Cell and Molecular Biology Section and 'Leukaemia Research Fund Centre, Institute of Cancer. Research, Chester Beatty Laboratories, 237 Fulham Rd, London ...
\.) 1993 Oxford University Press
Nucleic Acids Research, 1993, Vol. 21, No. 14 3335-3336
An EMSA-based method for determining the molecular weight of a protein - DNA complex Kim Orchard and Gillian E.Mayl * Cell and Molecular Biology Section and 'Leukaemia Research Fund Centre, Institute of Cancer Research, Chester Beatty Laboratories, 237 Fulham Rd, London SW3 6JB, UK Received March 15, 1993; Revised and Accepted May 28, 1993 We have developed an electrophoretic method to determine the size of a protein-DNA complex using the same binding and electrophoresis conditions as in an electrophoretic mobility shift assay (EMSA). The method is an adaptation of that of Ferguson (1), originally developed to determine the size of native proteins in non-denaturing gel systems. In SDS polyacrylamide gel electrophoresis, association of SDS gives all proteins equal charge per unit mass and also helps to unfold the protein. Thus the major determinant of protein mobility in an SDS gel is the length of the polypeptide chain. In contrast, in non-denaturing gels, mobility is influenced by molecular weight, conformation and charge. Ferguson analysis allows the molecular weight of a native protein to be determined indirectly by electrophoresing it, along with a number of standard proteins, on a series of non-denaturing tube gels of different polyacrylamide concentrations. Since the ionic conditions are identical in each experiment, the decrease in mobility observed as the polyacrylamide concentration is increased is due entirely to the sieving effects of the gel and hence is related to the species' size and shape but not its charge. The molecular weight of the unknown can be determined graphically from this change in mobility, as detailed below. We demonstrate here that the method can be successfully adapted to the analysis of protein-DNA complexes, using the same low ionic strength polyacrylamide slab gel that is used for the EMSA. The method offers a number of advantages over other techniques developed for this purpose: (i) there is no uncertainty as to whether the complex analysed is the same as that detected in the EMSA, because the binding and electrophoresis conditions are unchanged; (ii) uniike UV cross-linking methods, there is no denaturation step and so accessory proteins present in the complex which do not directly contact the DNA can also be detected (2); (iii) relatively little material is required since the position of the complex is still determined by autoradiography, and (iv) it can be performed on an unfractionated nuclear extract (3). In addition, we find that it yields reproducible results of high quality. The method is particularly useful for distinguishing complexes containing homodimeric and heterodimeric DNAbinding proteins from those containing a monomeric or mixture of monomeric proteins, respectively (3, 2). Like gel filtration and gradient centrifugation, Ferguson analysis measures molecular size rather than mass, and so assumes that the unknown species has a similar shape and partial specific volume as the standards. For this reason, short oligonucleotides rather than *
To whom
correspondence should be addressed
restriction fragments should be employed. We
use a
cellular
extract containing baculovirus-expressed human oestrogen receptor, known to bind as a dimer to its consensus DNA sequence (4), to illustrate the application of this technique to
DNA-binding proteins. A series of polyacrylamide gels containing 0.25 xTBE and covering a range of 4.5-10% acrylamide concentration (29:1 acrylamide:bis-acrylamide) are cast between glass plates (1Ox 13 cm) with 1.5 mm spacers and pre-run for 2 hours in 0.25 x TBE running buffer. Binding reactions are performed with 5 fmol 32P-end-labelled oligonucleotide and 1 11 protein (corresponding to 1-5 /tg protein in a crude extract or 1-10 ,ug purified protein) in 10 mM Tris-HCl (pH 7.8), 2 mM MgCl2, 4% Ficoll, 5 Ag poly(dI:dC).poly(dI:dC), in a final volume of 20 1tl. After 15 minutes incubation on ice, 2 ,ul loading buffer (20 mM Tris-HCI (pH 7.8), 0.05% bromophenol blue) is added and the samples immediately loaded onto the pre-run gel. 1 /g of each protein standard (non-denatured molecular weight markers; Sigma) is mixed with the same loading buffer and run alongside the binding reaction. The gels are run at 11 V/cm until the running dye just reaches the bottom of the gel. The lane containing the binding reaction is dried onto paper and subjected to autoradiography, with the position of the bromophenol blue marked with radioactive ink. The remainder of the gel is stained with Coomassie blue or silver. The distance migrated by the protein-DNA complex and by each standard is then measured and divided by the distance that the bromophenol blue in the same sample migrated, giving the relative mobility (Rf). Where the protein standards contain more than one band due to the presence of charge isomers, the Rf of the major isomer is used as described (5). The data is analysed according to Ferguson (1). First, a plot of the logarithm of the relative mobility against the percent gel concentration is constructed for each species (Figure 1). The gradient of such a plot is known as the retardation coefficient (Kr). Kr is then plotted against molecular weight for the protein standards, and the resulting linear plot used to determine the molecular weight of the protein-DNA complex from its Kr (Figure 2). Subtracting the contribution of the DNA oligonucleotide provides an estimate of the molecular weight of the protein component of the complex. For the human oestrogen receptor bound to a 35 bp oligonucleotide containing the oestrogen response element from the Xenopus vitellogenin A2 promoter, a value of 154 kD was obtained for the protein-DNA complex.
3336 Nucleic Acids Research, 1993, Vol. 21, No. 14 Subtracting 23 kD to allow for the contribution of the DNA yields an estimate of 131 kD for the molecular weight of the protein component of the complex. This is in good agreement with the value expected for binding of the homodimeric receptor, the molecular weight of the monomer being reported as 65 kD (6). The method has also been successfully applied to a variant of the heterodimeric DNA-binding Ku autoantigen (2) and a T lymphocyte protein which binds to site B of the human immunodeficiency virus type 1 (3).
100
ui
o
10.
010
n
ACKNOWLEDGEMENTS We are grateful to Dr M.Parker (Imperial Cancer Research Fund, London) for providing baculovirus-expressed human oestrogen receptor and the Xenopus vitellogenin A2 promoter oligonucleotide. We are also grateful to Hannah Gould, Mary Collins and Tariq Enver for helpful discussions. K.O. was supported by a Clinical Training Fellowship (MRC AIDSdirected Programme).
REFERENCES 1. Ferguson,K.A. (1964) Metab. Clin. Exp. 13, 985-1002. 2. May,G.E., Sutton,C. and Gould,H. (1991) J. Biol. Chem. 266, 3052-3059. 3. Orchard,K., Lang,G., Collins,M. and Latchman,D. (1992) Nucleic Acids Res. 20, 5429-5434. 4. Kumar,V. and Chambon,P. (1988) Cell 55, 145-156. 5. Sigmna Technical Bulletin MKR-137. 6. Walter,P., Green,S., Greene,G., Krust,A., Bomert,J.-M., Jeltsch,J.-M., Staub,A., Jensen,E., Scrace,G., Waterfield,M. and Chambon,P. (1985) Proc. Natl. Acad Sci. USA 82, 7889-7893.
200
180
0 a
160
N
v-
C m
140O
l0 S
0
o 120,
100.
80
GEL CONC.
Figure 1. Calibration curves. Logarithm of relative mobility (Rf) plotted against percentage acrylamide, showing the relationship between each species' mobility and the gel concentration: lactalbumin ( * ); carbonic anliydrase ([E); ovalbumin (A); BSA monomer (0); BSA dimer (M); BSA trimer ( o ); urease trimer (El) oestrogen receptor-DNA complex (A).
1
10
100 MOL. WEIGHT X 1000
1000
Figure 2. Ferguson Plot. The gradient of each line in Figure 1 (Kr) plotted against the molecular weight of the standards (0), generating a standard curve from which the molecular weight of the oestrogen receptor-DNA complex (0) can be determined.
