis I

109

without the need for a separate stacking gel. Above neutrality it is not necessary to have a discontinuity in pH [ 1771 and unstacking of the proteins will occur by changing the gel concentration [ 1771. It should also be noted that SDS will act as a separate ion boundary to the C1- ions [85l. Allen ef al. [1781 have described a system which results in stacking of proteins using a constant pH but with different ionic strengths. Possible problems that might be encountered with 2-D systems have not been investigated. Urea may act to increase the viscosity of the IEF gel thereby making it equivalent to a more restrictive gel (see section 6.2) and making protein elution more difficult. Adaptation of the discontinuous buffer system to obtain a steep voltage gradient [851 might be advantageous in obtaining rapid and complete elution of protein from the IEF gel. In addition, there are high levels of carrier ampholytes and protein so that the buffer concentration should be high enough to prevent these materials affecting pH. Non-ionic detergent will form mixed micelles with SDS, which will effectively reduce mobility of SDS due to the presence of non-ionic components. These features may change the stacking limits as compared to standard SDS gels. 8.2 Gel size Recently very large gels have been recommended for 2-D PAGE. Gel size was increased 2.5-fold over that used in the original O’Farrel system, producing an IEF gel length of 32 cm and an SDS gel of 40.64 cm in length [901. A three-fold increase in resolution was obtained as measured by the number of spots detected. This was attributed to two factors. First, by applying more radioactive sample, many minor proteins present at low concentration could be detected. Secondly, autoradiographic spreading was minimized by the large area of the slab gels. As the spots were separated over a larger area and there was consequently a greater distance between the spots, fewer small spots were obsured by the major spots 1901. In theory the resolution of 10 cm and 20 cm IEF gels can be made similar by increasing the field strength in the shorter gel 1179, 1801. Therefore, a gel run to equilibrium with a high field strength should give equivalent resolution to a longer gel run at a lower field strength. Resolution in the second dimension gel can be increased by lengthening the SDS gel. In this respect the SDS gel is analogous to DNA sequencing gels [181]. Thus an IEF gel of normal dimension coupled with a longer second dimension SDS gel should yield good resolution, and such systems have been used for 2-D PAGE [ 182, 1831. In small IEF gels run under high field strengths the bands will be more highly concentrated than in large IEF gels run under a lower field strength, i. e. showing equivalent theoretical resolution. Thus as the rate of protein diffusion is proportional to concentration, the small gel will exhibit greater diffusion of bands during equilibration and transfer to the SDS gel. Therefore, large 2-D gels run at moderate field strength will show higher resolution than gels using a small IEF gel at high field strengths. Some investigators have recommended micro systems [47, 1841. Poehling and Neuhoff [47] have described the running and casting of microgel systems using both gradient and non-

110

Electrophoresis 1983, 4, 97-1 16

M. J. Dunn and A. H. M. Burghes

gradient techniques. Riichel’s system [ 1841 is extremely small, literally the size of a “postage stamp”. These miniaturised gels have the advantage of using small quantities of reagents. In addition, when the gels are thin, staining and destaining times are considerably reduced. However, distances between spots become very small, thus potentially decreasing resolution [ 1751. In summary, then, the system which should posses the best resolution would be thin gels of large size. To handle such gels on a routine basis it would be necessary to have them firmly bound to a solid support. 8.3 Features of SDS gels Best results are obtained if the upper edges of the SDS gels are perfectly flat. This can be achieved by thoroughly cleaning the glass plates and by careful overlayering of the gels. Vertical streaks are quite often seen in 2-D gels when samples such as serum are applied, which contain proteins in high relative abundance [ 1851. It should be remembered that IEF gels have a higher loading capacity than SDS gels, and thus a sample which focuses well can subsequently overload the SDS gel resulting in vertical streaking. O’Farrell [11 found that carrier ampholytes migrated to the front of SDS gels much as small proteins would. NP-40 forms mixed micelles with SDS and also migrates to the gel front, apparently with no detrimental effects [ 11. Whether these mixed micelles interfere with the solubility of certain proteins has not been investigated, but in the presence of excess SDS solubilization does not appear to be impaired. We have observed that with alkaline carrier ampholytes in the presence of NP-40 there is considerable trailing (M. J. Dunn and A. H. M. Burghes, unpublished observation) which may be the result of the alkaline carrier ampholytes and detergents forming strong complexes [781. These complexes can be removed by prolonged washing of the gel in alcohol/acid mixtures. The use of high field strengths for SDS electrophoresis should decrease lateral diffusion of spots during electrophoresis. However, Joule heating may become a problem at high voltages and therefore more sophisticated cooling systems may be required. 8.4 Gradient engineering O’Farrell [ 1I noted that the distribution of proteins in the gel affects resolution. If proteins are clustered into small regions of the gel resolution will be degraded [ l l . Although proteins can differ greatly in their PI,studies on the distribution of protein prvalues showed that they tend to cluster around certain mean values [ l , 1861. O’Farrell [ l l tested a series of carrier ampholytes in order to optimize the pH gradient so that an even distribution of proteins over the gel area was obtained for E. coli. Due to the instability of the pH gradient in this system it was not possible to extend the gradients above pH 7 without severe loss of resolution. This is not ideal for mammalian cells which tend to show a greater number of alkaline proteins [1141. Also, published distributions of protein pZvalues [ 1861 may be biased towards acidic and neutral proteins. Subunit molecular weights for a large series of proteins have also been shown to cluster around a mean value [ 1861. A gel of a single polyacrylamide concentration is unlikely to produce a uniform distribution of protein spots [ 11 and will result in the loss of some proteins unless a series of

gels of different polyacrylamide concentrations is used [ 1871. A linear polyacrylamide gradient is also unlikely to distribute proteins uniformly since they are often clustered around a mean molecular weight. O’Farrell [ 1 I found that an exponential polyacrylamide gradient gave good resolution for E. coli proteins. Since O’Farrell’s original work [ 11, very few investigators have considered the distribution of proteins over the area of the gel when designing a 2-D PAGE procedure. The optimization of protein distribution will allow the maximum area of the gel to be used for effective separation of the proteins and thereby increase resolution. For most preparations it would be advantageous to maintain the two extremes of the pH gradient while flattening the region of the gradient where the majority of proteins lie. We have achieved this using a flat-bed IEF procedure, where the pH gradient extended to pH 1 1 in the cathode, by mixing narrow range carrier ampholytes to give the desired pH gradient shape [921. We have also described a method of reproducibly casting SDS polyacrylamide gradient gels with a flattened mid-region which evenly spreads proteins in the SDS PAGE dimension [921. Thus, by manipulating gradient shape in both dimensions, the distribution and resolution of proteins in the particular sample preparation can be optimized. A series of SDS gels of different concentrations can be run instead of a single gradient gel [611. This approach has the advantage of simplicity of gel casting. However, providing that care is taken while casting, reliable and reproducible gradients can be obtained and gradient gels can be used to examine high and low molecular weight proteins simultaneously [9, 10, 1881. We believe that for initial examination of a sample it is easier to use a single gel, optimized to produce a uniform distribution of spots, than a series of narrow-range carrier ampholytes and gels of different polyacrylamide concentrations. On the other hand, narrow range pH gradients and single concentration polyacrylamide gels can be very useful for analysing particular regions of interest and to increase resolution in that area. A potential problem with gradient gels is that the density of the polyacrylamide may affect the penetration of radioactivity or stain. This effect may need to be taken into account when using quantitative methods. 8.5 Gradient casting Various procedures have been described for casting gradient gels (reviewed in [ 1891). The Isodalt system [I881 for casting multiple slabs is shown in Fig. 3. The two acrylamide solutions are placed in the gradient maker, from which they flow into a mixing chamber. The volume of the chamber necessary to ensure complete mixing will depend upon a number of parameters: (1) the rate of flow of acrylamide, which should be slow enough to prevent mixing in the casting chamber; (2) the rate of mixing of the solution; and (3) the viscosity of the solution. One can test whether adequate mixing has occurred by incorporating Bromophenol Blue into one of the solutions and densitometrically scanning the polymerized gels to check that the appropriate gradient has been formed [47, 1261. So that slow flow rates can be used, the speed of polymerization of acrylamide can be adjusted by changing the concentration of the catalyst. A sealed mixing chamber is

Electrophoresis 1983, 4, 97-1 16

Figure 3. Schematic drawing of the Isodalt gel casting apparatus. (A) gradient former; (B) light acrylamide solution; ( C )dense acrylamide solution; (E) magnetic mixer; (F) line to vacuum, (G) inlet for underlay solution; (H) reservoir for underlay; (I) 10-exit manifold; (J) bubble outlet; (K) peristaltic pump; (L) rotatable gel casting chamber. From [ 1881.

Two-dimensional gel electrophoresisI

111

Figure 4 . Apparatus used for casting gradient polyacrylamide gels. An LKB Ultrograd 11 300 (a), fitted with the appropriate gradient template (b), controlled a three-way valve (e) for the dense (c) and light (d) polyacrylamide solutions. The solution in the mixing chamber (h) with its plunger (i) was mixed by a magnetic flea (8) and stirrer (0. Fluid flow was controlled by a three-channel peristaltic pump (n). All lines (k) were of narrow-bore tubing. The perspex casting tower (n) was fitted with three inlets controlled by three-way taps (I) allowing fluid to be run to waste (m). From 1921.

preferable [921 so that very high mixing rates can be used without the formation of air bubbles. The process of polymerization should proceed downwards from the top of tion of a microcomputer for the Ultrograd to control the 3the gel in order to prevent convective mixing 19, 10, 1901. way valve would make the system less expensive and more This can be achieved by suitably adjusting the catalyst con- versatile. centration in the light and dense solutions. Gradients can be cast from the bottom using either displacement by gravity or a peristaltic pump which gives better control of flow. The chamber containing the slabs has to be designed so as to distribute the gel solution evenly across the width of all the gels and there should be a mechanism for slowing the solution as it enters. In the Isodalt system [ 1881, the slab holders are maintained in a tilted position so as to form a V-shaped funnel in the lower corner of the holder. Alternatively a V-shaped cone can be placed at the bottom of the holder and mutiple inlets used [9, 10,921. After the gradient is poured, the dead volume can be displaced using a dense glycerol solution [1881. It is also possible to cast gradients from the top of the chamber [1261, but it is our feeling that this can lead to considerable disturbance of the gradient and does not allow several gels to be cast simultaneously. Thin gels possess several advantages, especially in autoradiographic procedures. However, the casting of thin gradient gels is complicated by the fact that the capillary action of glass plates can disturb gradient formation. Gorg et al. [ 1261 have described a method for casting such gcls, but only a single gel can be cast at one time. Poehling and Neuhoff [47] have described a method for casting multiple gradient gels with thicknesses from 0.1 to 1 mm. The gel cassette is initially filled with a solution of 0.1 % w/v SDS which eliminates capillary forces. The gradient is then poured into the chamber by upward displacement in the normal manner. The less concentrated acrylamide solution can contain a low concentration of a viscosity agent to prevent mixing with the SDS solution [471. Recently, an electronic gradient forming apparatus (Ultrograd; LKB) has been used for casting gradient gels [92, 119, 1911. This device produces more reliable and reproducible gradients [1921. Our system [921 is shown diagrammatically in Fig. 4. The Ultrograd consists of a device which scans a template of the desired gradient shape and controls a 3-way valve. This allows gradients of any shape to be cast repeatedly and reproducibly. The substitu-

8.6 Molecular weight standards

Molecular weight standards have been implemented in two ways. In the first method standard marker proteins are electrophoresed along one edge of the gel. Alternatively whole tissue homogenate is added to the agarose used to seal the first dimension gel in place [1931. These standards give reference bands which stretch all they way across the gel. In the case of gels of radioactive samples, it may be possible to calibrate the autoradiographs following exposure by staining the gel. It should be noted that non-ionic detergent interferes with the use of Bromophenol Blue as a marker of the gel front [1941. 8.7 Multiple gel techniques

Systems have been developed which allow multiple 2-D gels to be run simultaneously. These provide high output capacity for screening studies, and improve intergel comparisons as samples run within a batch of gels are likely to be more reproducible than those electrophoresed in different batches [89, 1881. In the system of Garrel’s [611 the IEF gels consist of 1.2 mm diameter tube IEF gels. Four first dimension gels are electrophoresed simultaneously on a large rectangular second dimension vertical slab gel which is cut into four pieces for subsequent processing and quantitation. Anderson and Anderson (89, 1881 have developed a multiple gel system termed ‘Isodalt’ in order to carry out screening studies. The Isodalt system is capable of running simultaneously a large number of tube IEF gels and SDS PAGE gradient slab gels. The IEF gels are cast twenty at a time by displacement of water into tubes (1.5 1 mm i. d.) using the ‘Iso’ apparatus, and run in the same apparatus. This procedure considerably simplifies casting of IEF gels in rods and allows more consistent control of gel length. A ‘Dalt’ apparatus for simultaneously casting ten SDS gradient gel slabs has also been designed [ 1881 (Fig. 2). The ten gels are electrophores-

112

M. J. Dunn and A. H.M. Burghes

ed simultaneously in a three-compartment tank. Recently an improved version of the Isodalt system was produced which uses less buffer [ 1951. It is also possible to use this type of apparatus with larger gels (30 x 30 cm) and to increase the number of gels cast simultaneously [ 1871.

8.8 Binding gels to supports A problem which arises with multiple gel techniques is the handling of large numbers of SDS PAGE slab gels. Recently, it has become possible to bind polyacrylamide to glass using silanes [lo31 or to a plastic backing, e. g. FMC GelBond PAG 11961, Serva Gel-Fix, or other treated plastics 168, 92, 1031. It appears that GelBond PAG is more reliable than Gel-Fix. The vinyl bonds on the surface of these plastics are heat and light sensitive and this can render the plastic unusable. In our hands, GelBond PAG was unsuitable for overnight fixation of gels in 20 % TCA as they became detached from the plastic. Under the same conditions gels remained adhered to silanized glass. A specific drying procedure must be followed with gels bound to supports in order to prevent cracking and curling of the gels [921. Anderson [1971 has used dehydration with ethanol, but it is not known whether this results in protein loss. Acetone can also be used (M. J. Dunn and A. H. M. Burghes, unpublished observation) and should prevent resolubilization of proteins. The binding of gels to supports, especially plastic, offers considerable advantages as standard photographic frames and tanks can be used to handle large numbers of gels [1971. An important advantage of having gels bound to supports is that no stretching of the gel can occur during the various procedures. This results in increased reproducibility and more symmetrical spots [921.

9 Estimates of resolution An important factor when considering resolution of a 2-D separation is the number of proteins to be separated in any given sample. It has been calculated [1981 that in a typical mammalian cell no more than 2000 proteins are expressed at a physiologically significant level. These proteins would be expected to result in 3000 to 4000 polypeptide spots in a 2-D separation. It can be argued that several proteins could have coincident positions on a 2-D gel. However, in practice this has not been found to be the case, for when two proteins are in similar locations they tend to displace each other and form separate spots [ 1991. How many of these proteins can actually be resolved by 2-D PAGE? Some theoretical estimates have been made concerning resolving power of two dimensional gels. O’Farrell [ 1I found using E . coli that 70 bands could be separated by IEF and 100 bands by SDS PAGE. Thus, a combination of the two procedures should be capable of resolving 7000 spots. Resolution in both dimensions was maximized by adjusting the distribution of the proteins over the whole gel area. However, the spots in the 2-D gel were 40 % greater in size than in the IEF gel, due to lateral diffusion which occurred during equilibration, transfer, and SDS PAGE [ 11. Hence, the resolution dropped to 70 % of the theoretical value yielding a maximum resolution of 5000 spots. Thus, theoretical estimates, such as that of Anderson and Anderson [2001 at 10 000 spots (simply the product of the number of bands observed in each dimesion run independently) should be treated with some caution.

Electrophoresis 1983, 4, 97-1 16

Another factor that affects resolution in 2-D gels is spreading of spots due to the detection system used. A band in a onedimensional gel can be composed of several proteins which can be resolved into several spots in a 2-D gel. The density of staining or length of autoradiographic exposure required to detect the single band on a one-dimensional gel will be less than that required to detect the multiple spots in a 2-D gel. This can result in overstainina or overexposure of the more major components which will lead to spreading of these spots. The magnitude of this effect is dependent on the particular sample as it will be more pronounced if some proteins are present at much higher relative abundancies than others. The effect could be estimated by comparing the width of protein bands in one-dimensional SDS PAGE with the diameter in the SDS dimension of the spots due to the same proteins in a 2-D separation. O’Farrell [ 1I also found that increasing autoradiographic exposure time to detect minor proteins caused a loss of resolution. Thus the ability to detect trace amounts of proteins is limited by spreading of spots rather than by the absolute sensitivity of the detection method used. The ideal detection system, in addition to being sensitive and possessing a high dynamic range, would not be prone to lateral spreading of the spots when saturation is approached. Other general factors which decrease estimates of resolution are streaking of high molecular weight proteins and insolubility of components in the first dimension. Considering these features it seems likely that the theoretical estimate of resolution would fall into the range actually obtained in practice. This would appear to be about 1000 to 2000 spots. Even using giant gels only 1750 spots were observed [901. Another feature that O’Farrell discussed is the “noise” or background level due to normal turnover of major proteins [2011. This may be difficult to distinguish from a protein present in very small amounts. Recently a computer programme has been developed to measure resolution [202l. However, the exact definition of resolution is not clear. In this review the term has been used in a general sense and refers to the number of protein species observed and the clarity of the separations. The programme [2021measures spot size relative to gel area and uses this as a criterion for assessing resolution. This approach can be criticised for, as a consequence of spreading of spots by the detection system, the area of the spot will vary with its relative abundance or with the autoradiographic exposure that has been used. The system assumes that proteins are spread evenly over the gradient, but this is not necessarily valid. Considering these factors, the number of spots that can actually be resolved in practice may be a better criterion of resolution.

10 Heterogeneity and artefacts There are various forms of translational modification of proteins that can occur which will introduce microheterogeneity. If these changes result in an alteration of charge then a single protein spot can be transformed into a row of spots 11. Such charge modifications occur due to phosphorylation, deamidation, acetylation and the addition of sialic acid residues. In addition, glycoproteins can show anomalous migration in SDS gels. Interestingly, this effect can be partially overcome using borate buffers in SDS PAGE as borate will bind to the cis-hydroxyl group of the Sugar thus compensating for the mobility change of the protein [1921.

Two-dimensional gel electrophoresis I

Electrophoresis 1983, 4, 97-1 16

Artefactual charge and molecular weight changes can also occur, but provided that care is taken these can usually be avoided. Deamidation of asparagine and glutamine residues can readily occur as can oxidation of cysteine residues. These effects are particularly problematical with treatments such as lyophilization [I]. Also some procedures used for radiolabelling can result in protein modification. For example, methylation procedures are carried out at alkaline pH in the presence of urea and in the absence of ampholytes, conditions which are optimal for carbamylation. It may be preferable to label the proteins in the absence of urea, i. e. in their native state or in the presence of SDS. Protein iodination procedures may well result in charge modification. The use of protease inhibitors can also result in protein modification [ 1141. Other artefacts could occur due to the presence of proteins in multiple states of aggregation or to binding of ampholytes such as occurs with cytochrome P450. It is not clear how many protein carrier ampholyte interactions persist in the presence of 8 M urea. It has been reported that interactions with sulphated dyes were not completely disrupted [69, 771, whereas protein interactions were disrupted under these same conditions (771. We have noticed when running 2-D gels of human skin fibroblast proteins that two major proteins focus at different positions in the gel according to the type of commercial carrier ampholyte used (M. J. Dunn and A. H. M. Burghes, unpublished observation). Also, significantly more of these two proteins enter the gel when SDS is included in the solubilization procedure. This change in position in the gel represents a very large shift in apparent p1. This effect is observed even after long focusing times (69 Vh/cm*) and does not appear to be a result of failure to attain equilibrium. We believe that the spots involved are the two tubulin polypeptides and published 2-D gels run using different carrier ampholytes appear to show the same effect [61, 1111. The reason for this phenomenon is unknown, but it is rather disturbing. The apparent p1 of tubulin has also been noted to shift on the addition of Triton X-100 [2031. Although the amino acid composition of the two tubulin polypeptides has been established 12031, the asparagine and glutamine content appear not to have been determined so that it is difficult to determine their true p1 values.

11 Conclusion The advances which have been made in 2-D PAGE methodology, described in this review, make this technique a unique research tool with the resolution capacity necessary for analysis of the many thousand gene products of a typical cell. In the second part of this review we will discuss how such complex protein mixtures can be analysed by 2-D PAGE and how the technique can be applied to a variety of biological problems.

We would like to thank Mrs. C. Trand for patiently and skillfully typing the manuscript, Mrs. L . J. White for her photographic expertise and Ms. A . Mapplebeck of the library staff for the literature search. We are grateful to Ms. S . E . CouIsonfor many helpful discussions and to colleagues who have allowed us to use reprints of their work. We acknowledge support by the Muscular Dystrophy Group of Great Britain (MJD) and the Medical Research Council (A HMB).

113

12 References O’Farrell, P. H., J. Biol. Chem. 1975, 250, 4007-4021. Smithies, O., Biochem. J. 1955, 61, 629-641. Smithies, 0.and Poulik, M. D., Nature(London)1956,177, 1033. Raymond, S. and Weintraub, L., Science 1959, 130, 711. Ornstein, L., Ann. N . Y. Acad. Sci. 1964, 121, 321-249. Davis, B. J., Ann. N . Y. Acad. Sci. 1964, 121, 404-436. Raymond, S. and Nakarnichi, M., Anal. Biochem. 1964, 7, 225-232. Slater, G. G., Fed. Proc. 1965, 24, 225. Margolis, J. and Kenrick, K. G., Nature (London) 1967 214, 1334-1336. Margolis, J. and Kenrick, K. G., Nature (London) 1969, 221, 1056-1057. Kaltschmidt, E. and Wittrnann, H. G.. Anal. Biochem. 1970,36, 401-4 12. Kaltschmidt, E. and Wittmann, H. G., Proc. Natl. Acad. Sci USA 1970,67, 1276-1282. Howard,G.A. andTraut,R. R.,FEBSLetters 1973,29, 177-180. Hoffrnann, N. L. and Dowben, R. M., Anal. Biochem. 1978,89, 540-549. Maizel, J. V., Science 1966, 151, 988-990. Shapiro, A. L., Scharff, M. O., Maizel, J. V. and Uhr, J. W., Proc. Natl. Acad. Sci. USA 1966,56,216-221. Shapiro, A. L. Vinuela, E. and Maizel, J. V., Biochem. Biophys. Res. Commun. 1967,28,815-820. Weber, K. andOsborn, M.,J. Biol. Chem. 1969,244,4406-4412. Laernrnli, U. K., Nature (London) 1970,227, 680-685. Martini, 0. H. W. and Gould, H. J., J. Molec. Biol. 1971, 62, 403-405. Mets, L. J. and Bogorad, L.,AnaZ. Biochem. 1974,57,200-210. Svensson, H., Acta Chem. Scand. 1961, IS, 325-341. Svensson, H., Acta Chem. Scand. 1962,16,456-466. Vesterberg, O., Acta Chem. Scand. 1969, 23, 2653-2666. Macko, V. and Stegemann, H., Hoppe-Seyler’s Z . Physiol. Chem. 1969,350,917-919. Dale, G. and Latner, A. L., Clin. Chim. Acta 1969, 24, 61-68. Emes, A. V., Latner, A. L. and Martin, J. A., Clin. Chim. Acta 1975,64, 69-78. Latner, A. L. and Emes, A. V., in: Righetti, P. G. (Ed.), Progress in Isoelectric Focusing and Isotachophoresis, North-Holland, Amsterdam 1975, pp. 223-233. Stegernann, H., Angew. Chem. 1970,82, 640. Barret, T. and Gould, H. J., Biochim. Biophys. Acta 1973, 294, 165- 170. Suria, D. and Liew, C. C., Can.J.Biochem. 1974,52,1143-1153. MacGillivray, A. J. and Rickwood, D., Eur. J. Biochem. 1974,41, 181-190. Bhakdi, S., Knufermann, M. and Wallach, D. F. H., Biochim. Biophys. Acta 1974,345,448-457. Bhakdi, S., Kniifermann, H. and Wallach, D. F. H., in: Righetti, P. G. (Ed.), Progress in Isoelectric Focusing and Is0tachophoresis, North-Holland, Amsterdam 1975,28 1-291. Bhakdi, S., Knufermann, H. and Wallach, D. F. H., Biochim. Biophys. Acta 1975,394,550-557. Chignell, D. A. and Wingfield, P. T., Fed. Proc. 1974, 33, 1283. Klose, J., Humangenetik 1975, 26, 231-243. Klose, J., in: Neubert, D. and Merkes, H. J. (Eds.), NeM Approaches to the Evaluation of Abnormal Embryonic Development, G. Thieme Verlag, Stuttgart 1975, pp. 375-387. Scheele, G. A., J. Biol. Chem. 1975,250, 5375-5385. Iborra, G. and Buhler, J. M., Anal. Biochem. 1976,74,503-5 11. Bray, D., Nature (London) 1977, 267, 481-482. Shackelford, D. A., Mann, D. L., Van Rood, J. J., Ferrara, G. B. and Strominger, J. L., Proc. Natl. Acad. Sci. USA 1981, 78, 4566-4570. Tuszyncki, G. P., Buck, C. A. and Warren, L., Anal. Biochem. 1979,85,224-229. Shackelford, D. A. and Strorninger, J. L.,J. Exp. Med. 1980,151, 144- 165.

114

M.J. Dunn and A. H. M. Burghes

[451 Siemankowski, R. F., Giambalvo, A. and Dreizen, P., Physiol. Chem. Phys. 1978,10,415-434. [461 Singer, B. S., Morrissett, H. and Gold, L., Anal. Biochem. 1978,85, 224-229. [471 Poehling, H. M. andNeuhoff, V., Electrophoresis 1980,1,90-101. [481 Jonsson, M., Electrophoresis 1980, 1, 141-149. [491 Agarose IEF - a supporting matrix f o r isoelectric focusing Pharmacia Fine Chemicals, Uppsala 1980. [501 Tanford, C., Adv. Prot. Chem. 1968,23, 121-282. [5 11 Ames, G. F. L. andNikaido, K., Biochemistry 1970,15,616-623. 1521 Willard, K. E., Giometti, C. S., Anderson, N. L., O’Connor, T. E. and Anderson, N. G., Anal. Biochem. 1979,100, 289-298. [531 Griffith, I. P., Biochem. J. 1972,126, 553-560. [541 Steck, T. L. and Fox, C. F., in Fox, C. F. and Keith, A. B. (Eds.), Membrane Molecular Biology, Sinauer Associates, Stamford 1972, pp. 27-75. [551 Hjelmeland, L. M., Nebert, D. W. and Chrambach, A., in: Catsimpoolas, N. (Ed.), Electrophoresis ’78, Elsevier, Amsterdam 1978, pp. 29-56. [561 Wilson, D., Hall, M. E., Stone, G. C. and Rubin, R. W., Anal. Biochem. 1977,83, 33-44. 1571 Giometti, C. S., Anderson, N. G. and Anderson, N. L., Clin. Chem. 1979,25, 1877-1884. [581 Novak-Hofer, I. and Siegenthaler, P. A., Biochim. Biophys. Acta 1977,468,461-471. [591 Minssen, M. and Munkries, K. D., Biochim. Biophys. Acta 1973, 291,398-410. [601 Booz, M. L. andTravis, R. L., Plant Physiol. 1980,66,1037-1043. I611 Garrels, J. I., Dev. Biol. 1979, 73, 134-152. [621 Rosenblum, B. B., Hanash, S. M., Yew, N. and Ned, J. V., Clin. Chem. 1982,28,925-931. [631 Rubin, R. W. and Milikowski, C., Biochim. Biophys. Acta 1978, 509, 100-110. [641 Harell, D. and Morrison, M., Arch. Biochem. Biophys. 1979,193, 158-168. 1651 Copeland, B. R., Todd, S. A. and Furlong, C. E., Am. J. Hum. Genet. 1982,34, 15-31. [661 Marchesi, V. T., Semin. Hematol. 1979,16, 3-20. 1671 Lux, S . E., Semin. Hematol. 1979, 16, 21151. (681 Burghes, A. H. M., Dunn, M. J., Statham, H. E. and Dubowitz, V., Electrophoresis 1982,3, 185-196. [691 Righetti, P. G., Gianazza, E., Brenna, 0. and Galante, E., J. Chromatogr. 1977,137, 171-181. [701 Klose, J. and Feller, M., Electrophoresis 1981, 2, 12-24. [711 Edwards, J. J., Tollaksen, S. L. and Anderson, N. G., Clin. Chem. 1981,27, 1335-1340. [721 Horst, M. N., Mahaboob, S.,Basha, M., Baumbach, G. A., Mansfield, E. H. and Roberts, R. M., Anal. Biochem. 1980, 102, 399-408. 1731 Basha, S . M. M., Plant Physiol. 1979,63, 301-306. [741 Hari, V., Anal. Biochem. 1981,113,332-335. [751 Kaderbhai, M. A. and Freedman, R. B., Biochim. Biophys. Acta 1980,601, 11-21. [761 Vlasuk, G. P. and Wok, F. G., Anal. Biochem. 1980, 105, 112- 120. [771 Gianazza, E. and Righetti, P. G., in: Radola, B. J. (Ed.), Electrophoresis ’79, de Gruyter, Berlin 1980, pp. 129-139. [781 Gianazza, E., Astorri, C. and Righetti, P. G.,J. Chromatogr. 1979, 171, 161-169. 1791 Gonenne, A. and Ernst, R., Anal. Biochem. 1978,87, 28-38. [Sol Hjelmeland, L. M., Nebert, D. W. and Chrambach, A., Anal. Biochem. 1979,95,201-208. [811 Booz, M. L. and Travis, R. L., Phytochemistry 1981, 20, 1773-1779. [821 Baron, C. and Thompson, T. E., Biochim. Biophys. Acta 1975, 382,276-285. [831 Wyman, J. Jr., J. Amer. Chem. Soc. 1933,55,4116. 1841 Gordon, J. A. and Jencks, W. P., Biochemistry 1963,2, 47-57. 1851 Booth, A. G., Biochem. J. 1977,163, 165-168. [861 Steinfeld, R. C. and Vidaver, G. A., Biophys. J. 1981, 33, 185. 1871 Delincee, H. andRadola,B. J.,Anal.Biochem. 1978,90,603-623.

Electrophoresis 1983, 4, 97-1 16

[881 Finlayson, G. R. and Chrambach, A., Anal. Biochem. 1971, 40, 292-3 11. [891 Anderson, N. L. and Anderson, N. G., Anal. Biochem. 1978,85, 331-340. [901 Voris, B. P. and Young, D. A., Anal. Biochem. 1980, 104, 478-484. [911 O’Farrell, P., Goodman, M. M. and O’Farrell, P. H., Cell 1977,12, 1133-1 142. 1921 Burghes, A. H. M., Dunn, M. J. and Dubowitz, V., Electrophoresis 1982,3, 354-363. 1931 Zechel, K., Ana6. Biochem. 1977,83,240-251. [941 Cuono, C. B. and Chapo, G. A., Electrophoresis 1982,2,65-75. [951 Creighton, I. E., J. Molec. Biol. 1979,129, 235-264. [961 Gelfi, C. and Righetti, P. G., Electrophoresis 1981, 2, 220-228. [971 Bode, H. J., in: Radola, B. J. (Ed.), Electrophoresis ’79, de Gruyter, Berlin 1980, pp. 39-52. [981 Cabral, F. and Schatz, G., Meths. Enzymol. 1979, 56, 602-613. [991 Goldsmith, M. R., Rattner, E. C., Macy, M., Koehler, D., Balikov, S. R. and Bock, S . C., Anal. Biochem. 1979,99,33-40. [I001 Olsson, I. and LABS, T., J. Chromatogr. 1981,215, 273-378. [ l o l l Righetti, P. G. and Drysdale, J. W., Isoelectric Focusing, NorthHolland, Amsterdam 1976. [lo21 Bianchi Bosisio, A., Loeherlein, C., Snyder, R. S. and Righetti, P. G., J. Chromatogr. 1980, 189, 317-330. 11031 Radola, B. J., Electrophoresis 1980, 1, 43-56. [I041 Rodbard, D., Levitov, C. and Chrambach, A., Separat. Sci. 1972, 7, 705-723. [I051 Gelfi, C. and Righetti, P. G., Electrophoresis 1981,2, 213-219. [lo61 O’Connell, P. B. H. and Brady, C. J., Anal. Biochem. 1976, 76, 63-73. [ 1071 Righetti, P. G. and Macelloni, G. J., J. Biochem. Biophys. Methods 1982, 5, 1-15. [lo81 Burghes, A. H. M., Dunn, M. J., Statham, H. E. and Dubowitz, V., Electrophoresis 1982,3, 177-185. [lo91 Saravis, C. A. and Zamcheck, N.,J. Immunol. Methods 1979,29, 91-96. [ l l O l Rosin, A., Ek, K. and Aman, P., J. Immunol. Methods 1979,28, 1-11. [I111 Hirabayashi, T., Anal. Biochem. 1981,117,443-451. 11121 Thompson, B. J., Dunn, M. J., Burghes, A. H. M. and Dubowitz, V., Electrophoresis 1982,3, 307-314. [ 1131 Serwer, P. and Hayes, S . J., Electrophoresis 1982, 3, 80-85. [1141 O’Farrell, P. H. and O’Farrell, P. Z., Methods Cell Biol. 1977,16, 407-420. [I151 Piperno, G., Huang, B. and Luck, D. J. L., Proc. Natl. Acad. Sci. USA 1977, 74, 1600-1604. [1161 Breithaupt, T. B., Nystrom, I. E., Hodges, D. H. and Babitch, J., Anal. Biochem. 1978,84, 579-582. [ I 171 Tracy, R. P., Currie, R. M., Kyle, R. A. and Young, D. S., Clin. Chem. 1982,28, 900-907. [1181 Burghes, A. H. M., Dunn, M. J., Statham, H. E. and Dubowitz, V., in: Allen, R. C. and Amaud, P. (Eds.), Electrophoresis ’81, de Gruyter, Berlin 1981, 295-308. [I 191 Sanders, M. M., Groppi, V. E. and Browning, E. T., Anal. Biochem. 1980,103, 157-165. [ 1201 Tullis, R. H. and Rubin, H., Anal. Biochem. 1980,107,260-264. 11211 Kuhn, 0. and Wilt, F. H., Anal. Biochem. 1980,105, 274-280. [122] Righetti, P. G., J. Chromatogr. 1979, 173, 1-5. [1231 Ferreira, A. and Eichinger, D., J. Zmmunol. Methods 1981, 43, 29 1-299. 11241 Hunter, L., Anal. Biochem. 1978,89, 279-283. [1251 Rangel-Aldao, R., Kupiec, J. W. and Rosen, 0. M., J . Biol. Chem. 1979,254,2499-2508. [1261 Gorg, A., Postel, W., Westermeier, R., Gianazza, E. and Righetti, P. G., J. Biochem. Biophys. Methods 1980,3,273-284. [I271 Valkonen, K., Gianazza, E. and Righetti, P. G., Clin.Chim. Acta 1980,107,223-229. [1281 Burghes, A. H. M., Dunn, M. J., Witkowski, J. A. and Dubowitz, V., in: Stathakos, D. (Ed.), Electrophoresis ’82, de Gruyter, Berlin 1983, (in press). [1291 Ui, N., Ann. N. Y. Acad. Sci. 1973,209, 198-209.

Electrophoresis 1983, 4, 97-1 16 [I301 Gelsema, W. J., de Ligny, C. L. and van der Veen, N. G., J. Chromatogr. 1978,151, 161-174. [ 13 I] Gelsema, W. J., de Ligny, C. L. and van der Veen, N. G., J. Chromatogr. 1979, 171, 171-181. [I321 Bravo, R. and Celis, J. E., J . Cell Biol. 1980, 84, 795-802. [I331 Bravo, R., Bellatin, J. and Celis, J., Cell Biol. Zntl. Rep. 1981, 5, 93-96. [ 1341 Steinberg, R. A., O'Farrell, P. H., Friedrich, U. and Coffino, P., Cell 1977, 10, 381-391. [I351 Anderson, N. L. and Hickman, B. J., Anal. Biochem. 1979, 93, 3 12-320. [ 1361 Hickman, B. J., Anderson, N. L., Willard, K . E. and Anderson, N. G., in: Radola, B. J. (Ed.), Electrophoresis '79, de Gruyter, Berlin 1980, pp. 341-360. [137] Frearson, N., Taylor, R. D. and Perry, S . V., Clin. Sci. 1981,61, 141-149. 11381 Anderson, N. L. and Anderson, N. G., Biochem. Biophys. Res. Commun. 1979,88,258-265. [ 1391 Tollaksen, S. L., Edwards, J. J. and Anderson, N. G., Electrophoresis 1981,2, 155-160. [1401 Edwards, J. J. and Anderson, N. G., Electrophoresis 1981, 2, 161-167. [141] Rilbe, H., Ann. N. Y. Acad. Sci. 1973, 209, 80-93. [ 1421 Righetti, P. G., Pagani, M. and Gianazza, E.,J. Chromatogr. 1975, 109,341-356. [I431 Allen, R. C., Christopher, J., Lorincz, L., Allen, R. C. Jr. and Liu, P. in: Radola, B. J. (Ed.), Electrophorese Forum '80, Technical University, Munich 1980, 117-125. [144] Taylor, J., Anderson, N. L. and Anderson, N. G., in: Allen, R. C. and Arnaud, P. (Eds.), Electrophoresis '81, de Cruyter, Berlin 1981, pp. 383-400. [I451 Cantrell, S. J., Babitch, J. A. and Torres, S., Anal. Biochem. 1981, 116, 168-173. [I461 Righetti, P. G., Tudor, G. and Gianazza, E., J. Biochem. Biophys. Methods 1982, 6, 219-227. 11471 Charlionet, R., Martin, J. P., Sesboiie, R., Madec, P. J. and Lefebvre, F., J. Chromatogr. 1979, 176, 89-101. [I481 Svensson, H., Prot. Biol. Fluids 1967, 15, 515-522. [1491 Patent USA, No. 3, 485, 736 (1969). [ 1501 Vinogradov, S. N., Lowenkrau, S., Andonian, M. R., Bagshaw, J., Felgenhauer, K . and Pak, S. J., Biochem. Biophys. Res. Comrnun. 1973,54,501-506. [I511 Lundblad, G., Vesterberg, O., Zimmermann, R. and Ling, J., Acta Chem. Scand. 1972,26, 1711-1713. [1521 Vesterberg, O., Acta Chem. Scand. 1973, 27, 2415-2420. [I531 Righetti, P. G. in: Righetti, P. G., van Oss, C. J. and Vanderhoff, J. W. (Eds.), Electrokinetic Separation Methods, North-Holland Press, Amsterdam 1979, pp. 389-441. [154] Gianazza, E., Pogani, M., Luzzana, M., and Righetti, P. G., J. Chromatogr. 1975,109, 357-364. [155] Grubhofer, N. and Borja, C., in: Radola, B. J. and Graesslin, D. (Eds.), Electrofocusing and lsotachophoresis, de Gruyter, Berlin 1977, pp. 1 1 1-120. I1561 Pogacar, P. and Jarecki, R., in: Allen, R. C. and Maurer, H. R. (Eds.), Electrophoresis and Isoelectric in Polyacrylamide Gels, de Gruyter, Berlin 1974, pp. 153-158. [1571 Righetti, P. G., Balzarini, L. and Gianazza, E., J. Chromatogr. 1977,134,279-284. [I581 Williams, K . W. and Soderberg, L., Zntl. Lab. Jan/Feb 1979, pp. 45-53. [I591 Righetti, P. G. and Gianazza, E., J . Chromatogr. 1980, 184, 4 15-456. [I601 Binion, S. B., Rodkey, L. S., Egen, N. B. and Bier, M., Electrophoresis 1982, 3, 284-288. [I611 Just, W. W., Anal. Biochem. 1980,102, 134-144. [1621 Gelsema, W. J., de Ligny, C. L. and yan der Veen, N. G., J. Chromotogr. 1979, 173, 33-41. I1631 Nguyen, N. Y. and Chrambach, A., Anal. Biochem. 1976, 74, 145-153, [1641 Pekkula-Flagan, A. and Comings, D. E., Anal. Biochem. 1982, 122, 295-297.

Two-dimensional gel electrophoresis I

115

[I651 Caspers, M. L., Posey,Y., andBrown, R. K.,Anal.Biochem. 1977, 79, 166-180. [1661 Babitch, J. A. and Benavides, L. A,, Neuroscience 1979, 4, 603-6 13. 11671 Pena, S. D. J. and Hughes, R. C., Biochem. Biophys. Acta 1979, 550, 100-109. [1681 Patent USA, No. 4, 130,470 (1978). [I691 Bjellqvist, B., Ek, K., Righetti, P. G., Gianazza, E., Gorg, A., Westermeier, R. J. and Postel, W., J . Biochem. Biophys. Methods 1982,6,317-339. [1701 Bjellqvist, B. and Ek, K., L K B Application Note 321, 1982. [I711 Tracy, R. P., Currie, R. M. and Young, D. S., Clin. Chem. 1982, 28, 908-914. [1721 Peterson, J. L. and McConkey, E. M., J. Biol. Chem. 1976, 251, 548-554. [I731 Rosenmann, E., Kreis, C., Thompson, R. G., Dobbs, M., Hamerton, J. L. and Wrogemann, K., Nature (London) 1982, 298, 563-565. 11741 Jackle, H., Anal. Biochem. 1979,98, 81-84. I1751 Gorg, A., Postel, W. and Westermeier, R., in: Allen, R. C. and Arnaud, P. (Eds.), Electrophoresis '81, de Gruyter, Berlin 1981, pp. 259-270. [1761 Giometti, C. S., Willard, K. E. and Anderson, N. L., Clin. Chem. 1982,28, 955-961. [1771 Wyckoff, M., Rodbard, D. and Chrambach, A., Anal. Biochem. 1977, 78,459-482. [1781 Allen, R. C. in: Allen, R. C. and Maurer, H. R. (Eds.), Electrophoresis and Isoelectric Focusing in Polyacrylamide Gels, de Gruyter, Berlin 1974, pp. 105-113. [1791 Allen, R. C., Electrophoresis 1980, 1, 32-37. [1801 L i b , T. and Olsson, Z., in: Allen, R. C. and Arnaud, P. (Eds.), Electrophoresis '81, de Gruyter, Berlin 1981, pp. 191-203. [ 1811 Ansorge, W. and Garoff, H., in: Allen, R. S.and Arnaud, P. (Eds.), Electrophoresis '81, de Gruyter, Berlin 1981, pp. 635-646. [ 1821 Nickodem, V. M., Trus, B. L. and Rall, J. E., Proc. Natf.Acad. Sci. USA 1981, 78,4411-4415. I1831 Anderson, D. W. and Peterson, C., in: Stathakos, D. (Ed.), Electrophoresis '82, de Gruyter, Berlin 1983, (in press). 11841 Riichel, R., J. Chromatogr. 1977,132,451-468. [I851 Anderson, N. L. and Anderson, N. G., Proc. Natl. Acad. Sci. USA 1977, 74, 5421-5425. [1861 Gianazza, E. and Righetti, P. G.,J. Chromatogr. 1980,293,l-8. [ 1871 Anderson, N. G. and Anderson, N. L., Behring Inst. Mittl. 1979, 63, 169-210. [I881 Anderson, N. L. and Anderson, N. G., Anal. Biochem. 1978, 85, 341-354. [1891 Gianazza, E. and Righetti, P. G., in: Righetti, P. G., Van Oss, C. J. and Vanderhoff, J. W. (Eds.), Electrokinelic Separation Methods, Elsevier/North-Holland, Amsterdam 1979, pp. 293-3 11. [1901 Altland, K. and Hackler, R., Electrophoresis 1981,2, 49-54. [I911 Groppi, V. E. and Browning, E. T., Molec. Pharmacol. 1980,18, 427-437. [1921 Poduslo, J. F., Anal. Biochem. 1981, 114, 131-139. 11931 Giometti, C. S., Anderson, N. G., Tollaksen, S.L., Edwards, J. J. and Anderson, N. L., Anal. Biochem. 1980,102,47-58. 11941 Ogita,Z.I.andMarkert,C.L.,Anal. Biochem. 1979,99,233-241. 11951 Jones, M. I., Massingham, W. E. and Spragg, S . P., Anal. Biochem. 1980,106,446-449. [1961 Nochumson, S. and Gibson, S.G., in: Stathakos, D. (Ed.), Electrophoresis '82, de Gruyter, Berlin 1983, (in press). [I971 Anderson, N. L., presented at: Technical Advances of Two-Dimensional Electrophoresis and Clinical Applications of the Technique, Argonne National Laboratory, August 1982. 11981 Duncan, R. and McConkey, E. H., Clin. Chem. 1982, 28, 749-755. [I991 McConkey, E. M., Anal. Biochem. 1979,96, 39-44. [200] Anderson, N. L. Taylor, J., Scandora, A. E., Coulter, B. P. and Anderson, N. G., Clin. Chem. 1981,27, 1807-1820. [201] Anderson, N. G., Nature (London) 1979,278, 122-123. [202] Taylor, J., presented at: Technical Advances of Two-Dimensional Electrophoresis and Clinical Applications of the Technique,

1 16

M.J. Dunn and A. H. M. Burghes

Argonne National Laboratory, August 1982. I2031 Nelles, L. P. and Bamburg, J. R., J. Neurochem. 1979, 32, 477-489. [2041 Wu, B. C., Spohn, W. H. and Busch, H., Cancer Res. 1979, 39, 116-122. [2051 Clemetson, K. J., Capitanio, A. and Luscher, E. F., Biochim. Biophys. Acta 1979, 553, 11-24. [2061 Pearson, T. W., Kar, S . K., McGuire, T. C. and Lundin, L. B., J . Immunol. 1981,126,823-828. [2071 Comings, D. E., Nature (London) 1979, 277, 28-32. [2081 Jackson, P. and Thompson, J. J., J. Neurol. Sci. 1981, 49, 429-438.

Electrophoresis 1983, 4, 97-1 16

[2091 Atkinson, B. G. and Atkinson, K. H., Exptl. Parasitol. 1982, 53, 26-38. [2101 Tyrell, D., Isachson, P. J. and Reeck, G. R., Anal. Biochem. 1982, 119,433-439. (2111 Cole, W. G. and Chan, D., Biochem. J. 1981, 197, 377-383. [2121 Ivalde, R. O., Baxter, J. 0. and Morris, J. A., J . Biol. Chem. 1981, 256,4520-4528. [2131 Hansen, E. J., Wilson, R. M. and Baseman, J. B., Infect. Zmmun. 1979,24,468-475. [2141 Reilly, E. B., Auditore-Hargreaves, K., Hammerung, V. and Gottlieb, P. D., J . Immunol. 1980, 125, 2245-2251. [215] Bravo, R. and Celis, J. E., Exptl. Cell Res. 1980,127, 249-260.