We have separated double stranded DNA fragments by electrophoresis using a membrane instead of a gel as separating matrix. Such a ... Several manual [7-91.
162
Electrophoresis 1993, 14, 162-164
C. Heller and S . Beck
Christoph Heller Stephan Beck
Membrane electrophoresis of DNA
Imperial Cancer Research Fund, London
We have separated double stranded DNA fragments by electrophoresis using a membrane instead of a gel as separating matrix. Such a membrane, containing separated DNA fragments, is suitable for subsequent hybridization analysis. The method combines electrophoretic separation and transfer onto a solid support and therefore eliminates the need for a transfer or blotting step.
Electrophoresis is a widely used technique to separate biological macromolecules. However, for many applications, such as the separation of DNA, a separation based on the mobilities in free solution is not possible [l]. Therefore, a separating matrix is needed,which has two functions: it serves as an anticonvective medium as well as a “sieving” matrix. The matrices most commonly used, in the electrophoretic separation of DNA, are agarose, which is extracted from several genera of red algae, and polyacrylamide, a synthetic polymer (see [2] forareview).Both substances form a gel in which the separation takes place. In recent years, the separation with these matrices has been pushed to the limit, i.e. from molecules as short as trinucleotides in high percentage polyacrylamide gels (e.g. [3]) to molecules greater than lo6bp in pulsed field agarose gels. In order to separate even bigger molecules, new matrices with controllable pore size will be needed. A further problem in the electrophoretic separation of biological molecules is the fact that they are “trapped” in the gel after the separation and therefore are not easily accessible for further analysis. This can be avoided by using liquid polymer media as a separation matrix [4-61, but in that case the identification of bands, which is usually done with lane-by-lane comparison or hybridization, is not possible. Therefore, in order to characterize the separated molecules it is often necessary to transfer them out of the gel onto a solid support (so-called “blotting”). Several manual [7-91 and automated [101blotting methods have been developed, but the process can still be inconvenient and time-consuming. As a blotting matrix, membranes made from, e. g., nitrocellulose, nylon, polyvinylidene difluoride or polysulfone are used. Microscopic pictures from such membranes [ l l ] show a similar porous structure as observed in gels [ 12-14]. Therefore, we investigated whether a single matrix can be used forboth the electrophoretic separation and the subsequent analysis of DNA (e.g. by hybridization), thus eliminating the need for blotting. Chemically modified paper supports (e.g.cellulose acetate) have been used in the past as separation matrices for proteins and RNA (see [2] for a review) but, to our knowledge, have never been applied to the separation of DNA.
TuqI which gives 4 fragments of 1444 bp, 736 bp, 476 bp and 30 bp in size. The pUC-18 digest was labeled in the same way as the 123 bp ladder. For the hybridization the 736 bp fragment was gel purified and labeled with [32p]dATP by random primer labeling [15].
Two different set-ups were tested for membrane electroph0resis.h the first set-up (Fig. la), the membrane was prewetted in electrophoresis buffer (1 X TBE, pH 8.8: 16.2 g/L Tris, 2.75 g/L boric acid, 0.92 g/LNa,EDTA) and then transferred into a horizontal electrophoresis chamber (e.g. KochLight, 70500-B). The buffer compartments were half-filled and the ends of the membrane immersed into the buffer. The membrane was supported by one sheet of 3MM paper (Whatman) to keep it moist. In the second set-up (Fig. Ib), the wet membrane was placed in a shallow electrophoresis chamber (e.g. Cambridge Electrophoresis, EM 100). Electrical contact with the electrodes was made with two strips of 1% agarose. This set-up is similar to the one used by Allen ef al. [16] for precast gels. Excess liquid was blotted off with 3 MM paper and the radioactive labeled samples A)
I I
Y
. P
B
*
B)
Two DNA standards were used. The 123 bp ladder was from Gibco-BRL. Labeling with [%]dATP was carried out by using two enzymes, Klenow fragment and T4 DNA polymerase, as recommended by the manufacturer. The second standard consisted of the plasmid pUC-18 digested with
Correspondence: C. Heller, Ecole Supkrieure de Physique et de Chimie Industrielles, Laboratoire de Physico-ChimieThBorique, 10 rue Vauquelin, F-75231 Paris cedex 05, France Abbreviations: bp, basepair; dATP, 2‘-deoxy-adenosine-S’-triphosphate 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1993
E Figure 1. Two schematic set-ups for membrane electrophoresis.M, membrane; P,3MM paper; E, electrode; B, buffer; A, agarose; L, lid. 0173-0835/93/0102-0162 $5.00+.25/0
Electrophoresis 1993, 14, 162-164
Membrane electrophoresis of DNA
were loaded onto the surface of the membranes with the help of a glass capillary. Typical electric fields were 13 V/cm and the runtime was about 2 h. After the run, the membrane was taken out of the chamber, dried at 80” C and exposed to X-ray film. In the case of a subsequent hybridization, the moist membrane was carefully transferred onto a sheet of 3MM paper, which had been soaked with denaturing solution (0.5 M NaOH, 1.5 M NaCl). After 10 min it was transferred onto a new 3MM paper soaked in neutralizing solution (1OX SSC, pH 7: 87.6 g/L NaC1,44.1 g/L sodium citrate). After another 10 min the membrane was dried at 80°C for30 min and hybridized in 0.5 M sodium phosphate (pH 7.2), 0.1% sodium dodecyl sulfate (SDS), 1 mM EDTA at 6S°C overnight. We have tested a variety of membranes consisting of different materials and different pore sizes. The best separation of double-stranded DNA molecules was achieved using a nitrocellulose membrane (Schleicher & Schuell, PH 70, Lot Nr. 7032/94 and 7010/13020) with a pore size of 0.025 pm. Under the conditions described here, three bands of the pUCl8-TaqI digest and 9 bands of the 123 bp ladder could be resolved (Fig. 2, lanes a and b). We can assume that the 30 bp fragment of the digest moves much faster than the others and has been run out of the membrane. Therefore, we can identify the three bands in lane (b) as being the 1444, 736 and 476 bp fragments. In a separate experiment using the same DNA markers we show that DNA fragments separated by membrane electrophoresis are suitable for subsequent hybridization analysis. Lanes (c) and (d) of Fig. 2, show the result of such an analysis using the 736 bp fragment as hybridization probe. The probe hybridizes to the middle band, giving the confirmation of the correct assignment of the bands (Fig. 2d). However, the equally strong
163
signal at the origin indicates that a substantial amount of the loaded DNA remained bound to the origin of loading during e1ectrophoresis.The strong signal in Fig. 2c is due to the fact that the 736 bp fragment (which was used as probe) also presents part of the vector construction of the 123 bp ladder [17]. Figure 3 shows a comparison of the migration distances of the same DNA fragments after membrane electrophoresis and 1% agarose gel electrophoresis. The same buffer and the same electric field strength was used in both experiments but the runtime for the membrane electrophoresis was four times longer. The membrane-separated DNA fragments show a smaller separation “window” than the gelseparated fragments but with larger interband spacing. In conclusion, we could show that membranes such as nitrocellulose are potentially useful as a separation matrix for the electrophoretic separation of DNA molecules. So far, it is not clear whether this separation is due to molecular sieving or to other interactions, such as adsorption. Parallel experiments with membranes of the same material but pore sizes of0.05 um (Schleicher & Schuell, PH 75) and 0.1 pm (Schleicher&Schuell, PH 79) resulted in a much poorer separation. However, the mobility of the bulk DNAwas higher with larger pore size (data not shown), which indicates that the pore size (and therefore a sieving mechanism) plays an important role. On the other hand, the runtime for membrane electrophoresis is currently longer, compared to agarose gel electrophoresis. This could mean that there might be a stronger interaction of DNA with nitrocellulose than with agarose. We have also shown that the separated DNA is suitable for subsequent hybridization analysis. Consequently, only one single matrix is necessary for separating and analyzing DNA molecules, a blotting step is therefore no longer necessary.
50
Compared to gels, membranes offer several advantages: (i) They have a higher physical and chemical stability. (ii) They are easier to handle, e.g. no gel casting is required. (iii) Because membranes are very thin (- 100 pm) high
E
f
Y
40
F
c
Q
L
.-m E
30
20
0
I
I
I
500
1000
1500
2000
Mol. Weight (bp) Figure 2, Aulorddiograph of DNA fragments separated by membrane
electrophoresis (lanes a and b). Autoradiograph of same DNA fragments but different membrane after hybridization with radioactive labeled 736 bp fragment (lanes c and d). Lanes (a) and (c) 123 bp ladder; (b) and (d), pUC18/7iuqIdigest.The30 bpbandofthepUClS/TaqIdigestwasoutside the separation window used here. In both experiments nitrocellulose with a pore size of0.025 pm and a thickness of 105 pm was used as separating matrix (Schleicher & Schuell, PH 70).
Figure 3. Comparison of the migration distances of double-stranded DNA fragments after membrane electrophoresis and 1% agarose gel electrophoresis. The conditions were: buffer, 1 X TBE, pH 8.8; electric field strength, 13 V/cm; runtime,30 min (gel) and 120 min (membrane; Schleicher & Schuell, PH 70).
164
C . Heller and S . Beck
electric fields can be used without temperature control requirements. However, as pointed out by Grossman et al. [18], the limitation of the electric field strength in the separation of larger DNA fragments is more due to the induced stretching of the molecules than to the dispersion effects caused by Joule heating. Thus, membranes might be most useful for separating small molecules. Membranes are also nontoxic and can be manufactured in a wide variety of pore sizes. Nitrocellulose, for example, is currentlyavailable in pore sizes from 0.025 pm up to 5.0 wm. It remains to be tested if separations with other pore sizes can be improved. A “fine tuning” of pore size might be possible by supplementing the buffer with dilute polymer solutions, such as cellulose derivatives or polyacrylamide, which are successfully used for the separation of DNA in tubes [4] and capillaries [5,6,19]. In this case the two functions mentioned above would be split, i.e. the “ske1eton”of the membrane fibers would act as an anticonvective media and the polymer solution would provide the sieving matrix. Clearly, there are also areas which still require improvement.The sample loading,for instance, which affects the reproducibility, requires some experience and is limited to very small amounts. However, we are confident that these problems can be overcome, e.g. by using sample applicators for loading. We thank Elmar Maier,for his help with the hybridization. Parts of this work were financialb supported by Amersham International. Received July 31, 1992
Electrophoresis 1993, 14, 162-164
References [ l ] Olivera, B.M., Baine,P. and Davidson, N., Biopolyrnerv 1964, 2, 245-257. [2] Righetti, P.G., J. Biochem. Biophys. Methods 1989, 19, 1-20. [3] Frank, R. and Koster, H. Nucleic Acids Res. 1979, 6, 2069-2086. [4] Garncr, M. and Chrambach, A., Electrophoresis 1992, 13, 176-178. [5] Heiger,D.N.,Cohen, A . S.and Kdrger,B. L.,.I. Chromatogr. 1990,5/6, 33-48. [6] Strege, hf. and Lagu, A., A n d . Chem. 1991, 63, 1233-1236. [7] Southern, E.M., J. Mol. Biol. 1975.98.503-517. (81 Peferoen, M.,Huybrechts, R. and Dc Loof. A , ,FEBSLctt. 1982,145, 369-372. [9] Towbin, H., Staehelin, T. and Gordon, J . , Proc. Natl. Acad. Sci. USA 1979, 76,3116-3120. [lo] Beck. S . and Pohl, F.M., EMBO J. 1984. 3,2905-2909. [ l l ] Gelman Sciences, The Filter Book. Ann Arbor 1991. [I21 Ruchcl, R. and Brager. M. D., Anal. Biuchem. 1975, 68,415-428. [13] Ruchel, R., Steere, R. L. and Erbe, E. F., Chrumufugr. 1978, 166, 563-575. [14] Atlwood,T.K., Nelmes, B. J. and Selien, D.B.. BiopolJimers 1988,27, 201-212. [15] Feinberg, A.P. and Vogelstein, B., Anal. Biochern. 1983, 132. 6-13. [I61 Allen, R. C., Graves, G. and Budowle, B.. Bio Z ~ h n / q i t e s1989, 7. 736-744. [I71 Hartley, J.L. and Gregori,T.J., Gene 1981, 113, 347-353. [IS] Grossman, P.D., Menchen, S., and Hershey, D., Genet. Anal. Techn. Applic. 1992, 9,9-16. [19] Grossman, P.D. and Soanc, D. S . Biopolyvnem 1991,31, 1221-1228.
