Journul of Nrurocbrmirtr~,1975 Vol 2% pp. 109-1 I 5 Pergamon Press. Prlnted ~n Great Brltam
RNA-DNA HYBRIDIZATION STUDIES WITH UNIQUE MOUSE DNA SEQUENCES M. WINTZERITH,~ K. Moru3and P . MANDEL Centre de Neurochimie du CNRS, and Unite de Recherches Fondamentales sur la Biochimie de la Cellule Candreuse de I'INSERM, Institut de Chimie Biologique. Faculte de Medecine, 67085 Strasbourg Cedex, France (Received 17 October 1974. Accepted 3 February 1975)
Abstract--Unique mouse DNA sequences (not reassociated at cot 200) were prepared and hybridized with RNA fractions extracted at different temperatures with phenol from brain, liver and plasmocytoma cells. The hybridizations were carried out in liquid media, and the RNA-DNA hybrids formed were separated on hydroxyapatite columns. The stringency of hybridization was measured by the sensitivity of the hybrids towards nuclease S , , and by the Tm"sof the RNA-DNA hybrids. A tissue specificity of RNA transcription has been found. Brain RNA's seemed to be more complex than those of liver. Brain RNA fraction I (primarily of cytoplasmic origin) hybridized to unique DNA sequences more efficiently than RNA fractions I1 + 111(enriched in nuclear messenger RNA's) whereas the reverse was true for liver and plasmocytoma cell RNA.
MOUSEbrain RNA has been reported to hybridize to a much greater extent with unique mouse DNA sequences than does liver RNA (HAHN & LAIRD, 1971; BROWN & CHURCH, 1971; GROUSE et a!., 1972). These studies were performed either on nuclear (BROWN& CHURCH,1971) or total RNA (GROUSE ef al., 1972). We have investigated this problem by further fractionating the RNA species, and by rneasuring their ability to hybridize with unique DNA sequences by separating the single stranded nucleic acids and the hybrids by hydroxyapatite chromatography. METHODS AND MATERIALS Animals
ton X-100 for 10s. The tumour nuclei were pelleted at 6009 for 10 min at O T , then repeatedly washed with fresh Tyrode buffer until the nuclei appeared pure by phase contrast microscopy. The method used to prepare DNA was essentially that of MARMUR(1961) with minor modifications. After ribonuclease treatment samples were treated with pronase (250 mg/ml) for I h at 37°C in order to completely destroy ribonuclease. D N A flagmen ta tion
DNA was sheared into fragments of approximately 500 nucleotide pairs a s previously described (MORIet al., 1972) by sonication using a MSE sonicator. Unique D N A sequences
The fragmented DNA samples were denatured in a boiling water bath for 10 min just before reassociation at 65°C in 0.12 M-sodium phosphate buffer pH 6.8 at cot 200. Sodium phosphate buffer (pH 6.8)refers to an equimolar ratio of mono- and dibasic phosphates of sodium with a Preparation of R N A pH between 6.7 and 6.8. Following BRITTEN& KOHKE Brains, livers and tumours were rapidly removed after (1968) the cot is defined as a product of DNA condecapitation of the mice and immediately immersed in centration (co) and the time of reassociation (t) expressed liquid nitrogen before lyophilisation. The RNA was frac- asmole nucleotides x s- I x lit- I. After incubation at 65'C. tionated by a previously described technique (KEMPF& the DNA was fractionated on hydroxyapatite columns MANDEL,1966) derived from those of SCHERRER & DAR- maintained at 60'C as previously described (MORIet a/.. NELL (1962) and GEORCIEV & MANTIEVA (1962). The pro- 1972) into single-stranded DNA. eluted with 0 1 2 M-sodium cedures are summmarized in Fig. I phosphate buffer (pH 6.8). DNA-DNA duplexes (here from the highly and moderately repetitive sequences) were eluted Preparation OJ D N A with 0.25 M-sodium phosphate buffer. The 012 M-sodium Organs were rapidly and immediately processed for nu- phosphate fraction of material which had not reassociated ec al. (1957) at cot 200 was taken as the unique DNA sequences. It clei preparation by the method of CHAUVEAU in the case of brain and liver. For isolating tumour nuclei contained approximately about 809., of total DNA (WINTthe tissue was homogenized with an Ultraturrax apparatus ZERITH et al.. 1973; MORIet al.. 1973). in Tyrode solution (IOml per tumour) containing 0.1% Tri-
Adult Balb/C strain mice were used in these experiments. Solid MPCI I plasmocytoma cells were transplanted 2 weeks before the animals were killed.
With the technical assistance of A. STAUBand M. C. MARFING. * Chargee de Recherche au CNRS. Present address: Department of Neuropsychiatry. Osaka City University Medical School. Osaka, Japan.
D N A labelling
The unique DNA sequences were labelled in ritru with [3H]sodium borohydride (Amersham TK47) by photoreduction according to the technique of LEE & GORWX (1971). Specific activities of 500&15.000 c.p.m. per pg were
M . WINTZERITH.K. MORIand P. MANDEL Tissue frozen in liquid nitrogen
I
Lyophilization
+ 3Oml acetate buffer-phenol/g lyophilized
+I
tissue (I vol 0 0 2 sodium ~ buffer pH 5.1; I vol rcdistilled phenol saturated with acetate buffer)
1
ISextraciion at O”C
I
Centrifugation Phenol phase discarded
Intermediate phase
phases pooled Deproteinization (Sew3
Interphase
1
+ acetate buffer-
Precipitation with 2.5 vol alcohol overnight
Centrifugation Phenol phase/l\ discarded
Intermediate phase
1
Aqueous phases poolcd
2 times r e - e x t r a c t e d H
Deproteinization I
&
Precipitation with 2.5 vol alcohol overnight
Centrifugation Phenol phase discarded
Intermediate phase
63°C
phases pooled
Deproteinization
1
2 times re-extracted
Precipitation with 2 5 vol alcohol
FIG.I. Scheme of the RNA extraction procedure (KEMPF& MANDEL,1966) used usually obtained. The radioactivity was counted in an Intertechnique SL 407 spectrometer using Instagel (Packard). No quench corrections were made since the radioactivity was always determined under the Same conditions.
R N A - D N A hybridizations assays RNA-DNA hybridization experiments were carried out in liquid media, in the presence of a vast excess of RNA. RNA, dissolved in distilled water, was pipetted into small
glass tubes of around 1.5d capacity and freeze-dried. Then about I2,M)o and 15,ooO counts of labelled unique DNA and Id of 0 7 5 M-sodium phosphate were added to each tube. The tubes were sealed. then heated for 10 min in a boiling water bath. With each set of hybridization assays 2 blanks were run containing the same quantity of DNA and sodium phosphate buBer that was used per hybridization assay in order to allow for the DNA selfreassociation. The tubes were incubated and stirred in a
RNA-DNA hybridizations with unique DNA sequences water bath at 65°C for 48h. After incubation the tubes were cooled in ice and diluted to a final concentration of 0.18 wsodium phosphate buffer. Concentrated formamide (pro analysi, Merck Germany) was also added to final concentration of 50% (v/v). The RNA-DNA mixtures were simultaneously chromatographed on hydroxyapatite columns. Hydroxyapatite was prepared according to the et al. (1956) as modified by LEVIN method of TISELIUS (1962).The hydroxyapatite was packed into water-jacketed columns, equilibrated with 0.03 M-sodium phosphate buffer (pH 6.8). The columns were 20mm high and l5mm in diameter. Samples were adsorbed at room temperature. The columns were then exhaustively washed with 200ml 0.18 M-sodium phosphate buffer (pH 6.8) at room temperature, and then at 40°C. in order to eliminate all singlestranded nucleic acids. The double-stranded material, RNA-DNA essentially, was eluted twice with 7ml of 025 M-sodium phosphate buffer (pH 6.8) at 100°C for direct counting of radioactivity or with 30ml 0.5 M-sodium phosphate buffcr (pH 6.8) at 40°C when S, enzyme treatment was subsequently carried out. In this case the hybrid RNA-DNA fractions were immediately dialysed to completely remove phosphate ions which inhibit S , enzyme activity.
S , nuclease action S , nuclease extract was prepared from commercial takadiastase (Koch-Light, G.B.) according to SUTTON(1971). Enzymatic action was performed in the medium of SUTTON (1971)at 37°C. After S, treatment the RNA-DNA in incubation mixture was diluted with 10 vol 0.03 M-sodium phosphate buffer (pH 6.8) and then adsorbed on a hydroxyapatite column to separate single-stranded nucleic acids, liberated by S , nuclease action, from the S , resistant double-stranded hybrid. The hydroxyapatite column chromatography was carried out as described above except that the extensive washings were not necessary.
Measurement of T,
T,'s of the RNA-DNA hybrid bound on hydroxyapatite column were determined by sequential thermal elution with 0.12 M-sodium phosphate buffer (pH 6.8) as already reported (MORIet a/., 1972). A final elution with 025 Msodium phosphate buffer at 100°C was performed to test total recovery from the column. The T, is defined as the temperature at which 50% of the RNA-DNA hybrid was released from hydroxyapatite column. Mathematical analysis
The amount of RNA-DNA hybridization in assays is expressed as the percent of unique DNA sequences hybridized to RNA after the substraction of the value of the blank. The results were expressed with the double reciprocal plot of I/(RNA/DNA) ratio against 1/H as proposed by BISHOP(1969) for RNA-DNA hybridizations with vast RNA excess. The intercept is the reciprocal of the percent H of DNA hybridized when the RNA/DNA ratio is infinite i.e at the saturation value. The saturation hybridization value & standard deviation was determined from the least squares regression lines l/H = f (DNA/RNA) computed on a Programma (Olivetti) on the basis of & I 1 experiments. The various experiments were carried out using several preparations of RNA and DNA. RESULTS
We hybridized RNA fractions extracted at different temperatures rather than total RNA with unique sequences of DNA. The RNA fraction I corresponded essentially to cytoplasmic ribosomal RNA and cytoplasmic messenger RNA, and secondarily to nucleolar RNA. whereas RNA fraction 111 seemed to be enriched in heterogenous nuclear RNA plus preribosoma1 RNA and DNA-like or messenger RNAs. RNA
M Sodium phosphate buffer pH 6.8
E 0 W N
kl
I I
RNA.
fractions
Ill
DNA,
fractions
FIG.2. Chromatography on hydroxyapatite columns of DNA and RNA. At left RNA. at right DNA. In (a) K N A or DNA were adsorbed in the native state onto hydroxyapatite columns. in (bl RNA or DNA were treated (heat denaturation. ionic strength. formamide) as described for RNA-DNA hybridization assays (see Methods and Material) prior to adsorption. The chromatographies were carried out similarly in (a) and (b): same load of nucleic acids (2.5-3.0 Ezhnunits). bed of hydroxyapatite. adsorption at room temperature and elution at 40°C. and stepwise elution with an increasing molarity of sodium phosphate buffer (10 fractions of 1.5ml per step).
M. WINTZERITH.K. MORIand P. MANDEL
II2
I
I o m
I
I w 3
I
0 L 0
0 005
DNA,
RNA
fraction I1 is an intermediate fraction. RNA fractions I, I1 and 111accounted respectively for on the average 70, 20 and 10% of the total cellular RNA. Before studying the hybridization properties of different RNA fractions from various tissues, the optimum conditions of separating single-stranded RNA and DNA from double-stranded RNA-DNA hybrids were established. Separate chror&tography of native RNA, native DNA and denatured RNA and DNA were performed as described in Methods and Materials (section RNA-DNA hybridization assays) (see Fig.2). As the hybridizations are done with a vast excess of RNA, virtually all the single-stranded nucleic acid material can be washed out from the hydroxyapatite column with 0. I6 M-sodium phosphate buffer. Sodium phosphate buffer of molarity above 0.25 Mcompletely eluted the hybrids at 40°C. The influence of incubation temperature on RNADNA hybridization was also studied. The optimum yield was obtained at 65°C. This temperature was the
FIG. 5. Hybridization of RNA fraction I and fractions 11 + 111 from mouse brain with 'H unique mouse DNA sequences. For the definition of H see legend to Fig. 3 and Methods and Materials (section Mathematical analysis). The dark bars on the ordinate axis represent the standard deviation of the intercept value. same used for DNA-RNA reassociations (MORI ~t a/.. 1972). The effect of incubation time was investigated. The hybridization reaction proceeded with time for a week and then slowed down. probably due to partial thermal degradation of the nucleic acids. Furthermore we have compared the method used in this study (Material and Methods, section R N A DNA hybridization assays) for the separation of RNA-DNA hybrids from the single-stranded material with that used in DNA-RNA reassociation studies (MORIet a/.. 1972): adsorption in 0.03 M-sodium phosphate buffer (pH 6.8) without formamide at W C and elution at the same temperature with 0 1 2 M-sodium phosphate buffer (PH 6.8) for the single-stranded DNA and with 025 M-sodium phosphate buffer (pH 6.8) for the double-stranded DNA. The two techniques gave practically the same result in terms of the amount of DNA hybridizable at infinite RNA concentration to a given R N A fraction (Fig. 3). Figures 4 - 6 report the results found for hybridvation of RNA fractions from liver. brain. tumour with 'H-labelled unique DNA scquenccs. Thc hybridimtion capacities of RNA fraction I and 111 wcrc analysed separately. Thc ca.w of RNA fraction 11. an inI
Tvmur
LIW
I
,t
0004
0 010
FIG.3. Hybridization of mouse brain RNA fraction I with 'H-labelled unique mouse DNA. The RNA-DNA hybrids were separated as described under Methods and Material (section RNA-DNA hybridization assays) (A A A) or as in DNA-DNA reassociation studies (MORI et al., 1972) (VVV).Brain RNA fraction 1 was used in this study. H is the percent of DNA hybridized to RNA at a given RNA-DNA ratio.
'4-4
0002
'
,
,
.
,
.,
Frodl;lU
04
0
U,oa
uooz
urn
WOI
m%Na
FIG.4. Hybridization of RNA fractions I and 111 from mouse liver "H-labelled unique DNA sequences. For the definition of H see legend to Fig.3 and Methods and Material (section Marhematical awlysis). The dark bars on the ordinate axis represent the standard deviation of the intercept value.
FIG. 6. Hybridization of RNA fractions I and 111 from mouse plasmocytoma cells with 3H unique mouse DNA sequences. For the definition of H see legend to-Fig. 3 and Methods and Materials (section Morhemarical analysis). The dark bars on the ordinate axis represent the standard deviation of the intercept value.
RNA-DNA hybridizations with unique DNA sequences
TABLE1. SATURATION HYBRIDIZATION VALUES
FOR RNA FRACTIONS FROM MOUSE LIVER, BRAIN A N 0 TUMOUR WITH UNIQVE DNA SEQUFNC‘ES
RNA
H
Tissue
fractions
(% unique DNA hybridized)
Liver
I I11
0.52 f 0.04 (10) 0.80f 0-I2 (6) 13.1 f 2.0(11) 1.0 f 0.12(6) 7.8 f 0.7(7) 14.7 8.6 (6)
___
Brain
I
I1
Tumour
+ I11 I 111
termediate of fraction I and 111, has not been considered except in brain. For brain RNA fractions I1 + 111 were hybridized together since in brain in contrast to other tissues, RNA fraction I11 is quantitatively minor. The H values a t saturation are shown in Table I . There are significant differences between the hybridizability of the tissues examined. In the brain the RNA fraction I hybridized more than RNA fractions 11 111. The tumour showed particularly high H values compared to liver. RNA fraction 111, enriched in mRNA, gave higher values than the RNA fraction I in liver and tumour. A higher dispersion of the results was observed in tumour. No differences in hybridization percent at saturation plateau were noted if 3H-labelled unique DNA sequences were prepared either from tumours, brain or liver. Some properties of the RNA-DNA hybrids were studied. The T,’s which gave information on the double-strandedness were only 4“-5”C lower than those of the native DNA (example in Fig. 7) indicating that there was a good base pairing. The mismatching responsible for the decreased Tm’sof the RNA-DNA hybrids as compared to native DNA was about 810% according to LAIRDet 01. (1969). The measurement of sensitivity of RNA-DNA hybrids to s, nuclease action was a complementary way to test the
+
Q
100-
E
e U
+
5
50-
a 0
s 40
50
60
70
80
90
00
mismatching, since this enzyme preferentially hydrolyses single-stranded DNA. Of the hybrids SO-SO% resisted S, enzyme action under conditions where denatured DNA was 70% digested. These results showed that RNA-DNA hybridization proceeded to a satisfactory level in our assays. DISCUSSION
H is expressed as the percent of DNA hybridized at saturation. followed by the standard deviation. The figures in brackets correspond to the number of determinations. The correction of DNA self renaturation was in average 0.24”/, k0-03 (21) in different RNA-DNA hybridization series.
a I
113
DO’
oc FIG. 7. T,’s of RNA-DNA hybrids and native DNA. T,”s were determined as described under Methods and Materials (section Measurement of T,,). x .~~x - .-x native DNA. MRNA-DNA hybrid. Tumour RNA fraction 111 was used in this study.
We did not use the nitrocellulose filter technique & SPIEGELMAN (1965) in our RNA-DNA of GILLESPE assays for several reasons. In our hands the DNA fragments were not retained satisfactorily by membranes and the retention varied from one batch to another, as has been previously noted by GILLESPIE & GILLESPIE (1971). Furthermore as reported by SCHMECKPEPER & SMITH(1972) and as we have observed the loss of DNA from filters increases with time and is therefore not adequate for use with 48h hybridization. Moreover it has been shown by MFLLI & BISHOP(1970) that the percent of hybridization was significantly diminished when sonicated DNA was used instead of unfragmented DNA, whereas in RNA-DNA hybridization assays carried out in liquid media no differences were noted under the experimental conditions of these authors. Moreover RNA-DNA hybridization in liquid medium is 20 times faster than that on filters (MCCARTHY,1967). The kinetics of RNA-DNA hybridization (SPIEGELMAN rt al.. 1973) have also stressed the complexity of the formation and dissociation of hybrids on membrane filters. The other technique used for RNA-DNA hybridization studies was the hydroxyapatite chromatography. first described by BERNARDI (1965) for nucleic acids and MIYAZAWA & THOMAS (1965) for DNA. However with these techniques a higher temperature is needed for the hybridization incubation and the hydroxyapatite chromatography to avoid non-specfic binding and interaction between nucleic acids. Many authors preferred using formamide throughout their experiments (GILLESPE & GILLESPIE, 1971; GROUSE et al., 1972; GOODMANet al., 1973; SCHMECKPEPER & SMITH, 1972) which allowed them to work at lower temperatures. We did not use formamide in the elution buffers (GOODMAN et al.. 1973) but only during the adsorption to the hydroxyapatite column. GOODMAWer al. (1973) obtained similar results with the high temperature method and their formamidecontaining elution buffers technique at room temperature. We obtained similar results with our technique using formamide only during adsorption and the high temperature technique. The use of formamide creates some problems since formamide from different sources may have high absorbances or conductivities and lead to high pHs (TIBBETTS et a/.. 1973) in aqueous solutions. especially in the SSC (standard saline citrate) often used in DNA studies. The pH of phosphate buffers (pH 6.8) commonly employed in hydroxyapatite chromatography was less influenced by formamide. T, and S , nuclease sensitivity measurements of RNA--DNA hybrids were used to assess the stringency of our hybridization conditions.
II 4
M.W I S T Z ~ K I TK. H . MOWand P. Mnsuft
Quantitative information dealing with the complcn- mouse brain RNA hybridvcs to :I much greater ity of RNA transcription in diflerent tissues was extent with unique mouse DNA scqueiiccs than does favoured by RNA hybridization with unique DNA h e r RNA ( H A W & LAIRD.1971; Bwowx & C'IN it. cf d..1972). sequences rather than with total DNA. Frag- 1971; GKOI'SI The hvbridtration per cent of plasmcq tima cell mentation of DNA was necessary to enable us to separate the reptitibe from the unique sequences RNA with unique mouse DNA sequence3 gave on ( B ~ i r n \& KOHNI. 1968). DNA sequences nut reas- average values intermediate k t u e m t h o v of h\cI o r sociatcd at cot IOI) have been uwd: since. a t thls cot. hrain RNA. This indicate3 the di\cr\it> 01 gene all satellite and redundant DNA sequences were eyxession in cancer cc.1I~due prohiihl! to the tr;iiiremoved under experimental conditions (MORIt'f ul.. scription of genes involved in cell prolikration and 1971; WINTZWITHef al.. 1973). As is pointed out by to structural and metabolic characteristics of plnsmot'r al. (1977) found a decreasc MCCARTHV & CHURCH (1970) "the categorization of cytoma cells. GROI'Y: nucleic acids sequences a s redundant or unique is in the hybridization saturation plateau of the turnour itself somewhat arbitrary". However other authors cell compared to the normal. but we could not check (GROUSE rf d.. 1972; BROWN& C H L R C H . I971 : H A H N this observation since we did not stud! the normal & LAIRD.1971) took respectively cot 200 and cot 220 counterpart of the plasmeqtoma cells. as the arbitrary point to separate the repetitive sequences from the unique sequences of mouse DNA. The RNA-DNA hybridization values at saturation were expressed. a s DNA hybridized in percent of unique DNA. after subtraction of the value due to selfreassociation of DNA ('Blank'). The hybridization values reported were only indicative, and probably underestimated (GFLDERMAN er ol.. 1971), since unique DNA represents 213 of the total genome and presumably only one strand is transcribed. As expected. RNA fractions extracted at different temperatures displayed different hybridization capacities with unique DNA sequences. Surprisingly brain RNA fraction I which is essentially of cytoplasmic origin. hybridized more efficiently with unique DNA sequences than RNA fraction I1 + 111. consisting of several types of large-size RNA species (preribosomal and heterogeneous nuclear RNA's with DNA-like base compositions probably messengers R NA (GEORG E ~ .1967). Brain RNA fractions were thus different to liver and tumour fractions. where fraction 111 hybridized more than fraction 1. This is a significant qualitative difference. The meaning of the different distribution of RNA species hybridi~ingwith unique DNA sequences is not yet known. However we do not think that technical ditierences can explain the different distributions. It may he that in adult brain. in contrast to the other tissues examined there werc more RNA messages available in the cytoplasm as ribonucleoproteins or as rnewnger R NA's hound to ribosomes than in the nuclei. This may hc linked to the passage of messenger RNA's from the nucleus into the cytoplasm o r to their stability. The high values of transcribed unique sequences found in hrain compared to liver showed that a significantly larger part of the mouse genome is expressed in adult brain than in liver, even if liver cells have a multiplicity of metabolic functions. These RNA-DNA hybridization results suggest that brain RNA is much more complex and may be in relation with the functions specific to nervous system. Studies by DNA-RNA hybridization of transcriptional diversity in human brain done by GROUSE er a/. (1973) demonstrated a great RNA complexity of higher neuronal function. O u r results are in agreement with reports demonstrating that
RNA-DNA hybridizations with unique DNA sequences MIYAZAWA Y. & THOMASC. A. (1965) J . molec. Biol. 11. 223237.
MORIK., WINTZERITH M . & MANDEL P. (1972) Biochimir 54, 1427-1434.
MORIK., WINTZERITH M. & MANDELP. (1973) FEBS Lett. 35, 7-10. SCHERRER K . & DARNELL J. E. (1962) Biochem. biophp. Res. Commun. 7 , 486490. SCHMECKPEPER B. J. & SMITHK. D. (1972) Biochemistry, Easton 11, 131%1326.
115
SPIECELMAN G. B., HAFJER J. E.& HALVORWN H. 0.(1973) Biochemistry, Easton 12. 1234-1242. SUTTONW. C. (1971) Biochim. hwphys. Acta 240, 522-531. TIBBE C.,~JOHANSWN K. & PHILIPSON L. (1973) J . Krol. 12. 218-225.
TISELIUS A., HEREN S. & LEVIN0. (1956) Archs Biochem. Biophys. 65. 132-1 55. W~NTZERITH M., MORIK. & MANDELP. (1973) J . Nrurochem. 21, 1341-1343.
RNA-DNA HYBRIDIZATION STUDIES WITH UNIQUE MOUSE DNA SEQUENCES M. WINTZERITH,~ K. Moru3and P . MANDEL Centre de Neurochimie du CNRS, and Unite de Recherches Fondamentales sur la Biochimie de la Cellule Candreuse de I'INSERM, Institut de Chimie Biologique. Faculte de Medecine, 67085 Strasbourg Cedex, France (Received 17 October 1974. Accepted 3 February 1975)
Abstract--Unique mouse DNA sequences (not reassociated at cot 200) were prepared and hybridized with RNA fractions extracted at different temperatures with phenol from brain, liver and plasmocytoma cells. The hybridizations were carried out in liquid media, and the RNA-DNA hybrids formed were separated on hydroxyapatite columns. The stringency of hybridization was measured by the sensitivity of the hybrids towards nuclease S , , and by the Tm"sof the RNA-DNA hybrids. A tissue specificity of RNA transcription has been found. Brain RNA's seemed to be more complex than those of liver. Brain RNA fraction I (primarily of cytoplasmic origin) hybridized to unique DNA sequences more efficiently than RNA fractions I1 + 111(enriched in nuclear messenger RNA's) whereas the reverse was true for liver and plasmocytoma cell RNA.
MOUSEbrain RNA has been reported to hybridize to a much greater extent with unique mouse DNA sequences than does liver RNA (HAHN & LAIRD, 1971; BROWN & CHURCH, 1971; GROUSE et a!., 1972). These studies were performed either on nuclear (BROWN& CHURCH,1971) or total RNA (GROUSE ef al., 1972). We have investigated this problem by further fractionating the RNA species, and by rneasuring their ability to hybridize with unique DNA sequences by separating the single stranded nucleic acids and the hybrids by hydroxyapatite chromatography. METHODS AND MATERIALS Animals
ton X-100 for 10s. The tumour nuclei were pelleted at 6009 for 10 min at O T , then repeatedly washed with fresh Tyrode buffer until the nuclei appeared pure by phase contrast microscopy. The method used to prepare DNA was essentially that of MARMUR(1961) with minor modifications. After ribonuclease treatment samples were treated with pronase (250 mg/ml) for I h at 37°C in order to completely destroy ribonuclease. D N A flagmen ta tion
DNA was sheared into fragments of approximately 500 nucleotide pairs a s previously described (MORIet al., 1972) by sonication using a MSE sonicator. Unique D N A sequences
The fragmented DNA samples were denatured in a boiling water bath for 10 min just before reassociation at 65°C in 0.12 M-sodium phosphate buffer pH 6.8 at cot 200. Sodium phosphate buffer (pH 6.8)refers to an equimolar ratio of mono- and dibasic phosphates of sodium with a Preparation of R N A pH between 6.7 and 6.8. Following BRITTEN& KOHKE Brains, livers and tumours were rapidly removed after (1968) the cot is defined as a product of DNA condecapitation of the mice and immediately immersed in centration (co) and the time of reassociation (t) expressed liquid nitrogen before lyophilisation. The RNA was frac- asmole nucleotides x s- I x lit- I. After incubation at 65'C. tionated by a previously described technique (KEMPF& the DNA was fractionated on hydroxyapatite columns MANDEL,1966) derived from those of SCHERRER & DAR- maintained at 60'C as previously described (MORIet a/.. NELL (1962) and GEORCIEV & MANTIEVA (1962). The pro- 1972) into single-stranded DNA. eluted with 0 1 2 M-sodium cedures are summmarized in Fig. I phosphate buffer (pH 6.8). DNA-DNA duplexes (here from the highly and moderately repetitive sequences) were eluted Preparation OJ D N A with 0.25 M-sodium phosphate buffer. The 012 M-sodium Organs were rapidly and immediately processed for nu- phosphate fraction of material which had not reassociated ec al. (1957) at cot 200 was taken as the unique DNA sequences. It clei preparation by the method of CHAUVEAU in the case of brain and liver. For isolating tumour nuclei contained approximately about 809., of total DNA (WINTthe tissue was homogenized with an Ultraturrax apparatus ZERITH et al.. 1973; MORIet al.. 1973). in Tyrode solution (IOml per tumour) containing 0.1% Tri-
Adult Balb/C strain mice were used in these experiments. Solid MPCI I plasmocytoma cells were transplanted 2 weeks before the animals were killed.
With the technical assistance of A. STAUBand M. C. MARFING. * Chargee de Recherche au CNRS. Present address: Department of Neuropsychiatry. Osaka City University Medical School. Osaka, Japan.
D N A labelling
The unique DNA sequences were labelled in ritru with [3H]sodium borohydride (Amersham TK47) by photoreduction according to the technique of LEE & GORWX (1971). Specific activities of 500&15.000 c.p.m. per pg were
M . WINTZERITH.K. MORIand P. MANDEL Tissue frozen in liquid nitrogen
I
Lyophilization
+ 3Oml acetate buffer-phenol/g lyophilized
+I
tissue (I vol 0 0 2 sodium ~ buffer pH 5.1; I vol rcdistilled phenol saturated with acetate buffer)
1
ISextraciion at O”C
I
Centrifugation Phenol phase discarded
Intermediate phase
phases pooled Deproteinization (Sew3
Interphase
1
+ acetate buffer-
Precipitation with 2.5 vol alcohol overnight
Centrifugation Phenol phase/l\ discarded
Intermediate phase
1
Aqueous phases poolcd
2 times r e - e x t r a c t e d H
Deproteinization I
&
Precipitation with 2.5 vol alcohol overnight
Centrifugation Phenol phase discarded
Intermediate phase
63°C
phases pooled
Deproteinization
1
2 times re-extracted
Precipitation with 2 5 vol alcohol
FIG.I. Scheme of the RNA extraction procedure (KEMPF& MANDEL,1966) used usually obtained. The radioactivity was counted in an Intertechnique SL 407 spectrometer using Instagel (Packard). No quench corrections were made since the radioactivity was always determined under the Same conditions.
R N A - D N A hybridizations assays RNA-DNA hybridization experiments were carried out in liquid media, in the presence of a vast excess of RNA. RNA, dissolved in distilled water, was pipetted into small
glass tubes of around 1.5d capacity and freeze-dried. Then about I2,M)o and 15,ooO counts of labelled unique DNA and Id of 0 7 5 M-sodium phosphate were added to each tube. The tubes were sealed. then heated for 10 min in a boiling water bath. With each set of hybridization assays 2 blanks were run containing the same quantity of DNA and sodium phosphate buBer that was used per hybridization assay in order to allow for the DNA selfreassociation. The tubes were incubated and stirred in a
RNA-DNA hybridizations with unique DNA sequences water bath at 65°C for 48h. After incubation the tubes were cooled in ice and diluted to a final concentration of 0.18 wsodium phosphate buffer. Concentrated formamide (pro analysi, Merck Germany) was also added to final concentration of 50% (v/v). The RNA-DNA mixtures were simultaneously chromatographed on hydroxyapatite columns. Hydroxyapatite was prepared according to the et al. (1956) as modified by LEVIN method of TISELIUS (1962).The hydroxyapatite was packed into water-jacketed columns, equilibrated with 0.03 M-sodium phosphate buffer (pH 6.8). The columns were 20mm high and l5mm in diameter. Samples were adsorbed at room temperature. The columns were then exhaustively washed with 200ml 0.18 M-sodium phosphate buffer (pH 6.8) at room temperature, and then at 40°C. in order to eliminate all singlestranded nucleic acids. The double-stranded material, RNA-DNA essentially, was eluted twice with 7ml of 025 M-sodium phosphate buffer (pH 6.8) at 100°C for direct counting of radioactivity or with 30ml 0.5 M-sodium phosphate buffcr (pH 6.8) at 40°C when S, enzyme treatment was subsequently carried out. In this case the hybrid RNA-DNA fractions were immediately dialysed to completely remove phosphate ions which inhibit S , enzyme activity.
S , nuclease action S , nuclease extract was prepared from commercial takadiastase (Koch-Light, G.B.) according to SUTTON(1971). Enzymatic action was performed in the medium of SUTTON (1971)at 37°C. After S, treatment the RNA-DNA in incubation mixture was diluted with 10 vol 0.03 M-sodium phosphate buffer (pH 6.8) and then adsorbed on a hydroxyapatite column to separate single-stranded nucleic acids, liberated by S , nuclease action, from the S , resistant double-stranded hybrid. The hydroxyapatite column chromatography was carried out as described above except that the extensive washings were not necessary.
Measurement of T,
T,'s of the RNA-DNA hybrid bound on hydroxyapatite column were determined by sequential thermal elution with 0.12 M-sodium phosphate buffer (pH 6.8) as already reported (MORIet a/., 1972). A final elution with 025 Msodium phosphate buffer at 100°C was performed to test total recovery from the column. The T, is defined as the temperature at which 50% of the RNA-DNA hybrid was released from hydroxyapatite column. Mathematical analysis
The amount of RNA-DNA hybridization in assays is expressed as the percent of unique DNA sequences hybridized to RNA after the substraction of the value of the blank. The results were expressed with the double reciprocal plot of I/(RNA/DNA) ratio against 1/H as proposed by BISHOP(1969) for RNA-DNA hybridizations with vast RNA excess. The intercept is the reciprocal of the percent H of DNA hybridized when the RNA/DNA ratio is infinite i.e at the saturation value. The saturation hybridization value & standard deviation was determined from the least squares regression lines l/H = f (DNA/RNA) computed on a Programma (Olivetti) on the basis of & I 1 experiments. The various experiments were carried out using several preparations of RNA and DNA. RESULTS
We hybridized RNA fractions extracted at different temperatures rather than total RNA with unique sequences of DNA. The RNA fraction I corresponded essentially to cytoplasmic ribosomal RNA and cytoplasmic messenger RNA, and secondarily to nucleolar RNA. whereas RNA fraction 111 seemed to be enriched in heterogenous nuclear RNA plus preribosoma1 RNA and DNA-like or messenger RNAs. RNA
M Sodium phosphate buffer pH 6.8
E 0 W N
kl
I I
RNA.
fractions
Ill
DNA,
fractions
FIG.2. Chromatography on hydroxyapatite columns of DNA and RNA. At left RNA. at right DNA. In (a) K N A or DNA were adsorbed in the native state onto hydroxyapatite columns. in (bl RNA or DNA were treated (heat denaturation. ionic strength. formamide) as described for RNA-DNA hybridization assays (see Methods and Material) prior to adsorption. The chromatographies were carried out similarly in (a) and (b): same load of nucleic acids (2.5-3.0 Ezhnunits). bed of hydroxyapatite. adsorption at room temperature and elution at 40°C. and stepwise elution with an increasing molarity of sodium phosphate buffer (10 fractions of 1.5ml per step).
M. WINTZERITH.K. MORIand P. MANDEL
II2
I
I o m
I
I w 3
I
0 L 0
0 005
DNA,
RNA
fraction I1 is an intermediate fraction. RNA fractions I, I1 and 111accounted respectively for on the average 70, 20 and 10% of the total cellular RNA. Before studying the hybridization properties of different RNA fractions from various tissues, the optimum conditions of separating single-stranded RNA and DNA from double-stranded RNA-DNA hybrids were established. Separate chror&tography of native RNA, native DNA and denatured RNA and DNA were performed as described in Methods and Materials (section RNA-DNA hybridization assays) (see Fig.2). As the hybridizations are done with a vast excess of RNA, virtually all the single-stranded nucleic acid material can be washed out from the hydroxyapatite column with 0. I6 M-sodium phosphate buffer. Sodium phosphate buffer of molarity above 0.25 Mcompletely eluted the hybrids at 40°C. The influence of incubation temperature on RNADNA hybridization was also studied. The optimum yield was obtained at 65°C. This temperature was the
FIG. 5. Hybridization of RNA fraction I and fractions 11 + 111 from mouse brain with 'H unique mouse DNA sequences. For the definition of H see legend to Fig. 3 and Methods and Materials (section Mathematical analysis). The dark bars on the ordinate axis represent the standard deviation of the intercept value. same used for DNA-RNA reassociations (MORI ~t a/.. 1972). The effect of incubation time was investigated. The hybridization reaction proceeded with time for a week and then slowed down. probably due to partial thermal degradation of the nucleic acids. Furthermore we have compared the method used in this study (Material and Methods, section R N A DNA hybridization assays) for the separation of RNA-DNA hybrids from the single-stranded material with that used in DNA-RNA reassociation studies (MORIet a/.. 1972): adsorption in 0.03 M-sodium phosphate buffer (pH 6.8) without formamide at W C and elution at the same temperature with 0 1 2 M-sodium phosphate buffer (PH 6.8) for the single-stranded DNA and with 025 M-sodium phosphate buffer (pH 6.8) for the double-stranded DNA. The two techniques gave practically the same result in terms of the amount of DNA hybridizable at infinite RNA concentration to a given R N A fraction (Fig. 3). Figures 4 - 6 report the results found for hybridvation of RNA fractions from liver. brain. tumour with 'H-labelled unique DNA scquenccs. Thc hybridimtion capacities of RNA fraction I and 111 wcrc analysed separately. Thc ca.w of RNA fraction 11. an inI
Tvmur
LIW
I
,t
0004
0 010
FIG.3. Hybridization of mouse brain RNA fraction I with 'H-labelled unique mouse DNA. The RNA-DNA hybrids were separated as described under Methods and Material (section RNA-DNA hybridization assays) (A A A) or as in DNA-DNA reassociation studies (MORI et al., 1972) (VVV).Brain RNA fraction 1 was used in this study. H is the percent of DNA hybridized to RNA at a given RNA-DNA ratio.
'4-4
0002
'
,
,
.
,
.,
Frodl;lU
04
0
U,oa
uooz
urn
WOI
m%Na
FIG.4. Hybridization of RNA fractions I and 111 from mouse liver "H-labelled unique DNA sequences. For the definition of H see legend to Fig.3 and Methods and Material (section Marhematical awlysis). The dark bars on the ordinate axis represent the standard deviation of the intercept value.
FIG. 6. Hybridization of RNA fractions I and 111 from mouse plasmocytoma cells with 3H unique mouse DNA sequences. For the definition of H see legend to-Fig. 3 and Methods and Materials (section Morhemarical analysis). The dark bars on the ordinate axis represent the standard deviation of the intercept value.
RNA-DNA hybridizations with unique DNA sequences
TABLE1. SATURATION HYBRIDIZATION VALUES
FOR RNA FRACTIONS FROM MOUSE LIVER, BRAIN A N 0 TUMOUR WITH UNIQVE DNA SEQUFNC‘ES
RNA
H
Tissue
fractions
(% unique DNA hybridized)
Liver
I I11
0.52 f 0.04 (10) 0.80f 0-I2 (6) 13.1 f 2.0(11) 1.0 f 0.12(6) 7.8 f 0.7(7) 14.7 8.6 (6)
___
Brain
I
I1
Tumour
+ I11 I 111
termediate of fraction I and 111, has not been considered except in brain. For brain RNA fractions I1 + 111 were hybridized together since in brain in contrast to other tissues, RNA fraction I11 is quantitatively minor. The H values a t saturation are shown in Table I . There are significant differences between the hybridizability of the tissues examined. In the brain the RNA fraction I hybridized more than RNA fractions 11 111. The tumour showed particularly high H values compared to liver. RNA fraction 111, enriched in mRNA, gave higher values than the RNA fraction I in liver and tumour. A higher dispersion of the results was observed in tumour. No differences in hybridization percent at saturation plateau were noted if 3H-labelled unique DNA sequences were prepared either from tumours, brain or liver. Some properties of the RNA-DNA hybrids were studied. The T,’s which gave information on the double-strandedness were only 4“-5”C lower than those of the native DNA (example in Fig. 7) indicating that there was a good base pairing. The mismatching responsible for the decreased Tm’sof the RNA-DNA hybrids as compared to native DNA was about 810% according to LAIRDet 01. (1969). The measurement of sensitivity of RNA-DNA hybrids to s, nuclease action was a complementary way to test the
+
Q
100-
E
e U
+
5
50-
a 0
s 40
50
60
70
80
90
00
mismatching, since this enzyme preferentially hydrolyses single-stranded DNA. Of the hybrids SO-SO% resisted S, enzyme action under conditions where denatured DNA was 70% digested. These results showed that RNA-DNA hybridization proceeded to a satisfactory level in our assays. DISCUSSION
H is expressed as the percent of DNA hybridized at saturation. followed by the standard deviation. The figures in brackets correspond to the number of determinations. The correction of DNA self renaturation was in average 0.24”/, k0-03 (21) in different RNA-DNA hybridization series.
a I
113
DO’
oc FIG. 7. T,’s of RNA-DNA hybrids and native DNA. T,”s were determined as described under Methods and Materials (section Measurement of T,,). x .~~x - .-x native DNA. MRNA-DNA hybrid. Tumour RNA fraction 111 was used in this study.
We did not use the nitrocellulose filter technique & SPIEGELMAN (1965) in our RNA-DNA of GILLESPE assays for several reasons. In our hands the DNA fragments were not retained satisfactorily by membranes and the retention varied from one batch to another, as has been previously noted by GILLESPIE & GILLESPIE (1971). Furthermore as reported by SCHMECKPEPER & SMITH(1972) and as we have observed the loss of DNA from filters increases with time and is therefore not adequate for use with 48h hybridization. Moreover it has been shown by MFLLI & BISHOP(1970) that the percent of hybridization was significantly diminished when sonicated DNA was used instead of unfragmented DNA, whereas in RNA-DNA hybridization assays carried out in liquid media no differences were noted under the experimental conditions of these authors. Moreover RNA-DNA hybridization in liquid medium is 20 times faster than that on filters (MCCARTHY,1967). The kinetics of RNA-DNA hybridization (SPIEGELMAN rt al.. 1973) have also stressed the complexity of the formation and dissociation of hybrids on membrane filters. The other technique used for RNA-DNA hybridization studies was the hydroxyapatite chromatography. first described by BERNARDI (1965) for nucleic acids and MIYAZAWA & THOMAS (1965) for DNA. However with these techniques a higher temperature is needed for the hybridization incubation and the hydroxyapatite chromatography to avoid non-specfic binding and interaction between nucleic acids. Many authors preferred using formamide throughout their experiments (GILLESPE & GILLESPIE, 1971; GROUSE et al., 1972; GOODMANet al., 1973; SCHMECKPEPER & SMITH, 1972) which allowed them to work at lower temperatures. We did not use formamide in the elution buffers (GOODMAN et al.. 1973) but only during the adsorption to the hydroxyapatite column. GOODMAWer al. (1973) obtained similar results with the high temperature method and their formamidecontaining elution buffers technique at room temperature. We obtained similar results with our technique using formamide only during adsorption and the high temperature technique. The use of formamide creates some problems since formamide from different sources may have high absorbances or conductivities and lead to high pHs (TIBBETTS et a/.. 1973) in aqueous solutions. especially in the SSC (standard saline citrate) often used in DNA studies. The pH of phosphate buffers (pH 6.8) commonly employed in hydroxyapatite chromatography was less influenced by formamide. T, and S , nuclease sensitivity measurements of RNA--DNA hybrids were used to assess the stringency of our hybridization conditions.
II 4
M.W I S T Z ~ K I TK. H . MOWand P. Mnsuft
Quantitative information dealing with the complcn- mouse brain RNA hybridvcs to :I much greater ity of RNA transcription in diflerent tissues was extent with unique mouse DNA scqueiiccs than does favoured by RNA hybridization with unique DNA h e r RNA ( H A W & LAIRD.1971; Bwowx & C'IN it. cf d..1972). sequences rather than with total DNA. Frag- 1971; GKOI'SI The hvbridtration per cent of plasmcq tima cell mentation of DNA was necessary to enable us to separate the reptitibe from the unique sequences RNA with unique mouse DNA sequence3 gave on ( B ~ i r n \& KOHNI. 1968). DNA sequences nut reas- average values intermediate k t u e m t h o v of h\cI o r sociatcd at cot IOI) have been uwd: since. a t thls cot. hrain RNA. This indicate3 the di\cr\it> 01 gene all satellite and redundant DNA sequences were eyxession in cancer cc.1I~due prohiihl! to the tr;iiiremoved under experimental conditions (MORIt'f ul.. scription of genes involved in cell prolikration and 1971; WINTZWITHef al.. 1973). As is pointed out by to structural and metabolic characteristics of plnsmot'r al. (1977) found a decreasc MCCARTHV & CHURCH (1970) "the categorization of cytoma cells. GROI'Y: nucleic acids sequences a s redundant or unique is in the hybridization saturation plateau of the turnour itself somewhat arbitrary". However other authors cell compared to the normal. but we could not check (GROUSE rf d.. 1972; BROWN& C H L R C H . I971 : H A H N this observation since we did not stud! the normal & LAIRD.1971) took respectively cot 200 and cot 220 counterpart of the plasmeqtoma cells. as the arbitrary point to separate the repetitive sequences from the unique sequences of mouse DNA. The RNA-DNA hybridization values at saturation were expressed. a s DNA hybridized in percent of unique DNA. after subtraction of the value due to selfreassociation of DNA ('Blank'). The hybridization values reported were only indicative, and probably underestimated (GFLDERMAN er ol.. 1971), since unique DNA represents 213 of the total genome and presumably only one strand is transcribed. As expected. RNA fractions extracted at different temperatures displayed different hybridization capacities with unique DNA sequences. Surprisingly brain RNA fraction I which is essentially of cytoplasmic origin. hybridized more efficiently with unique DNA sequences than RNA fraction I1 + 111. consisting of several types of large-size RNA species (preribosomal and heterogeneous nuclear RNA's with DNA-like base compositions probably messengers R NA (GEORG E ~ .1967). Brain RNA fractions were thus different to liver and tumour fractions. where fraction 111 hybridized more than fraction 1. This is a significant qualitative difference. The meaning of the different distribution of RNA species hybridi~ingwith unique DNA sequences is not yet known. However we do not think that technical ditierences can explain the different distributions. It may he that in adult brain. in contrast to the other tissues examined there werc more RNA messages available in the cytoplasm as ribonucleoproteins or as rnewnger R NA's hound to ribosomes than in the nuclei. This may hc linked to the passage of messenger RNA's from the nucleus into the cytoplasm o r to their stability. The high values of transcribed unique sequences found in hrain compared to liver showed that a significantly larger part of the mouse genome is expressed in adult brain than in liver, even if liver cells have a multiplicity of metabolic functions. These RNA-DNA hybridization results suggest that brain RNA is much more complex and may be in relation with the functions specific to nervous system. Studies by DNA-RNA hybridization of transcriptional diversity in human brain done by GROUSE er a/. (1973) demonstrated a great RNA complexity of higher neuronal function. O u r results are in agreement with reports demonstrating that
RNA-DNA hybridizations with unique DNA sequences MIYAZAWA Y. & THOMASC. A. (1965) J . molec. Biol. 11. 223237.
MORIK., WINTZERITH M . & MANDEL P. (1972) Biochimir 54, 1427-1434.
MORIK., WINTZERITH M. & MANDELP. (1973) FEBS Lett. 35, 7-10. SCHERRER K . & DARNELL J. E. (1962) Biochem. biophp. Res. Commun. 7 , 486490. SCHMECKPEPER B. J. & SMITHK. D. (1972) Biochemistry, Easton 11, 131%1326.
115
SPIECELMAN G. B., HAFJER J. E.& HALVORWN H. 0.(1973) Biochemistry, Easton 12. 1234-1242. SUTTONW. C. (1971) Biochim. hwphys. Acta 240, 522-531. TIBBE C.,~JOHANSWN K. & PHILIPSON L. (1973) J . Krol. 12. 218-225.
TISELIUS A., HEREN S. & LEVIN0. (1956) Archs Biochem. Biophys. 65. 132-1 55. W~NTZERITH M., MORIK. & MANDELP. (1973) J . Nrurochem. 21, 1341-1343.
