ANALYTICAL
BIOCHEMISTRY
197,59-64
(1991)
Gel Electrophoresis in the Presence of Soluble, Aqueous Polymers: Horizontal Sodium Dodecyl Sulfate-Polyacrylamide Gels Douglas Department
Received
M. Gersten,’ of Pathology,
February
Hervey Georgetown
Kimball, University
and Karen Medical
Gel electrophoresis of proteins in slabs of polyacrylamide is one of the more important techniques of modern biochemistry. Of the many applications of polyacrylamide gel electrophoresis, the most frequently used is that of Laemmli (l), in which the molecular weight of proteins can be estimated in the presence of SDS2-containing buffers. Laemmli-style SDS-polyacrylamide gel electrophoresis is almost always performed in slab gels 1 To whom correspondence ’ Abbreviations used: bis, SDS, sodium dodecyl sulfate. Copyright All
rights
D.C. 20007
27,199l
Because little is known about the use of aqueous polymers in polyacrylamide gel electrophoresis, we undertook a feasibility study that enables the discontinuous Laemmli-formulated system of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis to be performed in a horizontal format by the addition of largesized aqueous polymers (i.e., dextrans and methylcelluloses). We studied four parameters: the cross-linking agent (bisacrylamide vs AcrylAide) and the polymer concentration, nature, and size. Three concentrations of each polymer were used. The best differentiation between the standard markers and the sharpest bands were obtained using concentrations of 2.5 and 0.06% (w/v) for Dextran T-500 and methylcellulose 4000, respectively. There was no predictable pattern to the variation in the plots of log M, vs R, caused by varying the concentration and length of the dextrans; however, the methylcellulose patterns suggest that gel viscosity is important. The results suggest that the combination of 0.06% methylcellulose 4000 polymers with bisacrylamide is a convenient and inexpensive means of performing flatbed Laemmli SDS-polyacrylamide gel electrophoresis. 0 1991 Academic Press, Inc.
0003-2697/91
E. Bijwaard
Center, 3900 Reservoir Road N. W., Washington,
should be addressed. bisacrylamide; MC,
$3.00 0 1991 by Academic Press. of reproduction in any form
methylcellulose;
in a vertical format because difficulties have been encountered in the horizontal format for a variety of reasons. In our hands, the patterns obtained in horizontal SDS gels have never been reproducible. The causes have varied from separation of the seam between the stacking and separating gels, which in turn causes current arcing, to electrolyte channeling, to conductivity discontinuities. In the best cases, the bands are slightly wavy. In the worst cases, the patterns are totally uninterpretable. Nonetheless, the use of horizontal SDS gels would be desirable for many applications. Two different approaches have been taken to adapt SDS gels for flatbed use. The first is to eliminate the stacking gel from the Laemmli formulation and use only a separating gel containing a single buffer at the start of the experiment, rather than two connected gels having different buffers (2,3). This has been used primarily for two-dimensional applications. The second approach uses the discontinuous buffer system of Laemmli, but replaces the polyacrylamide matrix with agarose (4). We now report a third alternative, in which high-molecular-weight, soluble, aqueous polymers are included during the casting of SDS-polyacrylamide gels. Due to their large size, the polymers remain in solution, trapped in the matrix of the gel during electrophoresis. This approach, then, allows the use of discontinuous buffer systems and we illustrate the effects of different polymers and polymer lengths. MATERIALS
AND
METHODS
Materials. Acrylamide and bisacrylamide were electrophoresis grade and were purchased from Accurate Chemical Co. AcrylAide (Batch 10960) was from FMC. Dextrans T-500, T-70, and T-40 (Batches 3936, PCO8625, and 1600, respectively) were obtained from Pharmacia Fine Chemicals, as were the low-molecularweight protein standards. Polyethylene glycol (Carbo59
Inc. reserved.
60
GERSTEN,
KIMBALL,
AND
BIJWAARD
T500
AA
BIS
0.2
0.4
0.6
0.8
1.0
02
OU
06
0.8
1.0
RF
FIG. 1. cross-linker.
Log,, (molecular weight Right: bisacrylamide
X 10d3) vs R, of SDS-polyacrylamide cross-linker.
wax) 6000 (Batch 32832-6) was from Union Carbide, Methylcellulose 4000 (Batch 705755) was from Fisher. Methylcelluloses 1500 and 400 (Batches 88F0051 and 79F0028) were from Sigma Chemical Co., as was dextran sulfate 500 (batch 17F0401). Gel casting. Separating gels were cast according to the formulation of Laemmli (l), as 10% T, 2.66% C with the modification that 2.66% A&Aide was substituted for bisacrylamide as the cross-linker in some experiments. Stock solutions of dextran, dextran sulfate, methylcellulose, or polyethylene glycol were added to the acrylamide monomers and buffers, before addition of the initiators, to give the appropriate final concentrations (final concentrations are given as w/v). Methylcellulose was heated to dryness before weighing and then dissolved following the suggestion of Albertsson (5). The gel-casting cassette consisted of a silanized glass plate (Polifix 100, Serva Fine Biochemicals), a plexiglass plate to which 10 X 1.5 X 0.75-mm silicone rubber well-forming blocks had been glued, and a 0.75 mm U-form spacer. The separating gel was poured to 1 cm of the well-forming blocks. A 2-cm stacking gel, without added polymers, of 3% T, 2.66% C, cross-linked with bisacrylamide was cast atop the separating gel. The addition of the polymer solutions to the separating gel solution rendered it sufficiently viscous to allow the stacking gel to be poured with a sharp interface before the separating gel polymerized. This obviated the need to polymerize the gel in two stages using a water layer and effectively produced a “seamless” slab.
Electrophoresis. Horizontal electrophoresis was performed at 100 V, constant voltage, using a Peltier cooling device (MRA Coldfocus) having an interelectrode distance of 7.5 cm. Samples (10 ~1 of Pharmacia lowmolecular-weight markers in Laemmli sample buffer) were loaded into the preformed wells, onto filter paper wicks, or surface-loaded directly onto the stacking gel.
gel electrophoresis
in the presence
of Dextran
T-500.
Left:
AcrylAide
The data given in Figs. l-7 are for well-loaded samples. Electrode wicks (10 thicknesses of Whatmann No. l-Catalog lOOl-917), having dimensions of 18 X 100 mm, were saturated with Laemmli running buffer. The electrode wicks were observed during the electrophoresis, and, if necessary, running buffer was added to the wicks. Following electrophoresis, gels were stained with Coomassie (Serva) blue R-250. Gels cast using AcrylAide cross-linker were air-dried or microwave-dried (6) and then photographed, while those using bisacrylamide were photographed wet. Conductivity and temperature measurements. Four 10% T, 2.66% C SDS separating gels containing the Laemmli buffers were cast to have identical cylindrical geometry (30 mm diameter, 30 mm length). One had no added polymers, and the others had either 4% polyethylene glycol6000, 2.5% Dextran T-500, or 0.06% methylcellulose 4000. Conductivity measurements were taken
- 94-67-43-3o-2o-
A
B
FIG. 2. (A) Photograph of gel containing no polymers with AcrylAide cross-linker. (B) Photograph of gel containing 2.5% (w/v) Dextran T-500 in AcrylAide.
GEL
ELECTROPHORESIS
WITH
SOLUBLE,
AQUEOUS
61
POLYMERS
4. 7:
AA
\ I cl2
0.4
016
de
I
, I.0
0.2
OA
0.6
0.8
1.0
RF
FIG. 3. cross-linker.
Log,, (molecular weight Right: bisacrylamide
X 10m3) vs I$ of SDS-polyacrylamide cross-linker.
gel electrophoresis
in the presence
of Dextran
T-70.
Left:
A&Aide
tran sulfate 500 sometimes prevents complete polymerization and sticking to the supports. Hence, neither Dextran sulfate 500 nor Carbowax 6000 is appropriate for use in flatbed applications. The addition of the various dextrans increased the running time, but not in a predictable way. Doubling the voltage decreased the running time of the gels by 75%. At the higher voltages we exceeded the cooling capacity of our system and were therefore limited to using 100 V, constant voltage. A well size of 10 X 1.5 mm in a 0.75-mm-thick gel gave consistently sharp bands. Refilling the wells with Laemmli sample buffer and rewetting the wicking with running buffer when the dye front cleared the wells also increased the band sharpness and the consistency of the patterns. We also observed that surface loading is not entirely satisfactory in this system. Nor is wick loading satisfactory for all applications. However, wick loading on bisa&amide gels containing methylcellulose 4000 was
with a Wheatstone bridge, using an interelectrode distance of 24 mm. For temperature measurement, two slabs of 10% T, 2.66% C containing either 5 or 1.25% T-500 were prepared with stacking gels as above. Electrophoresis was performed as usual (100 V, constant voltage with the cooling device on). A surface thermistor was pushed into the slab and held for temperature reading.
RESULTS
General observations on polymer-containing polyacrylamide gels. We have studied the effects of four different parameters, nature of the polymer, polymer length, polymer concentration, and cross-linker, in conjunction with flatbed SDS-gel electrophoresis. In general, SDS gels cast with polyethylene glycol 6000 are unsuitable for horizontal electrophoresis because of streaking, although vertical gels are satisfactory (not shown). Dex-
T40 AA m
94’
i
67.
t;3
43 .
$
30.
g 1
20.
5.0 . 25 .
BIS
14. I
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
RF
FIG. 4. cross-linker.
Log,, (molecular weight Right: bisacrylamide
X 10w3) vs R, of SDS-polyacrylamide cross-linker.
gel electrophoresis
in the presence
of Dextran
T-40.
Left:
AcrylAide
62
GERSTEN, TABLE
Effects
of Polymer
Addition
KIMBALL,
1
on Protein
Standard
Mobilities
Cross-linker Polymer Dextran
MC
AC&Aide T-500 T-70 T-40
4000
Mobility in 0.03% MC Polyethylene glycol6000 Dextran-SO,
(%)
1.25 > 2.5 > 5.0 1.0 > 2.0 > 4.0 2.5 > 1.25 > 5.0 0.015 > 0.03 > 0.06 + AcrylAide
Bisacrylamide 2.5 r 2.0 > 5.0 > 0.015
(%)
5.0 > 1.25 1.0 > 4.0 2.5 > 1.25 > 0.03 z 0.06
MC 400 > MC 1500 > MC 4000 Unusable in horizontal system Unusable in horizontal system
comparable to well loading and is further discussed below. The increased viscosity of the dextran gel solutions limited us to a 0.75mm thickness in the present studies. Difficulty was encountered in casting thinner dextran gels. DextFum. Flatbed SDS-polyacrylamide gel electrophoresis performed in the presence of Dextran T-500 is shown in Fig. 1. The effect of varying T-500 concentration in the separating gel is given for AcrylAide crosslinker on the left and for bisacrylamide on the right. As expected, the standard marker proteins all had faster mobility in the AcrylAide gels than in the bisacrylamide gels, due to the larger pore size for AcrylAide crosslinkers. The linearity of the log molecular weight vs Rf relationship is generally maintained, the migration of the larger proteins appears to be uniformly retarded relative to that of lactalbumin (14K). For the bisacrylamide gels, the concentration of T-500 over the range 1.25-5% had no appreciable effect on the mobilities of phosphorylase b (94K), serum albumin (67K), ovalbumin (43K), carbonic anhydrase (30K), or trypsin inhibitor (20K). In contrast, the three concentrations studied had a noticeable effect on the mobility of all markers in AcrylAide gels. The mobilities for all markers in the 5% gels were intermediate between those in the 1.25 and 2.5% gels. This was an unexpected but consistent finding. In general for the dextrans, bisacrylamide gels containing 2.5% T-500 gave the sharpest bands. These were comparable to those of 5% T-500 with AcrylAide (Fig. 2b). Figure 2 also compares the horizontal SDS electrophoresis in the absence of any added polymers with that for 2.5% T-500, illustrating the difficulties encountered without polymers. These experiments were repeated using varying concentrations of Dextran T-70 (Fig. 3) and T-40 (Fig. 4). The effects of polymer addition are summarized in Table 1. Mobility differences between gels cross-linked with bisacrylamide and AcrylAide are maintained, but
AND
BIJWAARD
the extent to which the polymers retard the mobility of the proteins does not follow any predicted pattern. In addition to shifting the mobilities of the marker proteins, different areas of the pattern can be sharpened by the addition of different polymers. Figure 5 illustrates this point by comparing T-70 and T-40 in AcrylAide. Methylcelluloses. The effects of including methylcellulose 4000 in separating gels cross-linked with AcrylAide and bisacrylamide are shown in Figs. 6a and 6b, respectively. These results are also summarized in Table 1. AcrylAide gels containing 0.06% MC 4000 consistently gave the sharpest bands of any of the conditions we tested, including dextrans. The effect of polymer size was studied, with a comparison of MC 4000, MC 1500 and MC 400, at a concentration of 0.03% in AcrylAide. This is shown in Fig. 7. The rank order of mobilities for all markers was MC 400 > MC 1500 > MC 4000. For MC 400 and MC 1500, lactalbumin ran slightly ahead of the bromphenol blue dye marker. Finally, Fig. 8 demonstrates that, when 0.06% MC 4000&s is used, wick loading is comparable to well loading. This was a consistent finding for the MC 40001 bis combination, but inconsistent for other combinations. Thus wick loading is not recommended for all combinations. Conductivity and temperature measurements. The observation that some markers ran faster than the dye front led us to compare the conductivities of the gels. While the polymers should be electrically neutral given their structures, “reagent-grade” polymers were not used and it was not known whether conductive contaminants were present. The resistance of AcrylAide gels,
FIG. 6. Band Right: photograph
DYE FRONT
sharpening. Left: photograph of gel containing T-40.
of gel containing
T-70.
GEL
ELECTROPHORESIS
WITH
SOLUBLE,
AQUEOUS
63
POLYMERS
MC4000 AA m b
g4
R
67
BIS
0.06 . 0.03 . 015
.
E “r
43
6 2
3o
8 _I 2o 14 0.2
0.4
0.6
0.8
1.0
0.2
0.4
06
0.8
1.0
RF
FIG. 6. AcrylAide
Log,, (molecular weight x 10e3) vs R, of SDS-polyacrylamide cross-linker. Right: bisacrylamide cross-linker.
cast without polymers, with 4% Carbowax, with 2.5% T-500, and with 0.06% methylcellulose was measured to be 0.305, 0.280, 0.312, and 0.315 Mohm, respectively. These differences do not appear to be substantial. It has been suggested that conductivity fluctuations could be due to temperature changes with viscosity. Accordingly, temperature measurements were made on 5% T-500/his (the most viscous condition studied) and 1.25% T-500&s (intermediate viscosity). Under the standard running conditions, the cooling system held the 5% T-500 gel to 21°C and the 1.25% T-500 gel to 20°C.
DISCUSSION
Gel electrophoresis in the horizontal format has been advocated by many workers (for example, (7-9)). SDSpolyacrylamide gel electrophoresis of proteins, using the discontinuous buffer system of Laemmli (l), can
gel electrophoresis
in the presence
of methylcellulose
4000.
Left:
now be readily accomplished with the addition of polymers to the gel, which remain in solution during the electrophoresis. The problems previously encountered with flatbed SDS gels in our laboratories, primarily associated with electroendosmosis, seam separation between the stacking and separating gels, and electrolyte channeling, can be avoided by using the polymers to create a seamlesstwo-stage slab. Because little is known about the approach of polymer addition and the ways in which the separation pattern can be affected, the current study was necessary. Seamless gels can also be cast by including 30% glycerol in the separating gel (10). Two previous observations prompted us to consider this approach. First, Alien et al. (l&12) demonstrated DNA electrophoresis in flatbed polyacrylamide gels that had been preimpregnated with glycerol. In those experiments, the effective sieving range of the gel could be varied by changing the glycerol concentration. Presumably, this effect was due to changes in viscosity. The second observation was that Stracke (13) incorporated
L 3 43 g
30.
c7 0 _1
20. 14.
0.03%
MC+ AA
I 0.2
0.4
0.6
0.6
1.0
WELL FIG. ‘7. Log,, (molecular weight x 10-a) vs R, of SDS-polyacrylamide gel electrophoresis in the presence of 0.03% methylcelluloses. MC 4000, circles; MC 1500, triangles; MC 400, squares.
FIG. 8. trophoresis, ing.
WICK
Effect of well vs wick loading. SDS-polyacrylamide in 0.06% MC 4000. Left: well loading. Right:
gel elecwick load-
64
GERSTEN,
KIMBALL,
10% polyethylene glycol into vertical polyacrylamide gels in order to maintain the solubility of the proteins of interest. In the horizontal SDS gels cast with dextrans, the mobility of proteins was always greater in AcrylAide gels than in bisacrylamide gels, as expected. We attribute this to the pore size differences. The addition of dextrans in increasing concentrations or of different sizes did not displace the curves in any predictable way in either the AcrylAide or the bisacrylamide gels. For instance, the order of migration for T-500/AcrylAide is 1.25 > 5 > 2.5%, while that for T-70lAcrylAide is 1 > 2 > 4% and that for T-40 is 2.5 > 1.25 > 5%. In the absence of any predictable patterns, it is best to summarize our findings about dextrans as follows. The best patterns were generally obtained using 2.5% Dextran T-500 and bisacrylamide or 5% T-500 and AcrylAide. The higher or lower molecular weight bands can be preferentially sharpened by using 4% T-7O/bis or 2.5% T-40/his, respectively. Surface or wick loading was not entirely satisfactory for all the different combinations, but was generally acceptable for dextran concentrations at or below 2.5%. For methylcellulose, the shifting of the curves was predictable. That is, the rank order of mobilities for equal masses of different-sized methylcelluloses is 400 > 1500 > 4000 (Fig. 7). The rank order of mobilities for different concentrations of MC 4000 is 0.015 > 0.03 > 0.06% in both bisacrylamide and AcrylAide gels (Fig. 6). Changes in the relative mobilities of the marker proteins, caused by the addition of polymers, could be the result of several factors, the most likely being alteration of the ionic strength by contaminants, interaction of the proteins and polymers, alteration of viscosity, and effects of volume exclusion. Ionic strength considerations can be ruled out by the resistance measurements, and, in the presence of SDS, any interaction between the proteins and the polymers is probably not an appreciable factor. For the dextrans, it is unlikely that variations in viscosity alone can account for the differences in the curves. Two observations discount viscosity as the sole factor. First, for dextran solutions of equal mass, T-500 is approximately twice as viscous as T-70 and thrice as viscous as T-40 (14) (compare Figs. 1,3, and 5). Second, the rank order of mobilities is not the same in comparisons of AcrylAide with bisacrylamide (compare Fig. la with lb, Fig. 3a with 3b, and Fig. 5a with 5b). Thus both viscosity and volume exclusion probably participate. For methylcellulose, one would have predicted that shifting of the curves was based on viscosity alone because the size designations for methylcellulose polymers are determined by viscosity measurements. However, under two conditions (Fig. 7), the smallest marker
AND
BIJWAARD
ran ahead of the dye front. This must reflect volume exclusion by the polymer. Our experiments demonstrate the feasibility of this approach and offer the use of soluble polymers as a means for performing flatbed SDS-polyacrylamide gel electrophoresis in two-stage gels containing discontinuous buffer systems. Of the many combinations we tried, those that clearly do not work are bisacrylamide without polymers, AcrylAide without polymers, Dextran sulfate 500 with either cross-linker, and polyethylene glycol with either cross-linker; 5% T-40/his gels had a tendency to disintegrate if they were left in the fixing or staining solution longer than 2 days. The combinations that routinely gave the best results were 0.06% MC 4000 with bisacrylamide or AcrylAide and 5% T-500 with AcrylAide. Given the cost differential between bisacrylamide and AcrylAide, and between Dextran T-500 and MC 4000, we advocate the MC 4000&s combination as a very inexpensive combination. Undoubtedly, the use of other soluble polymers, cross-linkers, thicknesses, cooling systems, and power considerations, in different combinations, will ultimately lead to faster and better means of performing horizontal SDS-gel electrophoresis. ACKNOWLEDGMENTS We are grateful to J. Preuss for ments and to R. Lohr for resistance
performing preliminary measurements.
experi-
REFERENCES 1. Laemmli,
U. K. (1970)
2. Schiwara, and Gorg,
H. W., Hebell, T., Kirchherr, H., Postel, A. (1986) Electrophoresis 7, 496-505.
Nature
227,
680-685. W., Weser,
J.,
3. Gorg, A., Postel, W., Westermeier, R., Gianazza, E., and Righetti, P. G. (1981) in Electrophoresis ‘81 (Allen, R. C., and Arnaud, P., Eds.), pp. 259-270, de Gryuter, Berlin. 4. Coulson, S. E., and Cook, R. B. (1983) in Electrophoresis ‘82 (Stathakos, D., Ed.), pp. 363-370, de Gryuter, Berlin. 5. Albertsson, P. A. (1960) p. 31, Wiley, New York. 6. Gersten, D. M., Zapolski, phoresk 6, 191-192.
Partition
of Cells
and Macromolecules,
E. J., and Ledley,
R. S. (1985)
7. Allen, R. C., Budowle, B., Saravis, C. A., and Lack, Acta Hktochem. Cytochem. 19,637-645. 8. Radola, B. J. (1980) Electrophoresis 9. Allen, R. C. (1980) Electrophoresis
Electro-
P. M. (1986)
1, 43-56. 1,32-37.
10. Bjellqvist, B., Larsen, K., and Ostergren, K. (1991) Proceedings, International Electrophoresis Society Meeting, Washington, DC, p. 26. 11. Allen, R. C., Graves, G. M., and Budowle, B. (1989) Proceedings, Annual Meeting, American Electrophoresis Society, Washington, DC, p. 91. 12. Allen, R. C., Graves, G. M., and Budowle, B. (1989) Biotechniques 7,736-744. 13. Stracke, M., in Schiffmann, Suppl. 1 8, 7-8. 14. Dextran fractions technical Chemicals.
E.
(1990)
manual
Clin. (1971)
Exp.
Metastasis
Pharmacia
Fine
BIOCHEMISTRY
197,59-64
(1991)
Gel Electrophoresis in the Presence of Soluble, Aqueous Polymers: Horizontal Sodium Dodecyl Sulfate-Polyacrylamide Gels Douglas Department
Received
M. Gersten,’ of Pathology,
February
Hervey Georgetown
Kimball, University
and Karen Medical
Gel electrophoresis of proteins in slabs of polyacrylamide is one of the more important techniques of modern biochemistry. Of the many applications of polyacrylamide gel electrophoresis, the most frequently used is that of Laemmli (l), in which the molecular weight of proteins can be estimated in the presence of SDS2-containing buffers. Laemmli-style SDS-polyacrylamide gel electrophoresis is almost always performed in slab gels 1 To whom correspondence ’ Abbreviations used: bis, SDS, sodium dodecyl sulfate. Copyright All
rights
D.C. 20007
27,199l
Because little is known about the use of aqueous polymers in polyacrylamide gel electrophoresis, we undertook a feasibility study that enables the discontinuous Laemmli-formulated system of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis to be performed in a horizontal format by the addition of largesized aqueous polymers (i.e., dextrans and methylcelluloses). We studied four parameters: the cross-linking agent (bisacrylamide vs AcrylAide) and the polymer concentration, nature, and size. Three concentrations of each polymer were used. The best differentiation between the standard markers and the sharpest bands were obtained using concentrations of 2.5 and 0.06% (w/v) for Dextran T-500 and methylcellulose 4000, respectively. There was no predictable pattern to the variation in the plots of log M, vs R, caused by varying the concentration and length of the dextrans; however, the methylcellulose patterns suggest that gel viscosity is important. The results suggest that the combination of 0.06% methylcellulose 4000 polymers with bisacrylamide is a convenient and inexpensive means of performing flatbed Laemmli SDS-polyacrylamide gel electrophoresis. 0 1991 Academic Press, Inc.
0003-2697/91
E. Bijwaard
Center, 3900 Reservoir Road N. W., Washington,
should be addressed. bisacrylamide; MC,
$3.00 0 1991 by Academic Press. of reproduction in any form
methylcellulose;
in a vertical format because difficulties have been encountered in the horizontal format for a variety of reasons. In our hands, the patterns obtained in horizontal SDS gels have never been reproducible. The causes have varied from separation of the seam between the stacking and separating gels, which in turn causes current arcing, to electrolyte channeling, to conductivity discontinuities. In the best cases, the bands are slightly wavy. In the worst cases, the patterns are totally uninterpretable. Nonetheless, the use of horizontal SDS gels would be desirable for many applications. Two different approaches have been taken to adapt SDS gels for flatbed use. The first is to eliminate the stacking gel from the Laemmli formulation and use only a separating gel containing a single buffer at the start of the experiment, rather than two connected gels having different buffers (2,3). This has been used primarily for two-dimensional applications. The second approach uses the discontinuous buffer system of Laemmli, but replaces the polyacrylamide matrix with agarose (4). We now report a third alternative, in which high-molecular-weight, soluble, aqueous polymers are included during the casting of SDS-polyacrylamide gels. Due to their large size, the polymers remain in solution, trapped in the matrix of the gel during electrophoresis. This approach, then, allows the use of discontinuous buffer systems and we illustrate the effects of different polymers and polymer lengths. MATERIALS
AND
METHODS
Materials. Acrylamide and bisacrylamide were electrophoresis grade and were purchased from Accurate Chemical Co. AcrylAide (Batch 10960) was from FMC. Dextrans T-500, T-70, and T-40 (Batches 3936, PCO8625, and 1600, respectively) were obtained from Pharmacia Fine Chemicals, as were the low-molecularweight protein standards. Polyethylene glycol (Carbo59
Inc. reserved.
60
GERSTEN,
KIMBALL,
AND
BIJWAARD
T500
AA
BIS
0.2
0.4
0.6
0.8
1.0
02
OU
06
0.8
1.0
RF
FIG. 1. cross-linker.
Log,, (molecular weight Right: bisacrylamide
X 10d3) vs R, of SDS-polyacrylamide cross-linker.
wax) 6000 (Batch 32832-6) was from Union Carbide, Methylcellulose 4000 (Batch 705755) was from Fisher. Methylcelluloses 1500 and 400 (Batches 88F0051 and 79F0028) were from Sigma Chemical Co., as was dextran sulfate 500 (batch 17F0401). Gel casting. Separating gels were cast according to the formulation of Laemmli (l), as 10% T, 2.66% C with the modification that 2.66% A&Aide was substituted for bisacrylamide as the cross-linker in some experiments. Stock solutions of dextran, dextran sulfate, methylcellulose, or polyethylene glycol were added to the acrylamide monomers and buffers, before addition of the initiators, to give the appropriate final concentrations (final concentrations are given as w/v). Methylcellulose was heated to dryness before weighing and then dissolved following the suggestion of Albertsson (5). The gel-casting cassette consisted of a silanized glass plate (Polifix 100, Serva Fine Biochemicals), a plexiglass plate to which 10 X 1.5 X 0.75-mm silicone rubber well-forming blocks had been glued, and a 0.75 mm U-form spacer. The separating gel was poured to 1 cm of the well-forming blocks. A 2-cm stacking gel, without added polymers, of 3% T, 2.66% C, cross-linked with bisacrylamide was cast atop the separating gel. The addition of the polymer solutions to the separating gel solution rendered it sufficiently viscous to allow the stacking gel to be poured with a sharp interface before the separating gel polymerized. This obviated the need to polymerize the gel in two stages using a water layer and effectively produced a “seamless” slab.
Electrophoresis. Horizontal electrophoresis was performed at 100 V, constant voltage, using a Peltier cooling device (MRA Coldfocus) having an interelectrode distance of 7.5 cm. Samples (10 ~1 of Pharmacia lowmolecular-weight markers in Laemmli sample buffer) were loaded into the preformed wells, onto filter paper wicks, or surface-loaded directly onto the stacking gel.
gel electrophoresis
in the presence
of Dextran
T-500.
Left:
AcrylAide
The data given in Figs. l-7 are for well-loaded samples. Electrode wicks (10 thicknesses of Whatmann No. l-Catalog lOOl-917), having dimensions of 18 X 100 mm, were saturated with Laemmli running buffer. The electrode wicks were observed during the electrophoresis, and, if necessary, running buffer was added to the wicks. Following electrophoresis, gels were stained with Coomassie (Serva) blue R-250. Gels cast using AcrylAide cross-linker were air-dried or microwave-dried (6) and then photographed, while those using bisacrylamide were photographed wet. Conductivity and temperature measurements. Four 10% T, 2.66% C SDS separating gels containing the Laemmli buffers were cast to have identical cylindrical geometry (30 mm diameter, 30 mm length). One had no added polymers, and the others had either 4% polyethylene glycol6000, 2.5% Dextran T-500, or 0.06% methylcellulose 4000. Conductivity measurements were taken
- 94-67-43-3o-2o-
A
B
FIG. 2. (A) Photograph of gel containing no polymers with AcrylAide cross-linker. (B) Photograph of gel containing 2.5% (w/v) Dextran T-500 in AcrylAide.
GEL
ELECTROPHORESIS
WITH
SOLUBLE,
AQUEOUS
61
POLYMERS
4. 7:
AA
\ I cl2
0.4
016
de
I
, I.0
0.2
OA
0.6
0.8
1.0
RF
FIG. 3. cross-linker.
Log,, (molecular weight Right: bisacrylamide
X 10m3) vs I$ of SDS-polyacrylamide cross-linker.
gel electrophoresis
in the presence
of Dextran
T-70.
Left:
A&Aide
tran sulfate 500 sometimes prevents complete polymerization and sticking to the supports. Hence, neither Dextran sulfate 500 nor Carbowax 6000 is appropriate for use in flatbed applications. The addition of the various dextrans increased the running time, but not in a predictable way. Doubling the voltage decreased the running time of the gels by 75%. At the higher voltages we exceeded the cooling capacity of our system and were therefore limited to using 100 V, constant voltage. A well size of 10 X 1.5 mm in a 0.75-mm-thick gel gave consistently sharp bands. Refilling the wells with Laemmli sample buffer and rewetting the wicking with running buffer when the dye front cleared the wells also increased the band sharpness and the consistency of the patterns. We also observed that surface loading is not entirely satisfactory in this system. Nor is wick loading satisfactory for all applications. However, wick loading on bisa&amide gels containing methylcellulose 4000 was
with a Wheatstone bridge, using an interelectrode distance of 24 mm. For temperature measurement, two slabs of 10% T, 2.66% C containing either 5 or 1.25% T-500 were prepared with stacking gels as above. Electrophoresis was performed as usual (100 V, constant voltage with the cooling device on). A surface thermistor was pushed into the slab and held for temperature reading.
RESULTS
General observations on polymer-containing polyacrylamide gels. We have studied the effects of four different parameters, nature of the polymer, polymer length, polymer concentration, and cross-linker, in conjunction with flatbed SDS-gel electrophoresis. In general, SDS gels cast with polyethylene glycol 6000 are unsuitable for horizontal electrophoresis because of streaking, although vertical gels are satisfactory (not shown). Dex-
T40 AA m
94’
i
67.
t;3
43 .
$
30.
g 1
20.
5.0 . 25 .
BIS
14. I
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
RF
FIG. 4. cross-linker.
Log,, (molecular weight Right: bisacrylamide
X 10w3) vs R, of SDS-polyacrylamide cross-linker.
gel electrophoresis
in the presence
of Dextran
T-40.
Left:
AcrylAide
62
GERSTEN, TABLE
Effects
of Polymer
Addition
KIMBALL,
1
on Protein
Standard
Mobilities
Cross-linker Polymer Dextran
MC
AC&Aide T-500 T-70 T-40
4000
Mobility in 0.03% MC Polyethylene glycol6000 Dextran-SO,
(%)
1.25 > 2.5 > 5.0 1.0 > 2.0 > 4.0 2.5 > 1.25 > 5.0 0.015 > 0.03 > 0.06 + AcrylAide
Bisacrylamide 2.5 r 2.0 > 5.0 > 0.015
(%)
5.0 > 1.25 1.0 > 4.0 2.5 > 1.25 > 0.03 z 0.06
MC 400 > MC 1500 > MC 4000 Unusable in horizontal system Unusable in horizontal system
comparable to well loading and is further discussed below. The increased viscosity of the dextran gel solutions limited us to a 0.75mm thickness in the present studies. Difficulty was encountered in casting thinner dextran gels. DextFum. Flatbed SDS-polyacrylamide gel electrophoresis performed in the presence of Dextran T-500 is shown in Fig. 1. The effect of varying T-500 concentration in the separating gel is given for AcrylAide crosslinker on the left and for bisacrylamide on the right. As expected, the standard marker proteins all had faster mobility in the AcrylAide gels than in the bisacrylamide gels, due to the larger pore size for AcrylAide crosslinkers. The linearity of the log molecular weight vs Rf relationship is generally maintained, the migration of the larger proteins appears to be uniformly retarded relative to that of lactalbumin (14K). For the bisacrylamide gels, the concentration of T-500 over the range 1.25-5% had no appreciable effect on the mobilities of phosphorylase b (94K), serum albumin (67K), ovalbumin (43K), carbonic anhydrase (30K), or trypsin inhibitor (20K). In contrast, the three concentrations studied had a noticeable effect on the mobility of all markers in AcrylAide gels. The mobilities for all markers in the 5% gels were intermediate between those in the 1.25 and 2.5% gels. This was an unexpected but consistent finding. In general for the dextrans, bisacrylamide gels containing 2.5% T-500 gave the sharpest bands. These were comparable to those of 5% T-500 with AcrylAide (Fig. 2b). Figure 2 also compares the horizontal SDS electrophoresis in the absence of any added polymers with that for 2.5% T-500, illustrating the difficulties encountered without polymers. These experiments were repeated using varying concentrations of Dextran T-70 (Fig. 3) and T-40 (Fig. 4). The effects of polymer addition are summarized in Table 1. Mobility differences between gels cross-linked with bisacrylamide and AcrylAide are maintained, but
AND
BIJWAARD
the extent to which the polymers retard the mobility of the proteins does not follow any predicted pattern. In addition to shifting the mobilities of the marker proteins, different areas of the pattern can be sharpened by the addition of different polymers. Figure 5 illustrates this point by comparing T-70 and T-40 in AcrylAide. Methylcelluloses. The effects of including methylcellulose 4000 in separating gels cross-linked with AcrylAide and bisacrylamide are shown in Figs. 6a and 6b, respectively. These results are also summarized in Table 1. AcrylAide gels containing 0.06% MC 4000 consistently gave the sharpest bands of any of the conditions we tested, including dextrans. The effect of polymer size was studied, with a comparison of MC 4000, MC 1500 and MC 400, at a concentration of 0.03% in AcrylAide. This is shown in Fig. 7. The rank order of mobilities for all markers was MC 400 > MC 1500 > MC 4000. For MC 400 and MC 1500, lactalbumin ran slightly ahead of the bromphenol blue dye marker. Finally, Fig. 8 demonstrates that, when 0.06% MC 4000&s is used, wick loading is comparable to well loading. This was a consistent finding for the MC 40001 bis combination, but inconsistent for other combinations. Thus wick loading is not recommended for all combinations. Conductivity and temperature measurements. The observation that some markers ran faster than the dye front led us to compare the conductivities of the gels. While the polymers should be electrically neutral given their structures, “reagent-grade” polymers were not used and it was not known whether conductive contaminants were present. The resistance of AcrylAide gels,
FIG. 6. Band Right: photograph
DYE FRONT
sharpening. Left: photograph of gel containing T-40.
of gel containing
T-70.
GEL
ELECTROPHORESIS
WITH
SOLUBLE,
AQUEOUS
63
POLYMERS
MC4000 AA m b
g4
R
67
BIS
0.06 . 0.03 . 015
.
E “r
43
6 2
3o
8 _I 2o 14 0.2
0.4
0.6
0.8
1.0
0.2
0.4
06
0.8
1.0
RF
FIG. 6. AcrylAide
Log,, (molecular weight x 10e3) vs R, of SDS-polyacrylamide cross-linker. Right: bisacrylamide cross-linker.
cast without polymers, with 4% Carbowax, with 2.5% T-500, and with 0.06% methylcellulose was measured to be 0.305, 0.280, 0.312, and 0.315 Mohm, respectively. These differences do not appear to be substantial. It has been suggested that conductivity fluctuations could be due to temperature changes with viscosity. Accordingly, temperature measurements were made on 5% T-500/his (the most viscous condition studied) and 1.25% T-500&s (intermediate viscosity). Under the standard running conditions, the cooling system held the 5% T-500 gel to 21°C and the 1.25% T-500 gel to 20°C.
DISCUSSION
Gel electrophoresis in the horizontal format has been advocated by many workers (for example, (7-9)). SDSpolyacrylamide gel electrophoresis of proteins, using the discontinuous buffer system of Laemmli (l), can
gel electrophoresis
in the presence
of methylcellulose
4000.
Left:
now be readily accomplished with the addition of polymers to the gel, which remain in solution during the electrophoresis. The problems previously encountered with flatbed SDS gels in our laboratories, primarily associated with electroendosmosis, seam separation between the stacking and separating gels, and electrolyte channeling, can be avoided by using the polymers to create a seamlesstwo-stage slab. Because little is known about the approach of polymer addition and the ways in which the separation pattern can be affected, the current study was necessary. Seamless gels can also be cast by including 30% glycerol in the separating gel (10). Two previous observations prompted us to consider this approach. First, Alien et al. (l&12) demonstrated DNA electrophoresis in flatbed polyacrylamide gels that had been preimpregnated with glycerol. In those experiments, the effective sieving range of the gel could be varied by changing the glycerol concentration. Presumably, this effect was due to changes in viscosity. The second observation was that Stracke (13) incorporated
L 3 43 g
30.
c7 0 _1
20. 14.
0.03%
MC+ AA
I 0.2
0.4
0.6
0.6
1.0
WELL FIG. ‘7. Log,, (molecular weight x 10-a) vs R, of SDS-polyacrylamide gel electrophoresis in the presence of 0.03% methylcelluloses. MC 4000, circles; MC 1500, triangles; MC 400, squares.
FIG. 8. trophoresis, ing.
WICK
Effect of well vs wick loading. SDS-polyacrylamide in 0.06% MC 4000. Left: well loading. Right:
gel elecwick load-
64
GERSTEN,
KIMBALL,
10% polyethylene glycol into vertical polyacrylamide gels in order to maintain the solubility of the proteins of interest. In the horizontal SDS gels cast with dextrans, the mobility of proteins was always greater in AcrylAide gels than in bisacrylamide gels, as expected. We attribute this to the pore size differences. The addition of dextrans in increasing concentrations or of different sizes did not displace the curves in any predictable way in either the AcrylAide or the bisacrylamide gels. For instance, the order of migration for T-500/AcrylAide is 1.25 > 5 > 2.5%, while that for T-70lAcrylAide is 1 > 2 > 4% and that for T-40 is 2.5 > 1.25 > 5%. In the absence of any predictable patterns, it is best to summarize our findings about dextrans as follows. The best patterns were generally obtained using 2.5% Dextran T-500 and bisacrylamide or 5% T-500 and AcrylAide. The higher or lower molecular weight bands can be preferentially sharpened by using 4% T-7O/bis or 2.5% T-40/his, respectively. Surface or wick loading was not entirely satisfactory for all the different combinations, but was generally acceptable for dextran concentrations at or below 2.5%. For methylcellulose, the shifting of the curves was predictable. That is, the rank order of mobilities for equal masses of different-sized methylcelluloses is 400 > 1500 > 4000 (Fig. 7). The rank order of mobilities for different concentrations of MC 4000 is 0.015 > 0.03 > 0.06% in both bisacrylamide and AcrylAide gels (Fig. 6). Changes in the relative mobilities of the marker proteins, caused by the addition of polymers, could be the result of several factors, the most likely being alteration of the ionic strength by contaminants, interaction of the proteins and polymers, alteration of viscosity, and effects of volume exclusion. Ionic strength considerations can be ruled out by the resistance measurements, and, in the presence of SDS, any interaction between the proteins and the polymers is probably not an appreciable factor. For the dextrans, it is unlikely that variations in viscosity alone can account for the differences in the curves. Two observations discount viscosity as the sole factor. First, for dextran solutions of equal mass, T-500 is approximately twice as viscous as T-70 and thrice as viscous as T-40 (14) (compare Figs. 1,3, and 5). Second, the rank order of mobilities is not the same in comparisons of AcrylAide with bisacrylamide (compare Fig. la with lb, Fig. 3a with 3b, and Fig. 5a with 5b). Thus both viscosity and volume exclusion probably participate. For methylcellulose, one would have predicted that shifting of the curves was based on viscosity alone because the size designations for methylcellulose polymers are determined by viscosity measurements. However, under two conditions (Fig. 7), the smallest marker
AND
BIJWAARD
ran ahead of the dye front. This must reflect volume exclusion by the polymer. Our experiments demonstrate the feasibility of this approach and offer the use of soluble polymers as a means for performing flatbed SDS-polyacrylamide gel electrophoresis in two-stage gels containing discontinuous buffer systems. Of the many combinations we tried, those that clearly do not work are bisacrylamide without polymers, AcrylAide without polymers, Dextran sulfate 500 with either cross-linker, and polyethylene glycol with either cross-linker; 5% T-40/his gels had a tendency to disintegrate if they were left in the fixing or staining solution longer than 2 days. The combinations that routinely gave the best results were 0.06% MC 4000 with bisacrylamide or AcrylAide and 5% T-500 with AcrylAide. Given the cost differential between bisacrylamide and AcrylAide, and between Dextran T-500 and MC 4000, we advocate the MC 4000&s combination as a very inexpensive combination. Undoubtedly, the use of other soluble polymers, cross-linkers, thicknesses, cooling systems, and power considerations, in different combinations, will ultimately lead to faster and better means of performing horizontal SDS-gel electrophoresis. ACKNOWLEDGMENTS We are grateful to J. Preuss for ments and to R. Lohr for resistance
performing preliminary measurements.
experi-
REFERENCES 1. Laemmli,
U. K. (1970)
2. Schiwara, and Gorg,
H. W., Hebell, T., Kirchherr, H., Postel, A. (1986) Electrophoresis 7, 496-505.
Nature
227,
680-685. W., Weser,
J.,
3. Gorg, A., Postel, W., Westermeier, R., Gianazza, E., and Righetti, P. G. (1981) in Electrophoresis ‘81 (Allen, R. C., and Arnaud, P., Eds.), pp. 259-270, de Gryuter, Berlin. 4. Coulson, S. E., and Cook, R. B. (1983) in Electrophoresis ‘82 (Stathakos, D., Ed.), pp. 363-370, de Gryuter, Berlin. 5. Albertsson, P. A. (1960) p. 31, Wiley, New York. 6. Gersten, D. M., Zapolski, phoresk 6, 191-192.
Partition
of Cells
and Macromolecules,
E. J., and Ledley,
R. S. (1985)
7. Allen, R. C., Budowle, B., Saravis, C. A., and Lack, Acta Hktochem. Cytochem. 19,637-645. 8. Radola, B. J. (1980) Electrophoresis 9. Allen, R. C. (1980) Electrophoresis
Electro-
P. M. (1986)
1, 43-56. 1,32-37.
10. Bjellqvist, B., Larsen, K., and Ostergren, K. (1991) Proceedings, International Electrophoresis Society Meeting, Washington, DC, p. 26. 11. Allen, R. C., Graves, G. M., and Budowle, B. (1989) Proceedings, Annual Meeting, American Electrophoresis Society, Washington, DC, p. 91. 12. Allen, R. C., Graves, G. M., and Budowle, B. (1989) Biotechniques 7,736-744. 13. Stracke, M., in Schiffmann, Suppl. 1 8, 7-8. 14. Dextran fractions technical Chemicals.
E.
(1990)
manual
Clin. (1971)
Exp.
Metastasis
Pharmacia
Fine
