A modified primer extension procedure for specific detection of DNA-RNA hybrids on nylon membranes

ANALYTICALBIOCHEMISTRY

179,366-370

(1989)

A Modified Primer Extension Procedure for Specific Detection of DNA-RNA Hybrids on Nylon Membranes Peter

Kainz,’

Manfred

Seifriedsberger,

and Hans-Bernd

Strack

Department of Biochemistry, Institute of GeneralBiology, Biochemistry and Biophysics, University of Salzburg, A-5020 Salzburg, Austria

Received

August

31,1988

We have developed a modified primer extension procedure for specific detection of mRNA. Alkali-fragmented total cellular RNA or some RNA fraction is hybridized to single-stranded or double-stranded Ml3 DNA containing the insert of interest which is immobilized on nylon membranes. Hybridized RNA is then detected by incubation of membranes with Escherichia coli RNase H and DNA polymerase I. RNase H is used for nicking the RNA in the hybrids. The resulting 3’-OH groups can subsequently be used by DNA polymerase I to synthesize a labeled complementary strand. The method described is both relatively fast and sensitive and particularly useful for screening large numbers of DNA clones for their representation in RNA populations. Using total cellular RNA as hybridization probe and single-stranded Ml3 DNA as template as low as 0.25 ng of a specific mRNA was detected (2.5-fold background) when adding 1 &i [3H]dCTP or 2.5 &i [s2P]dCTP alternatively as radioactive precursor for the labeling reaction. The detection limit increased to 1 ng (a-fold background) with denatured replicative form o 1989 Academic double-stranded Ml3 DNA as template. Press,

Inc.

For analyzing differential gene expression several variants of filter hybridization methods are available. If several DNA clones are to be screened for their ability to hybridize with specific mRNAs, usually a purified mRNA population or the cDNA derived from it is end labeled by polynucleotide kinase. This fraction is subsequently hybridized with cloned recombinant cDNAs which are applied in dots to filter membranes (1). If the transcriptional state of one particular gene is investigated the method most commonly used involves the probing of northern blots or RNA dot blots with a cDNA 1 To whom 366

correspondence

should

be addressed.

which is labeled by nick translation (2,3), end labeling (4), primer extension (5,6), or random probe priming (7). In this communication we present a modified primer extension procedure for specific detection of DNA-RNA hybrids. Either single-stranded or denatured doublestranded Ml3 DNA containing an insert of interest is immobilized on nylon membranes. The formation of a specific DNA-RNA hybrid after hybridization with alkali-fragmented total cellular RNA or some RNA fraction is detected by an enzyme reaction analogous to the second strand cDNA synthesis by Okayama and Berg (8). The method described is both relatively fast and sensitive. As radioactive labeling is performed after specific DNA-RNA hybrid formation the assay is especially useful if large numbers of DNA clones are to be used as probes for specific mRNAs in a bulk population. Using existing methods, time consuming radioactive labeling of each DNA clone used would be necessary. MATERIALS

AND

METHODS

Enzymes and reagents. Escherichia coli DNA polymerase I (endonuclease free) was purchased from Boehringer-Mannheim. E. coli RNase H and restriction endonucleases HindIII, PstI, and EcoRI were obtained from Amersham (UK). Gene Screen nylon membranes were from New England Nuclear. Zeta probe and nitrocellulose membranes were purchased from Bio-Rad. Solutions. The 5X SET buffer (a fivefold concentrated stock solution) contains 5% SDS’ (Sigma), 50 mM Tris (pH 7.5), and 25 mM EDTA. The 20X SSC buffer contains 3 M NaCl and 0.3 M sodium citrate (pH 7.0). The 10X nick translation salts (10X NT) contain 200 mM Tris (pH 7.5), 100 mM MgC12, 10 mM dithiothreitol, and 250 pg/ml bovine serum albumin (endonuclease free, Sigma) and are stored at -20°C. Deoxyribonucleo’ Abbreviations used: SDS, sodium dodecyl sulfate, SSC, sodium rate-sodium chloride buffer; NT, nick translation salts.

cit-

0003-2697/89 $3.00 Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

ENZYMATIC

DETECTION

side triphosphate stock solutions (100 mM) were purchased from Pharmacia and diluted to 2 mM immediately before use. Formamide (Serva) is deionized by adding Bio-Rad AG 501-X8 (D) resin and stirring for 2 h at 20°C. The resin is removed by filtration through a Whatman No. 1 filter. Formamide stocks are stored at -20°C. For preparation of buffers, diethyl pyrocarbonate treated and autoclaved water (Millipore Q-grade) and reagents of the highest quality commercially available were used. Phenol (nucleic acid grade) was purchased from Bethesda Research Laboratories. Isolation of target DNA. The plasmid pMPa21 (a generous gift of U. Schibler) is a recombinant form of the plasmid pBR322 containing a 1.5kb fragment of cDNA coding for pancreatic cu-amylase (9,lO). This insert was cut out by digestion with restriction endonuclease PstI. The larger of the resulting two fragments (0.25 and 1.25 kb, the latter containing the 3’ end with poly(A) sequences) was subcloned into the Ml3 mp19 vector in order to prepare single-stranded Ml3 clones containing sense strand or antisense strand (complementary to mRNA) sequences, respectively (11). Because a restriction map of the cloned DNA had been determined (lo), a clone containing the antisense strand DNA sequences (Ml3 as-amyl) was identified by restriction mapping of the replicative form of the virus using the restriction enzyme EcoRI. A clone containing the sense strand DNA sequence (Ml3 s-amy2) as insert was identified by using the complementarity test (11). Ml3 DNA in its replicative form was isolated using the alkali method of Birnboim and Dolly (12). It was linearized by digestion with the restriction enzyme HindIII. DNAs were purified by agarose electrophoresis and subsequent electroelution. Preparations of membrane filters containing DNA. One-half microgram DNA of one of the single-stranded Ml3 clones was dissolved in 150 ~1 of 25 mM sodium phosphate buffer (pH 6.5) and adsorbed to Gene Screen nylon membranes or Zeta probe membranes, respectively, using a commercially available dot blot apparatus (Bio-Rad). Air-dried membranes were incubated at 37°C overnight or baked at 80°C for 2 h, respectively. One-half microgram of linearized RF Ml3 DNA was dissolved as described above and denatured by boiling at 100°C for 5 min before adsorption to the filter membrane. Preparation of RNA. Total RNA was isolated from pancreas and liver of adult mice by acid guanidinium thiocyanate-phenol-chloroform extraction (13). The RNA containing pellet was dissolved in water and fragmented by alkali treatment to enlarge the efficiency of hybridization: RNA was diluted 1:l with 0.2 N NaOH and put on ice for 1 h. This treatment produces RNA fragments 100 to 300 nucleotides long as established by glyoxal agarose gel electrophoresis. The solution containing RNA fragments was neutralized with 0.2 N HCI and passed twice through a column filled with poly(U)

OF

DNA-RNA

HYBRIDS

367

Sepharose to adsorb fragments containing poly(A) as described by Jacobson (14). In our case this step was necessary to avoid a background signal resulting from poly(T) sequences present in the Ml3 clone used containing antisense strand DNA. The ethanol precipitate of the eluate was dissolved in water and RNA concentration was determined by uv spectrophotometry at 260 nm. RNA amounts as low as 150 ng dissolved in 40 ~1 can be easily measured by using a commercially available microcell (Beckman Instruments). Aliquots were used as hybridization probes. Hybridization procedures. Membrane areas representing individual dots of immobilized single-stranded DNA were cut out, put into 1.5-ml Eppendorf tubes, and prehybridized in 0.5 ml of 50% formamide, 5X SET, 200 pg/ml yeast tRNA at 37°C for 1 h. Membrane areas with denatured, double-stranded DNA attached were subjected to a stringent wash step (10 min in boiling water) prior to prehybridization. The prehybridization solution was removed by aspiration. RNA probe was added to the hybridization solution (50% formamide, 5X SET) and incubated at 70°C for 10 min. Membranes were added and hybridization was performed in a total volume of 40 ~1 at 37°C for 18 to 24 h. Hybridized membranes were washed four times in 1 ml of 0.2X SET, 50% formamide at 37°C and twice in 1 ml of 0.2X SET at 37°C. Prior to the enzyme reaction washed membranes were soaked in 1 ml NT each and left on ice for 20 min. Enzyme reactions. One microcurie of [3H]dCTP (50 Ci/mmol, Amersham, UK) was evaporated to dryness under a gentle stream of nitrogen. Four microliters of 10X NT and 1 nmol of dATP, dGTP, and dTTP each were added followed by 2 U of DNA polymerase I and 0.04 U RNase H. In reactions with [32P]dCTP used as precursor (800 Ci/mmol, Amersham, UK) 2.5 &i was added. The final volume was adjusted to 40 ~1 with water. Membranes were added and incubated at 12°C for 1 h and subsequently at 20°C for 1 h. Membranes were washed twice in 1 ml of 0.1X SSC for 10 min at room temperature and air dried. The amount of radiolabeled DNA was determined by liquid scintillation counting in Beckman Ready-Solv HP, a high efficiency cocktail which is very effective for solubilizing biological precipitates from filters. RESULTS

AND

DISCUSSION

Figure 1 shows the incorporation of radioactivity after hybridization of different amounts of alkali-fragmented total cellular RNA from mouse pancreas with 0.5 pg of single-stranded Ml3 containing antisense strand DNA immobilized on Gene Screen nylon membranes and subsequent enzyme reactions as described above. Amounts of amylase mRNA as low as 0.25 ng were easily detected (2.5-fold background). In mouse pancreas, mRNA constitutes between 0.5 and 0.85% of the total cellular RNA

368

KAINZ,

P 7 0.2 0.4

0.8

SEIFRIEDSBERGER,

0 1 4

1.8 Tota

I

RNA

(Ng,

FIG. 1. Incorporation of radioactivity after hybridization of different amounts of total cellular pancreas RNA to 0.5 pg of immobilized single-stranded Ml3 DNA containing amylase antisense strand DNA or to 0.5 pg of double-stranded replicative form (RF) Ml3 DNA alternatively and subsequent enzyme reactions as described under Material and Methods, [3H]dCTP has been used as radioactive precursor. (0) Hybridization to antisense strand DNA; (0) control, hybridization to sense strand DNA, (A) hybridization to RF DNA; (A) control, total RNA from mouse kidney (amylase mRNA not expressed) was hybridized to RF DNA. Total cellular pancreas RNAs of 4, 1.6, 0.8, 0.4 and 0.2 pg contain 5,2,1,0.5 and 0.25 ng of cu-amylase RNA, respectively.

while amylase mRNA constitutes about 24% of total polyadenylated RNA (9). As background controls membranes either without target DNA or with 0.5 pg of Ml3 sense strand DNA were hybridized with 40 pg of pancreatic RNA. No incorporation of radioactivity above the blank value of the scintillation counter (30 cpm) was obtained in controls without target DNA, whereas the labeling reaction with Ml3 sense strand DNA as target resulted in an incorporation of as low as 2% of the values obtained from membrane bound antisense strand DNA and 40 pg of pancreatic RNA in the hybridization solution. As a further control, membranes with antisense strand DNA were hybridized with 40 pg of alkali-fragmented total mouse liver RNA after removal of poly(A) sequences on a poly(U) Sepharose column. In mouse liver a-amylase mRNA concentration is lower by about 3 orders of magnitude than in pancreas (9). The radioactivity incorporated was 5% of the values obtained from the same amount of pancreas RNA. If poly(U) Sepharose chromatography was omitted, incorporation of radioactivity increased to about 40% of the values obtained from the same amount of poly(U) Sepharose treated pancreas RNA (Table 1). In order to study whether complementary DNA interferes with the specific detection of DNA-RNA hybrids by increasing the background signal, in a further experiment 100 ng of Ml3 containing sense strand DNA was hybridized to 0.5 pg of immobilized Ml3 containing anti-

AND

STRACK

sense DNA. When hybridized membranes were incubated immediately before the final enzyme reaction with 5 U DNA polymerase I in the presence of 2 nmol dCTP, dGTP, dATP, and dTTP each for 2 h at 12”C, the background signal rose from 1000 to 2500 cpm. These results indicate that the possible presence of homologous DNA impurities in RNA preparations would not significantly interfere with the specific detection of DNA-RNA hybrids. Although the method described was first established by using a single-stranded Ml3 vector as template, the results shown in Fig. 1 indicate that the double-stranded replicative form of the Ml3 vector is also suitable as template for a specific RNA detection. However, due to enhanced background, the sensitivity of the reaction is thereby decreased to about one-fifth of that obtained by using single-stranded target DNA. Background radioactivity depends on the purity of DNA polymerase I and on the solid support material chosen. DNA polymerase I with contaminating DNase I activity produces a several times higher background than DNA polymerase I (endonuclease free) from Boehringer. Nitrocellulose, Gene Screen, and Zeta probe membranes were tested for their suitability as solid support for the RNA detection method described. Best results were obtained with Gene Screen membranes. The use of other membranes resulted in two- to threefold higher background radioactivity. As shown in Fig. 2 the concentration of RNase H has to be adjusted for an optimal signal-to-noise ratio. If the baselength of RNA is around 200 bases, the use of 0.04 U of RNase H will yield best results. If more RNase H is used, incorporation of radioactivity will decrease rapidly presumably due to the instability of the DNA-RNA hybrids. Among various incubation conditions tested (data not shown) incubation at 12°C for 1 h and subsequently at 20°C for 1 h was found to give the best results. The initial incubation at low temperature is important because of the decreased DNA-RNA hybrid stability under low salt conditions.

TABLE Effect

Total

of Poly(U) RNA

RNA

40 wg

Liver Pancreas

Sepharose

1

Chromatography

on Incorporation Poly(U) Sepharose chromatography

of Total

Without poly(U) Sepharose chromatography

(cpm) 5.8 X lo3

1.2 x lo5

Liver

of Radioactivity

kpm)

x lo4 Not determined

4.8

Note. The table shows the amount of radioactivity bound to 0.5 c(g of immobilized amylase antisense Ml3 DNA following hybridization of RNA and primer extension.

ENZYMATIC

DETECTION

3r

Units

of

RNase

FIG. 2. Effect of different concentrations poration of radioactivity after hybridization pancreas RNA to 0.5 pg of immobilized subsequent enzyme reactions.

Ii

of RNase H on the incorof 4 fig of total cellular antisense strand DNA and

Given the background of control hybridization 1 &i [3H]dCTP is more than adequate as precursor for the labeling reaction. With the background attainable at present, 200 nCi would suffice to exploit the sensitivity of the method. Unincorporated radioactive triphosphates were removed by two brief wash steps. The sensitivity of the reaction cannot be significantly improved when replacing tritiated precursors by 32P-labeled triphosphates, as long as background values are above the blank values of the scintillation counter (Fig. 3). If 3Hlabeled triphosphates are used for detection of hybrid formation localized on solid supports, quenching may cause serious departures from linearity. To obtain reliable data it is therefore very important to use liquid scintillation cocktails which effectively solubilize filter bound macromolecules. With 32P, conventional scintillation cocktails also provide reliable results. As the 1.25-kb insert of Ml3 antisense strand DNA contains poly(T) sequences of about 100 bases (10) it was necessary to remove these sequences from the hybridization mixture in order to relate incorporated radioactivity to a specific amount of amylase mRNA hybridized. Those amylase fragments containing oligo and poly(A) sequences at their 3’ end were adsorbed on poly(U) Sepharose. The concentration of amylase transcripts was thereby lowered insignificantly due to previous alkali fragmentation of total RNA to an average length of 200 bases. If alternatively RNase H in the presence of oligo(dT) (12-18) is used to cleave genuine poly(A) tails a purification step including phenol extraction and ethanol precipitation will be necessary to remove RNase H (15). In the method described, fragmented total RNA or some RNA fraction can be used as hybridization probe.

OF

DNA-RNA

369

HYBRIDS

Using 32P-end-labeled RNA for probing DNA clones immobilized on filter membranes it is necessary to purify and label the mRNA in order to obtain adequate sensitivity (16). If end-labeled mRNA is hybridized to an excess of membrane-bound cDNA a clone must be represented to a level of about 0.5% of the mass of mRNA to be detected (17). Using our method, the limit of detection is estimated to be about one order of magnitude lower: Total RNA from mouse kidney (amylase mRNA not expressed) has been treated the same way as pancreas RNA (alkali fragmentation, removal of poly(A) sequences). Hybridization of 100 pg of this RNA to 0.5 pg of immobilized, antisense strand DNA followed by primer extension resulted in a radioactivity of 5 X lo3 cpm. To answer the question whether the nonhomologous mRNA or the bulk of rRNAs and tRNAs is responsible for the increased background values, in a further experiment we hybridized various amounts of mRNA (purified on poly(U) Sepharose, then alkali fragmented and poly(U) Sepharose adsorbed) to 0.5 pg of immobilized antisense strand Ml3 DNA. The results indicate that the increase of the background values derive from mRNA, as hybridization with 1 pg of purified nonhomologous mRNA also resulted in incorporation of 5 X lo3 cpm (data not shown). As depicted in Fig. 1 0.4 pg of total pancreas RNA (containing 0.5 ng of amylase mRNA), after hybridization to 0.5 pg of amylase antisense strand DNA and subsequent primer extension, resulted in an incorporation of 5 X lo3 cpm. Therefore 1 pg of a mRNA population containing 0.5 ng of a specific mRNA (~0.05%) is estimated to produce a twofold background signal.

-

/

/ 0

0.2 0.4

0.8

1.6 Total

4 RNA

(.ug)

FIG. 3. Incorporation of radioactivity after hybridization of different amounts of total cellular pancreas RNA to 0.5 pg of immobilized antisense strand DNA and subsequent enzyme reactions as described under Material and Methods. [32P]dCTP was used as radioactive precursor. (0) Hybridization to antisense strand DNA, (0) control, hybridization to sense strand DNA. For concentration of amylase mRNA see legend to Fig. 1.

370

KAINZ,

SEIFRIEDSBERGER,

Although the method is not strictly quantitative there is a linear response of radioactivity incorporated when 0.25 to 2 ng of amylase RNA is hybridized to 0.5 pg of either single- or double-stranded DNA, respectively (Fig. 1). The departure from linearity in the RNA range above 2 ng can be attributed to differences in hybridization kinetics when increasing amounts of RNA are hybridized to a constant amount of filter bound DNA (18). The procedure most commonly used for studying the transcriptional state of a certain gene involves the probing of Northern blots or RNA dot blots with a cDNA labeled to high specific activity. It has been shown that as low as 50 pg of a specific RNA has been detected, according to extrapolation even 10 pg should be detectable (19). While the quantitative reliability of Northern blotting is limited by the extent of transfer to either nitrocellulose or nylon membranes, results from RNA dot blotting can be misinterpreted if the RNA samples are contaminated with DNA. If several DNA clones are to be used as probes for specific RNA, the dot blot method will also require labeling of each individual clone. In the study presented we have described a relatively fast and simple method for qualitative detection of specific mRNAs. The main difference between this method and published ones is that detection of DNA-RNA hybrids is accomplished by an enzyme reaction primed by hybridized mRNA fragments. Background signals are low enough for the total RNA input to be enlarged for detection of rare mRNAs with a detection limit of about 0.05%.

ACKNOWLEDGMENT We are grateful to Mrs. and for editorial work.

V. Kainz

for excellent

technical

assistance

AND

STRACK

REFERENCES 1. Dworkin,

M., and Dawid,

J. B. (1980)

2. Rigby, P. W. J., Dieckmann, J. Mol.Biol. 113.237-251.

M.,

Deu. Biol.

Rhodes,

76,435-443.

C., and Berg,

P. (1977)

3. Meinkoth, J., and Wahl, G. (1984) Anal. Biochem. 138,267-278. 4. Van de Sande, J. H., Kleppe, K., and Khorana, H. G. (1973) Biochemistry 12,5050-5055. 5. Ricca, G. A., Taylor, J. M., and Kalinyak, J. E. (1982) hoc. N&l. Acad. Sci. USA 79,724-728. 6. Hu, N., and Messing, J. (1982) 7. Feinberg, A. P., and Vogelstein, 17.

Gene 17,271-277. B. (1983) Anal.

Biochem.

132,6-

8. Okayama, H., and Berg, P. (1982) Mol. Cell. Biol. 2.161-170. 9. Schibler, U., Tosi, M., Pittet, A. C., Fabiani, L., and Wellauer, P. K. (1980) J. Mol. Biol. 142,93-116. 10. Schihler, Gellman, 266.

U., Pitted, A. C., Young, A., Hagenbiichle, O., Tosi, S., and Wellauer, P. K. (1982) J. Mol. Biol. 155,

11. Miller, H. (1987) in Guide (Berger, S. L., and Kimmel, Press, New York.

to Molecular A. R., Eds.),

M.,

247-

Cloning Techniques pp. 162-164, Academic

12. Birnboim, H. C., and Dolly, J. (1979) Nucleic Acid Res. 7, 15131521. 13. Chomczynsky, P., and Sacchi, N. (1987) Anal. Biochem. 162, 156-159. 14. Jacobson, A. (1987) in Guide to Molecular Cloning Techniques (Berger, S. L., and Kimmel, A. R., Eds.), pp. 257-258, Academic Press, New York. 15. Vournakis, J. N., Efstratiadis, A., and Kafatos, Natl. Acad. Sci. USA 72,2959-2966.

F. C. (1975)

16. Kafatos, F. C., Jones, Acids Rex 7,1541-1552.

A. (1979)

C. W., and Efstratiadis,

17. Anderson, M. L. M., and bridization-A Practical S. J., Eds.), pp. 103, IRL M. L., Sells, B. 18. Birnstiel, 63,21-41. 19. Robins, D. M., Pack, J., 29,623-631.

Proc. Nucleic

Young, B. D. (1985) in Nucleic Acid HyApproach (Hames, B. D., and Higgins, Press, Washington, DC. H., and Purdom, J. F. (1972) J. Mol. Biol. Seeburg,

P. H., and Axel,

R. (1982)

Cell