J. Mol. Biol. (1975) 98, 503-517 Detection of Specific Sequences Among DNA Fragments Separated by Gel Electrophoresis "E,. 1~. SOUTHERN Medical Research Council Mammalian Genome Unit DeIaartment of Zoology University of Edinburgh West Mains Road, Edinburgh, Scotland (Received 3 March 1975, and in revised form 26 June 1975) This paper describes a method of transferring fragments of DNA from agarose gels to cellulose nitrate filters. The fragments can then be hybridized to radio- active RNA and hybrids detected by radioautography or fluorography. The method is illustrated by analyses of restriction fragments complementary to ribosomal RNAs from Esoherichia coli and Xenopus lacvis, and from several mammals. 1. Introduction Since Smith and his colleagues (Smith & Wilcox, 1970; Kelly & Smith, 1970) showed that a restriction endonuclease from HaemoThilns influenzae makes double-stranded breaks at specific sequences in DNA, this enzyme and others with similar properties have been used increasingly for studying the structure of DNA. Fragments produced by the enzymes can be separated with high resolution by electrophoresis in agarose or polyacrylamide gels. For studies of sequences in the DNA that are transcribed into RNA, it would clearly be helpful to have a method of detecting fragments in the gel that are complementary to a given RNA. This can be done by slicing the gel, eluting the DNA and hybridizing to RNA either in solution, or after binding the DNA to filters. The method is time consuming and inevitably leads to some loss in the resolv- ing power of gel electrophoresis. This paper describes a method for transferring fragments of DNA from strips of agarose gel to strips of cellulose nitrate. After hybri- dization to radioactive RNA, the fragments in the DNA that contain transcribed sequences can be detected as sharp bands by radioautography or fluorography of the cellulose nitrate strip. The method has the advantages that it retains the high resolv- ing power of the gel, it is economical of RNA and cellulose nitrate filters, and several electrophoretograms can be hybridized in one day. The main disadvantage is that fragments of 500 nucleotide pairs or less give low yields of hybrid and such fragments will be under-represented or even missing from the analysis. 2. Materials, Methods and Results (a) Restriction endonuclease~ EcoRI prepared according to the method of u (1971) was a gift of K. Murray. HaeIII prepared by a modification of the method of Roberts (unpublished data) was a gift of H. J. Cooke. 503
`
`

`

`504 E.M. SOUTHERN (b) Gel dearophoresis Gels were cast between glass plates (de Waehter & Fiers, 1971). The plates were separated by Perspex ~de pieces 3 rnrn thick and along one edge was placed a "comb" of Perspex, which moulded the sample wells in the gel. The Perspex pieces were sealed to the glass plates bwith silicone grease and the plates clamped together with Bulldog clips. The assembly was stood with the comb along the lower edge. Agarose solution (Sigma electrophoresis grade agarose) was prepared by dissolving the appropriate weight in boiling electrophoresis buffer (E buffer of Loenlng, 1969). The solution was cooled to 60 to 70~ and poured into the assembly, where it was allowed to set for at least an hour. The assembly was then inverted, the comb removed and the wells filled with electrophoresis buffer. Samples made 5 % with glycerol were loaded from a drawn-out capillary by inserting the tip below the surface and blowing gently. Electrophoresis buffer was layered carefully to fill the remaining space and a filter-paper wick inserted between the glass plates along the top edge. The lower end of the assembly was immersed in a tray of eleetrophoresis buffer containing the plati- num anode, and the paper wick dipped into a similar cathode compartment. Electro- phoresis was at 1.0 to 1.5 mA/cm width of gel for a period of about 18 h. Bromo- phenol blue marker travels about 3/4 the length of the gel under these conditions, but it should be noted that small DNA fragments move ahead of the bromophenol blue, especially in dilute gels. Cylindrical gels were cast in Perspex tubes 9 mm i.d. and either 12 or 24 cm long. These were run at 3 to 5 mA/tube in standard gel electro- phoresis equipment. Dr J. Spiers donated ribosomal DNA that had been purified on actinomycin/ caesium chloride gradients from DNA made from the pooled blood of several animals, and also "qH-labelled 18 S and 28 S RNAs prepared from cultured Xenopns laevis kidney cells. Escherichia cell DNA was prepared by Marmur's (1961) procedure from strain iVIRE600. 82P-labelled E. coli RNA was prepared from cells grown in low phosphate medium with 82Pi at a concentration of 50 t, Ci/ml and frac- tionated by electrophoresis on 10% acrylamide gels. 82P-labelhd rat DNA was a gift of M. S. Campo. DNA .from human placenta was a gift of H. J. Cooke, DNA from rat liver was a gift of A. R. Mitchell, DNA from mouse and rabbit livers were gifts of M. White. Caff thymus DNA was purchased from Sigma Biochemicals. For digestion with restriction endonucleases, the DNAs were dissolved in water to a concentration of approximately 1 mg/ml. One-tenth volume of the appropriate buffer was added and sufficient enzyme to give a complete digestion overnight at 37~ Enzyme activity was checked on phage )~ DNA and digests of this DNA were also used as size markers in gel electrophoresis, using the values given by Thomas & Davis (1975). (c) Method of transfer This section describes the method finally adopted: preliminary experiments and controls are described in later sections. After eleetrophoresis, the gel is immersed for 1 to 2 h in electrophoresis buffer con- talning ethidium bromide (0.5 ~g/ml), and photographed in ultraviolet light (254 rim) with a red filter on the camera. A rule laid alongside the gel aids in matching the photograph of the fluorescence of the DNA to the final radioautograph of the hybrids. Strips to be used for transfer from flat gels are cut from the gel using a flamed blade. The strips should be 0-5 cm to 1 cm wide and normally extend from the origin to the
`
`

`

`HYBRIDIZATION TO RESTRICTION FRAGMENTS 505 anode end of the gel. The gels used in this laboratory are 3 mm thick, and the length from the origin to the anode end is 18 cm but the method can be adapted to gels with different dimensions and to cylindrical gels. Strips of gel are then transferred to measuring cylinders containing 1"5 M-NaCI, 0.5 ~-NaOH for 15 mln and this solution is then replaced by 3 ~-NaC1, 0.5 M-Tris.HC1 (pH 7) and the gel is left for a further 15 mln. The depth of liquid in the cylinders should be greater than the length of the gel strips and the cylinders should be inverted from time to time. For cylindrical gels (9 mm diam.), the times required for denaturation and neutralization are 30 and 90 mln. Each gel transfer requires: One piece of thick filter paper 20 cm (cid:141) 18 cm, soaked in 20 (cid:141) SSC (SSC is 0.15 ~- NaC1, 0.015 ~-sodinm citrate). Two pieces of thick filter paper 2 cm (cid:141) 18 em soaked in 2 (cid:141) SSC. One strip of cellulose nitrate filter (e.g. Milipore 25 HAWP), 2.2 cm (cid:141) 18 cm, soaked in 2 (cid:141) SSC. These strips are immersed first by floating them on the surface of the solution; otherwise air is trapped in patches, which leads to uneven transfer. Three pieces of glass or Perspex, 5 cm (cid:141) 20 cm and the same thickness as the gel. Four or five pieces of thick, dry filter paper, 10 cm (cid:141) 18 cm. Transfer of the denatured I)NA fragments is carried out as follows. The large filter paper soaked in 20 (cid:141) SSC is laid on a glass or plastic surface, care being taken to avoid trapping air bubbles below the paper. 20 (cid:141) SSC is poured on so that the surface is glistening wet. One of the glass or Perspex sheets is laid on top of the wet paper. The gel strip is taken from the neutralizing solution and laid parallel to the glass or Perspex sheet, 2 to 3 mm away from it. The second glass or Perspex sheet is laid 2 to 3 mm away from the other side of the gel (Fig. l(a)). The cellulose nitrate strip is then laid on top of the gel with its edges resting on the sheets of Perspex or glass, so that it bridges the two air spaces (Fig. l(b)). The two narrow pieces of filter paper, moistened with 2 (cid:141) SSC are laid with their edges overlapping the cellulose nitrate strip by about 5 mm (Fig. l(c)) and the dry filter paper is then placed on top of these (Fig. l(d)). For cylindrical gels, the arrangement is similar, but in this case, the Perspex that supports the Millipore filter may be in contact with the gel because an air space is retained over the top of the gel. Several cylindrical gels can be transferred at the same time using the apparatus shown in Fig. 2 and similar arrangements can be used for fiat gels. 20 x SSC passes through the gel drawn by the dry filter paper and carries the )NA, which becomes trapped in the cellulose nitrate. The minimum time required for complete transfer has not been measured: it depends on the size of the fragments and probably also depends on the gel concentration. A period of 3 h is enough to transfer completely all HaeIII fragments of ~. coli DNA from 2~ agarose gels 3 mm thick. But even after 20 h, transfer of large EcoRI fragments of mouse DNA from 9 mm diam. cylindrical gels is not complete. DNA remaining in the gel can be seen by the fluorescence of the ethidium bromide, which is not completely removed during treatment of the gel. During the period of the transfer, it is necessary occasionally to add more 20 (cid:141) SSC to the bottom sheet of filter paper. If the paper dries too much, the gel shrinr~ against the cellulose nitrate strip and liquid contact is broken. The paper may be flooded, but care must be taken that liquid does not fill the air spaces between the gel and the side-pieces and soak the paper, bypassing the gel. It may be found convenient to leave the cellulose nitrate in position overnight: ff the supply of
`
`

`

`506 E. M. SOUTHERN / Perspex~x,~ (a) Ge,~" //Ce.~lose ~ st,'ip Moisl filter paper strips Dry filter paper FIG. 1. Steps in the procedure for transferring DNA from agarose gels to cellulose nitrate strips. 20 X SSC has dried up it will be found that the gel has shrunk against the cellulose nitrate, but this does not impair the transfer. At the end of the transfer period the cellulose nitrate strip is lifted carefully so that the gel remains attached to its under- side. It is turned over and the outline of the gel marked in pencil by a series of dots. The gel is peeled off the cellulose nitrate, the area of contact cut out with a flamed blade, and immersed in 2 x SSC for 10 to 20 mln. The strip is then baked in a vacuum oven at 80~ for 2 h. (d) Hybridization Radioactive RNAs are usually available in small quantities only and iris important to keep the volume of the solution used for hybridization as small as possible so that the RNA has a reasonable concentration. Two procedures can be used for hybridizing the cellulose nitrate strips after transferring the restriction fragments. The procedure that uses the smallest volume is carried out by moistening the strip in hybridization mixture and then immersing it in paraffin oil. A drop of RNA solu- tion (0.3 ml for a strip 1 cm (cid:141) 18 era) is placed on a plastic sheet. One end of the
`
`

`

`HYBRIDIZATION TO RESTRICTION FRAGMENTS 507 DDDD(cid:1)89 Fro. 2. Apparatus for transferring DNA from a number of cylindrical gels. The apparatus is constructed of Perspex. The uprights which separate the gels and suppor~ the sheet of cellulose nitrate should be about 0.5 mm higher than the diameter of the gels, so that the cellulose nitrate sheet dips down to touch the gel. Thus an air gap is left between the cellulose nitrate sheet and the filter paper, above the line of contact between the gel and cellulose nitrate sheet. The apparatus is laid in a shallow tray containing 20 X SSC and the gels are then inserted into the troughs, care being taken to avoid trapping air bubbles beneath the gel. The cellulose nitrate sheet, wet with 2 X SSC, is laid over the gels and one piece of wet filter paper is laid over this. A stack of dry filter paper is then placed over the whole assembly. If necessary, a glass plate can be used to weigh down the filter papers. The depth of 20 (cid:141) SSC in the tray should be enough to cover the lower part of the gels, but not so much that the air space between the Perspex and the cellulose nitrate becomes flooded. cellulose nitrate strip is floated on the drop and when liquid is seen to soak through, the strip is drawn slowly over the surface of the drop. When it is completely wetted from one side, it is turned over and any remaining liquid is used to wet the other side. The strip is then immersed in paraffin oil saturated with the hybridization solution at the hybridization temperature. It should be borne in mind that baking the strip in 2 (cid:141) SSC introduces salt, which must be taken into account when deciding on a solvent for the RNA if this method of hybridization is used. For example, if hybridization is to be carried ont in 6 (cid:141) SSC the RNA should be dissolved in 4 (cid:141) SSC. Though this method can give good results (see Plate I) it often leads to high and uneven back- ground. Kourilsky et al. (1974) found that this problem is removed if the hybridiza- tion is carried out in 2 (cid:141) SSC, 40% formamide at 40~ I have not tried this method, because this solvent removed DNA from the filters (see later section). It may well be the best method for hybridization to large fragments. I have found it convenient to carry out the hybridization in a vessel designed to hold the strip in a small volume of ~qnid. The vessel (Fig. 3), which is easily made from Perspex, has internal dimensions of 0.8 mm deep by 2 cm high and about 1 cm longer than the strip to be hybridized. The vessel is filled with the solvent to be used for hybridization and the strip is fed in through the narrow opening in the top. The solvent is then drained off and the RNA solution introduced. Around 1 ml of solution is needed for a strip 1 cm (cid:141) 18 cm. The wide sheets of cellulose nitrate used for transferring several gels (e.g. using the apparatus shown in Fig. 2) are too wide to be hybridized in this type of vessel. They can be hybridized in a small volume by wrapping them around a cylinder of Perspex, which is then inserted into a close-fitting tube. In this way, it is possible to hybridize a sheet 24 cm (cid:141) 8 cm with about 4 ml of solution. If hybridization is carried out in a water-bath, it is not necessary to seal the top of the vessel provided the water-bath
`
`

`

`508 E. M. SOUTHERN FIG. 3. Vessel used for hybridization of narrow strips. itself is covered. The liquid in the vessel evaporates very slowly and can be replenished by small additions of water. A further advantage of this method of hybridization is that the RNA can be recovered and used again. The period allowed for hybridization depends on the RNA concentration, its sequence complexity, its purity, and on the conditions of hybridization (see for example Bishop, 1972). After the appropriate period, strips are removed from the solution or paraf~n oil, blotted between sheets of filter paper and washed, with stirring, for 20 to 30 min in a large volume of the hybridization solvent at the hybridization temperature. If the background is high, they may then be treated with a solution of RNAase A (20 ~gml in 2 (cid:141) SSC for 30 rain at 20~ After a final rinse in 2 (cid:141) SSC they are dried in air. So far the method has been tested with 3~p, 8H, 35 S and 125i.labelled RNAs. 82PRNAs have been detected by radioautography. For this the cellulose nitrate strips are laid on X-ray ~lm and flattened against it with light pressure. 3H, 1~5I, 35S and 14C may be detected by fluorography. The cellulose nitrate strip is dipped through a solution of PPO in toluene (200/o, wv) dried in air, laid against X-ray film (Kodak RP-Royal Xomat) and kept at --70~ (e) Completeness of transfer and retention of DNA PreJmlnary experiments showed that loading of DNA on to cellulose nitrate filters in 6 (cid:141) SSC, conditions widely used in hybridization work, did not give complete retention of small fragments and a systematic study was made of the effect of salt concentration on retention. 8H-labelled X. laevi~ DNA was sonicated to a single- strand molecular weight of 104 and denatured by boiling in 0.1 (cid:141) SSC. Samples were made up to various salt concentrations and 0.1-ml portions of these solutions were pipetted on to cellulose nitrate filters, previously moistened with 2 (cid:141) SSC, which were resting on glass-fibre filters. The solution that passed through the cellu- lose nitrate filter was thus collected in the glass-fibre filter. Both filters were then ~mmersed in 5~ trichloroacetic acid for 10 mln, dried for 30 rnln in & vacuum oven at 80~ and counted. It can be seen (Fig. 4) that the fraction of DNA retained by the cellulose nitrate increases with the salt concentration, and at concentrations above 10 (cid:141) SSC the DNA is almost completely retained. Losses of DNA at various stages of the transfer procedure were measured using 32P-labelled E. co~i DNA. The DNA was digested with EcoRI to give fragments in
`
`

`

`PLATE I. HaeIII digest of E. coli MRE600 DNA analyzed by electrophorosis on 2% agarose gel. DNA was then transferred to cellulose nitrate and hybridized with z~P-iabelled, high mole- cular weight RNA. (a) and (d) Photographs of ethidium bromide fluorescence. (b) and (e) Radio- autographs of hybrids. fSac/r,g ~o. ~o8
`
`

`

`PLATE II. EcoRI digest of purified X. la6vis ribosomal DNA analyzed by electrophoresis on 1% agarose gel. DNA was transferred to a cellulose nitrate strip, which was then cut longitudinally in two. The left-hand side was hybridized to 18 S RNA and the right-hand side to 28 S RNA (spec. act. of RNAs, 1.5 x I0 e e.p.m, per/~g}. Hybridization was done in 1 (cid:141) SSC at 65~ using the vessel shown in Fig. 3. A large excess of cold 28 S RNA was added to the labelled 18 S RNA to compete out any 28 S contamination. After hybridization, the strips were washed in 1 x SSC at 65~ for 1.5 h, and dried. They were then dipped through a solution of PPO in toluene (20%, w]v) dried in air and placed against Kodak RP Royal X-ray film at --70~ for 2 months. Photograph of ethidium bromide fluorescence (e). Fluorograph of 18 S hybrids (a). Fluorograph of 28 S hybrids (b).
`
`

`

`}r " 5~.,. F" o (cid:12)9 " .' " " '5;~'~:'~t;~ (cid:12)9 " . "., (cid:12)9 ."~. 5!~z,~,:~,' (cid:12)9 ' (cid:12)9 ,., .;,,,,: ~I~ ,.'. , .... (cid:12)9 .,~ '.. (cid:12)9 (cid:12)9 '~-I~,- . ~-. ~ta~.. ";i,_ ~ ' 4 ~-~..'-:. ,,F'., : + ~: ;~;,F ~ ~!~,;'..::-;~:--t, I ;5,,~-~,~ "~ ,~( :~=~,, ~:.- ;.: ,,! PLATE III. EcoRI digests of five mammalian DNAs, hybridized to 28 S RNA. Calf (a), human (b), mouse (c), rabbit (d) and rat (c) DNAs were digested to completion with EcoRI and separated by electrophoresis on 1% agarose gels ( 9ram x 12 cm, approx. 40 t~g DNA per tube, 3 mA/tube for 16 h). The gels were pretreated as usual and the DNA fragments transferred to a single sheet of cellulose nitrate filter (12 cm (cid:141) 8 era) using the apparatus shown in Fig. 2. The top end of each gel was carefully aligned with one edge of the cellulose nitrate sheet. After 20 h, traces of DNA could still be seen, by ethidium bromide fluorescence, in the high molecular weight region of the gel. The filter was hybridized with 28 S RNA and radioautographed as described in the legend to Fig. 8.
`
`

`

`HYBRIDIZATION TO RESTRICTION FRAGMENTS 509 100 80 6O Z 4O 20 IO / -- O / -- O / -- O 10 0 I I I I I-3 4x 8x 12x I6x 2.3,x SSC concenIFo~ion FIG. 4. Effect of salt concentration on efficiency of binding sonicated DNA to cellulose nitrate filters. the large size range and with HaeIII to give small fragments. The fragments were then separated on a fiat 1 ~ agarose gel and transferred in the usual way. The solu- tions, the gel and the cellulose nitrate strip were counted. It can be seen (Table 1) that, whereas a small proportion of the DNA is leached out into the solutions during denaturation and neutralization, only traces remain in the gel after transfer. TABLE 1 Losses of DNA at stages of the procedure EooRI HaeIII fragments fragments DNA lost (%) Denaturing solution 2.1 4.8 Neutralizing solution 1.3 4.4 Remaining in gel after 0.21 0.31 transfer Two samples of E. co//DNA (0.1 ~g; spee. act. approx. 106 e.p.m, per ~g) were digested with EcoRI and HaeIII. The fragments were separated by eleetrophoresis on 1~o gels in 1-om wide slots, and then transferred to cellulose nitrate strips as described in Materials and Methods. The transfer was left overnight. The radioactivity leached out of the gel by the denaturing and neutral- izing solutions, that remaining in the gel, and that which had been trapped on the cellulose nitrate filter were measured in a liquid scintillation counter (Cerenkov radiation). (f) Effect of D_~A size on yield of hybrid lYlelli & Bishop (1970) have shown that hybridization by the filter method gives low yields with low molecular weight DNA. Their results were obtained using a single set of hybridization conditions and it seemed possible that losses might be reduced by using high salt concentrations. The effect of salt concentration on loss of
`
`

`

`510 E. M. SOUTHERN DNA from the filters was examined by loading filters with radioactive X. laevis DNA, single-strand~molecular weight about 104, and incubating them in various salt , / solutions at different temperatures. Increasing the salt concentration does improve the retention of the DNA at any given temperature (Table 2) but the gain does not appear to be useful, because with increasing salt concentration it is necessary to use higher temperatures for hybridization, and this cancels the advantage of the high salt concentration. For example, the loss in 2 (cid:141) SSC at 65~ is the same as that in 6 (cid:141) SSC at 80~ and these are both typical hybridization conditions. Further experiments showed that it is disadvantageous to perform hybridization at high salt concentrations, below the optimum temperature. The optimum temperature for rate of hybridization of X. laevis 28 S RNA is around 80~ in 6 (cid:141) SSC but the rate at 70~ is still appreciable (Fig. 5). Below 70~ the rate fails rapidly. 28 S RNA was hybridized TABLE 2 Effects of temTerature and solvent on retention of sonicated DNA on cellulose nitrate filters Temperature Solvent 50~ 65~ 80~ 90~ DNA retained (%) 2 (cid:141) SSC 77 62 48 6(cid:141) 97 76 56 10 (cid:141) SSC 95 83 73 20 (cid:141) SSC 97 88 81 6 (cid:141) SSC in 50% formamide 58 50 3H-labelled X. laev/s DNA (spec. act. approx. 5 (cid:141) 105 c.p.m, per/~g) was dissolved in ice-cold 0.1 (cid:141) SSC and sonicated in six 15-s bursts. Between each treatment the solution was cooled in ice for 1 mln. The solution was boiled for 5 min, made to 20 (cid:141) SSC and cooled. Samples of this solution were pipetted on to 13-turn circles of cellulose nitrate, which were then washed in 2 (cid:141) SSC at room temperature. Approximately 650 c.p.m, were loaded on each filter, and there was no loss caused by washing in 2 (cid:141) SSC. The filters were dried, baked at 80~ for 2 h in a vacuum oven and ~rnmersed in 10 ml of the solvent equilibrated at the temperature used for incubation. After 90 rnln; the filters were removed, washed in 2 (cid:141) SSC at room temperature, dried under vacuum and counted in a liquid scintillation counter. to high molecular weight and sonicated DNA in 6 (cid:141) SSC at 70 and 80~ (Fig. 6). As expected, the rate of hybridization at 70~ was lower than the rate at 80~ but against expectation, both the rate and the final extent of hybridization were lower at the lower temperature, for the sonicated but not for the high molecular weight DNA. This result was unexpected because Melli & Bishop did not find an effect of DNA size on the rate of hybridization. They suggested that the decrease in yield for low molecular weight DNA is due to a loss of hybrid from the filter and it would be expected that such losses would increase with temperature. The lower yield for low molecular weight DNA at low temperature remains unexplained, but shows that there is no advantage to be gained in using high salt concentrations and low tempera- tures to retain small fragments of DNA during hybridization reactions. The advan- tage of using 6 (cid:141) SSC at optimum temperature is that the rate is greatly increased over the rate with, say, 2 (cid:141) SSC. A disadvantage is that the background of RNA that sticks to filters that have no DNA, increases with increasing salt concentration.
`
`

`

`HYBRIDIZATION TO RESTRICTION FRAGMENTS 511 (g) Me~ho~ o$ de~i~ and raea~rCng hybrid: adva~e~ of fdr~ ddeaion Radioactive RNA may be detected and measured either by radioautography (or fluorography for weak fl-emitters) or by cutting the strip into pieces, which can be counted in a scintillation counter. Film detection methods have the advantages over 8o .r (cid:12)9 ~ 60 ~' 40 ~20 7 n- O \o 50 60 70 80 90 Temperature (=C) FxQ. 5. Temperature dependence of hybridization of 28 S rRNA to X. ~v~s DNA. X. ~ev~s DNA was loaded on cellulose nitrate filters (17 pg DNA/13-mm diameter disc), which were pretreated as usual for hybridization. 8H-labelled 28 S RNA from X. b~ev~s kidney cels (spec. act. 1.5 X 10 e o.p.m./pg) was dissolved in 6 X SSC (0-28 pg/ml) and warmed to the temperature used for hybridization. Two filters loaded with DNA and 2 blank filters were introduced into the solutions and left for 30 min. They were washed in 2 1 of 2 x SSC at room temperature, treated with 200 ml of RNAase A (20 pg/ml in 2 X SSC) at room temperature for 20 m;nj washed in 200 ml of 2 X SSC for 10 mln; dried under vacuum and counted. Hybridization is expressed as a percentage of that obtained after 5 h at 80~ ~oo -! 8o "~ 60- N < 20- o --~ Z ~6Co'-~ ------ I I I 0 o 20 40 60 80 I00 "600 Time (rain) FIG. 6. Time course of hybridization of 28 S RNA to sonieated and high molecular weight DNA at 70 and 80~ Filters were loaded as described in the legend to Fig. 5. Two sets were loaded: one with high molecular weight DNA and one with DNA sonioated as" described in the legend to Table 2. Hybridization and subsequent treatment of the filters was carried out as described in the legend to Fig. 6 and filters removed at the times indicated. 6 X SSC at 80~ high molecular weight DI~A ((cid:12)9 6 x SSC at 70~ high molecular weight DNA ( (cid:12)9 ) : 6 x SSC, 80~ sonicated DNA (O): 6 x SSC at 70~ sonicated DNA (/k). 34
`
`

`

`512 E.M. SOUTHERN counting that they are more sensitive, give higher resolution, and can reveal artifacts not seen by counting. The high sensiti~ty is illustrated by the analysis of E. coli rDNA (Plato I(b)). None of the bands that is clearly visible in the radioautograph contained more than 10 e.p.m. The strip of cellulose nitrate was cut into 150, l-ram pieces and the pieces counf~l in a liquid scintillation counter. None of the pieces gave counts more than twice background and none of the features visible in the radioautograph was discern- ible from the counts. Around 100 c.p.m, of 82p in a single band 1 em wide can be detected with an overnight exposure. The radi0autograph shown in Plate I was exposed for 1 week. Fluorography of 3H is not so sensitive; about 3000 d.p.m, in a 1-em band are needed to give a visible exposure overnight. The fluorograph shown in Plate II was exposed for 2 months. The greater resolution of Alto detection is illustrated by a comparison of Plate II with Figure 7(c). Plate II is a fluorograph of the strip and Figure 7(c) shows the pattern of counts obtained by cutting the strip into l-ram pieces. Many of the bands seen in the fluorograph are not discernible in the pattern of counts (compare also the tracing of the fluorograph (Fig. 7(b)) with (o)). For ionizing radiation, blackening of the X-ray film is proportional to the amount of incident radiation, up to the limit whm'e a high proportion of silver grains are exposed. The relative amount of radioactivity in bands can therefore be compared by tracing radioautegraphs in a densitometer and comparing peak areas. However, like all other photosensitivie materials, X-ray Alms suffer from "reciprocity failure" at low intensities of illnmlnation by non-iouizing radiation and it is likely that bands which contain only a few counts of 8H will not be detected by fiuorography even after long exposures. I have not determined the lower limit of detection. Bonner & Laskey (1974) found that 500 d.p.m, of 3H in a band 1 cm (cid:141) 1 rnm could be detected in one week and in my own experience, less than 20 d.p.m, can be detected with longer exposure. Reciprocity failure could affect quantitation of fluorographs by densito- merry but comparison of Figure 7(b) and (e) suggests that the response of the Rim is linear within the limits of,this experiment. Clearly, quantitation of "~2P by densito- merry can be accurate and more sensitive than counting, but Alto response to 8H may not be linear for low amounts. An additional advantage of Aim detection is that non-specific binding of RNA to the cellulose nitrate is more easily distinguished from bands of hybrid. Plate HI illustrates this point. In this radioautograph, non-specific binding can be seen as dots and streaks with an appearance clearly different from that of a band. Had this strip been analysed by counting, non-specific binding would not have been dis- tinguishable from the hybrids. (h) Analysis of ribosomal DNA in X. laevis A total of 0-6/~g of purified X. laevis rDNA was digested with EcoRI and the fragments separated by electrophoresis in 1% agarose gels (Plate II(c)). The pattern of fragments is similar to that described by Wellauer et a/. (1974). They compared the secondary structures of the denatured DNA fragments with those of the ribo- somal RNAs and showed that the fastest intoning fragment (Mr approx. 3(cid:141) 108) contained most of the DNA coding for 28 S RNA, all of the transcribed spacer, and a small portion of the DNA coding for 18 S RNA. The larger fragments (Mr 4 to 6x 106) contained most of the DNA coding for 18 S RNA, all of the non-transcribed
`
`

`

`HYBRIDIZATION TO RESTRICTION FRAGMENTS 613 (a) 100 | ! t 4 5 6 l i i 7 8 9 120 (c) 20 E d. 80 -1D >. .>. 4O I I T 8 i (b) ~ o :>, Pt i 4 5 6 Migration (cm) I I ! 4 5 6 ? m~e ! 9 ! ! s Fia. 7. (a) Miorodensitometer tracing of the negative of PlatetII(c). (b) Microdensitometer tracing of Plate II(a). (o) Distribution of counts in the Millipore strip which on fluorography gave Plate H(a). The strip was cut into 1 mm pieces, which, were counted in a liquid scintillation counter at an efficiency of 40%. spacer, and a small portion of the DNA coding for 28 S RNA. Different lengths of non-transcribed spacer I)NA accounted for the variation in size of the longer frag- ments. The digest shown in Plate H(c) was transferred to cellulose nitrate as described previously. The strip was cut longitudinally inte 2 parts and 1 part was hybridized with 18 S RNA and the other with 28 S RNA. Hybrids were detected by fluorography of the 3H-labelled RNA (Plate II(a) and (b)). Comparison of Plate II(a) and (c)
`
`

`

`514 E.M. SOUTHERN shows that the resolution of the fine bands containing the 18 S coding sequence is not as high in the fluorograph as it is in the photograph of the gel. Whereas 9 bands can be distinguished in the photograph, only 7 can be distinguished with confidence in the fluorograph. From this analysis it is possible to locate the EcoRI site within the DNA coding for 18 S RzNA. As Wellauer et al. (1974) showed, 1 of the 2 breaks in the rDNA occurs towards one end of the 18 S region and the other is close to the distal end of the 28S region. The 3(cid:141) tool. wt fragment accounts for virtually all of the hybridization to 28 S RNA and for about 30~ of the hybridization to the 18 S RNA (27~ measured from the tracing of the fluorograph (Fig. 7(b)) and 31~ from the counts). Only traces of 28 S RNA hybridize to the heterogeneous collection of fragments with molecular weights between 4 and 6(cid:141) whereas about 70~ of the 18 S hybridization is accounted for in these fragments. Thus the break in the 28 S region of the DNA is very close to the end of the coding sequence and the break in the 18 S region is about one-third of the way into the coding sequence. (i) Analysis of mouse and rabbit ribosomal DNAs: evidence for long, non-transcribed spacer DNA An EcoRI digest of total